University of Groningen. Improving safety culture in health care Listyowardojo, Tita Alissa

Transcription

1 University of Groningen Improving safety culture in health care Listyowardojo, Tita Alissa IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below. Document Version Publisher's PDF, also known as Version of record Publication date: 2012 Link to publication in University of Groningen/UMCG research database Citation for published version (APA): Listyowardojo, T. A. (2012). Improving safety culture in health care: implications of individual and institutional variability Groningen: s.n. Copyright Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons). Take-down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim. Downloaded from the University of Groningen/UMCG research database (Pure): For technical reasons the number of authors shown on this cover page is limited to 10 maximum. Download date:

2 Improving Safety Culture in Health Care: Implications of Individual and Institutional Variability Tita Alissa Listyowardojo

3 Cover design by Briyan B. Hendro (brianos.daportfolio.com) and Myrna Herawati Layout by Briyan B. Hendro (brianos.daportfolio.com) Printed by Off Page ( 2012, Tita Alissa Listyowardojo, Groningen, The Netherlands ISBN (book): ISBN (ebook):

4 Improving Safety Culture in Health Care: Implications of Individual and Institutional Variability Proefschrift ter verkrijging van het doctoraat in de Gedrags- en Maatschappijwetenschappen aan de Rijksuniversiteit Groningen op gezag van de Rector Magnificus, dr. E. Sterken, in het openbaar te verdedigen op donderdag 19 januari 2012 om uur door Tita Alissa Listyowardojo geboren op 14 februari 1981 te Jakarta, Indonesia

5 Promotor: Prof. dr. A. Johnson Copromotor: Dr. R. E. Nap Beoordelingscommissie: Prof. dr. K.A. Brookhuis Prof. dr. T.S. van der Werf Prof. dr. P.T.W. Hudson

6 We should have the courage to change ourselves in response to changes in life, to reach perfection without waste, all at our own pace. (Sugandhi Listyowardojo & Eka Andriani-Listyowardojo) Dedicated to my dad in heaven and my mom on earth

7

8 Contents Preface Introduction Variations in Hospital Worker Perceptions of Safety Culture Demographic Differences between Health Care Workers who did or did not. respond to a Safety and Organizational Culture Survey Perceptions of Personal Health Risks by Medical and Non-Medical Workers in a University Medical Center: A Survey Study Using Social Network Analysis to Identify Sub-Groups in the Operating Room The Importance of Frequency of Working Together in the Operating Room General Discussion Summary Reference List Acknowledgement...129

9

10 Preface Betsy Lehman, a knowledgeable health reporter for the Boston, died from an overdose of cyclophospha, during her chemotherapy in The medical team who treated her failed to react to the excessive reaction (e.g., swelling, vomiting, and abnormal blood tests) she and her husband complained of after the treatment. Willie King had the wrong leg amputated in 1995 because of the decentralized information about patient information. The error on the patient file was found by a nurse who talked to Willie King earlier, and corrected approximately 3 hours before the surgery, but the corrected information never reached the surgeon. Eight year-old Ben Kolb died in 1995 during a minor surgery after receiving an injection of adrenaline for external use only instead of an anesthetic. This happened because the containers of adrenaline and anesthetic were wrongly labeled. A two-month old Jose Eric Martinez died in 1996 because of a misplaced decimal when his digitalis dose was written as.90 instead of.09. Jesica Santillan, a 17 year old girl, died in year 2003 after receiving donated heart and lungs with the wrong blood type. The surgeon in charge did not double check the match of the blood types because he assumed that when the organs were offered by the institution, they were the matched organs for Jesica. The mistake was found 1 1/2 hours after the organ transplant was completed and Jesica was diagnosed brain dead immediately. These are just a few of the stories to make headlines in the past years, all raising the question: Isn t health care supposed to be safe? Unfortunately, health care is not always safe. The Institute of Medicine (1999) has reported that human errors in health care were the eighth leading cause of death in the United States. In the Netherlands, it is reported that 1 of 20 deaths in hospitals is preventable (Zegers et al., 2009). It may seem reasonable to blame the health care workers (HCWs) for the preventable deaths because they are the ones who treat the patients. Ironically, the patients and their families are not the only victims of medical errors. Medical errors also cause deep scars to those who commit them. Scott et al. (2009) describe those who commit errors as the second victims. The HCWs may be traumatized by the event and feel personally responsible for the patient outcome. Furthermore, they may feel that they have failed the patient, causing them to question their clinical skills and knowledge base. Scott et al. give an example of the emotional distress that a HCW experienced after an error occurred. In the words of the HCW:

11 I remember feeling horribly sad that I couldn t do more for this child. This hit me harder than most of them. For some reason I m really related with this family. I guess one reason is that the child was the age of my oldest daughter and I guess that I felt that this could have been my family. They were a nice family and didn t deserve to have this outcome. I cried a lot over this case and I guess I still cry when I think about her. The emotional distress for HCWs who struggle with fallibility at work can also last for a long time. Doctor Wendy Levinson (Levinson & Dunn, 1989) has publicly shared her personal story about coping with a misdiagnose she made years ago that caused the death of a patient: It has been 12 years since my internship, but I frequently think about a mistake I made one night when I was on call..[..] the patient died and I had to tell his wife. Although I realized that many factors contributed to the patient s demise, I felt sick about my judgment error and ashamed the next day when the chief of medicine reprimanded me. The emotional distress of HCWs involved in medical errors is said to be caused by the pressure for HCWs to be infallible in doing their work (Bosk, 2003). However, as acknowledged by the Institute of Medicine (1999), to err is human. The question is not who to blame for an error but why errors occur and what we can do to prevent them. To understand and deal with these problems, we need to look not only at factors in health care that may contribute to error occurrences, such as miscommunication, but also at factors that may prevent errors, such as how HCWs work together in delivering patient care and whether a culture of health care supports or endangers patient safety.

12

13

14 Introduction

15 Medical Errors and a Blame-Free Culture To err is human, but for those who provide medical care, the consequences of a minor error can be catastrophic. Medical error is defined as the failure of a planned action to be completed as intended (i.e., an error of execution) or the use of a wrong plan to achieve an aim (i.e., an error of planning) whether or not the failure results in harm (Institute of Medicine, 2000). Medical error can happen in all stages in the process of care, from diagnosis, to treatment, to preventive care. Medical errors are said to cause preventable adverse events or errors in management due to the failure to follow accepted practice at an individual or system level in managing a condition in question (Zegers et al., 2007). Some adverse events are not considered preventable because the events may be new or unanticipated because of, for example, changing technologies. An example of an unavoidable adverse event is an unanticipated allergic reaction of a patient to an antibiotic (Baker et al., 2004). Regardless of the cause (e.g., errors in management or unknown drug mechanisms), all unintended injuries or complications that are caused by health care management rather than by the patient s underlying disease, which lead to death, disability at the time of discharge or prolonged hospital stay are defined as adverse events (Baker et al., 2004; de Vries, Ramrattan, Smorenburg, Gouma, & Boermeester, 2008). Adverse events are recognized as the leading cause of in-patient morbidity and mortality. The Harvard medical practice study (Brennan et al., 1991), which investigated adverse event rates using a retrospective hospital record review, showed that adverse events occurred in 3.7% of hospitalizations in New York State in which 2.6% of the total adverse events led to permanently disabling injuries and 13.6% led to death. Of the total adverse events, 27.6% were judged to be preventable adverse events. The retrospective hospital record review has been used widely, with modifications, to calculate adverse event rates in other countries such as in Australia (Wilson et al., 1995), Colorado and Utah (Thomas et al., 2000), the United Kingdom (Vincent, Neale, & Woloshynowych, 2001), New Zealand (Davis et al., 2002), Denmark (Schiøler T et al., 2002), France (Michel, Quenon, de Sarasqueta, & Scemama, 2004), Canada (Baker et al., 2004), Sweden (Soop, Fryksmark, Koster, & Haglund, 2009), the Netherlands (Zegers et al., 2009) and Tunisia (Letaief, El Mhamdi, El-Asady, Siddiqi, & Abdullatif, 2010). These studies have estimated that approximately 2.9% to 16.6% of hospitalized patients experience at least one adverse event (Zegers et al., 2009) of which 27.6% to 70% of adverse events were considered preventable (Letaief et al., 2010; Zegers et al., 2009; Soop et al., 2009). In the Netherlands, Zegers et al. (2009) have estimated that the national incidence of preventable 14 Improving Safety Culture in Health Care: Individual and Institutional Variability

16 adverse events among hospitalized patients is 2.3%. Among deceased patients, 4.1% had experienced preventable adverse events that contributed to their deaths. The adverse event rates estimated in the Harvard medical practice studies (Brennan et al., 1991; Thomas et al., 2000) led to a report by the Institute of Medicine (2000) suggesting that medical errors resulted in up to 100,000 deaths per year in the United States. Although this number has provoked some debates (Leape, 2000; McDonald, Weiner, & Hui, 2000), the report has raised awareness that medical errors are a serious problem and need to be reduced. Reducing medical errors requires an understanding of why errors occur. Reason (2000) has suggested that errors occur because of active failures and latent failures. Active failures are errors and violations committed by those in direct contact with the patient or system. Their consequences are visible and almost immediately or at least within a few hours. Errors fall here include forgetfulness, inattention, poor motivation, carelessness, negligence and recklessness of the individuals. Latent failures are the delayed-action consequences of decisions taken in the upper echelons of an organization or system. These failures relate to the design and construction of equipments, the structure of an organization, organizational planning and scheduling, training and selection of workers, budget distribution, allocating resources and so forth. The adverse safety effects of these decisions may lie latent for a very long time. The adverse safety effects of latent failures can create an accident when they are combined with active failures to create an accident opportunity. Latent failures are suggested to cause most adverse events (Institute of Medicine, 2000; Reason, 2000). However, blaming and punishing the individuals who commit active failures is the most common approach to adverse events (Dekker, 2007). This approach in dealing with errors is called a person approach (Reason, 2000). Errors in a person approach are treated as moral issues and those who commit errors are viewed as bad people who need to be punished. The Institute of Medicine (2000) has emphasized the importance of shifting a culture of blaming and punishing HCWs who commit errors to the one that focuses on improving the safety systems in order to build a safer health care organization. This shift is significant because blaming and punishing individuals who commit errors is said to worsen the quality of care (Dekker, 2007). For example, blaming and punishing HCWs who commit errors can cause emotional distress (e.g., sense of shame, self-doubt, fear, guilt and depression) to HCWs and can significantly affect their work (Levinson & Dunn, 1989; Schwappach & Boluarte, 2009; Vincent, 2003). As importantly, blaming and punishing HCWs who commit errors is likely to create a culture that discourages error-reporting (Dekker, 2007; Institute of Medicine, 2000; Institute of Medicine, 2001; Institute of Medicine, 2004). Introduction 15

17 The Institute of Medicine report (2004) makes the point that a culture that encourages the sharing rather than the hiding of errors and near misses is needed to promote patient safety. This is necessary because error traps and pitfalls can only be revealed and learned from by detailed analyses of incidents and accidents. Without error-reporting, error occurrences are expected regardless of who commits errors. Achieving safer health care needs an approach which seeks out and removes the error provoking properties within the system instead of within the individuals (Dekker, 2007). Reason (2000) has proposed a system approach that is based on the assumption that although we cannot change the human condition, we can change the conditions under which humans work. Who commits errors is not the focus in a system approach. The focus is on understanding how and why the systems fail. This approach, nevertheless, does not merely blame the system when an error occurs. This approach emphasizes developing safety systems by continuously involving the individuals within the system in safety improvements (Dekker, 2007). Encouraging error-reporting. Health care workers may be reluctant to report or disclose errors because they fear that the information shared may fall into the wrong hands (e.g., the press). They fear that reporting errors may lead to legal trials, damage to one s reputation or loss of their licenses. To encourage error-reporting, leaders in organizations should provide a secure and trusting environment. As importantly, leaders should ensure that the information shared is highly valued and used to improve safety systems. Leape (2002) has suggested several characteristics of effective reporting systems. First and foremost, reporting systems should be nonpunitive in that they protect reporters from retaliation or punishment from others. The reporting systems should also be highly confidential in that the identities of patient, reporter and institution are protected from third parties. The reporting systems should also be independent from any authority with power to punish reporters or organizations. Every reported error should be evaluated by experts who understand the clinical circumstances and are trained to recognize underlying systems. The analysis of reports and the recommendations made should be disseminated immediately to those who need to know. The recommendations made from the error-reporting systems should also emphasize improvement in processes instead of in individual performance. Implementing error-reporting systems needs to be accompanied by balancing safety and accountability. Dekker (2007) introduced the concept of a just culture, or a culture in an organization where a collective understanding is possessed of where the line should be drawn between blameless and blameworthy actions. A just culture aims to satisfy demands for accountability and to still contribute to learning from errors and improvement of safety sys- 16 Improving Safety Culture in Health Care: Individual and Institutional Variability

18 tems. In promoting a just culture, building relationships and trust between HCWs and of the organization leaders are fundamental. This can be done by abolishing all financial and professional penalties following an error occurrence, trying to prevent stigmatization of HCWs involved in an incident or accident, and building and reviewing the effectiveness of debriefing, critical incident or stress management programs to support HCWs after an incident or accident. Promoting a just culture is an essential early step in creating a culture that supports patient safety. Safety culture. It has become clear that the culture in health care environments is significant for patient safety. For example, a blame-free culture and a just culture are said to improve patient safety because such cultures encourage error-reporting. Improving the culture in health care needs an understanding of how a culture is created so that the culture can be changed if necessary. Antonsen (2009) describes the process of culture creation in three stages based on the model of Berger and Luckmann (1966). A culture starts to be created when people project their interpretation of the world into the world through their verbal and non-verbal actions. This projection is the first stage of culture creation and is called externalization. When the process of projecting interacts with other people s projections, the interactions form the basis of habitualized actions (i.e., recurrent patterns of action and interaction). Over time, these habitualized actions may be considered as the only right way to do things by the people involved. When ways of doing things become the objective reality around the people involved, the creation of a culture reaches the second stage, which is the institutionalization. Through the process of socialization, the institutionalized reality is projected back into the individual s consciousness. This third stage is called internalization. Culture can influence safety when the people involved consider no alternatives for habitualized actions even when a behavioral modification is needed to prevent an incident. To illustrate the role of culture in safety, Turner and Pidgeon (1997) introduced the man-made disaster model. They use this model to explain the complex process of events that accumulate over time and result in accidents. Turner and Pidgeon introduced a phase called incubation in their model. During this phase, relatively small hazards or problems receive more attention than do danger signals that will lead to real hazards or problems such as repeated violations of procedures. This phase can be compared to the latent failures in Reason s theory (Reason, 2000). The ability of people in health care to notice signals leading to disasters determines the organizational level of vulnerability in terms of dealing with errors. A cultural readjustment is needed when the people involved fail to detect danger signals prior to an event that triggers a disaster. The failure to detect danger signals prior to a disaster is caused Introduction 17

19 by rigidities of perception and beliefs (Turner & Pidgeon 1997, p. 47). These perceptions and beliefs comprise the frames of reference through which organizational members interpret situations in their surroundings. The frames of reference include shared perceptions about what is considered to be safe and what is considered to be dangerous. These frames of reference can cause a collective blindness towards specific hazards if something falls outside their frames of reference. An illustration of collective blindness as a result of rigidities of perceptions and beliefs is the events leading to the establishment of the Josie King foundation. Josie King was an 18 month-old toddler who was admitted to a Pediatric Intensive Care Unit at Johns Hopkins hospital because of first and second degree burns resulting from climbing into a hot bath. Shortly before her planned discharge, Josie s mother became concerned, noticing that Josie screamed every time she saw a drink and that she sucked vigorously on the washcloth when she was bathed. The mother s concern, however, was not shared by the HCWs who treated Josie. Instead, the HCWs reassured the mother that children often do this kind of thing and that Josie s vital signs were considered normal. Josie died two days before she was planned to return to home; a factor contributing to her death was severe dehydration. The culture in health care thus influences patient safety. The culture that supports patient safety, or, safety culture can be defined as the compilation of HCWs attitudes, values, beliefs, perceptions and assumptions towards organizational practices that directly or indirectly support patient safety (Huang et al., 2007; Nieva & Sorra, 2003; Pronovost et al., 2003b; Pronovost & Sexton, 2005; Reason, 2000; Sexton et al., 2006a; Singer et al., 2003; Singer et al., 2009b; Singer, Falwell, Gaba, & Baker, 2008; Antonsen, 2009). Apart from a commitment of hospital management to promote and protect error-reporting and to use the reports for safety improvement, characteristics of a strong safety culture include a commitment of all HCWs - ranging from the managers, administrative workers, laboratory workers, and so forth, who have limited contact with patients, to nurses, surgeons and others who come into direct contact with patients regularly - to make patient safety the main priority. The presence of safety culture is a potentially critical determinant of successes of other activities to improve patient safety (McCarthy & Blumenthal, 2006; O Connor, 2007; Zohar, Livne, Tenne-Gazit, Admi, & Donchin, 2007). For example, sustaining teamwork skills acquired through team training is unlikely to be effective if at the same time a strong safety culture is not promoted (Morey, Simon, Jay, & Rice, 2003). Another example is that critical care units with a strong safety culture were found to have a relatively low rate of subsequent blood-stream infection (Sexton et al., 2006a). 18 Improving Safety Culture in Health Care: Individual and Institutional Variability

20 Assessing and Promoting Safety Culture Understanding how different groups perceive safety culture is an important step in determining what and for whom institutional safeguards should be implemented to enhance patient safety. Surveying HCWs attitudes, values and perceptions towards organizational practices with regards to patient safety can be seen as a tool for identifying problematic areas (Nieva & Sorra, 2003; Sexton et al., 2006a). Information gained by assessing safety-related perceptions or attitudes can be used by hospital leaders to design institutional safeguards or educational programs to promote safety culture (Edmondson, 2003; Sexton et al., 2006a; Singer et al., 2009b). For example, Pronovost et al. (2003b) evaluated safety culture in a hospital and found that frontline HCWs in general perceived that their direct supervisors were more committed to patient safety than were senior hospital leaders. The study also found that physicians were less aware than were nurses of the proper channels to use to report adverse events. The results of the study were used to recommend that senior leaders efforts to improve patient safety be more visible than before to frontline HCWs, and that more should be done to involve and educate physicians about patient safety in the hospital. Variations in safety culture between groups within a hospital are commonly found. For example, differences in safety culture have been found between intensive care units (ICUs; (Huang et al., 2007) and between clinicians and senior managers (Pronovost et al., 2003b). These variations can, on the one hand, be disadvantageous for patient safety if they lead to false expectations and communication breakdowns (Singer et al., 2003; Huang et al., 2007). On the other hand, variations in safety culture can foster organizational learning about safety because learning is mediated by differences of perspectives among organizational members (Gherardi, Nicolini, & Odella, 1998; Antonsen, 2009). The focus of safety culture promotion thus should not be on achieving a homogenous and conflict-free organization, but on facilitating understanding between groups (Antonsen, 2009). Results of safety culture assessment can be used as outcome measures for the effectiveness of intervention programs to improve patient safety (Sexton et al., 2006a). For example, the effectiveness of executive walk rounds, in which hospital executives make visits to patient care areas to discuss patient safety issues with frontline HCWs, was measured using the Safety Climate Survey (Thomas, Sexton, Neilands, Frankel, & Helmreich, 2005). The Safety Climate Survey assesses the perceptions of frontline HCWs about patient safety practices in their clinical area and managerial commitment to patient safety. The executive walk rounds intervention program was found to improve safety-related attitudes and perceptions Introduction 19

21 of HCWs. The effectiveness of a comprehensive unit-based safety program ( CUSP ), a program aimed at the work unit level to engage and empower HCWs to identify and eliminate patient safety hazards (Pronovost et al., 2005), was also measured by assessing safety-related perceptions before and after the implementation of the CUSP program. The safety-related perceptions were measured using the Safety Climate Scale (Sexton & Klinect, 2001), a survey that assesses the extent to which HCWs perceive a strong and proactive organizational commitment to patient safety. Fourteen of the 21 items of the Safety Climate Scale were rated more positively after the implementation of CUSP, indicating that CUSP was effective in improving HCW perceptions of organizational commitment to patient safety. Nieva and Sorra (2003) suggested several criteria that need to be considered in choosing safety culture assessment surveys. First, the reliability and validity of a survey in question are important. The Safety Attitudes Questionnaire (SAQ), for example, is a validated survey with strong reliability developed by Sexton et al. (2006a). The SAQ measures six dimensions of safety-related attitudes (e.g., perceptions of strong organizational commitment to patient safety, job satisfaction, teamwork and stress recognition) and is focused on identifying problems in clinical areas such as in the ICU and the emergency department. The second criterion in evaluating a safety culture survey is to consider the settings in which the survey was developed. The SAQ, for example, may not be suitable for surveying non-clinical HCWs such as hospital managers. Considering the target population is the third criterion of evaluating a survey. The SAQ, for example, has several versions to measure safety culture in specific clinical areas such as the ICU and the operating room. The fourth criterion concerns the domain of culture (i.e., specific or broad cultural domains) that is to be assessed. For example, hospital leaders interested in measuring the effectiveness of team training programs to improve team skills of HCWs in the ICU can use the Intensive Care Unit Management Attitudes Questionnaire (ICUMAQ) developed by Thomas et al (2003). The ICUMAQ elicits attitudes suggested to predict performance in teams. Health care senior managers interested in capturing differences in safety culture between hospitals and professional groups, on the other hand, might use the Stanford Patient Safety Center of Inquiry (Singer et al., 2003). This survey is designed to capture differences in safety culture between groups. Important safety-related perceptions and attitudes of HCWs such as management and institutional commitment to safety, communication openness and teamwork are the focus of the inquiry. Adopting Crew Resource Management (CRM) concepts. Health care workers often need to work in teams consisting of members from different professional groups, not only because different expertise is needed to carry out tasks but because teams are said to accom- 20 Improving Safety Culture in Health Care: Individual and Institutional Variability

22 plish most safety-critical tasks better than do individuals (Kendall, 2003). Effective teamwork in patient care has also been known to reduce medical errors (Harris, Treanor, & Salisbury, 2006; Morey et al., 2003). Hall and Weaver (2001) distinguished three forms of health care teams consisting of members from different professional groups. The first form is the multidisciplinary team in which team members from different professional groups contribute their particular expertise in treating a patient. Direct communication between team members, however, is minimal. The only direct communication between team members is between a physician in charge and team members. The second form is the transdiciplinary team in which team members roles are overlapped and interchangeable. Every team member in the transdisciplinary team should be familiar with the concepts and approaches of the other members to enable the team to focus on a problem. The third form is the interdisciplinary team in which team members from different professional groups work together closely and communicate directly to each other to deliver care for a patient. Holistic management focuses on treating the patient and every team member contributes equally and considers each other s contributions. The interdisciplinary team is the focus in this chapter because this form is said to have benefits that may be lacking in the other forms, including enhanced patient compliance and greater patient satisfaction, increased efficiency and reduced costs, and decreases in hospitalization, cost of care and use of physicians (Walker et al., 1998). For the interdisciplinary team to work effectively in delivering patient care, every team member should possess teamwork skills such as effective communication and the ability to coordinate individual efforts to achieve a collective goal. However, possessing teamwork skills is not natural for HCWs. This concern is rooted in the way HCWs are trained. Acquiring teamwork skills is not part of their formal education; HCWs from different professional groups are trained separately (Hall & Weaver, 2001). This can lead to discrepant perceptions between professional groups regarding interdisciplinary teamwork (Sexton et al., 2006b). Makary et al. (2006), for example, found that operating room surgeons and anesthesiologists were more satisfied with physician-nurse collaboration than were nurses. Nurses perceived surgeons as being unapproachable and thus nurses were often reluctant to express their concerns to surgeons about patient care. Intervention programs such as team training may be needed to improve teamwork skills of HCWs. Crew resource management (CRM) intervention programs have been proposed to be the most effective safety programs in aviation to improve teamwork of cockpit crewmembers (Helmreich, Merritt, & Wilhelm, 1999). It is thus unsurprising that the success of CRM has attracted health care organizations to adopt CRM-concepts. Crew resource management is Introduction 21

23 defined as a set of instructional strategies designed to improve teamwork in the cockpit by applying well-tested training tools (e.g., performance measures, exercises, feedback mechanisms) and appropriate training methods (e.g., simulators, lectures, videos) targeted at a specific content including teamwork, knowledge, skills and attitudes (Salas et al., 1999). To train team members to use all the available resources (e.g., supporting facilities, crewmembers and airplane systems) to achieve a safety goal is the purpose of CRM team training. Crew resource management was first introduced by the NASA (National Aeronautic and Space Administration) in Crew resource management has been developed for more than 25 years and said to significantly improve safety in the aviation. For example, the United States fatality rates were less than one-third the fatality rate in mid-century. The safety peak was achieved in 1998 when there were no deaths in the United States commercial aviation (Institute of Medicine, 2000). Until the air crash in Lexington, Kentucky, on August 27, 2006, the National Transportation Safety Board had reported only two fatalities on air carriers in the United States since January 2003 (Doucette, 2006). The obvious improvement in aviation safety has led to the question of whether the safety techniques used in aviation can be applied to medical settings. Crew resource management intervention programs are improved from time to time by investigating incidents and accidents in aviation to discover error pitfalls. This is done by investigating objective records of actions and communications from cockpit voice and flight data recorders ( the black boxes ). The absence of the black boxes in health care can make the applications of CRM intervention programs a challenge (Guerlain et al., 2005). Morey et al. (2003) also listed differences in work environments of aviation and health care that affect the applications of CRM intervention programs in health care. For example, aviation usually practices a unified command structure (i.e., a team consisting of a pilot as the commander of cockpit crew members), whereas health care has separate departments for medicine, nursing and administration. The aviation industry has also incorporated CRM team training in aviator proficiency evaluation, whereas team training is not central to HCW evaluation. Aviation and medical settings also share similar goals such that both settings stand in complex systems that require best possible interactions between the actors to achieve safety. Stressful environments, the need for highly functioning teams, the importance of accurate and precise communication and the high cost associated with system failures are some similarities between aviation and medical settings. The comparable high-risk environments of aviation and health care have led to appeals to implement CRM-intervention programs in medical practices (Rivers, Swain, & Nixon, 2003; Morey et al., 2003; Morey & Salisbury, 2002). 22 Improving Safety Culture in Health Care: Individual and Institutional Variability

24 The use of checklists in the operating room (OR). Studies investigating adverse event rates using retrospective hospital record review have found that approximately half of all adverse events in the hospital are related to surgical procedures (Leape et al., 1991; Soop et al., 2009; Brennan et al., 1991; Thomas et al., 2000; Baker et al., 2004; Davis et al., 2002; Vincent et al., 2001; Letaief et al., 2010; Zegers et al., 2009). Errors associated with deaths are most likely to occur in the OR (Calland et al., 2002b). The reason that the OR is the most common site of adverse events in hospitals is due in part to the large number of groups that must coordinate their individual efforts in order to administer effective patient care (Lingard et al., 2004). Ineffective communication between HCWs has been mentioned in previous studies as the cause of errors in the OR (Lingard et al., 2004; Lyndon, 2006; Awad et al., 2005; Shendell- Falik, Feinson, & Mohr, 2007; Davies, 2005). Lingard et al. (2004) documented 421 procedural communication events over a 3-month period in the OR and found that 30% of the events resulted in communication failures that affected team performance. Lingard et al. categorized communication failures into errors of occasion (i.e., inappropriate situations or contexts of the communication event), content (i.e., insufficient or inaccuracy of the shared information), audience (i.e., missing appropriate individuals in the communication) and purpose (i.e., missing goals in communication events). Halverson et al. (2011) added errors of omission (i.e., absence of appropriate communication) and inappropriate communication (i.e., presence of inappropriate communication such as offensive remarks) into the categorization of communication failures. Halverson et al. also found that communication between HCWs about OR equipment or instruments and progress reports (i.e., updating other team members of the progress or changes related to operations or patients) most commonly led to communication errors. Team briefings before an operation using a structured checklist are said to reduce communication failures in the OR (Lingard et al., 2008; Halverson et al., 2011). A checklist is defined as a list that highlights essential actions or criteria arranged in a systematic manner to allow users to record the presence or absence of individual items or actions to ensure that all are considered or completed (Hales & Pronovost, 2006). The use of checklists in aviation has become part of flight protocol and they are completed before, during and after flights (Hales & Pronovost, 2006). An absence of checklists completion in aviation is considered a violation. In health settings, Haynes et al. (2009) have developed a peri-operative checklist aimed at promoting people to people transfer of information to establish a common understanding between team members. The checklist is conducted before an induction of anesthesia, before a skin incision and before a patient leaves the OR. The checklist includes 19 questions about the patient s identity, such as the name of the patient and the side of the patient s body that Introduction 23

25 needs to be operated upon, and procedural issues such as patient allergies and anticipated blood loss. The implementation of the checklist in several countries has been associated with reductions of death and complication rates (Haynes et al., 2009). The peri-operative checklist, however, is argued to be insufficient to reduce all errors related to surgery (de Vries, Boermeester, & Gouma, 2008). De Vries et al. (2009) found that more than 50% of errors in surgical processes happen before or after the peri-operative pathways such as when a patient is just admitted to a hospital or when a patient is back at the ward after an operation. De Vries et al. (2009) developed and validated the Surgical Patient Safety System checklist targeting the entire surgical pathway that consist of four stages (i.e., preoperative ward, OR, recovery or ICU and postoperative ward). The checklist is focused on patient transfers including admission to and discharge from the hospital. All HCWs involved in treating a patient are responsible for completion of the checklist. For example, ward doctors and nurses should complete the checklist when the patient is admitted to and discharged from the hospital, and OR surgeons, anesthesiologists and nurses should complete the checklist before and after an operation. The implementation of the checklist in six Dutch hospitals was associated with a significant reduction in surgical complications and mortality. The use of checklists in the ICU. High workload and time constraints in the ICU may increase the risk of missing essential information during patient transfers or handoffs between HCWs. Performing multiple tasks under time pressure increases HCWs levels of stress and fatigue, and compromises their level of cognitive function (Bourne & Yaroush, 2003). Hales and Pronovost (2006) emphasized the importance of a checklist to reduce HCWs cognitive workload and the risk of errors in the ICU. Using checklists may also facilitate and encourage clinician-to-clinician communication (Hales, Terblanche, Fowler, & Sibbald, 2008). Reader et al. (2007), for example, found that nurses reported less openness in communication than did doctors in an ICU in the United Kingdom and that trainee doctors showed more negative perceptions of communication openness than did senior doctors. Less communication openness may lead to missing essential information or actions. Pronovost et al. (2003a) developed a daily goals sheet for use in the ICU that lists specific tasks of HCWs for each patient. Examples of items in the daily goals sheet were What needs to be done for the patient to be discharged from the ICU?, What is this patient s greatest safety risk? How can we reduce that risk?, Medication changes (can any be discontinued?) and Can catheters/tubes be removed? The implementation of the daily goals sheet was said to significantly increase HCWs understanding of their specific tasks for the day. After approximately one year of implementing the daily goals sheet in an ICU, the length of stay 24 Improving Safety Culture in Health Care: Individual and Institutional Variability

26 of the patients decreased from a mean of 2.2 days to a mean of 1.1 days. The use of the daily goals sheet was also found to contribute to decreased rates of catheter-related bloodstream infection in 108 ICUs in Michigan, The United States (Pronovost et al., 2006). Although the use of checklists seems to be useful in reducing HCWs workload, an overwhelming number of required checklists can easily impede the quality and speed of care delivery. The use of required checklists thus should be monitored based on their suitability to a unit, quality and necessity (Hales & Pronovost, 2006). Simulation-based team training. Simulation-based CRM team training has been found to improve teamwork skills of cockpit crew members (Helmreich et al., 1999). In health settings, CRM-based team training in a simulated work environment was initiated in anesthesiology (Anesthesia Crisis Resource Management; (Gaba, Howard, Fish, Smith, & Sowb, 2001). Applications of simulator-based team training have since been adapted to other areas such as the OR, labor and delivery units for neonatal resuscitation, and the emergency department (Helmreich & Merritt, 2001; Risser et al., 1999; Halamek et al., 2000; Morey & Salisbury, 2002). The obvious benefit of simulation-based team training is that team members can be trained in various crisis scenarios to gain or retain specific teamwork skills without having to endanger patients (Gaba, 2004). Developing team training, nevertheless, requires a training-needs analysis. Salas et al. (1999) suggested a systematic methodology as a guide to develop CRM-based team training. The methodology includes thorough analysis of procedures in a domain and their impact on teamwork demands, identifying specific tasks that need to be completed using team elements, using the specific tasks to create scenarios for team training, developing measures of team performance that meet training objectives, and using the results of measurable team performance for giving effective feedback. Salas et al. suggested that the training sessions should begin with a lecture explaining important teamwork skills followed by presenting videotaped models of effective and ineffective teamwork behaviors. After the didactic sessions, trainees should be given the opportunity to practice effective teamwork behaviors in a simulated environment in which their performance is tracked. Feedback should be given at the end of training. In health care, a combination of the didactic sessions and simulation-based team training is found to improve HCWs attitudes towards the importance of teamwork skills and effective team performance for patient safety (Morey et al., 2002; Morey et al., 2003; Morey & Salisbury, 2002). Grogan et al. (2004), for example, developed CRM-based team training for health settings that is focused on improving six key areas for effective teamwork including fatigue man- Introduction 25

27 agement, team building, communication, recognizing adverse events, team decision making and performance feedback. Team training was implemented in several departments such as a trauma unit and an emergency department. The 8-hour team training included lectures on CRM concepts and principles to educate HCWs about explicit behaviors that were important for teamwork, and role-playing in simulated scenarios to give HCWs opportunities to practice the behaviors. The team training was found to increase HCWs attitudes towards the six key areas. For example, HCWs perceived that speaking up was imperative for patient safety, and that good communication and teamwork skills were as important as technical proficiency. Health care workers strongly agreed that the team training could reduce errors. Saphiro et al. (2004) developed CRM-based team training to improve interdisciplinary team performance in an emergency department at a Level 1 trauma center. The CRM-based team training developed was a combination of the Emergency Team Coordination Course (ETCC), an 8-hour didactic course and a daylong simulation-based teamwork training. The ETCC was taught by a physician and nurse pair, and was focused on five team dimensions including maintaining team structure and climate, applying problem solving strategies, communicating with the team, executing plans and managing workload, and improving team skills. The simulation-based teamwork training consisted of three patient care scenarios of increasing complexity. Interdisciplinary team members showed a trend towards improvement in the quality of team behaviors and rated simulation-based training as a useful method to improve teamwork skills in the emergency department. Morey and Salisbury (2002) suggested that team training needs to be tailored to the operational characteristics of a department in question. This is necessary because there may be differences in error patterns (i.e., frequencies of occurrence for individual behaviors attributed to teamwork failures) and teaming structures (i.e., single vs. multiple teams that are involved in treating a patient) between departments. These differences may affect the content of team training. Morey and Salisbury also emphasized that implementing team training may require a change in infrastructure in the hospital in question. If the successful implementation of team training is to be achieved, the safety culture in a hospital in question should also be continuously promoted to sustain the teamwork skills acquired during team training. Otherwise, time and efforts spent in training HCWs in the simulator are likely to resemble time spent in a theme park (O Connor, 2007). 26 Improving Safety Culture in Health Care: Individual and Institutional Variability

28 A View from the Top : Bridging the Gap between Safety Procedures and Practices In January 1983, the Challenger space shuttle broke apart only 73 seconds after it was launched. The technical failure of the shuttle s solid rocket boosters was said to cause the explosion (Presidential Commission on the Space Shuttle Challenger Accident, 1986). Engineers involved in the construction of the shuttle s solid rocket boosters acknowledged the technical problem before the launch and reported their safety concern to their direct managers (Winsor, 1988). Although the managers reported the technical problem to the senior managers, the senior managers eventually made the management decision to launch the space shuttle as scheduled regardless of the problem perceived by the engineers (Winsor, 1988; p. 106). The resulting accident may have been prevented if the engineers had not failed to convince the senior managers of gravity of the problem prior to the launch and if senior managers had involved engineers during the decision-making processes (Winsor, 1988). The Challenger space shuttle accident illustrates that different perceptions between senior managers and frontline workers of what is considered dangerous and what is considered safe can lead to erroneous decision-making of senior managers that can compromise safety (Winsor, 1988). Different perceptions of safety practices between senior managers and frontline workers are also found in health care. Previous studies found that frontline HCWs continue to report concerns about organizational practices with regards to patient safety although many senior managers maintain that patient safety is an organizational priority (Singer et al., 2003; Singer et al., 2008; Pronovost et al., 2003b). Singer et al. (2008) have suggested that frontline HCWs firsthand experiences in dealing with patient care may cause them to identify more problems with patient safety practices than do those who have limited contact with patients (e.g., senior managers) because, for example, frontline HCWs often receive complaints from patients. The discrepant perceptions of patient safety practices between senior managers and frontline HCWs may happen because frontline workers are often reluctant to share negative information with their supervisors (Winsor, 1988; Vaughan, 1997) because of, for example, the fear of being blamed for the problems (Vogelsmeier, Scott-Cawiezell, Miller, & Griffith, 2010). In addition, negative information that reaches senior managers may be interpreted in optimistic ways that confirm established organizational programs (Winsor, 1988; Singer et al., 2008). Discrepant perceptions of patient safety practices between senior managers and frontline HCWs imply that upward-downward communication in health care may not always be effective. Ineffective communication between senior managers and frontline HCWs may Introduction 27

29 lead to senior managers falsely sensing that their organizations are actively managing patient safety issues by implementing effective safety rules and procedures (Vogelsmeier et al., 2010; Antonsen, 2009). Senior managers have to rely on reported information by frontline HCWs regarding problems related to patient safety practices in order to determine changes needed and to assess their attempts to create and maintain safety culture. It is thus important for frontline HCWs to report problems about patient care because the organization s actual safety performance is more closely reflected in frontline HCWs perceptions of safety practices than in senior managers (Vogelsmeier et al., 2010). Through effective communication between managerial and frontline workers, shared perceptions of what is considered dangerous and what is considered safe can be achieved (Parker et al., 2009; Morey & Salisbury, 2002; O Connor, 2007). It is suggested that frontline HCWs need to speak up more than do senior managers, whereas senior managers need to listen more than do HCWs if effective communication is to be achieved (Edmondson, 2003; Tucker & Edmondson, 2003; Singer et al., 2008; Pronovost et al., 2003b). Senior managers need to involve frontline HCWs in the designation, modification and implementation of safety rules and procedures. Open communication with senior managers can also be perceived by frontline HCWs as reflecting the effort of senior managers to show their managerial commitment to patient safety. This is important because perceived managerial commitment to patient safety is found to be a significant predictor of safety culture (Feng, Acord, Cheng, Zeng, & Song, 2011). Outline of the Thesis This thesis aims to give recommendations to health care senior managers in their efforts to design intervention programs to improve safety culture. Chapter 2 reports the results of a safety culture survey investigating how different professional groups in a university hospital perceive organizational practices that can influence patient safety. Understanding different perceptions of organizational practices that can influence patient safety between professional groups is important in determining what and for whom intervention programs should be implemented to improve safety culture (Nieva & Sorra, 2003). Because achieving a high response rate is a documented problem in surveying HCWs (Asch, Jedrziewski, & Christakis, 1997; Asch, Connor, Hamilton, & Fox, 2000; Cartwright, 1978; Cummings, Savitz, & Konrad, 2001; Ward, 1994; VanGeest, Johnson, & Welch, 2007), 28 Improving Safety Culture in Health Care: Individual and Institutional Variability

30 Chapter 3 is focused on investigating possible non-response bias. This is done by comparing demographic characteristics of HCWs who did or did not respond to the safety culture survey in Chapter 2. Evidence that variables such as professional group, age, years of working in an institution and supervisory responsibilities influence perceptions of safety culture (Singer et al., 2009a; Cooper & Phillips, 2004; Lin, Tang, Miao, Wang, & Wang, 2008a) give reasons to investigate differences in response rate as a function of these demographic characteristics. This chapter documents differences in response rate that should be taken into account to reduce the risk of non-response bias. The implementation of intervention programs to improve safety culture can only be effective if HCWs comply with the safety programs. The extent to which HCWs comply with safety regulations is likely to be related to their perceptions of the personal risks involved with the behaviors being regulated (Sjöberg, 2003; Sjoberg, 2003). Chapter 4 thus investigates different perceptions of personal health risks between medical and non-medical HCWs. Compared to non-medical workers who have limited contact with infected patients, medical workers are exposed to various occupational health hazards which can result in serious longterm adverse health outcomes. Understanding how different groups perceive personal health risks is important to inform and complement future research on safety compliance. This study also has implications for the design of intervention programs to increase the level of compliance of HCWs. Chapter 5 and 6 are focused on investigating the benefits of frequently working together with particular OR team members. This is important because the OR is considered a high-risk area in which disproportionately more harm may result from errors than elsewhere in the hospital (Baker et al., 2004; Entin, Lai, & Barach, 2006; Soop et al., 2009; Thomas et al., 2000; Wilson et al., 1995; Zegers et al., 2009). This is said to be due in part to the large number of workers that must coordinate their individual efforts to administer effective patient care as a team (Harrison, Mohammed, McGrath, Florey, & Vanderstoep, 2003). The frequency with which OR team members work together can influence patient safety. Teams whose members are familiar to each other are likely to work more efficiently and create higher quality products than are teams whose members are strangers to each other (Harrison et al., 2003). Identifying OR team members who frequently work together is a first step in investigating characteristics of team coordination that can develop only if team members work together frequently. Identifying OR team members who frequently work together, however, is difficult because of the complexity of OR environments. The density of the data (high number of operations and OR team members) and ad-hoc scheduling practices in which any available OR Introduction 29

31 team members are assigned to operations are some of the barriers in identifying sub-groups by hand. Chapter 5 thus describes social network analysis techniques that are used to identify sub-groups in the OR whose members frequently work together. Chapter 6 reports findings of the interviews with the members of the sub-groups identified in Chapter 5. Participants are asked to recall successes in preventing an incident from becoming an accident in the OR in which in that operation they work together with familiar team members. Participants are asked the key factors that lead to the successful outcomes. Chapter 6 describes these key factors that highlight the importance of frequency of working together in the OR. 30 Improving Safety Culture in Health Care: Individual and Institutional Variability

32

33

34 Variations in Hospital Worker Perceptions of Safety Culture A shorter version of this chapter has been published as: Listyowardojo T.A.,Nap R.E.,Johnson A. Variations in hospital worker perceptions of safety culture. International Journal for Quality in Health Care (2011); doi /intqhc/mzr069

35 Abstract Objective. To compare the attitudes towards and perceptions of institutional practices that can influence patient safety between all professional groups at a university medical center. Design. A questionnaire measuring 9 dimensions of organizational and safety culture was distributed to all hospital workers. Each item was rated on a 1 ( strongly disagree ) to 5 ( strongly agree) scale. Participants professionals, grouped as physicians (16.6%), nurses (40.3%), clinical workers (e.g., psychologists; 21.7%), laboratory workers (e.g., technicians; 11%) and nonmedical workers (e.g., managers; 10.4%). Main Outcome Measures. One-way ANOVAs carried out separately on each dimension with professional group as the independent variable of interest. Results. Differences in ratings of organizational and safety culture were found across professional groups. Physicians and non-medical workers tended to rate the dimensions of organizational and safety culture more positively than did nurses, clinical workers and laboratory workers. For example, physicians gave more positive ratings of institutional commitment to safety than did nurses, clinical workers and laboratory workers (mean=3.71 vs. 3.62, 3.61 and 3.58, respectively, p < 0.01) and non-medical workers gave more positive ratings than did physicians, nurses, clinical workers and laboratory workers to perceptions towards the hospital (mean= 3.69 vs. 3.39, 3.36, 3.49 and 3.47, respectively, p < 0.001). Conclusions. Interventions to promote safety culture should be tailored to the target group as attitudes and perceptions may differ among groups. 34 Improving Safety Culture in Health Care: Individual and Institutional Variability

36 Introduction The implementation of institutional safeguards to enhance patient safety has become a major focus of health care organizations, especially since the landmark reports of the Institute of Medicine (IOM) in (Institute of Medicine, 2001; Institute of Medicine, 1999). The IOM estimated that 44,000 to 98,000 people in the United States alone die in hospitals each year because of preventable medical errors, making medical errors the eighth leading cause of death. Importantly, the IOM pinpointed failed systems and procedures, and not simply the negligence of health care workers, as the cause of 90% of these deaths. The 2004 IOM report (Institute of Medicine, 2004) makes the point that a culture that encourages the sharing rather than the hiding of errors and near misses is needed to promote patient safety. Such a shift from a culture in which workers are discouraged from reporting errors to one in which they are encouraged to report errors or failures may be accomplished by stopping the practice of focusing blame on the health care workers at the sharp-end and focusing instead on processes and procedures to improve patient safety that cut across individual units or hospital functions (Dekker, 2007). Improving patient safety requires the institution of an organizational culture that supports patient safety, or, safety culture (Antonsen, 2009). The safety culture of a hospital can be defined as the compilation of hospital workers attitudes, values, beliefs, perceptions and assumptions towards organizational practices that directly or indirectly influence patient safety (Reason, 2000). Characteristics of a strong safety culture include a commitment of hospital management to promote and protect error reporting and to use the reports for safety improvement, a commitment of leaders to flatten hierarchies by, for example, decentralizing authority for each unit (Singer et al., 2009a), and, especially, a commitment of all hospital workers - ranging from the managers, administrative workers, laboratory workers, and so forth, who have limited contact with patients to nurses and surgeons, and so forth. who come into direct contact with patients regularly - to make patient safety the main priority. Variations in safety culture, such as between ICU units (Huang et al., 2007) or between clinicians and senior managers (Pronovost et al., 2003b), may compromise patient safety because variations in safety culture may lead to unmet expectations and communication breakdowns (Singer et al., 2003; Huang et al., 2007). Understanding how different groups perceive safety culture is thus an important step in determining what and for whom institutional safeguards should be implemented to enhance patient safety. Assessment of attitudes, values and perceptions towards organizational practices known to promote safety culture can be seen as Variations in Hospital Worker Perceptions of Safety Culture 35

37 a tool for identifying problematic areas with regards to patient safety (Nieva & Sorra, 2003; Sexton et al., 2006a). Information gained by assessing safety-related perceptions or attitudes can be used by hospital leaders to design institutional safeguards or educational programs to improve safety culture, such as assertiveness training for nurses who find speaking up difficult (Edmondson, 2003; Sexton et al., 2006a; Singer et al., 2009b). Few prior studies have compared perceptions of safety culture variables of all professional groups within a hospital (e.g., Singer et al., 2009b; Singer et al., 2007; Singer et al., 2003). The current study thus aims to gain insight in institutional practices with regards to patient safety from all professional groups within a large university medical center in the Netherlands in order to identify areas for improvement. Methods Setting The study was conducted in the University Medical Center Groningen (UMCG) in the Netherlands, a hospital with approximately 1300 beds including 53 surgical and medical adult intensive care beds, and 46 neonatal and pediatric intensive care beds. The UMCG is the only university medical center in the northern part of the country. Therefore, it is the major hospital of referral for patients with many types of illness and an important center for all organ transplants. Participants From April to June 2009, invitations to participate in the survey study were sent electronically to all 5609 employees (including part-timers) of the UMCG on the basis of employment records using hospital addresses. The groups of interest were physicians and physician assistants, medical specialists (e.g., oncologists), nurses (nurses in all departments at the UMCG including Intensive Care, Emergency and Surgical Departments), laboratory workers (e.g., laboratory technologists and technicians), clinical consultants (e.g., dieticians, psychologists and pharmacists), clinical support workers (e.g., pharmacist assistants, perfusionists and biologists), facility management workers, secretarial and administrative workers, research and teaching assistants and managers (i.e., non-medical heads of department whose main task is to ensure business continuity). For the purposes of analysis, medical specialists, physicians and physician assistants 36 Improving Safety Culture in Health Care: Individual and Institutional Variability

38 were treated as one professional group ( physicians ). Nurses were treated as one professional group regardless of department ( nurses ). Similarly, clinical consultants and clinical support workers were both considered to be clinical workers. Facility management workers, secretarial and administrative workers, research and teaching assistants and managers were included in the group non-medical workers because these workers did not have direct contact with patients and the number of workers in each group was too small to be analyzed separately. Members of any professional group were considered to hold an executive function if they led a sector, department, unit, sub-unit or clinic and were involved directly in organizational decision making or policy making. Participation was on a voluntary basis and informed consent was given by respondents. The questionnaire was filled in on-line and anonymously. Responses were thus confidential. Anonymity was ensured by treating the responses to the demographic questions (other than professional group) separately from the questionnaire responses. The Medical Ethical Testing Committee of the University of Groningen waives Institutional Research Board approval for anonymous questionnaires conducted under personnel. Instrument The questionnaire was created by a professional testing center with experience measuring organizational and safety culture variables. The decision was made to create the questionnaire by adapting questions from existing questionnaires (Hackman & Oldham, 1975; Sexton et al., 2006a) and theories (Solomon, 1980; Saari & Judge, 2004; Salancik & Pfeffer, 1978; Hackman & Oldham, 1976; Locke, 1969) in order to cover the range of topics in which the organization was interested with a reasonable number of questions. The questionnaire consisted of 99 items representing nine dimensions of safety and organizational culture and an additional five demographic questions (professional group, gender, age group, years of working in the hospital and executive or non-executive function). Variations in Hospital Worker Perceptions of Safety Culture 37

39 Table 1. Overview of Organizational and Safety Culture Dimensions and Sample Items Dimensions Institutional commitment to safety (8 items): perceptions of a strong and proactive organizational commitment to safety. Sample items I would feel safe being treated here as a patient, The culture in my department makes it easy to learn from the errors of others. Teamwork climate (6 items): perceived quality of collaboration between hospital workers in delivering patient care. Disagreements in the department here are resolved appropriately (i.e., not who is right but what is best for the patient), The physicians and nurses here work together as a wellcoordinated team. Team performance (12 items): perceived quality of team performance. My team delivers high quality results, I experience good teamwork between different teams. Work satisfaction (16 items): positivity about work experiences. I enjoy my work, I have sufficient opportunities to carry out tasks that I am good at doing. Working conditions (10 items): perceived quality of work environment and logistical support (e.g., schedules, workload, and equipment). I am satisfied with my working hours, I am equipped with appropriate instruments and materials to carry out my work. Collegiality (8 items): perceived quality of collegial atmosphere. Relations with supervisors (13 items): perceived quality of supervision. I have pleasant colleagues, I am motivated by my colleagues. My supervisors give clear objectives, I receive sufficient feedback from my supervisors. Perceptions towards the hospital (17 items): perceptions towards the organization s objectives, structures, and practices. I am proud of the hospital, I support the objectives of the hospital. Relations with supervisors (13 items): perceived quality of supervision. Career perspectives (9 items): perceived quality of self-development and career perspectives. My supervisors give clear objectives, I receive sufficient feedback from my supervisors. At work, I can gain new knowledge sufficiently, I have opportunities to develop my career in this hospital. 38 Improving Safety Culture in Health Care: Individual and Institutional Variability

40 Table 1 gives an overview of the dimensions and sample items within each dimension. The dimensions institutional commitment to safety and teamwork climate were nearly identical to the dimensions safety climate and teamwork climate, respectively, of the Safety Attitudes Questionnaire (SAQ) Intensive Care Unit-version (Sexton et al., 2006a). The items were adapted so that they were not specific to any particular departmental context (e.g., High levels of workload are common in this ICU was adapted to read High levels of workload are common in my department ) and one item from the safety climate dimension ( Briefings (e.g., morning briefings and pre- and post-operative briefings) are important for patient safety ) was replaced by My team pays much attention to patient aftercare and handoffs in the dimension institutional commitment to safety. Of the other SAQ dimensions (i.e. job satisfaction, perceptions of management and working conditions ), only stress recognition was not represented in our questionnaire. All items were phrased as statements and rated according to the amount of agreement or disagreement ranging from 1 (strongly disagree) to 5 (strongly agree). Cronbach s α calculated for each dimension revealed acceptable levels of reliability for each dimension (Cronbach s α = 0.82 for institutional commitment to safety, 0.80 for teamwork climate, 0.88 for team performance, 0.87 for work satisfaction, 0.71 for working conditions, 0.89 for collegiality, 0.95 for relations with supervisors, 0.91 for perceptions towards the hospital and 0.81 for career perspectives). Data analysis Scores on each of the dimensions were computed by averaging the ratings for the items within a dimension. Scores for the five professional groups (physicians, nurses, clinical workers, laboratory workers and non-medical workers) were compared for each dimension using one-way ANOVAs. Because of the need to preserve the anonymity of the respondents, it was not possible to couple demographic information to questionnaire responses. Pearson correlation coefficients were computed between all dimensions. A significance level of p < 0.05 was used for all analyses, Bonferroni corrected for pairwise comparisons between professional groups within each dimension. Variations in Hospital Worker Perceptions of Safety Culture 39

41 Results Out of 5609 invitations sent, 2995 employees filled out the questionnaire (response rate 53.4%). Non-medical workers (n = 311) had the highest response rate (67.5%) followed by nurses (n = 1208, response rate 58.1%), clinical workers (n = 649, response rate 56.1%), laboratory workers (n = 331, response rate 51.0%) and physicians (n = 496, response rate 46.5%). The majority of the participants were female (72.2%). Five age groups were used as response categories: years old (4.1% of the respondents), years old (28.6%), years old (24.5%), 45 to 54 years old (28.6%) and over 54 years old (14.2%). Time spent working in the hospital was categorized as less than 5 years (35.4%), 5-9 years (27.5%), years (9.8%) or more than 20 years (17.3%). Only 5.6% of respondents had an executive function. The mean scores for all dimensions were higher than 3 (the neutral point) out of 5 possible points, implying that, in general, perceptions of organizational and safety culture were positive. Although most differences were small, significant differences between professional groups were found for all dimensions except for team performance (see Table 2). Physicians and non-medical workers tended to rate the dimensions of organizational and safety culture more positively than did nurses, clinical workers and laboratory workers. Physicians gave more positive ratings of institutional commitment to safety than did nurses, clinical workers and laboratory workers. Teamwork climate was rated more positively by physicians and non-medical workers than by clinical workers. Non-medical workers gave more positive ratings of work satisfaction than did physicians, nurses, clinical workers and laboratory workers. Working conditions was rated more positively by laboratory workers and non-medical workers than by physicians, nurses and clinical workers; clinical workers gave higher ratings than did physicians and nurses. Physicians and nurses rated collegiality more positively than did laboratory workers. Physicians and non-medical workers gave more positive ratings of relations with supervisors than did nurses, clinical workers and laboratory workers; nurses gave higher ratings than did clinical workers. Perceptions towards the hospital was rated more positively by non-medical workers than by physicians, nurses, clinical workers and laboratory workers. Perceptions towards the hospital was also rated more positively by clinical workers than by physicians and nurses, and by laboratory workers than by nurses. Physicians and non-medical workers gave more positive ratings of career perspectives than did nurses, clinical workers and laboratory workers. Nurses and clinical workers did, however, rate career perspectives more positively than did laboratory workers. 40 Improving Safety Culture in Health Care: Individual and Institutional Variability

42 Table 2. Mean ratings of safety and organizational culture dimensions across five pro fessional groups (standard deviation in parentheses) Dimension Physicians (n=496) Nurses (n=1208) Clinical workers (n=649) Laboratory workers (n=331) Non-medical workers (n=311) p-value Institutional commitment to safety 3.71 (0.54) 3.62 (0.49) 3.61 (0.55) 3.58 (0.56) 3.66 (0.60) <0.01 Teamwork climate 3.87 (0.51) 3.81 (0.48) 3.74 (0.53) 3.85 (0.47) 3.89 (0.54) <0.001 Team performance 3.77 (0.54) 3.75 (0.47) 3.74 (0.52) 3.73 (0.48) 3.83 (0.56) <0.05 Work satisfaction 3.72 (0.52) 3.73 (0.44) 3.76 (0.51) 3.77 (0.45) 3.97 (0.48) <0.001 Working conditions 3.37 (0.49) 3.31 (0.45) 3.49 (0.49) 3.70 (0.44) 3.65 (0.50) <0.001 Collegiality 3.98 (0.57) 3.96 (0.47) 3.92 (0.56) 3.84 (0.52) 3.94 (0.62) <0.01 Relations with supervisors 3.72 (0.78) 3.55 (0.77) 3.44 (0.81) 3.50 (0.75) 3.77 (0.78) <0.001 Perceptions towards the hospital 3.39 (0.57) 3.36 (0.51) 3.49 (0.50) 3.47 (0.48) 3.69 (0.54) <0.001 Career perspectives 3.54 (0.57) 3.39 (0.54) 3.37 (0.60) 3.25 (0.55) 3.62 (0.59) <0.001 Significant positive correlations ranging from 0.27 (between the dimensions collegiality and working conditions ) to 0.68 (between the dimensions work satisfaction and career perspectives ) were found between all dimensions (see Table 3). Correlations between dimensions for each professional group were all positive and significant at p Correlations above 0.59 were found between the dimensions institutional commitment to safety, teamwork climate and team performance, suggesting common variance attributable to shared awareness of safety procedures. The team performance dimension was highly correlated with the dimensions collegiality and work satisfaction. The work satisfaction dimension was also highly correlated with working conditions. Correlations above 0.59 were also found between the dimensions perceptions towards the hospital, work satisfaction and career perspectives, Variations in Hospital Worker Perceptions of Safety Culture 41

43 suggesting that this part of the questionnaire measured underlying organizational climate. In order to test the reliability of the findings, a bootstrapping method was used by which separate ANOVAs were carried out on just 10% of the data. For this analysis, 10% of the data was randomly sampled, and ANOVAs were carried out on each dimension with professional group as a between-subjects factor. This procedure was repeated 100 times. Five of the eight dimensions (work satisfaction, working conditions, relations with supervisors, perceptions towards the hospital and career perspectives) consistently showed significant results despite the greatly reduced number of participants included in the analysis. The remaining three dimensions (institutional commitment to safety, teamwork climate and collegiality) that were significant in the overall analysis showed a clear tendency towards significance (distribution skewed towards p = 0.000) in the bootstrapping method (see Appendix 1). Table 3. Correlations between safety and organizational dimensions (all correlations are significant at p < 0.001) 1 Teamwork climate Team performance Work satisfaction Working conditions Collegiality Relations with supervisors Perceptions towards the hospital Career perspectives Institutional commitment to safety Teamwork climate Team performance Work satisfaction Working conditions Collegiality Relations with supervisors Perceptions towards the hospital A Pearson correlation coefficient was computed to assess the relationships between safety and organizational dimensions. 42 Improving Safety Culture in Health Care: Individual and Institutional Variability

44 Discussion A questionnaire study was conducted to determine how different professional groups perceive safety culture. In general, ratings of organizational and safety culture were positive. However, in an institution where few had strong negative responses, we still detected small, but statistically significant differences between professional groups. Physicians and nonmedical workers tended to give more positive ratings of dimensions of organizational and safety culture than did nurses, clinical workers and laboratory workers. In particular, physicians evaluated institutional commitment to safety as well as relations with supervisors and career perspectives more positively than did most of the other professional groups, rated collegiality more positively than did laboratory workers and rated teamwork climate more positively than did clinical workers. Non-medical workers gave more positive ratings than did all or most other professional groups for work satisfaction, working conditions, relations with supervisors, perceptions towards the hospital and career perspectives, and rated teamwork climate more positively than did clinical workers. On the other hand, nurses were relatively negative regarding working conditions and perceptions towards the hospital. Clinical workers gave more negative ratings to teamwork climate than did physicians and non-medical workers and were relatively negative regarding relations with supervisors. Laboratory workers gave more negative ratings to collegiality than did physicians and nurses and gave the most negative ratings of career perspectives. The fact that nurses and clinical workers perceived less institutional commitment to safety than did physicians may suggest that they are more likely to observe deficiencies in the organizational infrastructure related to patient safety than are physicians (Singer et al., 2009b). Nurses and clinical workers often spend more time with patients than do physicians (Corser, 2000; Singer et al., 2009b), and thus may receive complaints and hear opinions from the patients perspective which influence their own perceptions of safety procedures (Aiken et al., 2001; Singer et al., 2009b). Laboratory workers, on the other hand, may not feel directly involved in patient care practices and this may influence their ratings of institutional commitment to safety. Physicians and nurses are also likely to differ in their perceptions of the usefulness of safety rules and guidelines for patient safety and clinical practice. McDonald et al. (McDonald, Waring, Harrison, Walshe, & Boaden, 2005) suggested that compliance with safety rules and guidelines plays a greater role in nurse clinical practice than in physician practice. Physicians tend to ignore safety rules and guidelines and use the non-routine nature of events (i.e., that each patient needs different clinical treatment) as an argument against conforming Variations in Hospital Worker Perceptions of Safety Culture 43

45 to safety rules and guidelines. It may be this greater emphasis on safety rules and guidelines that is perceived as part of nurse professionalism and safe clinical practice that makes nurses more critical than physicians of institutional practices with regards to patient safety. The relatively negative nurse ratings of the dimensions working conditions and perceptions towards the hospital are unsurprising given that work dissatisfaction and high turnover are well-documented problems in the nursing profession (Aiken et al., 2001; Adams & Bond, 2000; Lu, While, & Barriball, 2005). Inflexible working schedules, overwhelming workloads and an unsupportive health-care environment have been shown to lead to burnout among nurses. This is especially problematic when the hospital management focuses on improving productivity (Aiken et al., 2001) rather than patient safety. Improving working conditions of nurses, for example, scheduling more reasonable working hours and providing better ergonomic and psychological supports, can improve nurses work satisfaction (Lu et al., 2005; Tzeng, 2002; Adams & Bond, 2000) and lead to better patient safety outcomes (Laschinger & Leiter, 2006). The relatively negative perceptions of teamwork and relations with supervisors of clinical workers may be due to the fact that these workers often play a supporting role in delivering patient care and may be expected to report to physicians (e.g., to complement the process of patient diagnosis) rather than taking part in the whole course of diagnosis and treatment. The relative independence of the clinical workers in performing tasks and dependent position with respect to physicians in the whole process of patient care may influence clinical worker perceptions of teamwork. Laboratory workers, whose primary tasks are examining and analyzing body fluids and cells (2009) and who tend to work independently rather than in cooperation with coworkers (Wolf, 1999) gave more negative ratings than did physicians and nurses for the collegiality of the atmosphere and gave the most negative ratings of career perspectives. Because laboratory workers are used to working in a more solitary working environment than are physicians and nurses, their relatively low ratings of collegiality may reflect a low involvement in collaborative activities. Laboratory workers are increasingly playing an integral role in patient care (e.g., advising physicians about what samples to take, developing clinical guidelines, and validating laboratory results). Improving the collaborative aspects of laboratory work is therefore a point of concern (Pearson, 1999). The fact that laboratory workers were more negative than were the other professional groups about perceived opportunities for career development also suggests that future research should concentrate on this potentially vulnerable group. The positive correlations found between the safety culture dimensions, although moderate to high, were lower than those found in a study using the SAQ (Sexton et al., 2006a). One 44 Improving Safety Culture in Health Care: Individual and Institutional Variability

46 explanation for the differences is that the SAQ was developed as a diagnostic tool aimed at those who work in clinical areas (i.e., patient care areas) as the group-level of interest, whereas the current study compared safety culture between all professional groups. Given differences between professional groups and departments (Pronovost et al., 2003b; Singer et al., 2009b), the finding of less shared variance between dimensions is not surprising. Nonetheless, the correlations between dimensions do suggest that organizational efforts to improve clinical and laboratory workers collaborative skills can also improve their perceptions of organizational commitment to safety and teamwork. Also, organizational efforts to raise nurses work satisfaction may also raise their appreciation of the hospital s objectives and policies, just as organizational efforts to give laboratory workers more opportunities for career development may improves their work satisfaction. Limitations of the current study include that use of the questionnaire has thus far been restricted to the Netherlands. However, as the items in the questionnaire cover all of the dimensions in the SAQ except for stress recognition, it should measure most aspects of safety culture. Moreover, although the number of participants in each group in the study fits the profile of the target population, we cannot preclude non-response bias. Conclusions and Recommendations Professional groups vary in how they perceive different dimensions of safety culture, and this finding suggests that group-specific interventions should be a part of any campaign to improve safety culture. The findings also suggest that some interventions should be expanded to include groups other than the group for which they were developed. For example, intervention programs such as executive walk rounds (Thomas et al., 2005), in which hospital executives make visits to patient care areas to discuss patient safety issues with frontline workers, should include not only physicians and nurses, but also clinical and laboratory workers. Involving clinical and laboratory workers in collaborative rounds (Kendall, 2003; Pearson, 1999), in which hospital workers from different disciplines conduct rounds together to discuss current and future plans of care and any patient care problems with patients, may also raise awareness about the importance of the team in patient care processes. Improving safety culture must involve those who set policy and those directly involved in patient care. Hospital leaders need to present and discuss the results of safety culture studies with all staff to raise awareness of safety culture and to break down barriers between managers, team leaders and all workers who play an integral role in patient care (Pronovost & Sexton, 2005; Singer et al., 2003; Pronovost et al., 2003b). Variations in Hospital Worker Perceptions of Safety Culture 45

47 Appendix 1 Bootstrapping results for each safety and organizational dimension Dimension Skewness S.E. of Skewness Kurtosis S.E. of Kurtosis Institutional commitment to safety Teamwork climate Team performance Work satisfaction Working conditions Collegiality Relations with supervisors Perceptions towards the hospital Career perspectives , , , , , , , , , Improving Safety Culture in Health Care: Individual and Institutional Variability

48

49

50 Demographic Differences between Health Care Workers who did or did not respond to a Safety Culture Survey A shorter version of the chapter has been published as: Listyowardojo T.A.,Nap R.E.,Johnson,A. (2011). Demographic differences between health care workers who did or did not respond to a safety and organizational culture survey. BMC Research Notes,4:328

51 Abstract Introduction Areas for institutional improvement to enhance patient safety are commonly identified by surveying health care workers (HCWs) attitudes, values, beliefs, perceptions and assumptions regarding institutional practices. An ideal response rate of 100% is rarely achieved in such surveys, and non-response bias can occur when non-respondents differ from respondents on a dimension likely to influence survey conclusions. The conditions for non-response bias to occur can be detected by comparing demographic characteristics of respondents and non-respondents and relating any differences to findings in the literature of differences in the construct of interest as a function of these demographic characteristics. The current study takes this approach. Findings All 5,609 HCWs at a university medical center were invited to participate in a survey measuring safety and organizational culture (response rate=53.40%). Respondents indicated their professional group, gender, age group, years of working in the hospital and executive function. Because all HCWs were invited, the demographic composition of the group who did not respond was known. Differences in the demographic composition of respondents and nonrespondents were compared using separate Pearson s chi-square tests for each demographic characteristic. Nurses and clinical workers were generally more likely to respond than were physicians, laboratory workers and non-medical workers. Male HCWs were less likely to respond than were females, HCWs aged younger than 45 years old had a lower response rate than did HCWs aged 45 to 54 years old, HCWs who had worked in the hospital for less than 5 years were less likely to respond than were those who had worked in the hospital for 5 years or more and HCWs without an executive function were less likely to respond than were executives. Conclusions Demographic characteristics can be linked to response rates and need to be considered in conducting surveys among HCWs. The possibility of non-response bias can be reduced by conducting analyses separately as a function of relevant demographic characteristics, sampling a higher percentage of groups that are known to be less likely to respond, or weighting responses with the reciprocal of the response rate for the respective demographic group. 50 Improving Safety Culture in Health Care: Individual and Institutional Variability

52 Introduction Patient safety in the hospital depends on health care workers (HCWs) commitment to safety (i.e., safety culture) (Singer et al., 2003; Singer et al., 2009a; Singer et al., 2009b; Nieva & Sorra, 2003; Pronovost & Sexton, 2005; Pronovost et al., 2003b; Huang et al., 2007; Sexton et al., 2006a; Antonsen, 2009). Characteristics of a strong safety culture include a commitment of all health care workers (HCWs) to give patient safety priority and a commitment of hospital management to promote and protect error reporting and to use error reports to improve patient safety (Singer et al., 2009a). Instilling or improving safety culture requires an assessment of HCWs attitudes, values, beliefs, perceptions and assumptions towards institutional practices with regards to patient safety. This is necessary in order to know who and what to transform (Nieva & Sorra, 2003; Singer et al., 2003; Pronovost et al., 2003b; Pronovost & Sexton, 2005; Huang et al., 2007; Antonsen, 2009; Singer et al., 2009b; Sexton et al., 2006a). Safety culture is typically assessed with surveys whose results are then used to design patient safety improvement programs, to evaluate the effectiveness of intervention programs and to track transformations of safety culture over time (Nieva & Sorra, 2003; Sexton et al., 2006a; Singer et al., 2009b; Singer et al., 2003). A challenge in surveying HCWs is to achieve a high response rate. Physicians, in particular, often show relatively low response rates (Asch et al., 1997; Asch et al., 2000; Cartwright, 1978; Cummings et al., 2001; Ward, 1994; VanGeest et al., 2007; Singer et al., 2003). For example, in a study by Singer et al. (2003), the response rate of physicians was 33%, considerably lower than the response rate of 60% that is recommended for achieving sufficient reliability in the measurement of safety culture (Pronovost & Sexton, 2005). Although some previous safety culture surveys have achieved overall response rates of 64 to 74% (Kho, Carbone, Lucas, & Cook, 2005; Huang et al., 2007), these surveys had relatively small samples ( responses). Safety culture studies in which thousands of HCWs were sampled have typically achieved response rates of only 47 to 52% (Singer et al., 2003; Singer et al., 2009b), with the exception of a Sexton et al. study that achieved a response rate of 67% (Sexton et al., 2006a). A major problem with low response rates is that they may result in under- or over-representation of particular groups and, thus, in non-response bias (Cull, O Connor, Sharp, & Tang, 2005; Nieva & Sorra, 2003; de Vaus, 1996; Guadagnoli & Cunningham, 1989). Non-response bias is said to occur when a significant number of those who do not respond differ in terms of relevant characteristics (i.e., characteristics that can influence survey outcomes and conclusions) from those who do respond. Non-response bias is a real concern in surveys of HCWs. For example, McFarlane and colleagues (McFarlane, Olmsted, Murphy, & Hill, 2007) Demographic Differences between Health Care Workers who did or did not Respond to a Safety Culture Survey 51

53 found gender bias in a mail survey that asked physicians to nominate the best American hospitals in their specialty. Female physicians were less likely to respond than were males, and, more importantly, female physicians were more likely than were males to nominate hospitals in the region where they have direct experience. Underrepresentation of females could thus lead to biased survey outcomes. Another example is a study that surveyed physicians views of a new primary care prescribing group created as a part of changes in the NHS (Armstrong & Ashworth, 2000). Here it was found that responding physicians tended to be younger than did non-respondents and that these younger physicians were more willing to reveal their personal prescribing practices than were older physicians. Furthermore, they were more optimistic, in general, with regard to changes in the NHS than were older physicians. In this case, age bias in the sample could lead to an inflated assessment of proposed changes to the NHS. Age bias was also found in a study by Templeton and colleagues who surveyed physicians opinions of alcohol misuse training (Templeton, Deehan, Taylor, Drummond, & Strang, 1997). Templeton and colleagues found that those who did not respond tended to be older than those who did, and, as revealed in post-survey interviews, tended to be more negative about the training than did younger physicians. One way to detect non-response bias is to contact non-respondents personally in order to investigate whether non-respondents opinions differ substantially from those of respondents (Armstrong & Ashworth, 2000; Templeton et al., 1997). This way of dealing with non-response cannot be applied when surveys are anonymous, as is likely to be the case when HCWs are surveyed (Huang et al., 2007; Asch et al., 1997; VanGeest et al., 2007). However, when the demographic characteristics of the surveyed population are known, even anonymous surveys can benefit from comparing the demographic composition of respondents with that of non-respondents (Cartwright, 1978; McFarlane et al., 2007; Armstrong & Ashworth, 2000; Templeton et al., 1997; Bjertnaes, Garratt, & Botten, 2008; Cull et al., 2005). If demographic characteristics of non-respondents differ from those of respondents, the generalizability of the survey results is called into question. Although some safety culture surveys conducted in the past have considered the effects of demographic variables on how safety culture is evaluated (Singer et al., 2009a; Singer et al., 2009b; Pronovost et al., 2003b; Huang et al., 2007; Johnson & Bakker, 2007), most studies of safety culture fail to consider the role of demographic characteristics in the evaluation of safety culture or make only vague reference to such differences (Singer et al., 2009a). An explanation for this might be that the goal of many studies of safety culture is to capture organizational (and not individual) factors that influence patient safety and that the importance of demographic differences in safety culture research has not been recognized (Cooper & Phillips, 2004). We argue that demographic characteristics should be considered in understanding whether HCWs 52 Improving Safety Culture in Health Care: Individual and Institutional Variability

54 will be likely to respond to safety culture surveys and that they should be taken into account in reducing the risk of non-response bias for anonymously conducted surveys. The purpose of this study was thus to compare the demographic composition of the groups of HCWs who did or did not respond to a survey measuring safety and organizational culture. Differences in demographic characteristics of respondents and non-respondents that should be taken into account to reduce the risk of non-response bias are documented and suggestions are made for understanding the responding behavior and increasing the response rate of HCWs. Methods Participants In April 2009, invitations to participate in a safety and organizational culture survey were sent electronically (via intranet) to 5,609 HCWs (out of a total of approximately 8,000 HCWs) involved in patient care (including managers) at the University Medical Center Groningen (UMCG), The Netherlands. The UMCG is a large university medical center that has approximately 1,300 beds, including 53 surgical and medical adult intensive care beds and 46 neonatal and pediatric intensive care beds. Reminders were sent in May and again in June. Participation was on a voluntary basis and no incentives to respond (financial or otherwise) were offered. Anonymity was ensured and informed consent was given by those who responded. Instrument Nine dimensions of safety and organizational culture were assessed in a survey containing 99 items. Because the survey addressed general organizational concerns such as work satisfaction, working conditions and perceptions towards the hospital, in addition to aspects of safety culture, it was relevant to all HCWs who were invited to participate. Additional questions addressed department, gender, age, years of working in the hospital and whether one worked in an executive or non-executive function. Because some HCWs could belong to more than one department (e.g., a medical specialist who also teaches at the university belongs to both medical specialist and teaching departments), HCWs were asked to indicate the department where they spent most of their time as stated on their work contract. The ten departments were combined into five professional groups because of similar job descriptions (e.g., the medical specialist and physician assistant departments were combined). The five professional groups thus obtained were physicians (e.g., medical specialists and physician assis- Demographic Differences between Health Care Workers who did or did not Respond to a Safety Culture Survey 53

55 tants), nurses (e.g., nurse practitioners and intensive care nurses), clinical workers (e.g., dieticians, psychologists and pharmacists), laboratory workers (e.g., laboratory technologists and technicians) and non-medical workers (e.g., facility management workers, secretarial and administrative workers and managers). The age groups used were years old, years old, years old, years old and older than 54 years old. The categories used for years of working in the hospital were less than 5 years, 5-9 years, years and longer than 20 years. Members of any professional group were considered to hold an executive function if they led a sector, department, unit, sub-unit or clinic; data were coded according to the number of respondents who worked in an executive or non-executive function. Data analyses Because of the need to preserve anonymity, data were available only for the demographic characteristics of the group and not for the conjunctions of the variables for each individual (i.e., we had available to us the data regarding how many, e.g., nurses responded, but not their age distribution, gender, etc., as this would uniquely identify individuals). It was therefore not possible to carry out multivariate analysis to identify which characteristics differed between groups of respondents and non-respondents while taking other characteristics into account. Instead, differences in the demographic characteristics of respondents and non-respondents were tested using Pearson s chi-square tests separately for each demographic characteristic (i.e., professional group, gender, age group, years of working in the hospital, and executive function), with group (respondents vs. non-respondents) and category (e.g., for the characteristic gender the categories were male and female) as factors. If interactions of group and category were found (i.e., if the demographic composition of respondents and non-respondents differed), follow-up pairwise Pearson s chi-square tests were conducted between sub-groups. Odds ratios (where an odds ratio of 1 indicates that the demographic composition of respondents and non-respondents did not differ) were used to calculate effect size. A significance level of p <.05, Bonferroni correction for each family of comparisons, was used where necessary. Results Out of 5,609 invitations sent, 2,995 were responded to (response rate = 53.40%). Response rates as a function of demographic characteristic are summarized in Table 1. Chi-square analyses (n = 5609) revealed interactions between group (respondents vs. 54 Improving Safety Culture in Health Care: Individual and Institutional Variability

56 non-respondents) and each of the demographic characteristics. The Group x Professional Group interaction (X 2 (4) = 53.54, p <.001) reflects that nurses and clinical workers had significantly higher response rates than did physicians and non-medical workers, and that nurses also had a significantly higher response rate than did laboratory workers (see Table 2 for all follow-up comparisons and odds ratios). The Group x Gender interaction (X 2 (1) = 6.33, p <.05) reflects that female HCWs had a higher response rate than did males. The Group x Age Group interaction (X 2 (4) = 30.07, p <.001) reflects that HCWs who were younger than 45 years old had significantly lower response rates than did those aged 45 to 54 years old. The Group x Years of Working in the Hospital interaction (X 2 (3) = 41.31, p <.001) reflects that HCWs who had worked in the hospital for less than five years had a significantly lower response rate than did those who had worked in the hospital for five years or more. The Group x Executive Function interaction (X 2 (1) = 24.77, p <.001) reflects that HCWs with an executive function had a significantly higher response rate than did those without an executive function. Table 1. Response rates as a function of demographic characteristic Category Respondents (n=2995) Non-respondents (n=2614) Response rate (%) Professional groups Physicians Nurses Clinical workers Laboratory workers Non-medical workers Gender Male Female Age range years old years old years old years old >54 years old Years of working in the hospital <5 years years years >20 years Executive function Executive Non-executive Demographic Differences between Health Care Workers who did or did not Respond to a Safety Culture Survey 55

57 Table 2. Chi-square results and odds ratios per sub-group Category X 2 Odds Ratio 95% CI Professional group: 1 Physicians vs. Nurses ** Physicians vs. Clinical workers ** Physicians vs. Laboratory workers Physicians vs. Non-medical workers Nurses vs. Clinical workers Nurses vs. Laboratory workers 10.15* Nurses vs. Non-medical workers 23.78** Clinical workers vs. Laboratory workers Clinical workers vs. Non-medical workers 13.12** Laboratory workers vs. Non-medical workers Gender: Male vs. female Age group: years old vs years old years old vs years old years old vs years old ** years old vs. >54 years old years old vs years old years old vs years old ** years old vs. >54 years old years old vs years old 10.99* years old vs. >54 years old years old vs. >54 years old Years of working in the hospital: 1 <5 years vs. 5-9 years ** <5 years vs years ** <5 years vs. >20 years * years vs years years vs. >20 years years vs. >20 years Executive function: Executive 1 vs. Non-executive ** Note. Bonferroni correction was used as appropriate for all analyses. **p <.001. *p < Had significantly higher response rate than the other sub-group. 56 Improving Safety Culture in Health Care: Individual and Institutional Variability

58 Discussion The demographic composition of groups who did or did not respond to a survey of safety and organizational culture was analyzed and significant differences between groups of respondents and non-respondents were found. Response rate was found to depend on professional group, gender, age, years of working in the hospital and executive function. The survey study on which the current study of demographic differences in response rates is based revealed not only differences in response rates, but differences in how aspects of safety and organizational culture are perceived [Unpublished data of T. A. Listyowardojo, R. E. Nap and A. Johnson]. The existence of differences in response rates as a function of demographic characteristics makes it important to consider whether non-response bias is likely to have influenced the interpretation of the survey results. In our study of safety and organizational culture, the data were analyzed and reported per group, and the major finding of the study was that perceptions of safety and organizational culture differed significantly across professional group, with physicians and non-medical workers tending to give more positive ratings of safety and organizational culture than did nurses, clinical workers and laboratory workers. The key question addressed in this paper is whether non-response bias due to unequal representation of professional groups influences study results. If group composition is not taken into account when analyzing survey results, under- or over-representation of some groups can influence the conclusions that are drawn. For example, it might be that physicians and non-medical workers, who were relatively positive in their ratings of safety and organizational culture, are less likely to respond to safety culture surveys than are nurses and clinical workers because they feel that changes in institutional practices are not urgently needed. Basing conclusions on the response group as a whole could then lead to an overly negative evaluation of safety culture. More specific aspects of safety culture, such as fear of shame and blame (Singer et al., 2009b), have been shown to be evaluated differently by nursing professionals than by physicians, and their evaluation may thus also be subject to non-response bias. The relatively low response rate of physicians may also, in part, be due to the perception that they are too often asked to respond to surveys and that their time is too valuable to be spent completing them (Sudman, 1985; Armstrong & Ashworth, 2000; Kaner, Haighton, & Mcavoy, 1998; Cartwright, 1978). Whereas physicians have been found to complain about being asked to participate in surveys too often, nurses and clinical workers have reported that they are not asked for their professional views often enough (Cartwright, 1978). Nurses Demographic Differences between Health Care Workers who did or did not Respond to a Safety Culture Survey 57

59 and the clinical workers may thus embrace the opportunity to voice their points of view by responding to surveys. The fact that gender affected response rate, with female HCWs being more likely to respond than their male colleagues may be tied to the fact that nurses are predominantly female, and nurses have a higher response rate. Because the effect of one demographic variable could not be statistically isolated from the others, we cannot say that females, in general, were more likely to respond. The finding that HCWs in the 45 to 54 year old age group were more likely than were younger HCWs to participate contrasts with previous findings of higher response rates for younger HCWs (Cartwright, 1978; Armstrong & Ashworth, 2000; Cull et al., 2005; Templeton et al., 1997). The fact that these previous studies sampled only physicians (in contrast to the current study, in which all professional groups were surveyed) may be responsible for this difference in findings. Alternatively, it may be that HCWs in this age group were more likely to be senior staff with administrative duties. The lower response rate of HCWs who had worked in the hospital for less than 5 years as compared to those who had worked for 5 years or more may be related to professional commitment to the hospital, which can be expected to increase as one works longer (Sudman, 1985). A final difference in response rates was that those who had an executive function were twice as likely to participate as were HCWs without an executive function. Those with executive functions may be more willing to participate because they are among those who will use the survey results to develop or defend patient safety intervention programs (Pronovost et al., 2003b). Limitation The main limitation of the current study is clearly the survey anonymity. The hospital would release only the demographic characteristics of the group as a whole and those of the group of respondents, and not those of the individuals. Furthermore, because we could not contact non-respondents, we were unable to investigate whether there was a difference in the evaluation of safety and organizational culture between respondents and non-respondents. However, previous studies have reported a strong link between relevant demographic characteristics (i.e., professional groups, age, years of work experience and executive function) and safety attitudes (Cooper & Phillips, 2004; Lin, Tang, Miao, Wang, & Wang, 2008b; Singer et al., 2009a; Sexton et al., 2006b; Pronovost et al., 2003b; Singer et al., 2003; Singer et al., 2009b), making it possible to draw tentative conclusions about the possibility of non-response bias. 58 Improving Safety Culture in Health Care: Individual and Institutional Variability

60 Conclusion Demographic characteristics can be linked to response rates and thus need to be taken into account in conducting surveys among HCWs. The possibility of non-response bias can be reduced by conducting analyses separately as a function of relevant demographic characteristics or by sampling a higher percentage of members of groups that are known to be less likely to respond (Singer et al., 2003; Singer et al., 2009a; Singer et al., 2009b). Another approach to reducing the possibility of non-response bias is to weight responses with the reciprocal of the response rate for the respective demographic group (Holt & Elliot, 1991; Singer et al., 2003). Demographic Differences between Health Care Workers who did or did not Respond to a Safety Culture Survey 59

61

62 Perceptions of Personal Health Risks by Medical and Non-Medical Workers in a University Medical Center: A survey study Published as: Listyowardojo,T.A.,Nap,R.E.,Johnson,A. (2010). Perceptions of personal health risks by medical and non-medical workers in a university medical center: A survey study. BMC Public Health,10: 681.

63 Abstract Introduction Health care workers (HCWs) are faced with many work-related choices which may depend on how they perceive risk, such as whether or not to comply with safety regulations. Little research has investigated risk perception in medical workers in comparison with non-medical workers and the extent to which risk perception differs in these groups. The current study thus investigates risk perception of medical and non-medical workers to inform and complement future research on safety compliance. The study has implications for the design of intervention programmes to increase the level of compliance of HCWs. Methods A survey study was conducted in which questionnaires were distributed to 6380 HCWs. The questionnaire asked for ratings of risk perception for cold, annual influenza, pandemic influenza, cancer, heart attack and food poisoning. Of 2495 returned questionnaires (response rate: 39%), 61.40% were from medical workers (24.1% of these were from physicians, 39.7% from nurses and 36.2% from paramedics) and 38.60% were from non-medical workers. Results Medical workers gave lower risk perception ratings than did non-medical workers for cancer, but not for other health risks. Within the medical workers, physicians rated the risk of getting a cold as higher, but of having a heart attack as lower than did nurses and paramedics; physicians also rated their risk of getting cancer as lower than did nurses. Perceived risk was higher as a function of age for pandemic influenza, cancer and heart attack, but lower for cold and annual influenza. HCWs who lived with a partner and children rated the risk of getting a cold or annual influenza higher than those who lived alone or with a partner only. Full-time HCWs gave lower ratings for annual influenza than did part-time HCWs. Conclusions Different base levels of risk perception between medical and non-medical workers need to be taken into account for successful implementation of safety regulations. Intervention programmes to improve compliance with safety regulations may need to be customized for different groups as a function of how they perceive risk. 62 Improving Safety Culture in Health Care: Individual and Institutional Variability

64 Introduction HCWs are faced with work-related choices such as whether to participate in voluntary immunization programmes or to comply with safety regulations. Compared to non-medical workers who have limited contact with infected patients, medical workers are exposed to various occupational health hazards which can result in serious long-term adverse health outcomes (Sepkowitz, 1996a; Sepkowitz, 1996b; Rogers, 1997). This makes compliance with safety regulations especially important for medical workers personal health. Previous studies have reported low rates of compliance of HCWs with hospital regulations and recommendations such as universal precautions (UPs) (Gershon et al., 1995; Becker et al., 1990) and vaccination programmes (Nichol & Hauge, 1997). The extent to which HCWs comply with safety regulations is likely to be related to their perceptions of the personal risks involved with the behaviours being regulated (Sjöberg, 2003; Sjoberg, 2003). For example, compliance with UPs is lower among those who rate their personal risk of infection lower (Gershon et al., 1995), and HCWs who perceive their risk of contracting an infection as higher are also more likely to participate in a pre-pandemic influenza vaccination programme than those who perceive their risk as lower (Chor et al., 2009). Previous work on compliance has focused primarily on HCWs perceptions of occupational risks such as influenza (Nichol & Hauge, 1997; Wicker, Rabenau, Doerr, & Allwinn, 2009; Esposito et al., 2008; Smedley, Palmer, Baird, & Barker, 2002) and those related to exposure to blood-borne viruses (Leliopoulou, Waterman, & Chakrabarty, 1999), and not on their perceptions of general health risks such as heart attack. Understanding how HCWs perceive health-related risks, in general, can help to understand factors involved in compliance with safety regulations. Research on risk perception has shown that people tend to rate their own personal risk related to general health conditions lower than they rate risks for others (Sjoberg, 2003). This is a concern because if people are unrealistically optimistic about their health, they will tend to feel less susceptible to diseases and be less likely to change their behaviour to reduce risks by (Weinstein, 1987), for example, complying with safety regulations. The current research compared the ratings of perceived risk of medical and non-medical workers to draw conclusions that may inform future research on compliance and risk perception. The study has significant implications for the design of intervention programmes to increase the level of compliance with safety regulations for different groups (i.e., medical and non-medical workers). Perceptions of Personal Health Risks by Medical and Non-Medical Workers in a University Medical Center: A survey study 63

65 Methods Participants Stratified sampling across professional groups was used to select medical (i.e., physicians, nurses and paramedics) and non-medical (e.g. financial services, board of directors, human resource management) workers of the University Medical Center Groningen (UMCG), the Netherlands, for invitation to participate in the study. The UMCG has approximately 1,300 beds, including 53 surgical and medical adult intensive care beds and 46 neonatal and pediatric intensive care beds. The UMCG is the only university medical center in the northern part of the country and as such is the major hospital of referral for patients with many types of illness. In October and November 2008, invitations to participate in an on-line survey were sent electronically to medical and non-medical workers in the group of interest. Because the participants of the study were not patients and the study was conducted anonymously and based on voluntary participation, approval of the medical ethical committee was not necessary. Questionnaire The questionnaire used was administered as a part of a larger study about compliance of HCWs with guidelines for controlling pandemic influenza (Nap, 2009). The demographic information asked for in the questionnaire included function in the hospital (medical or nonmedical worker, physician or nurse or paramedic for medical workers), gender, age, type of work contract (i.e., full-time or part-time) and family status (i.e., lives alone or with a partner and/or children). The risk perception questions were of the form What is the likelihood that you will have or get in the next one year? This question was completed with a cold, annual influenza, pandemic influenza, cancer, a heart attack and food poisoning. Responses were made using a Likert scale ranging from 1 (very unlikely) to 5 (very likely). Statistical Analyses Differences in demographic characteristics between medical and non-medical workers and between roles within the medical group (physicians, nurses and paramedics) were tested using Pearson s chi-square tests, except for age, which was tested using a t-test (for medical vs. non-medical groups) and one-way ANOVA (within medical group). ANCOVAs were conducted to determine whether demographic variables interacted with group and role to determine risk perception. Separate ANCOVAs were conducted for each of the health risks. A 64 Improving Safety Culture in Health Care: Individual and Institutional Variability

66 significance level of p <.05, Bonferroni corrected for multiple comparisons where necessary, was used for all analyses. Results A total of 2495 questionnaires were returned out of the 6380 questionnaires sent, for a response rate of 39%. Of the returned questionnaires, 61.40% were from medical workers (n = 1532) and 38.60% were from non-medical workers (n = 963). Within the medical workers (n = 1532), 369 (24.1 %) respondents were physicians, 608 (39.7%) were nurses and 555 (36.2%) were paramedical health care workers. The demographic characteristics are presented in Table 1. Table 1. Demographic characteristics of health care workers by group and role Category Medical group Total Physicians Nurses Paramedics Medical group (n=369) (n=608) (n=1532) (n=555) Total Non-medical group (n=963) Age (mean, with standard deviation in parentheses) Gender (%) Male Female Family status (%) Live alone Live with a partner only Live with children only Live with a partner and children Type of work contract (%) Full-time (40 hrs/week) Part-time (mean = 26.5 hrs/week; SD = 7.5 hrs/week) a ** a ** b ** (11.01) (10.47) (10.56) (10.72) (10.03) c ** c ** c ** c ** d** e ** e ** e * Note. a Significantly older than physicians. b Significantly older than medical group. c Significantly outnumbered male counterparts. d Significantly outnumbered part-timers. e Significantly outnumbered full-timers. **p <.001. *p <.01. Perceptions of Personal Health Risks by Medical and Non-Medical Workers in a University Medical Center: A survey study 65

67 The average age was (SD = 10.54) years old. Non-medical workers were older than were medical workers (t( ) = 6.23, p <.001; see Table 1). Within the medical group, oneway ANOVA conducted on age with role (i.e., physicians, nurses, paramedics) as a betweensubject factor revealed a main effect of role (F(2, 1526) = 14.19, p <.001; see Table 1). Chi-square analyses revealed interactions between gender and group (medical vs. non-medical workers; X 2 (1, n = 2495) = 10.66, p <.05; see Table 1) and between gender and role within the medical group (X 2 (2, n = 1532) = , p <.001; see Table 1). Chi-square analyses also revealed interactions between type of work contract and group (X 2 (1, n = 2495) = 5.61, p <.05; see Table 1) and between type of work contract and role (X 2 (2, n = 1532) = , p <.001; see Table 1). No significant differences were found for family status between medical and non-medical groups or between physicians, nurses and paramedics. Table 2. The mean risk perception ratings by group and role for each health risk (standard error of the mean in the parentheses) Risk perception Physicians (n=369) Nurses (n=608) Paramedics (n=555) Total Medical workers (n=1532) Total Non-medical workers (n=963) For cold For annual influenza For pandemic influenza For cancer For heart attack For food poisoning a *** 3.86 a * (.05) (.05) (.05) (.03) (.04) (.05) (.04) (.04) (.02) (.03) (.04) (.03) (.03) (.02) (.03) b ** c * (.04) (.03) (.04) (.02) (.03) b ** 1.68 b ** (.03) (.03) (.03) (.02) (.03) (.05) (.04) (.04) (.02) (.03) Note. a Significantly lower than physicians ratings. b Significantly higher than physicians ratings. c Significantly higher than medical workers ratings. ***p <.001. **p <.01. *p <.05. The mean risk perception ratings of the medical and non-medical workers are given in Table 2. To investigate whether demographic characteristics interacted with group to deter- 66 Improving Safety Culture in Health Care: Individual and Institutional Variability

68 mine risk perception, ANCOVAs were conducted with age as a covariate and group (medical vs. non-medical), gender (female vs. male), type of work contract (full-time vs. part-time) and family status (live alone, live with a partner only, live with children only, live with a partner and children) as between-subject variables. The ANCOVAs showed that age as the covariate was significantly related to risk perception for all health risks except for food poisoning. Parameter estimates showed that risk perception increased with age for pandemic influenza (β =.01, p <.001), cancer (β =.02, p <.001) and heart attack (β =.03, p <.001), but decreased with age for cold (β = -.02, p <.001) and annual influenza (β = -.01, p <.001). The analyses showed a main effect of group for risk perception for cancer (F (1, 2458) = 5.65, p <.05; see Table 2). No other main effects of group were significant. The main effect of type of work contract was significant only for annual influenza (F (1, 2458) = 4.89, p <.05). Full-time workers rated the risk for annual influenza lower than did part-time workers (mean = 2.40, SE =.05 vs. mean = 2.57, SE =.06, p <.05). The main effect of family status was significant for cold (F (3, 2458) = 9.47, p <.001) and annual influenza (F (3, 2458) = 4.34, p <.01). HCWs who lived alone rated the risk for cold lower than did HCWs who lived with a partner and children (mean = 3.69, SE =.07 vs. mean = 4.03, SE =.05, p <.001). HCWs who lived with a partner only rated the risk for annual influenza lower than did HCWs who lived with a partner and children (mean = 2.39, SE =.04 vs. mean = 2.58, SE =.04, p <.01). ANCOVAs were also conducted within the medical group with age as a covariate and role (physician, nurse or paramedic), gender (female vs. male), type of work contract (fulltime vs. part-time) and family status (live alone, live with a partner only, live with children only, live with a partner and children) as between-subject variables. The ANCOVAs showed that age as a covariate was significantly related to all health risks, except for food poisoning. Parameter estimates showed that risk perception increased with age for pandemic influenza (β =.01, p <.01), cancer (β =.01, p <.001) and heart attack (β =.02, p <.001), but decreased with age for cold (β = -.02, p <.001) and annual influenza (β = -.01, p <.001). The analyses showed a main effect of role on risk perception for cold (F (2, 1482) = 5.81, p <.01), cancer (F (2, 1482) = 4.37, p <.05) and heart attack (F (2, 1482) = 6.39, p <.01; see Table 2). The main effect of family status was significant for cold (F (3, 1482) = 6.19, p <.001) and annual influenza (F (3, 1482) = 2.77, p <.05). Posthoc tests using Bonferroni correction showed that medical workers who lived with a partner and children gave higher ratings to the risk of getting a cold (mean = 4.08, SE =.09 vs. mean = 3.70, SE =.07, p <.001) or an annual influenza (mean = 2.51, SE =.05 vs. mean = 2.29, SE =.06, p <.05) than did medical workers who lived with a partner only. Perceptions of Personal Health Risks by Medical and Non-Medical Workers in a University Medical Center: A survey study 67

69 Discussion Compliance with safety regulations can be explained, at least in part, by how HCWs perceive risks. If perceived risk is low, the incentive to comply may be lacking. In this study we looked for and found differences in how different groups of HCWs perceive risk. Most importantly, both type of function and demographic characteristics of HCWs were found to influence risk perception. Effects of function type Medical workers perceived their risk of getting cancer as lower than did non-medical workers. The lower risk perception for cancer of medical workers is largely due to the relatively low ratings given by physicians. A number of factors may contribute to the lower risk ratings of physicians for cancer. The fact that physicians may have to diagnose cancer patients and are directly involved in treating them may lead physicians to feel that they have more control over cancer than do nurses and paramedics. Both this perceived control and familiarity with risk may influence their risk perception (Sjoberg & Drottz-Sjoberg, 2008). It is also possible that physicians define risks differently than do nurses and paramedics. The current study found that physicians tend to perceive their risks related to more serious health risks (i.e., cancer and heart attack) as lower than do nurses and paramedics, but those related to a less serious health risk (i.e., cold) as higher. Physicians may define risk based on the probability of occurrence of the health hazard, whereas nurses and paramedics may be influenced by the severity of the disease in their perceptions of risk (Sjoberg, 1999). That is, physicians may have been more able than nurses or paramedics to do what was asked in this study, namely to rate the likelihood of suffering a health condition in the next year. Finally, the fact that physicians may need to communicate health risks more often to patients (or the general public) than do nurses and paramedics (Smith, 2003; Alaszewski & Horlick-Jones, 2003) may influence how physicians perceive health risk. For example, the expectation for physicians to sympathize with patients conditions in communicating risk (Paling, 2003) may lead physicians to slightly emphasize the benefits of medical treatments and minimize the severity of serious health risks (Smith, 2003). Effects of age and family status The current study also found that aging is correlated with higher risk perception for pandemic influenza, cancer and heart attack but lower risk perception for annual influenza 68 Improving Safety Culture in Health Care: Individual and Institutional Variability

70 and cold. HCWs are probably fully aware that aging is correlated with increased health risks such as cancer (DePinho, 2000), cardiovascular disease (Booth, Kapral, Fung, & Tu, 2006) and pandemic influenza (Glezen, 1996), making it unsurprising that these risks are rated higher by older HCWs. Younger HCWs may realize that they are not in the risk group of getting or having cancer, cardiovascular disease and pandemic influenza, thus leading to lower ratings in this age group. Age tends to be confounded with family status, with younger workers being more likely to live with children. Given that children who still live with their parents may be young and susceptible to cold and annual influenza, it stands to reason that HCWs who live with a partner and young children rate their chances of contracting a cold or annual influenza higher than those who live alone or with a partner only. Our findings of decreased risk perception with increasing age for annual influenza should be interpreted with caution considering the lack of vaccination status data in this study (the overall uptake rate for influenza vaccination at the UMCG in the years was 21% to 34%). If the older HCWs were vaccinated for annual influenza or were planning to be vaccinated, this could lead them to rate their risk as lower. Limitation The relatively low response rate of 39% is the main limitation of the study. Although the number of participants in each group in the study fits the profile of the target population, we cannot preclude non-response bias. Conclusions Different base levels of risk perception between medical and non-medical workers and among medical workers need to be taken into account for successful implementation of safety regulations. Intervention programmes to improve compliance with safety regulations may need to be customized for different groups as a function of how they perceive risk. Acknowledgements We thank the Board of Directors of the University Medical Center Groningen and Mrs. Laura de Jong, director of Personnel and Organization, in preparing the questionnaire. Perceptions of Personal Health Risks by Medical and Non-Medical Workers in a University Medical Center: A survey study 69

71

72 Using Social Network Analysis to Identify Sub-Groups in the Operating Room Published as: Listyowardojo,T.A.,Steglich,C.E.G.,Peuchen,S.,Johnson,A. (2009). Using social network analysis to identify sub-groups in the operating room. In D. DeWaard,H. Godthelp,F. Kooi,& K. Brookhuis (Eds), Human factors,security and safety (pp ). Maastricht,the Netherlands: Shaker Publishing.

73 Abstract The frequency with which operating room (OR) staff work together can impact patient safety because staff who often work together share a set of experiences which may enable them to anticipate each other s actions and reactions in the future. Identifying sub-groups of staff who frequently work together is thus a significant step in investigating team skills and the knowledge needed to prevent mishaps. Here, social network analysis (a set of statistical techniques for analyzing networks of interactions) is used to quantify the frequency with which individual OR staff members in a large university hospital worked together on three types of operation over a period of 2.6 years. Details of the specific techniques used are given. It is concluded that social network analysis is a viable method to identify sub-groups in the OR. 72 Improving Safety Culture in Health Care: Individual and Institutional Variability

74 Introduction Operating room (OR) staff who frequently work together may be more effective in preventing mishaps because these staff share a set of experiences which enables them to anticipate each other s actions and reactions, especially in emergency situations where communication may otherwise be ineffective. An example of such tacit communication is described in a study by Friedman and Bernell (2006): A perfusionist was being interviewed when she suddenly excused herself and walked into the OR where her team was performing a coronary artery bypass graft surgery. She immediately prepared the heart-lung machine before hooking the patient s heart up into the machine without being told to do so by anybody. When asked about this afterwards, she told the interviewer that when she saw the monitor which provided the patient s vital signs, she knew that the procedures to isolate a small part of the patient s heart without using the heart-lung machine (an off-pump procedure) had gone wrong and that the surgeon, whom she had worked together with for many years, would want to put the patient on the heart-lung machine (an on-pump procedure) to circulate and oxygenate the blood while the heart was stopped and being operated. As a result, without having any explicit communication between the perfusionist and the surgeon, a potential mishap was prevented. Ineffective communication has been identified as the cause of many medical errors (Awad et al., 2005; Lingard et al., 2004), including wrong-site operations (Makary et al., 2007; Brennan et al., 1991), and has resulted in near-misses (Brennan et al., 1991; Leape et al., 1991) and deaths (Calland et al., 2002a; Kawashima et al., 2003). The aforementioned example suggests that staff who frequently work together may develop unique work styles and habitual routines and become better able to use the information conveyed by body language and facial expressions as a result of repeated exposure to each other (Gersick & Hackman, 1990). The tacit communication which may thus be enabled may improve performance when explicit communication breaks down. An investigation of the ability to anticipate actions and reactions in the OR requires the identification of staff who frequently work together. Once these staff members are identified, an interview study can be conducted with, for example, the critical incident technique. In this method of probing interaction styles, participants are asked to recall a potential mishap which they successfully anticipated and prevented in the past. The details on how they anticipated the potential mishaps are then probed to uncover the roles of different types of communication. Using Social Network Analysis to Identify Sub-Groups in the Operating Room 73

75 Identifying OR staff who frequently work together is a problem in itself because of the complexity of OR environments. The density of the data (high number of operations and OR staff) and ad hoc scheduling practices in which any available OR staff are assigned to operations are some of the barriers to identifying sub-groups by hand. Specific analysis techniques are thus needed to identify these sub-groups. A general method with specific techniques that may be able to capture sub-groups in the OR environment is social network analysis. Social network analysis (SNA) can be used to analyze various types of relational data (Scott, 1991) and has been extensively utilized. For example, SNA has been used as an analytic tool to capture decision-making patterns in health care practices (Scott et al., 2005). In this case, participants were asked to identify whom they referred to when they had to make important decisions in the work setting. Networks of decision making patterns were captured that allowed the identification of significant differences between health care workers and which suggested interventions to promote practice organizational change. Another example of the use of SNA is in describing the structure of multidisciplinary long-term care teams by identifying the relationships that develop among staff as they do their work (Cott, 1997). Keating et al. (2007) showed that SNA can also be used to capture factors that affect influential discussions among physicians in a primary care practice, such as the accessibility of colleagues based on location and schedule. This information can be used to organize practices to promote more rapid dissemination of medical information. Social network analysis differs from other analysis techniques by focusing on patterns of relationships and by emphasizing empirical observation of these relationships. The current study aims to identify sub-groups in the OR whose staff members work together the most frequently. This data will be used in subsequent research to select groups for an interview study to investigate how staff who frequently work together prevent mishaps in the OR. Method Data The data were drawn from the OR staffing records of a large university hospital in the Netherlands. To reduce noise in the data, medical students in training were excluded from the data analysis, as were staff who were involved only at the beginning (e.g., less than 1.5 years) or at the end (e.g., started to be involved in the operations at the second year) of the 2.6 year period, or who were involved in only a few operations. All appendectomies, esophageal resections, and liver resections that took place between 74 Improving Safety Culture in Health Care: Individual and Institutional Variability

76 June 2, 2005 and December 31, 2007 were included, for a total of 1,423 operations with 63 staff. Of the 63 staff, 33 were female and 25 were male; 5 were not identified. The average age was 44 years old (SD=9.71; range from 25 to 62 years old). In the period of approximately 2.6 years, each staff member performed a minimum of 32 operations and a maximum of 238 operations (M=78 operations; SD=40). Three types of operation were compared because the fact that they differ in difficulty level may affect staff interactions. Liver resections were the most difficult of these operations and appendectomies the simplest. Liver resections were the most common operations in the dataset (N=627; 44.10%) followed by esophageal resections (N=413; 29%) and appendectomies (N=383; 26.90%). Almost all of the staff took part in all types of operation: appendectomies (N=60), esophageal resections (N=62), and liver resections (N=63). The roles in the OR were assistant surgeon/surgeon, scrub nurse, nurse anesthetist, and anesthetist (see Table 1). Six staff performed both as anesthetist and assistant surgeon/surgeon and one as both scrub nurse and nurse anesthetist (see Table 1). Table 1. Number of Staff per Role Staff Roles Value Anesthetist 4 Assistant surgeon/surgeon 10 Anesthetist and Assistant surgeon/surgeon 6 Scrub nurse 30 Nurse anesthetist 12 Scrub nurse and nurse anesthetist 1 Total 63 Using Social Network Analysis to Identify Sub-Groups in the Operating Room 75

77 Procedure The data were first put in the form of an adjacency matrix (see Figure 1) for analysis in UCINET VI (2002), a computer programme developed for performing SNA. For confidentiality purposes, staff identities were converted into individual IDs. The following codes were used: A for anesthetist, S for surgeon or assistant surgeon, SN for scrub nurse, NA for nurse anesthetist, A/S for anesthetist/assistant surgeon/surgeon, and NA/SN for nurse anesthetist/scrub nurse. Each OR staff member was assigned to a column and a row in the matrix. The number of operations performed together for each pair of staff was entered in each matrix cell. For example, in Figure 1, surgeon 1 (S1) worked together for 5 operations with scrub nurse 1 (SN1), therefore, 5 is entered in the matrix cell indexed by S1 and SN1. The matrix cells below the diagonal were identical to those above the diagonal. Figure 1. An example of an adjacency matrix. A clique at level C analysis was then used to identify sub-groups whose members worked together frequently. A clique is defined as the maximum number of actors who share all the possible ties among themselves (Wassermann & Faust, 1994). That is, it is a group within which all members interact with each other. A clique at level C can be defined as a sub-graph in which the relationships between all pairs of staff have values of C or greater. In other words, the frequency with which staff within a clique work together is given by the parameter C. For example, if C is a set at 10, only staff who work together at least 10 times will be included in a clique. Although Wassermann and Faust define a clique as having a minimum of three members, in this study, cliques of only two members were allowed because the minimum number of staff to perform an operation is two. One person can belong to many cliques. However, because clique analysis in its most 76 Improving Safety Culture in Health Care: Individual and Institutional Variability

78 general form (C = 1) simply connects all people who have worked together at least once, staff who have performed ten operations together cannot be distinguished from those staff who have performed only one operation. To identify sub-groups whose members have worked together at a certain level of frequency, C, a value for C must first be determined. In order to define a threshold level for C, hierarchical clustering is performed to determine the frequency with which pairs of staff work together (Reffay & Chanier, 2003). Hierarchical clustering is based on the strength of the relationship (i.e., similarity) between actors (here, the number of operations performed together). Clusters are produced by joining the most similar pairs of staff based on their proximity. For example, if actor A and actor B work together 10 times, and not more than ten times with anyone else, A and B will be joined in a single cluster (A/B at level 10). If both A and B work together 8 times with actor C, then cluster A/B is joined with C to create cluster A/B/C at level 8. However, if A only works together with D for 7 times and 7 times is the highest level of interaction for D, then A and D are joined into cluster A/D at level 7 (see Borgatti, (1994). Results Figure 2, 3, and 4 show the results of the hierarchical clustering analysis. Each group of three or more Xs represented the clusters involving the staff of the corresponding columns. For example, for appendectomies (see Figure 2), staff member A/S5 and SN28 performed 23 operations together. They performed six operations with SN20, and three with SN20 and S9. The highest level of interaction was reached for liver resections, with 37 operations performed together by S3 and SN28 (see Figure 4). The frequency of working together by any one pair was the lowest for appendectomies, with 23 operations performed together by A/S5 and SN28 (see Figure 2). On the basis of frequency of working together, a threshold level of C = 10 was selected for the clique at level C analysis. This resulted in the selection of nine staff for appendectomies, fourteen for esophageal resections, and eighteen for liver resections. Using Social Network Analysis to Identify Sub-Groups in the Operating Room 77

79 Figure 2. Hierarchical clustering of appendectomies. Figure 3. Hierarchical clustering of esophageal resections. Figure 4. Hierarchical clustering of liver resections. 78 Improving Safety Culture in Health Care: Individual and Institutional Variability

80 Clique at level C analyses. NetDraw (2002) was used to visualize the figures: five subgroups were found for appendectomies (see Figure 5), eight for esophageal resections (see Figure 6) and sixteen for liver resections (see Figure 7 ). The members of each sub-group per type of operation are given in Table 2. Table 2. Sub-groups found for each type of operation Type of Operation Appendectomies Esophageal resections Liver resections Sub-groups Found S1, S5, SN22; S1, S5, SN28; A/S5, SN28; S6, SN10; S7, SN16 S3, S5, SN20, SN28; A/S6, S3, S5, SN28; S1, S5, SN11; S5, SN11, SN29; S10, S4, SN10; S10, S4, SN25; A/S6, SN12; SN26, SN29 A/S5, S1, S5, SN28; S1, S5, SN20, SN28; S5, SN11, SN28; A/S6, S5, SN28; A/S5, S5, SN24; S5, SN11, SN24; A/S5, S3, SN28; A/S6, S3, SN28; A/S6, S8, SN28; S5, SN29; S10, SN10; S10, SN25; S4, SN10; S4, SN25; S7, SN16; SN26, SN29 Figure 5. The five sub-groups found for appendectomies. Using Social Network Analysis to Identify Sub-Groups in the Operating Room 79

81 Figure 6. The eight sub-groups found for esophageal resections. Figure 7. The sixteen sub-groups found for liver resections. 80 Improving Safety Culture in Health Care: Individual and Institutional Variability

82 The clique at level C analysis also produces a so-called hierarchical clustering of overlap matrix. The hierarchical clustering of overlap matrix shows how many sub-groups were shared between two or more staff (see Figures 8, 9 and 10). In the hierarchical clustering of overlap matrix for appendectomies, S5 and S1 were seen to share two sub-groups (see Figure 8). Because no other sub-group shared two or more subgroups, this sub-group (S5 and S1) was considered as the most central one in appendectomies. For esophageal resections, the hierarchical clustering of overlap matrix revealed two subgroups as the most central ones, both sharing members of two other subgroups: {S4 and S10} and {S5, S3, and SN28} (see Figure 9). Staff S5 and SN28 shared four sub-groups and were thus considered as the most central sub-group in liver resections (see Figure 10) Figure 8. The hierarchical clustering of overlap matrix for appendectomies. Using Social Network Analysis to Identify Sub-Groups in the Operating Room 81

83 Figure 9. The hierarchical clustering of overlap matrix for esophageal resections. Figure 10. The hierarchical clustering of overlap matrix for liver resections. 82 Improving Safety Culture in Health Care: Individual and Institutional Variability

84 Discussion The aim of this preliminary study was to identify sub-groups whose members frequently work together. Techniques from social network analysis hierarchical clustering and clique at level C analysis successfully identified these sub-groups. How each technique complements the other and the implications for further study are discussed here. In hierarchical clustering, two items which are joined in a cluster at an earlier stage cannot belong to different clusters at a later stage: Only the highest frequency of working together between a pair of staff is included. For example, if A and B work together for 10 times, but B also work together with C for 12 times, and A also works together with D for 13 times, hierarchical clustering will only identify the relationships between B and C and between A and D, and the relationship between A and B will not be documented. The relationship between A and B is made visible using clique at level C analysis (C=10). As a result, these two techniques are complementary. The results showed that most of the sub-groups consist of staff members with diverse roles. This can be considered as particularly beneficial for the planned interview study which will focus on how staff who frequently work together prevent mishaps in the OR. The communication between disciplines (i.e., surgeons, nurses, anesthetists) in the OR has been suggested to be often ineffective which may lead to mishaps (Lingard, Reznick, Devito, & Espin, 2002; Awad et al., 2005). We hope to learn how staff from different disciplines prevent mishaps, for example, by investigating how they improvise and communicate during the crisis, and how these behaviours depend on frequency of working together. In conclusion, clique at level C analysis is a viable technique to highlight the most central sub-groups whose members frequently work together in the operating room. The techniques should lend themselves to more relational studies to identify sub-groups in complicated environments. Using Social Network Analysis to Identify Sub-Groups in the Operating Room 83

85

86 The Importance of Frequency of Working Together in the Operating Room: An interview study Listyowardojo,T.A.,Hoekstra,T.,Nap,R.E.,Johnson,A. The importance of frequency of working together in the operating room: An interview study. (Manuscri pt in preparation)

87 Abstract Introduction. Operating room (OR) teams whose members are familiar to each other are likely to work more efficiently and perform better than are teams whose members are less familiar to each other. Familiarity with coworkers working habits as a result of working together frequently may result in the development of characteristics that are important for team coordination such as efficiency in conducting procedures and effective communication. These characteristics can be especially important for OR teams under time pressure. Methods. Social network analysis (a set of statistical techniques for analyzing social network interactions) was used to identify OR team members who worked together the most frequently with other coworkers. Four surgeons and four anesthesiologists identified as frequently working together were asked to recall an operation in which an unexpected incident happened and which they had to prevent from becoming an accident. Participants were asked to base their recall on operations for which they worked together with the OR team members most familiar to them. The critical incident technique was used to probe participants about the key factors important in preventing incidents from becoming accidents. Results. Participants perceived that working with familiar team members was beneficial for teams under time pressure. The participant responses of factors important for preventing incidents from becoming accidents were categorized into two groups: Factors related to working with familiar team members and general factors such as staying calm. Factors related to working with familiar team members included mutual understanding of a situation, efficient team coordination, effective communication and trust. General factors included medical experience, knowledge and skills, communication and the presence of interdisciplinary team members and personal attributes. Conclusions. The frequency of working together with particular OR team members can be considered an important factor for teams under time pressure. Complicated, rare or difficult operations, which involve high-risk procedures, may need to be conducted only by OR teams whose members are familiar to each other. 86 Improving Safety Culture in Health Care: Individual and Institutional Variability

88 Introduction Familiarity with coworkers working habits as a result of working together frequently may be a significant factor for operating room (OR) teams working under time pressure. Teams whose members are familiar to each other are likely to work faster and create higher quality products than are teams whose members are less familiar to each other (Harrison et al., 2003). Working together frequently with particular team members may result in the development of characteristics that are important for team coordination such as mutual understanding (Friedman & Bernell, 2006), trust (Schoorman, Mayer, & Davis, 2007; Wilson, Straus, & McEvily, 2006; Dirks, 1999) and feelings of belonging to the team (Gersick & Hackman, 1990). These characteristics can be especially important for OR teams under time pressure. Friedman and Bernell (2006) suggested that the more frequently OR team members perform operations together, the more familiar they become with each other s working styles (e.g., preferred glove size for surgeons, preference of lighting in the OR), body language and facial expressions, thus enabling team members to anticipate each other s reactions. The ability of OR team members to anticipate each other s reactions is important because the OR is a high-risk area in which disproportionately more harm may result from errors than elsewhere in the hospital (Baker et al., 2004; Entin et al., 2006; Soop et al., 2009; Thomas et al., 2000; Wilson et al., 1995; Zegers et al., 2009). Espinosa et al. (2004), who investigated team coordination in student and employee teams, have suggested that team members who have worked together for many years may coordinate their actions based on implicit coordination mechanisms. Implicit coordination mechanisms are defined by Espinosa et al. (2004) as those mechanisms that are available to team members from shared cognition, which enable them to explain and anticipate task states and member actions, thus helping them manage task dependencies (p.10). These mechanisms may become important for maintaining effective team coordination under time pressure when explicit (verbal and gestural) communication is not effectively used. Xiao et al. (1998) found that explicit communication was rarely used during trauma patient resuscitation because explicit communication adds team workload. Team members often used implicit communication to coordinate their actions including monitoring teammates or a leader s activities, understanding the significance of patient physiological events and scanning the workspace to perceive tasks needed to be carried out. Xiao et al. found that team coordination was compromised when there was no explicit communication and implicit communication was not interpreted accurately. The study of Xiao et al., however, did not consider The Importance of Frequency of Working Together in the Operating Room: An interview study 87

89 the degree of familiarity between team members. The data in the study of Xiao et al. were collected from 100 cases of real-life patient resuscitation over a period of three years. The data thus covered a wide variety of cases and team compositions. When explicit communication can not feasibly be used effectively due to time pressure, teams whose members are familiar with each other may predict more accurately what other team members are likely to do next and what to expect from each other than those whose members are less familiar with each other. Friedman and Bernell (2006) suggested that it is the shared experiences and history of working together for many years that may enable OR team members to prevent incidents (e.g., near misses) from becoming accidents. They reported an example of how a potential mishap was prevented by a team whose members had worked together for years: A perfusionist excused herself from an interview and walked into the OR where her team was performing a coronary artery bypass graft surgery. She immediately prepared the heart-lung machine before hooking the patient s heart up to the machine without being told to do so. When asked about this afterwards, she told the interviewer that when she saw the monitor which provided the patient s vital signs, she knew that the procedure to isolate a small part of the patient s heart without using the heart-lung machine (an off-pump procedure) had gone wrong. She immediately understood that the surgeon, whom she had worked together with for many years, would want to put the patient on the heart-lung machine (an on-pump procedure). The on-pump procedure was needed to circulate and oxygenate the blood while the heart was stopped and being operated. This example illustrates the benefit of working with familiar team members for patient safety: When explicit communication was not feasible, team coordination between familiar team members was not impaired. Working with familiar team members may become a safety net during emergency situations. A high degree of familiarity between team members, however, may have unwanted effects. First, when a modification is needed while performing an operation, team members who too often work together may fail to recognize this need (Gruenfeld, Mannix, Williams, & Neale, 1996). Second, a high degree of familiarity between team members can impede the likelihood that team members will acquire new knowledge or skills and ultimately can contribute to stagnation of team performance (Gersick & Hackman, 1990). A high degree of familiarity thus may lead to complacency and a sense that even difficult operations are routine, potentially endangering patient safety. A high degree of familiarity between team members may be a rarity in the OR because of staff rotation (Friedman & Bernell, 2006; Morey et al., 2003; Mackenzie, Jeffcott, & Xiao, 88 Improving Safety Culture in Health Care: Individual and Institutional Variability

90 2009). The fact that OR team members tend to be part of interchangeable teams makes it difficult to distinguish those who more frequently work together from those who less frequently work together (Friedman & Bernell, 2006). A high number of operations and a high degree of staff rotation add to the complexity of identifying OR team members who frequently work together. The complexity of the OR environment may be the reason why there is still little research on the influence of the frequency of working together on OR team coordination under time pressure. In order to deal with the complexity of the OR, we used social network analysis (SNA; a set of statistical techniques for analyzing networks of interactions) to quantify the frequency with which individual OR team members in a university medical center worked together (Listyowardojo, Steglich, Peuchen, & Johnson, 2009b). Using SNA we highlighted the central sub-groups whose members most frequently worked together in the OR. Several of these OR team members were interviewed in the current study to explore possible key factors that contributed to the team successes in preventing incidents from becoming accidents. The present study is focused on the benefits of working with familiar team members under time pressure. Method Setting The study was conducted in the University Medical Center Groningen (UMCG) in the Netherlands, a hospital with approximately 1,300 beds including 53 surgical and medical adult intensive care beds, and 46 neonatal and pediatric intensive care beds. The UMCG is the only university medical center in the northern part of the country. Therefore, it is the major hospital of referral for patients with many types of illness and an important center for all organ transplants. Participants The data on the frequency of working together between OR team members were drawn from the OR staffing records of the UMCG. To reduce noise in the data, medical students in training were excluded from the data analysis, as were staff who were involved only at the beginning (e.g., less than 1.5 years) or at the end (e.g., started to be involved in the operations at the second year) of the 2.6 year period (the data were drawn 2.6 years after the implementation of a new schedules recording system), or who were involved in only a few operations. The Importance of Frequency of Working Together in the Operating Room: An interview study 89

91 Social network analysis was used to quantify the frequency with which individual OR team members worked together in different operations over a period of 2.6 years. Social network analysis identified sub-groups whose members worked together the most frequently with the same coworkers. A detailed description of the frequency of working together analysis is given in Listyowardojo et al. (2009a). The 20 identified OR team members who most frequently worked together with particular team members were invited to participate in this interview study. Of these 20 OR team members, eight (Response rate = 40%) team members agreed to be interviewed. Eight OR team members (4 surgeons, 4 anesthesiologists) were interviewed individually in the period from May to July The four surgeons were men; three of the anesthesiologists were men and one was a woman. Participation was on a voluntary basis. The confidentiality of the participants identities was ensured by replacing participants names with coded names such as Surgeon 1 and Anesthesiologist 2 throughout the analysis processes. Participants gave informed consent. The Medical Ethical Testing Committee of the University of Groningen waives Institutional Research Board approval for interviews conducted under personnel. Procedure The data were collected during a semi-structured interview based on the Critical Incident Technique (CIT; developed by Flanagan; 1954). Flanagan defines the CIT as a technique that consists of a set of procedures for collecting direct observations of human behavior in such a way as to facilitate their potential usefulness in solving practical problems and developing broad psychological principles (p. 327). The CIT is a retrospective method that requires participants to recall one or more critical incidents that they experienced personally in the specific context being studied. A critical incident is an activity that has an important impact, positive or negative, on the outcome of an event. We asked participants to recall an emergency situation in which they successfully prevented an incident from becoming an accident in the OR. Participants were asked to base their recall on emergency situations for which they worked together with the OR team members with whom they most often worked together. The interviewer asked participants to name key factors that influenced the successful outcomes (i.e., incidents that prevented from becoming accidents). Participants were asked how they performed as a team during the emergency situations and probed using the CIT (Graneheim & Lundman, 2004). For example, participants were probed by asking, How did you and your team members know what to do at that 90 Improving Safety Culture in Health Care: Individual and Institutional Variability

92 specific moment?, Can you tell more about variable A?, What causes variable A? and How did that happen? Participants were also asked to compare the situation during the incident to that of situations when planned procedures went well by asking What was changed during the incident compared to that of operations when planned procedures went well in general? In responding to this question, participants were told that they need not necessarily base their recall on operations for which they worked with their familiar team members. Except for one anesthesiologist, participants reported that they had a fixed team whose members frequently worked together. These seven participants had worked together within their fixed teams for six to fifteen years (Mean = 10.4 years). These participants based their recall on emergency situations for which they worked within their fixed teams, except for the anesthesiologist who did not have a fixed team; he based his recall on an operation for which he worked together with the most familiar team members. Participants were interviewed individually, in interviews ranging from 30 to 60 minutes each. The interviews were held in participants private offices and digitally tape-recorded. The interviews were transcribed and the recordings of interviews were destroyed six months after the dates of the interviews to ensure confidentiality. Data analysis A total of 313 minutes of interviews were transcribed verbatim and coded. The codes were developed from the literature (deductive codes) about team familiarity and team coordination (Friedman & Bernell, 2006; Espinosa, Lerch, & Kraut, 2004; Espinosa, Slaughter, Kraut, & Herbsleb, 2007; Hinds, Carley, Krackhardt, & Wholey, 2000). The codes were also developed directly from the interviews (inductive codes) by noting issues raised by participants (Hennink, Hutter, & Bailey, 2011). The coded interviews were analyzed using qualitative techniques and showed team members perceptions of the experiences of working together with familiar team members in an emergency situation. A section of texts could be coded with more than one way (see Appendix 1). Quotes from the interview data in this paper were translated from Dutch to English. Results Participants perceived that working with familiar team members was beneficial in preventing incidents from becoming accidents in the OR. This is because working with familiar team The Importance of Frequency of Working Together in the Operating Room: An interview study 91

93 members saved time in conducting procedures. Time was repeatedly stressed to be very limited during emergency situations. The perceived key factors for successful outcomes were categorized into two groups: Factors related to working with familiar team members and general factors (see Table 1). General factors were not necessarily related to working with familiar team members. Table 1. The perceived key factors for the success in preventing incidents from becoming accidents in the OR Factors related to working with familiar team members Mutual understanding of a situation Team members knew what to do next Team members knew what they were doing Team members knew what the others were doing Team members could anticipate different situations Knowing team members emotional reactions to stress Knowing how team members usually tackled problems Knowing how to work with each other Efficient team coordination Working with familiar team members saved time Conducting procedures were relatively smooth Participants did not need to double check every procedure Participants did not need to control every task of others Effective communication Team members felt free to express opinions or concerns Quick communication and proactive reactions from every team member There was no need to be very explicit in communication Things were prepared without being asked Gestures were enough to express what was needed Team members knew each other s languages Trust Participants trusted familiar team members Participants trusted that team members would do their tasks properly Participants felt that they could rely on team members Participants trusted that team members would get the right instruments quickly General factors Medical experiences, knowledge and skills Team members had experiences in a specific medical area Team members had experienced similar problems Team members had knowledge and skills on a specific medical area Team members knew what instruments were needed Team members were specialized in the same medical area The way team members communicated Directive and short communication Drew attention from every team member Team members listened to each other Team members updated each other with the development of the situation Personal attributes Staying calm Being creative Respect each other 92 Improving Safety Culture in Health Care: Individual and Institutional Variability

94 Factors related to working with familiar team members Participants reported that working with familiar team members was one of the key factors contributing to the successful outcomes. Working together was perceived to facilitate members in observing working behaviors of the other team members. Participants reported that by working together frequently, team members encountered and solved problems together. These shared experiences were perceived to connect team members and lead to the understanding of each other s personalities and the development of trust. Participants responses were categorized into four benefits of working with familiar team members: Mutual understanding of a situation, efficient team coordination, effective communication and trust. Mutual understanding of a situation. Participants reported that familiar team members shared understanding of the next steps of a procedure and their individual tasks. Familiar team members also knew how each team member usually tackled problems and each other s emotional reactions to stress. This mutual understanding was said to reduce the team s workload, enabling team members to focus on finding and conducting solutions to prevent the incidents from becoming accidents. A surgeon described the mutual understanding of what the next steps would be between him and his team members during the incident. In the words of the surgeon: [My team members] knew that he is going to do this, and if that does not work, the ultimate step is to put in the knife. And if at that moment [the first procedure] did not work, without me asking for it, the next step could be done very quickly. You did not need to ask first. They just knew that it was necessary. Efficient team coordination. Participants emphasized the importance of efficiency of team coordination during emergency situations. Working with familiar team members was perceived as an important factor enabling team members to conduct procedures smoothly and efficiently. Participants also reported that working with familiar team members saved time in conducting procedures because participants did not need to double-check every procedure and control every task of other team members. Participants often used their experiences of working with less familiar team members in general to emphasize the efficiency of working with familiar team members. Participants said that they tended to control and check the work of less familiar team members more often than the work of familiar team members. This tendency was perceived as costing time and adding to team workload in conducting The Importance of Frequency of Working Together in the Operating Room: An interview study 93

95 procedures. Effective communication. Participants suggested that verbal communication often became ineffective during emergency situations. Verbal communication between team members was said to become very short and directive. Limited time to communicate verbally was not seen as a problem because verbal communication was perceived to add to team workload during emergency situations. Participants, however, perceived that every team member would not hesitate to express concerns or problems when necessary. Participants also suggested that familiar team members understood each other using only a half word and gestures. Gestures were said to support verbal communication under time pressure as well as the mutual understanding of a situation. Participants also suggested that instruments needed during emergency situations were often prepared or immediately taken from outside the OR without being asked. Effective communication between familiar team members was perceived to be achieved because team members understood each other s languages. Working with less familiar team members, on the other hand, was seen to create a problem when verbal communication became ineffective. An anesthesiologist compared his experiences of working with familiar team members to that of working with less familiar team members when verbal communication became ineffective. In the words of the anesthesiologist: Here we have nurse anesthetists who are highly qualified and always work together with specific anesthesiologists. For example, if I need to give an adrenaline shot, I will say adrenaline 55 and I will get what I want. If I work in the emergency department here in this hospital, and there are also nurse anesthetists, then I will get an adrenaline ampoule pressed on my hand. I say what is this, I am asking for adrenaline. I cannot inject the ampoule because the adrenaline needs to be sucked out up [before being injected to the patient]. This is the reason why, if we are about to do a [cardiopulmonary] resuscitation, I want my own nurses and nurse anesthetists [because] you know each other s languages. Trust. Participants reported that they trusted familiar team members much more than less familiar team members. Familiar team members were trusted that they would do their tasks properly such as getting the right instruments quickly during emergency situations. Participants also trusted that their team members would understand what should be done although verbal communication became ineffective. Trusting familiar team members made 94 Improving Safety Culture in Health Care: Individual and Institutional Variability

96 participants feel that they could rely on each other. A surgeon explicitly said that trust was one of the most important factors for preventing the incident from becoming an accident. The surgeon summarized this in his words: The interviewer: What factors do you think were the most important for the successful outcome? The surgeon: Trust, that if you ask for something, the person immediately grabs what is needed. In emergency situations, something always needs to be picked up from outside the OR. You then can trust your team members that they will pick up the instrument quickly, and that they will not walk back and forth three times with a wrong instrument. The way I see it, conducting an operation is certainly a stressful situation, when you must feel safe in your environment and with the people around you. You should be able to trust that even if you say things in a half word, things still happen and are implemented in order to achieve a good end. General factors Participants also reported general factors that were also perceived to contribute to the successful outcomes. These general factors were not necessarily related to working with familiar team members. Participant responses were categorized into medical experiences, knowledge and skills of team members, communication and personal attributes. Medical experiences, knowledge and skills in a specific area. Participants perceived that experiences, knowledge and skills in a specific medical area such as lever surgery were ones of the key factors for the successful outcomes. Experiences in a specific medical area were perceived to be important because every medical area had its own specific protocols in conducting procedures. Firsthand experiences in a specific area were perceived as important because these experiences were needed to build medical knowledge and skills that could not be learned otherwise. Through working in a specific area, team members might also experience common problems in the area. Having experienced problems made team members know what to do and what was needed in emergency situations. The way team members communicated. Participants reported that communication during emergency situations should be made directive and short, focusing on finding and implementing solutions. Directive and short communication was also perceived as a way to The Importance of Frequency of Working Together in the Operating Room: An interview study 95

97 draw team members attention to focus on a problem. The key factors to effective communication under time pressure were to listen to each other and update the development of a situation within and between disciplines. It was perceived to be vital to involve team members from different disciplines in finding solutions for a problem. A surgeon described the importance of interdisciplinary communication for patient safety. In the words of the surgeon: Communication is very important. You have to update your anesthesiologist, your operating assistant, and at that moment you have to say what you are going to do, what kind of solution that you choose, because there may be consequences that you do not know. Personal attributes. Participants reported that staying calm was important in dealing with emergency situations as well as being creative in finding solutions for a problem. A surgeon summarized this in his words: If you stay calm, you will be prepared for any procedures, that you can be creative in dealing with things, and that you do not keep thinking in the same standard framework, but that you can see a broader view, you see more solutions. Participants also reported that respect was important to be able to work together. Respect was perceived as trusting team members of their capabilities in carrying out their tasks such as leading the team. Discussion The key finding of this study is that working with familiar team members is crucial during emergency situations. Participants reported that familiar team members share mutual understanding of a situation, are efficient in conducting procedures, communicate effectively and share mutual trust. These benefits of working with familiar team members are perceived as being very important for teams in preventing incidents from becoming accidents in the OR. Participants reported that teams whose members are familiar to each other share mutual understanding about what to do during emergency situations. Team members may learn to 96 Improving Safety Culture in Health Care: Individual and Institutional Variability

98 synchronize their individual actions by working together frequently. Over time, this process is likely to make familiar team members become more efficient in conducting procedures, which can be especially important for teams under time pressure (Mackenzie et al., 2009). The mutual understanding of a situation is also perceived as a factor that maintains team coordination under time pressure when verbal communication becomes ineffective. Participants reported that familiar team members use non-verbal and implicit communication more often than do verbal communication to coordinate their actions during emergency situations. Familiar team members may maintain their team coordination when verbal communication becomes ineffective using implicit coordination mechanisms (Friedman & Bernell, 2006; Espinosa et al., 2004). Participants also reported that they trusted their familiar team members regarding their competence in carrying out their individual tasks, so that collective goals can still be achieved during emergency situations. This finding is in line with a study by Costa et al. (2001) who found that perceived trustworthiness (i.e., the extent to which individuals expect others to be and to behave according to their implicit or explicit claims) explains the highest variance of the total variance of trust (83%). Although participants reported that working with familiar team members is important for teams under time pressure, they also suggested that it is not necessary to work with familiar team members for every operation. Working with familiar team members was perceived to be particularly important for rare, difficult or complicated operations involving high-risk procedures. This is because protocols in conducting standard or routine operations are perceived to be very clear and understandable, in comparison to high-risk procedures, so that any OR team members can be assigned to these operations. An anesthesiologist described the importance of working with familiar team members for rare operations: For example, we do around 50 liver transplants per year. There are five anesthesiologists who do this procedure, thus approximately an anesthesiologist does it 10 times per year. [For procedures that need] such specialized knowledge then you have to assign fixed teams, because such procedures are not standard operations and they will never be [standard operations]. If you look at kidney transplants that are done so often so that it really becomes a routine in which [the procedure] is so clear, then you can almost assign anybody to [conduct] this procedure. The Importance of Frequency of Working Together in the Operating Room: An interview study 97

99 Participants thus reported other factors contributing to the successful outcomes that are not necessarily related to working with familiar team members. Medical experiences, knowledge and skills in a specific area (e.g., Ear, Nose and Throat specialization) are perceived as important for team members to know what to do in emergency situations. The presence of interdisciplinary team members is also perceived as vital in finding solutions for problems during emergency situations as well as staying calm in the process of stabilizing emergency situations. This study is not without limitations. First, the absence of nurses among the interviewees may limit the generalizability of the findings. Second, because OR team members were only asked to recall one incident that was prevented from becoming an accident, the current study may ignore other cases when team coordination of OR fixed teams failed to prevent an incident from becoming an accident. Third, information regarding specific clinical outcomes such as morbidity and mortality could not be obtained. Accordingly, a causal association between the frequency of working together in the OR and clinical outcomes cannot be drawn. In conclusion, the frequency of working together with particular OR team members can be considered as an important factor for team coordination under time pressure. Turnovers of team members in the OR may need to be reduced. Complicated, rare or difficult operations involving high-risk procedures may need to be conducted only by OR fixed teams whose members have worked together frequently. Research on teams also needs to account for the level of familiarity between team members as a factor that may influence team coordination and performance. Acknowledgement We thank dr. Dina Burkolter of the University of Groningen, the Netherlands, for her helpful comments on the manuscript. 98 Improving Safety Culture in Health Care: Individual and Institutional Variability

100 Appendix 1. Examples of coded text segments for the key factors related to working with familiar team members Examples of text segments Codes Sub-codes [Working with familiar team members] saves time, and that you can rely on each other. If I am going to do this, the other knows what he should do. (A surgeon) In this case, I was working with a [familiar] surgeon and we were making decisions very quickly. You only need a half word to say to each other I choose for this solution. Then you both immediately know what to do. (A surgeon) Mutual understanding of a situation Efficient team coordination Trust Mutual understanding of a situation Effective communication Team members knew what to do next Working with familiar team members saved time Participants felt that they could rely on each other Team members knew what to do next Quick communication and proactive reactions from every team member I think [that] if you let people work together more frequently, in which they communicate more to each other and they cooperate more to each other, you automatically create a healthy environment, at least in my opinion, where teamwork can develop. (A surgeon) Mutual understanding of a situation Knowing how to work with each other I think it matters a great deal whether or not you know someone in your team personally, that you know how someone is going to react in certain situations and that you can anticipate one another. This kind of knowledge makes things run more smoothly. (A surgeon) Mutual understanding of a situation Efficient team coordination Knowing team members emotional reactions to stress Working with familiar team members saved time If you work with someone you are not so familiar with, then yes I think I have to check more often for things I am responsible for. And if I work with someone I have worked together a lot, I control his or her work less. (An anesthesiologist) [If you work with less familiar team members] you have the tendency to try to control everything. You also have to give more general instructions. (A surgeon) Efficient team coordination Participants did not need to control every task of others If you are working within your fixed team, a gesture is enough to say what you need. Without saying anything, things are prepared for me. I did not even have to ask, but they knew that I would need these things. (A surgeon) Effective communication Effective communication Mutual understanding of a situation Things were prepared without being asked Gestures were enough to express what was needed Team members knew what to do next The Importance of Frequency of Working Together in the Operating Room: An interview study 99

101 Body language is always more important [than verbal communication]. One word should be enough [to communicate], and [communication] is supported by your body language. (A surgeon) Effective communication There was no need to be very explicit in communication During stressful situations, communication between our team members becomes shorter but is adequate and clear in understanding what is going on and what is each other s task. (A surgeon). Mutual understanding of a situation Effective communication Team members knew what the others were doing Quick communication and proactive reactions from every team member If you know someone so well, then you trust him that he will perform as [good as] usual, and that nothing unusual will happen. (An anesthesiologist) Trust Participants trusted that team members would do their tasks properly I think that trust plays an important role because you have someone who knows what the next step is, to help me anticipating things, because she knows, she has been standing next to me, so she knows it all by now. (A surgeon) Trust Trust Participants trusted familiar team members Participants felt that they could rely on team members 100 Improving Safety Culture in Health Care: Individual and Institutional Variability

102

103

104 General Discussion

105 Variations in safety culture Organizational learning with regards to safety is mediated by the different perspectives, interpretations and framing of situations or problems related to patient safety taken by organizational members (Gherardi et al., 1998; Antonsen, 2009). The focus of safety culture research thus should not be on diminishing differences in safety culture, but on understanding these differences to create an approach that facilitates understanding between groups (Antonsen, 2009). Understanding how different groups perceive safety culture is thus an important step in determining what and for whom institutional safeguards should be implemented to enhance patient safety (Nieva & Sorra, 2003; Antonsen, 2009; Singer et al., 2003; Pronovost et al., 2003b; Pronovost & Sexton, 2005; Huang et al., 2007; Singer et al., 2009b; Sexton et al., 2006a). Differences in safety culture between groups can be understood by measuring important aspects of safety culture using surveys (Nieva & Sorra, 2003). Flin et al. (2006) suggest including in such surveys ten most common and important safety culture dimensions. The ten dimensions including management commitment to safety, safety systems, risk perception, job demands, reporting/speaking up, safety attitudes/behaviors, communication/feedback, teamwork, personal resources and organizational factors. Except for the safety systems dimension, all of the dimensions are discussed in this thesis. Low ratings on these dimensions suggest areas for improvement. Safety culture can be assessed periodically in order to evaluate the effectiveness of patient safety intervention programs and to track progress in safety culture transformation over time (Nieva & Sorra, 2003; Sexton et al., 2006a; Singer et al., 2009b; Singer et al., 2003). This thesis aimed to investigate individual and institutional variability in safety and organizational culture in health care that should be considered in efforts to improve patient safety. Chapter 2 was aimed at gaining insights in institutional practices with regards to patient safety from all professional groups within a large university medical center in order to identify areas for institutional improvement. A questionnaire measuring nine dimensions of safety culture was distributed to all professional groups within a large university hospital. The professional groups included physicians (e.g., medical specialists), nurses (nurses in all departments of the university medical center including Intensive Care, Emergency and Surgical Departments), clinical workers (e.g., psychologists), laboratory workers (e.g., laboratory technicians) and non-medical workers (i.e., HCWs who do not have direct contact with patients such as managers). The results showed that professional groups varied in their perceptions of the different 104 Improving Safety Culture in Health Care: Individual and Institutional Variability

106 aspects of safety culture. Physicians and non-medical workers tended to rate the dimensions of safety culture more positively than did nurses, clinical workers and laboratory workers. For example, physicians evaluated institutional commitment to safety as well as relations with supervisors and career perspectives more positively than did most of the other professional groups. Nurses were relatively negative regarding working conditions and perceptions towards the hospital. The findings suggest that group-specific interventions should be a part of any campaign to improve safety culture. The findings also suggest that some interventions should be expanded to include groups other than the group for which they were developed. For example, intervention programs such as executive walk rounds (Thomas et al., 2005), in which hospital executives make visits to patient care areas to discuss patient safety issues with frontline workers, should include not only physicians and nurses, but also clinical and laboratory workers. Involving clinical and laboratory workers in collaborative rounds (Kendall, 2003; Pearson, 1999), in which hospital workers from different disciplines conduct rounds together to discuss current and future plans of care and any patient care problems with patients, may also raise awareness about the importance of the team in patient care processes. Improving safety culture must involve both those who set policy and those directly involved in patient care. The results presented in Chapter 2 suggest that hospital senior managers should present and discuss the results of safety culture studies with all staff to raise awareness of safety culture and to break down barriers between managers, team leaders and all workers who play an integral role in patient care (Pronovost & Sexton, 2005; Singer et al., 2003; Pronovost et al., 2003b). Demographic differences between responding and non-responding HCWs in the safety culture survey A challenge in surveying HCWs is to achieve a high response rate. In general, a response rate of 60% is considered necessary for achieving sufficient reliability in the measurement of safety culture (Pronovost & Sexton, 2005). Low response rates may result in under- or overrepresentation of particular groups and, thus, in non-response bias (Cull et al., 2005; Nieva & Sorra, 2003; de Vaus, 1996; Guadagnoli & Cunningham, 1989). Non-response bias is said to occur when a significant number of those who do not respond differ in terms of relevant characteristics (i.e., characteristics that can influence survey outcomes) from those who do respond. Non-response bias can be detected by investigating whether non-respondents opinions differ from those of respondents (Armstrong & Ashworth, 2000; Templeton, Deehan, Taylor, Drummond, & Strang, 1997). In principle, methods such as telephone interviews or personal visits General Discussion 105

107 can be used to contact and encourage non-respondents to participate in a survey. (Armstrong & Ashworth, 2000; Templeton et al., 1997). This way of dealing with non-response, however, cannot be applied when surveys are anonymous. Detecting non-response bias in anonymous surveys, nevertheless, can be done by comparing the demographic composition of respondents with that of non-respondents (Cartwright, 1978; McFarlane et al., 2007; Armstrong & Ashworth, 2000; Templeton et al., 1997; Bjertnaes et al., 2008; Cull et al., 2005). If demographic characteristics of non-respondents differ from those of respondents, the generalizability of the survey results may be called into question, especially if differences in the construct of interest as a function of the demographic variables have been documented in the literature. Although some safety culture surveys conducted in the past have considered the effects of demographic variables on responses and, to a lesser extent, on response rates, most studies of safety culture fail to discuss differences in safety culture as a function of demographic characteristics or make only a vague reference to such differences (Singer et al., 2009a). We argue that demographic characteristics should be considered in understanding whether HCWs will be likely to respond to safety culture surveys and that they should be taken into account in reducing the risk of non-response bias for anonymously conducted surveys. Chapter 3 is a report of demographic differences between HCWs who did or did not respond to the safety culture survey discussed in Chapter 2. The safety culture survey included five demographic questions (i.e., professional group, gender, age group, years of working in the hospital and whether one worked in an executive or non-executive function). The compositions of demographic characteristics were then compared between responding and non-responding HCWs to investigate the possible non-response bias. The results showed that demographic characteristics of responding and non-responding HCWs differed significantly in this survey. For example, nurses and clinical workers in general were found to be more likely to respond than were physicians, laboratory workers and nonmedical workers. Health care workers who had worked in the hospital for less than five years had a lower response rate than those who had worked for five years or more. It is thus concluded that demographic characteristics can be linked to response rates and need to be taken into account in conducting surveys among HCWs. The possibility of non-response bias can be reduced by conducting analyses separately as a function of relevant demographic characteristics or by sampling a higher percentage of members of groups that are known to be less likely to respond (Singer et al., 2003; Singer et al., 2009a; Singer et al., 2009b). Another approach to reducing the possibility of non-response bias is to weight responses with the reciprocal of the response rate for the respective demographic group (Holt & Elliot, 1991; Singer et al., 2003). 106 Improving Safety Culture in Health Care: Individual and Institutional Variability

108 Risk perception of general health hazards Compliance with safety regulations is arguably more important for medical workers than for non-medical workers because medical workers are exposed to various occupational health hazards which can result in serious long-term adverse health outcomes (Sepkowitz, 1996a; Sepkowitz, 1996b; Rogers, 1997). Risk perception is said to influence the level of complying behaviors with safety regulations (Sjöberg, 2003; Sjoberg, 2003). Research on risk perception has shown that people tend to rate their personal risk related to general health conditions lower than they rate risks for others (Sjoberg, 2003). This is a concern because if people are unrealistically optimistic about their health, they will tend to feel less susceptible to diseases and be less likely to change their behavior to reduce risks by (Weinstein, 1987), for example, complying with safety regulations. Chapter 4 compared the ratings of perceived risk of medical and non-medical workers to draw conclusions that may inform future research on compliance and risk perception. A questionnaire was administered as a part of a larger study about compliance of HCWs with guidelines for controlling pandemic influenza (Nap, 2009). Participants were asked about their likelihood of getting or having a cold, annual influenza, pandemic influenza, cancer or a heart attack in the next one year. Stratified sampling across professional groups was used to select medical (i.e., physicians, nurses and paramedics) and non-medical (e.g. financial services, board of directors, human resource management) workers within a university medical center to invite for participation in the study. The results showed that medical workers in general perceived their risk of getting health hazards as lower than did non-medical workers. Medical workers gave significantly lower ratings of getting cancer than did non-medical workers. The lower risk perception for cancer of medical workers is largely due to the relatively low ratings given by physicians. Physicians tended to perceive their risks of getting cancer or having a heart attack as lower than do nurses and paramedics, but having cold as higher. Age was also found to influence risk perception. Aging is correlated with higher risk perception for more serious health risks (i.e., pandemic influenza, cancer and heart attack) but lower risk perception for less serious health risks of annual influenza or cold. The findings reported in Chapter 4 suggest that different base levels of risk perception between medical and non-medical workers and among medical workers need to be taken into account for successful implementation of safety regulations. Intervention programs to improve compliance with safety regulations need to be customized for different groups as a function of how they perceive risk. General Discussion 107

109 Working together in the operating room (OR) Adverse events in the OR compose the highest number of all adverse events in hospitals (Baker et al., 2004; Wilson et al., 1995; Zegers et al., 2009; Thomas et al., 2000; Soop et al., 2009). This may be due in part to the relatively large number of HCWs who must coordinate their individual efforts to administer effective patient care as a team (Calland et al., 2002c).In treating a patient, an OR team is likely composed by several groups such as anesthesia, surgical, circulating or support and perfusion groups. Communication between OR team members is thus crucial for patient safety. The frequency with which OR team members work together may influence patient safety because the more OR team members work together, the more they are familiar with each other s working habits (Friedman & Bernell, 2006). Teams whose members are familiar to each other are likely to work more efficiently and create higher quality products than are teams whose members are strangers to each other (Harrison et al., 2003). Identifying OR team members who frequently work together is a first step in investigating characteristics of team coordination that can develop only if team members work together frequently. Chapter 5 describes a study in which social network analysis techniques were used to identify sub-groups in the OR whose members frequently work together. The techniques were able to highlight OR sub-groups whose members worked together the most frequently. Chapter 6 contains the findings of an interview study involving the members of the sub-groups identified in Chapter 5. The interview study was focused on investigating the importance of familiarity between OR team members for OR team coordination under time pressure. In order to elicit the benefits of working with familiar team members for team coordination under time pressure, team members were interviewed using the Critical Incident Technique (CIT; developed by Flanagan; 1954). The CIT is a retrospective method that requires participants to recall one or more critical incidents that they experienced personally. A critical incident is defined as an activity that has an important impact, positively or negatively, on an outcome being studied. In this study, team members were asked to recall their successes in preventing an incident from becoming a mishap in the OR. Team members were asked to base their recall on operations for which they worked together with the OR team members with whom they most often worked together. The interviewer asked how they performed as a team and probed OR team members using the CIT to identify factors that influenced success in preventing incidents from turning into accidents (Graneheim & Lundman, 2004). The interviews revealed that OR team members in general perceived that working together frequently with particular team members enabled them to coordinate their actions 108 Improving Safety Culture in Health Care: Individual and Institutional Variability

110 together efficiently in preventing incidents from becoming accidents. The responses of participants were categorized into five benefits of working with familiar team members: mutual understanding, effective communication, efficient team coordination, effective teamwork and trust. It was concluded that working together with particular OR team members frequently can be considered as an important factor for successful team coordination under time pressure. Therefore, [need a transition here] turnover of team members in the OR may need to be reduced so that team members are able to work together with each other more frequently. OR schedulers may also need to arrange OR team members to work more frequently with each other. Complicated, rare or difficult operations, which involve high-risk procedures, may need to be conducted only by OR teams whose members are familiar to each other. General conclusion Differences in HCWs perspectives, interpretations and framing of situations or problems related to patient safety can influence their compliance with intervention programs to improve safety culture. One size does not fit all because different groups are likely to perceive institutional practices differently. As importantly, different groups are also likely to have different problematic areas. It is very important to tailor intervention programs for different groups in order to achieve safer health care. General Discussion 109

111

112 Summary Organizational learning with regards to safety is mediated by the different perspectives, interpretations and framing of situations or problems taken by organizational members (Gherardi et al., 1998; Antonsen, 2009). The focus of safety culture research thus should not be on diminishing differences between groups, but on understanding these differences to create an approach that facilitates understanding between groups (Antonsen, 2009). This thesis aimed to investigate individual and institutional variability in safety and organizational culture in health care that should be considered in efforts to improve patient safety. Investigating how different groups perceive safety and organizational culture is an important step in determining what and for whom institutional safeguards should be implemented to enhance patient safety (Nieva & Sorra, 2003; Singer et al., 2003; Pronovost et al., 2003b; Pronovost & Sexton, 2005; Huang et al., 2007; Antonsen, 2009; Singer et al., 2009b; Sexton et al., 2006a). Chapter 2 thus aimed at gaining insights in institutional practices with regards to patient safety from all professional groups (i.e., physicians, nurses, clinical workers, laboratory workers and non-medical workers) within a large university medical center. A questionnaire measuring nine dimensions of safety and organizational culture was distributed to 5,609 HCWs. The results showed that professional groups varied in their perceptions of the different aspects of safety and organizational culture. Physicians and non-medical workers tended to rate the dimensions more positively than did nurse, clinical workers and laboratory workers. The findings suggest that group-specific interventions should be a part of any campaign to improve safety and organizational culture. The findings also suggest that some interventions should be expanded to include groups other than the group for which they were developed. Because surveying HCWs, especially physicians, has been documented as a problem (Asch et al., 1997; Asch et al., 2000; Cartwright, 1978; Cummings et al., 2001; Ward, 1994; VanGeest et al., 2007; Singer et al., 2003), Chapter 3 investigated demographic differences between HCWs who did or did not respond to the safety and organizational culture survey in Chapter 2. This is important because low response rates may result in under- or overrepresentation of particular groups and, thus, in non-response bias (Cull et al., 2005; Nieva & Sorra, 2003; de Vaus, 1996; Guadagnoli & Cunningham, 1989). The findings show that response rate depended on professional group, gender, age, years of working in the hospital and executive function. Demographic characteristics thus need to be taken into account in conducting surveys among HCWs. The possibility of non-response bias can be reduced by conducting Summary 111

113 analyses separately as a function of relevant demographic characteristics, sampling a higher percentage of groups that are known to be less likely to respond (Singer et al., 2003; Singer et al., 2009a; Singer et al., 2009b), or weighting responses with the reciprocal of the response rate for the respective demographic group (Holt & Elliot, 1991; Singer et al., 2003). Understanding how HCWs perceive risks is important if safety culture is to be improved because risk perception is said to influence the level of complying behaviors with safety regulations (Sjöberg, 2003; Sjoberg, 2003). Research on risk perception has shown that people tend to rate their personal risk related to general health conditions lower than they rate risks for others (Sjoberg, 2003). This is a concern because if people are unrealistically optimistic about their health, they will tend to feel less susceptible to diseases and be less likely to change their behavior to reduce risks by (Weinstein, 1987), for example, complying with safety regulations. Chapter 4 compared the ratings of perceived risk of medical and non-medical workers to draw conclusions that may inform future research on compliance and risk perception. Compliance with safety regulations is arguably more important for medical workers than non-medical workers, because medical workers are exposed to various occupational health hazards which can result in serious long-term adverse health outcomes (Sepkowitz, 1996b; Sepkowitz, 1996a; Rogers, 1997). A questionnaire was administered to 6,380 HCWs as a part of a larger study about compliance of HCWs with guidelines for controlling pandemic influenza (Nap, 2009). Participants were asked about their likelihood of getting or having a cold, annual influenza, pandemic influenza, cancer or a heart attack in the next one year. The results showed that professional group, age and family status influenced risk perception. For example, medical workers in general perceived their risk of getting health hazards as lower than did non-medical workers. The findings suggest that different base levels of risk perception between medical and non-medical workers and among medical workers need to be taken into account for successful implementation of safety regulations. Intervention programs to improve compliance with safety regulations need to be customized for different groups as a function of how they perceive risk. Chapter 5 and 6 were focused on investigating the importance of frequency of working together in the operating room (OR). Adverse events in the OR compose the highest number of adverse events in hospitals (Baker et al., 2004; Wilson et al., 1995; Zegers et al., 2009; Thomas et al., 2000; Soop et al., 2009). This may be due in part to the relatively large number of HCWs who must coordinate their individual efforts to administer effective patient care as a team (Calland et al., 2002d). In treating a patient, an OR team is likely composed by several groups such as anesthesia, surgical, circulating or support and perfusion groups. The frequency with 112 Summary

114 which OR team members work together may influence patient safety. Teams whose members are familiar to each other are likely to work more efficiently and create higher quality products than are teams whose members are strangers to each other (Harrison et al., 2003). Chapter 5 describes a study in which social network analysis techniques were used to identify sub-groups in the OR whose members frequently work together. The techniques were able to highlight OR sub-groups whose members worked together the most frequently. Chapter 6 discusses the results of the interviews conducted with the members of the sub-groups identified in Chapter 5. Chapter 6 was focused on investigating the importance of working together with familiar OR team members to prevent an incident from becoming an accident. The findings show that OR team members generally perceived that working with familiar team members enabled the team to coordinate actions efficiently in preventing incidents from becoming accidents. The responses of participants were categorized into four benefits of working with familiar team members: Mutual understanding, effective communication, efficient team coordination and trust. Working together frequently with particular OR team members can be considered as an important factor for successful team coordination under time pressure. Complicated, rare or difficult operations, which involve high-risk procedures, may need to be conducted only by OR teams whose members are familiar to each other. Summary 113

115

116 Samenvatting Het verbeteren van de veiligheidscultuur in de gezondheidszorg: Implicaties van individuele en institutionele variabiliteit Translated to Dutch by Leon G. Faber Organisationeel leren op het gebied van veiligheid wordt gemedieerd door de verschillende perspectieven, interpretaties en inkadering van situaties of problemen door deelnemers van organisaties (Gherardi et al., 1998; Antonsen, 2009). De focus van onderzoek naar veiligheidscultuur zou dus niet moeten liggen op het verkleinen van verschillen tussen groepen, maar op het begrijpen van deze verschillen, om tot een benadering te komen die het begrip tussen groepen faciliteert (Antonsen, 2009). Dit proefschrift is gericht op het onderzoeken van individuele en institutionele variabiliteit binnen de organisatorische en veiligheidscultuur in de gezondheidszorg, die in acht genomen zouden moeten worden bij het verbeteren van de veiligheid voor patiënten. Onderzoeken hoe verschillende groepen naar organisatie- en veiligheidscultuur kijken is een belangrijke stap om te bepalen voor wie en welke institutionele veiligheidsmaatregelen geïmplementeerd zouden moeten worden om patiëntveiligheid te verbeteren (Nieva & Sorra, 2003; Singer et al., 2003; Pronovost et al., 2003b; Pronovost & Sexton, 2005; Huang et al., 2007; Antonsen, 2009; Singer et al., 2009b; Sexton et al., 2006a). Hoofdstuk 2 is daarom gericht op het verkrijgen van inzichten in institutionele routines met betrekking tot patiëntveiligheid van alle professionele groepen (bijv. artsen, verpleegsters, klinisch medewerkers, laboranten en niet-medisch medewerkers) binnen een groot universitair medisch centrum. Een vragenlijst die negen dimensies van organisatie- en veiligheidscultuur mat, werd gedistribueerd over 5,609 gezondheidszorgmedewerkers (GZM s). De resultaten lieten zien dat de percepties van de verschillende aspecten van organisatie- en veiligheidscultuur varieerden over de professionele groepen. Artsen en niet-medisch medewerkers neigden de dimensies positiever in te schatten dan verpleegkundigen, klinisch medewerkers en laboranten. De bevindingen duidden erop dat groepsspecifieke interventies deel zouden uit moeten maken van iedere campagne om de organisatorische en veiligheidscultuur te verbeteren. Tevens lieten de bevindingen zien dat sommige interventies uitgebreid zouden moeten worden om andere groepen te betrekken dan de groepen waarvoor ze in eerste instantie ontwikkeld zijn. Samenvatting 115

117 Omdat het ondervragen van GZM s, en met name artsen, vaak problematisch blijkt te zijn (Asch et al., 1997; Asch et al., 2000; Cartwright, 1978; Cummings et al., 2001; Ward, 1994; VanGeest et al., 2007; Singer et al., 2003), wordt er in hoofdstuk 3 gekeken naar demografische verschillen tussen GZM s die wel of niet geantwoord hebben op de organisatie- en veiligheidscultuur-vragenlijst van hoofdstuk 2. Het belang hiervan is dat als weinig mensen geantwoord hebben op de vragenlijst, dit kan resulteren in onder- of overrepresentatie van bepaalde groepen en dus in een non-response bias (Cull et al., 2005; Nieva & Sorra, 2003; de Vaus, 1996; Guadagnoli & Cunningham, 1989). De bevindingen lieten zien dat de responspercentages afhingen van professionele groep, geslacht, leeftijd, aantal arbeidsjaren in het ziekenhuis en het al dan niet hebben van een leidinggevende positie. Demografische karakteristieken moeten daarom meegenomen worden bij het doen van onderzoek onder GZM s. De kans op non-response bias kan verkleind worden door analyses onafhankelijk uit te voeren als een functie van relevante demografische karakteristieken, een hoger percentage mensen te ondervragen in groepen waarvan bekend is dat ze minder geneigd zijn om te responderen (Singer et al., 2003; Singer et al., 2009a; Singer et al., 2009b), of de responsen te wegen met het omgekeerde van de respons percentages van de betreffende demografische groepen (Holt & Elliot, 1991; Singer et al., 2003). Begrip van hoe GZM s risico s ervaren is van belang bij het verbeteren van de veiligheidscultuur, omdat van risicoperceptie bekend is dat het bepaalt in welke mate men veiligheidsregels naleeft (Sjöberg, 2003; Sjoberg, 2003). Onderzoek naar risicoperceptie heeft laten zien dat mensen hun eigen risico, betreffende hun algemene gezondheidstoestand, lager inschatten dan risico s voor anderen (Sjoberg, 2003). Dit leidt tot zorg, omdat als mensen onrealistisch optimistisch zijn over hun gezondheid, ze het gevoel hebben minder vatbaar te zijn voor ziekten en ze minder geneigd zijn om hun gedrag aan te passen om risico s te verkleinen (Weinstein, 1987), bijvoorbeeld door het naleven van veiligheidsregels. Hoofdstuk 4 vergelijkt de inschattingen van risicoperceptie van medisch en niet-medisch personeel, om hieruit conclusies te trekken die gebruikt kunnen worden in toekomstig onderzoek naar naleving van regels en risicoperceptie. Naleving van veiligheidsregels is mogelijk belangrijker voor medisch dan voor niet-medisch personeel, omdat medisch personeel blootgesteld wordt aan verscheidene werkgerelateerde gevaren die kunnen leiden tot ernstige gezondheidsproblemen op de lange termijn (Sepkowitz, 1996b; Sepkowitz, 1996a; Rogers, 1997). Er werd een vragenlijst uitgegeven aan 6,380 GZM s, als onderdeel van een groter onderzoek naar naleving van richtlijnen door GZM s om een grieppandemie te beheersen (Nap, 2009). Deelnemers werd gevraagd naar de waarschijnlijkheid dat ze gedurende het eerstkomende 116 Samenvatting

118 jaar een verkoudheid, de jaarlijkse griep, pandemische griep, kanker of een hartaanval zouden krijgen. De resultaten lieten zien dat professionele groep, leeftijd en gezinsstatus risicoperceptie beïnvloedden. Medisch personeel schatte bijvoorbeeld hun risico op gezondheidsrisico s lager in dan niet-medisch personeel. De bevindingen lijken erop te wijzen dat er voor een succesvolle implementatie van veiligheidsregels, er rekening moet worden gehouden met verschillende basisniveaus van risicoperceptie tussen medisch en niet-medisch personeel. Interventieprogramma s om de naleving van veiligheidsregels te bevorderen zouden aangepast moeten worden aan de verschillende doelgroepen, op basis van hoe ze risico s inschatten. Hoofdstukken 5 en 6 zijn gericht op het onderzoeken van het belang van de frequentie van samenwerking in de operatiekamer (OK). Het grootste aantal ongelukken in ziekenhuizen bestaat uit ongelukken in de OK (Baker et al., 2004; Wilson et al., 1995; Zegers et al., 2009; Thomas et al., 2000; Soop et al., 2009). Dit kan ten dele liggen aan het relatief grote aantal GZM s die hun individuele inspanning moeten coördineren om als team effectieve zorg te leveren (Calland et al., 2002c). Tijdens de behandeling van een patiënt bestaat een OK-team meestal uit verschillende groepen, zoals anesthesie-, chirurgie-, circulatie- of ondersteuningen perfusiegroepen. Hoe vaak OK-teamgenoten samenwerken zou patiëntveiligheid kunnen beïnvloeden. Teams, waarvan leden elkaar kennen zullen waarschijnlijk efficiënter werken en een hogere kwaliteit werk leveren dan teams die bestaan uit mensen die vreemden voor elkaar zijn (Harrison et al., 2003). Hoofdstuk 5 beschrijft een onderzoek waarin sociale netwerkanalysetechnieken gebruikt zijn om subgroepen in de OK te identificeren, waarvan de leden het vaakst hebben samengewerkt. Hoofdstuk 6 bespreekt de resultaten van de interviews die uitgevoerd werden met de leden van de subgroepen die in hoofdstuk 5 geïdentificeerd werden. Hoofdstuk 6 was erop gericht om te onderzoeken wat het belang is van het samenwerken met bekenden om te voorkomen dat incidenten ongelukken worden. De bevindingen laten zien dat OK-teamleden over het algemeen ervoeren dat het samenwerken met teamleden die ze kenden hen in staat stelden om acties efficiënter te coördineren bij het voorkomen dat incidenten ongelukken werden. De responses van deelnemers gecategoriseerd in vier voordelen van werken met bekende teamgenoten: wederzijds begrip, effectieve communicatie, efficiënte teamcoördinatie en vertrouwen. Regelmatig samenwerken met specifieke OK-teamgenoten kan opgevat worden als een belangrijke factor voor succesvolle teamcoördinatie onder tijdsdruk. Gecompliceerde, uitzonderlijke of moeilijke operaties met riskante procedures, zouden mogelijk alleen door OK-teams waarvan de teamgenoten elkaar kennen uitgevoerd moeten worden. Samenvatting 117

119

120 Reference List Occupational Outlook Handbook, Edition. (2009). Bureau of Labor Statistics,U.S.Department of Labor. Retrieved from Adams, A., & Bond, S. (2000). Hospital nurses job satisfaction, individual and organizational characteristics. J Adv Nurs, 32(3), Retrieved from ISI: Aiken, L. H., Clarke, S. P., Sloane, D. M., Sochalski, J. A., Busse, R., Clarke, H. et al. (2001). Nurses reports on hospital care in five countries. Health Aff, 20(3), Retrieved from ISI: Alaszewski, A., & Horlick-Jones, T. (2003). How can doctors communicate information about risk more effectively? Br Med J, 327(7417), Retrieved from ISI: Antonsen, S. (2009). Safety Culture: Theory, Method, and Improvement. Surrey: Ashgate Publishing Limited. Armstrong, D., & Ashworth, M. (2000). When questionnaire response rates do matter: a survey of general practitioners and their views of NHS changes. Br J Gen Pract, 50(455), Retrieved from ISI: Asch, D. A., Jedrziewski, M. K., & Christakis, N. A. (1997). Response rates to mail surveys published in medical journals. J Clin Epidemiol, 50(10), Retrieved from ISI:A1997YA Asch, S., Connor, S. E., Hamilton, E. G., & Fox, S. A. (2000). Problems in recruiting community-based physicians for health services research. J Gen Intern Med, 15(8), Retrieved from ISI: Awad, S. S., Fagan, S. P., Bellows, C., Albo, D., Green-Rashad, B., De la, G. M. et al. (2005). Bridging the communication gap in the operating room with medical team training. Am Lit Hist, 190(5), Retrieved from PM: Baker, G. R., Norton, P. G., Flintoft, V., Blais, R., Brown, A., Cox, J. et al. (2004). The Canadian Adverse Events Study: the incidence of adverse events among hospital patients in Canada. Can Med Assoc J, 170(11), Retrieved from ISI: Becker, M. H., Janz, N. K., Band, J., Bartley, J., Snyder, M. B., & Gaynes, R. P. (1990). Noncompliance with Universal Precautions policy - why do physicians and nurses recap needles. Am.J.Infect.Control, 18(4), Retrieved from ISI:A1990DV Berger, P. L., & T.Luckmann. (1966). The social construction of reality: a treatise in the sociology of knowledge. Garden City, NY: Anchor Books. Bjertnaes, O. A., Garratt, A., & Botten, G. (2008). Nonresponse bias and cost-effectiveness in a Norwegian survey of family physicians. Eval Health Prof, 31(1), Retrieved from ISI: Booth, G. L., Kapral, M. K., Fung, K., & Tu, J. V. (2006). Relation between age and cardiovascular disease in men and women with diabetes compared with non-diabetic people: a population-based retrospective cohort study. Lancet, 368(9529), Retrieved from ISI: Borgatti, S. P. (1994). How to Explain Hierarchical Clustering. Borgatti, S. P., Everett, M. G., & Freeman, L. C. (2002). UCINET 6 for Windows: Software for Social Network Analysis Harvard, MA: Analytic Technologies. Bosk, C. (2003). Forgive and remember: Managing medical failure (2nd ed.). Chicago: University of Chicago Press. Bourne, L. E., & Yaroush, R. A. (2003). Stress and cognition: A cognitive psychological perspective. Unpublished manuscript, NASA grant NAG Retrieved from Citeseer. Brennan, T. A., Leape, L. L., Laird, N. M., Hebert, L., Localio, A. R., Lawthers, A. G. et al. (1991). Incidence of adverse events and negligence in hospitalized patients. Results of the Harvard Medical Practice Study I. New Engl J Med, 324(6), Retrieved from PM: Calland, J. F., Adams, R. B., Benjamin, D. K., Jr., O Connor, M. J., Chandrasekhara, V., Guerlain, S. et al. (2002a). Thirtyday postoperative death rate at an academic medical center. Ann Surg, 235(5), Retrieved from PM: Calland, J. F., Guerlain, S., Adams, R. B., Tribble, C. G., Foley, E., & Chekan, E. G. (2002b). A systems approach to surgical safety. Surg Endosc-ultras, 16(6), Retrieved from ISI: Reference List 119

121 Calland, J. F., Guerlain, S., Adams, R. B., Tribble, C. G., Foley, E., & Chekan, E. G. (2002d). A systems approach to surgical safety. Surg.Endosc., 16(6), Retrieved from PM: Cartwright, A. (1978). Professionals as responders - variations in and effects of response rates to questionnaires, Br Med J, 2(6149), Retrieved from ISI:A1978FW Chor, J. S. Y., Ngai, K. L. K., Goggins, W. B., Wong, M. C. S., Wong, S. Y. S., Lee, N. et al. (2009). Willingness of Hong Kong healthcare workers to accept pre-pandemic influenza vaccination at different WHO alert levels: two questionnaire surveys. Br Med J, 339. Retrieved from ISI: Cooper, M. D., & Phillips, R. A. (2004). Exploratory analysis of the safety climate and safety behavior relationship. J Safety Res, 35(5), Retrieved from ISI: Corser, W. D. (2000). The contemporary nurse-physician relationship: Insights from scholars outside the two professions. Nurs Outlook, 48(6), Retrieved from ISI: Costa, A. C., Roe, R. A., & Taillieu, T. (2001). Trust within teams: The relation with performance effectiveness. EJWOP, 10(3), Cott, C. (1997). We decide, you carry it out : A social network analysis of multidisciplinary long-term care teams. Soc Sci Med, 45(9), Retrieved from PM: Cull, W. L., O Connor, K. G., Sharp, S., & Tang, S. F. S. (2005). Response rates and response bias for 50 surveys of pediatricians. Health Serv Res, 40(1), Retrieved from ISI: Cummings, S. M., Savitz, L. A., & Konrad, T. R. (2001). Reported response rates to mailed physician questionnaires. Health Serv Res, 35(6), Retrieved from ISI: Davies, J. M. (2005). Team communication in the operating room. Acta Anaesth Scand, 49(7), Retrieved from PM: Davis, P., Lay-Yee, R., Briant, R., Ali, W., Scott, A., & Schug, S. (2002). Adverse events in New Zealand public hospitals I: occurrence and impact. N.Z.Med J, 115(1167), U271. Retrieved from PM: de Vaus, D. A. (1996). Finding a sample. In Surveys in social research (Fourth ed., pp ). NSW: Allen and Unwin Pty Ltd. de Vries, E. N., Boermeester, M. A., & Gouma, D. J. (2008). WHO s checklist for surgery: don t confine it to the operating room. Lancet, 372(9644), Retrieved from ISI: de Vries, E. N., Hollmann, M. W., Smorenburg, S. M., Gouma, D. J., & Boermeester, M. A. (2009). Development and validation of the SURgical PAtient Safety System (SURPASS) checklist. Qual Saf Health Care, 18(2), Retrieved from ISI: de Vries, E. N., Ramrattan, M. A., Smorenburg, S. M., Gouma, D. J., & Boermeester, M. A. (2008). The incidence and nature of in-hospital adverse events: a systematic review. Qual Saf Health Care, 17(3), Retrieved from ISI: Dekker, S. (2007). Just Culture. Aldershot: Ashgate. DePinho, R. A. (2000). The age of cancer. Nature, 408(6809), Retrieved from ISI: Dirks, K. T. (1999). The effects of interpersonal trust on work group performance. J Appl Psycho, 84(3), Retrieved from ISI: Doucette, J. N. (2006). View from the cockpit: what the airline industry can teach us about patient safety. Nursing, 36(11), Retrieved from PM: Edmondson, A. C. (2003). Speaking up in the operating room: How team leaders promote learning in interdisciplinary action teams. J manage stud, 40. Entin, E. B., Lai, F., & Barach, P. (2006). Training teams for the perioperative environment: a research agenda. Surg Innov, 13(3), Retrieved from PM: Espinosa, J. A., Slaughter, S. A., Kraut, R. E., & Herbsleb, J. D. (2007). Familiarity, complexity, and team performance in geographically distributed software development. Organ Sci, 18(4), Retrieved from ISI: Espinosa, J. A., Lerch, F. J., & Kraut, R. E. (2004). Explicit versus implicit coordination mechanisms and task dependencies: One size does not fit all. In E. Salas & Fiore S.M. (Eds.), Team cognition: Understanding 120 Reference List

122 the factors that drive process and performance (pp ). Washington, DC, US: American Psychological Association. Esposito, S., Bosis, S., Pelucchi, C., Tremolati, E., Sabatini, C., Semino, M. et al. (2008). Influenza vaccination among healthcare workers in a multidisciplinary university hospital in Italy. Bmc Public Health, 8. Retrieved from ISI: Feng, X. Q., Acord, L., Cheng, Y. J., Zeng, J. H., & Song, J. P. (2011). The relationship between management safety commitment and patient safety culture. Int Nurs Rev. Retrieved from Wiley Online Library. Flanagan, J. C. (1954). The critical incident technique. Psychol Bull, 51(4), Retrieved from PM: Flin, R., Burns, C., Mearns, K., Yule, S., & Robertson, E. M. (2006). Measuring safety climate in health care. Qual Saf Health Care, 15(2), Retrieved from ISI: Friedman, L. H., & Bernell, S. L. (2006). The importance of team level tacit knowledge and related characteristics of high-performing health care teams. Health Care Manage Rev, 31(3), Retrieved from PM: Gaba, D. M. (2004). The future vision of simulation in health care. Qual Saf Health Care, 13, I2-I10. Retrieved from ISI: Gaba, D. M., Howard, S. K., Fish, K. J., Smith, B. E., & Sowb, Y. A. (2001). Simulation-based training in anesthesia crisis resource management (ACRM): A decade of experience. SAGE: A Scholarly Journal on Black Women 32, Gershon, R. R., Vlahov, D., Felknor, S. A., Vesley, D., Johnson, P. C., Delclos, G. L. et al. (1995). Compliance with universal precautions among health care workers at three regional hospitals. Am.J.Infect.Control, 23(4), Retrieved from PM: Gersick, C. J. G., & Hackman, J. R. (1990). Habitual Routines in Task-Performing Groups. Organ Behav and Hum Decis Process, 47(1), Retrieved from ISI:A1990EF Gherardi, S., Nicolini, D., & Odella, F. (1998). What Do You Mean By Safety? Conflicting Perspectives on Accident Causation and Safety Management in a Construction Firm. JCCM, 6(4), doi: / Retrieved from Retrieved from Blackwell Publishers Ltd. Glezen, W. P. (1996). Emerging infections: pandemic influenza. Epidemiol Rev, 18(1), Retrieved from ISI:A1996VF Graneheim, U. H., & Lundman, B. (2004). Qualitative content analysis in nursing research: concepts, procedures and measures to achieve trustworthiness. Nurse Educ Today, 24(2), Retrieved from ISI: Grogan, E. L., Stiles, R. A., France, D. J., Speroff, T., Morris, J. A., Jr., Nixon, B. et al. (2004). The impact of aviationbased teamwork training on the attitudes of health-care professionals. J.Am.Coll.Surg., 199(6), Retrieved from PM: Gruenfeld, D. H., Mannix, E. A., Williams, K. Y., & Neale, M. A. (1996). Group composition and decision making: How member familiarity and information distribution affect process and performance. Organ Behav and Hum Decis Process, 67(1), Retrieved from ISI:A1996VB Guadagnoli, E., & Cunningham, S. (1989). The effects of nonresponse and late response on a survey of physician attitudes. Eval Health Prof, 12(3), Retrieved from ISI:A1989AP Guerlain, S., Adams, R. B., Turrentine, F. B., Shin, T., Guo, H., Collins, S. R. et al. (2005). Assessing team performance in the operating room: development and use of a black-box recorder and other tools for the intraoperative environment. J Am Coll Surg, 200(1), Retrieved from PM: Hackman, J. R., & Oldham, G. R. (1975). Development of the Job Diagnostic Survey. Appl Psychol, 60(2), Hackman, J. R., & Oldham, G. R. (1976). Motivation Through Design of Work - Test of A Theory. Organ Behav Hum Perform, 16(2), Retrieved from ISI:A1976CA Halamek, L. P., Kaegi, D. M., Gaba, D. M., Sowb, Y. A., Smith, B. C., Smith, B. E. et al. (2000). Time for a new paradigm in pediatric medical education: teaching neonatal resuscitation in a simulated delivery room environment. Pediatrics, 106(4), E45. Retrieved from PM: Hales, B., Terblanche, M., Fowler, R., & Sibbald, W. (2008). Development of medical checklists for improved quality of Reference List 121

123 patient care. Int J Qual Health C, 20(1), Retrieved from ISI: Hales, B. M., & Pronovost, P. J. (2006). The checklist- a tool for error management and performance improvement. J Crit Care, 21(3), Retrieved from ISI: Hall, P., & Weaver, L. (2001). Interdisciplinary education and teamwork: a long and winding road. Med Educ, 35(9), Retrieved from ISI: Halverson, A. L., Casey, J. T., Andersson, J., Anderson, K., Park, C., Rademaker, A. W. et al. (2011). Communication failure in the operating room. Surgery, 149(3), Retrieved from PM: Harris, K. T., Treanor, C. M., & Salisbury, M. L. (2006). Improving patient safety with team coordination: Challenges and strategies of implementation. J Obstet Gynecol Neonatal Nurs, 35(4), Retrieved from ISI: Harrison, D. A., Mohammed, S., McGrath, J. E., Florey, A. T., & Vanderstoep, S. W. (2003, FAL). Time matters in team performance: Effects of member familiarity, entrainment, and task discontinuity on speed and quality. Pers Psychol, 56(3), Retrieved from ISI: Haynes, A. B., Weiser, T. G., Berry, W. R., Lipsitz, S. R., Breizat, A. H. S., Dellinger, E. P. et al. (2009). A Surgical Safety Checklist to Reduce Morbidity and Mortality in a Global Population. New Engl J Med, 360(5), Retrieved from ISI: Helmreich, R. L., & Merritt, A. C. (2001). Culture at Work in Aviation and Medicine: National, Organizational and Professional Influences. Brookfield, VT: Ashgate Publishing. Helmreich, R. L., Merritt, A. C., & Wilhelm, J. A. (1999). The evolution of crew resource management training in commercial aviation. Int J Aviat Psychol, 9(1), Retrieved from ISI: Hennink, M., Hutter, I., & Bailey, A. (2011). Qualitative Research Methods (1 ed.). London: Sage Publisher, Ltd. Hinds, P. J., Carley, K. M., Krackhardt, D., & Wholey, D. (2000). Choosing work group members: Balancing similarity, competence, and familiarity. Organ Behav Hum Decis Process, 81(2), Retrieved from ISI: Holt, D., & Elliot, D. (1991). Methods of weighting for unit nonresponse. Statistician, 40(3), Retrieved from ISI:A1991GB Huang, D. T., Clermont, G., Sexton, J. B., Karlo, C. A., Miller, R. G., Weissfeld, L. A. et al. (2007). Perceptions of safety culture vary across the intensive care units of a single institution. Am J Respir Crit Care Med, 35(1), Retrieved from PM: Institute of Medicine. (1999). To Err Is Human: Building a Safer Health System. Washington, D.C.: National Academy Press. Institute of Medicine. (2000). To err is human: Building a safer health system. Washington, D.C.: National Academy Press. Institute of Medicine. (2001). Crossing the Quality Chasm: A New Health System for the 21st Century. Washington, D.C.: National Academy Press. Institute of Medicine. (2004). IOM report: Patient safety - Achieving a new standard for care (12 ed.) Retrieved from ISI: Johnson, A., & Bakker, S. (2007). Cooperation and safety culture in intensive care units and safety and rescue teams. Tijdschrift voor Ergonomie, 32(2), Kaner, E. F., Haighton, C. A., & Mcavoy, B. R. (1998). So much post, so busy with practice - so, no time! : a telephone survey of general practitioners reasons for not participating in postal questionnaire surveys. Br J Gen Pract, 48(428), Retrieved from ISI: Kawashima, Y., Takahashi, S., Suzuki, M., Morita, K., Irita, K., Iwao, Y. et al. (2003). Anesthesia-related mortality and morbidity over a 5-year period in 2,363,038 patients in Japan. Acta Anaesthesiol Scand, 47(7), Retrieved from PM: Keating, N. L., Ayanian, J. Z., Cleary, P. D., & Marsden, P. V. (2007). Factors affecting influential discussions among physicians: A social network analysis of a primary care practice. J Gen Intern med, 22(6), Retrieved from PM: Kendall, E. M. (2003). Improving patient care with collaborative rounds. Am J Health Syst Pharm, 60(2), Reference List

124 Retrieved from ISI: Kho, M. E., Carbone, J. M., Lucas, J., & Cook, D. J. (2005). Safety climate survey: reliability of results from a multicenter ICU survey. Qual Saf Health Care, 14(4), Retrieved from ISI: Laschinger, H. K. S., & Leiter, M. P. (2006). The impact of nursing work environments on patient safety outcomes - The mediating role of burnout/engagement. J Nurs Adm, 36(5), Retrieved from ISI: Leape, L. L. (2000). Institute of Medicine Medical Error Figures Are Not Exaggerated. JAMA-J Am Med Assoc, 284(1), Leape, L. L. (2002). Reporting of adverse events. New Engl J Med, 347(20), Retrieved from ISI: Leape, L. L., Brennan, T. A., Laird, N., Lawthers, A. G., Localio, A. R., Barnes, B. A. et al. (1991). The nature of adverse events in hospitalized patients. Results of the Harvard Medical Practice Study II. New Engl J Med, 324(6), Retrieved from PM: Leliopoulou, C., Waterman, H., & Chakrabarty, S. (1999). Nurses failure to appreciate the risks of infection due to needle stick accidents: a hospital based survey. J Hosp Infect, 42(1), Retrieved from ISI: Letaief, M., El Mhamdi, S., El-Asady, R., Siddiqi, S., & Abdullatif, A. (2010). Adverse events in a Tunisian hospital: results of a retrospective cohort study.int J Qual Health Care, 22(5), Retrieved from ISI: Levinson, W., & Dunn, P. M. (1989). Coping with Fallibility. JAMA-J Am Med Assoc, 261(15), Retrieved from ISI:A1989U Lin, S. H., Tang, W. J., Miao, J. Y., Wang, Z. M., & Wang, P. X. (2008a). Safety climate measurement at workplace in China: A validity and reliability assessment. Saf Sci, 46(7), Retrieved from ISI: Lin, S. H., Tang, W. J., Miao, J. Y., Wang, Z. M., & Wang, P. X. (2008b). Safety climate measurement at workplace in China: a validity and reliability assessment. Saf Sci, 46(7), Retrieved from ISI: Lingard, L., Espin, S., Whyte, S., Regehr, G., Baker, G. R., Reznick, R. et al. (2004). Communication failures in the operating room: an observational classification of recurrent types and effects. Qual Saf Health Care, 13(5), Retrieved from PM: Lingard, L., Regehr, G., Orser, B., Reznick, R., Baker, G. R., Doran, D. et al. (2008). Evaluation of a preoperative checklist and team briefing among surgeons, nurses, and anesthesiologists to reduce failures in communication. Arch Surg-Chicago, 143(1), Retrieved from ISI: Lingard, L., Reznick, R., Devito, I., & Espin, S. (2002). Forming professional identities on the health care team: discursive constructions of the other in the operating room. Med.Educ., 36(8), Retrieved from PM: Listyowardojo, T. A., Steglich, C. E. G., Peuchen, S., & Johnson, A. (2009b). Using social network analysis to identify sub-groups in the operating room. In Dick de Waard, Hans Godthelp, Frank Kooi, & Karel Brookhuis (Eds.), Human Factors, Security and Safety (pp ). Maastricht, The Netherlands: Shaker Publishing. Locke, E. A. (1969). What is job satisfaction. Organ Behav Hum Perform, 4(4), Retrieved from ISI:A1969Y Lu, H., While, A. E., & Barriball, K. L. (2005). Job satisfaction among nurses: A literature review. Int J Nurs Stud, 42(2), Retrieved from ISI: Lyndon, A. (2006). Communication and teamwork in patient care: how much can we learn from aviation? J Obstet gynecol Neonatal Nurs, 35(4), Retrieved from PM: Mackenzie, C. F., Jeffcott, S. A., & Xiao, Y. (2009). Measuring the impact of time pressure on team task performance. In R. Flin & L. Mitchell (Eds.), Safer Surgery: Analysing behaviour in the operating theatre (pp ). Farnham: Ashgate Publishing Limited. Makary, M. A., Mukherjee, A., Sexton, J. B., Syin, D., Goodrich, E., Hartmann, E. et al. (2007). Operating room briefings and wrong-site surgery. J Am Coll Surg, 204(2), Retrieved from PM: Makary, M. A., Sexton, J. B., Freischlag, J. A., Holzmueller, C. G., Millman, A., Rowen, L. et al. (2006). Operating room Reference List 123

125 teamwork among physicians and nurses: teamwork in the eye of the beholder. J Am Coll Surg, 202(5), Retrieved from ISI: McCarthy, D., & Blumenthal, D. (2006). Stories from the sharp end: Case studies in safety improvement. Milbank Q, 84(1), Retrieved from ISI: McDonald, C. J., Weiner, M., & Hui, S. L. (2000). Deaths Due to Medical Errors Are Exaggerated in Institute of Medicine Report. JAMA-J Am Med Assoc, 284(1), McDonald, R., Waring, J., Harrison, S., Walshe, K., & Boaden, R. (2005). Rules and guidelines in clinical practice: a qualitative study in operating theatres of doctors and nurses views. Qual Saf Health Care, 14(4), Retrieved from ISI: McFarlane, E., Olmsted, M. G., Murphy, J., & Hill, C. A. (2007). Nonresponse bias in a mail survey of physicians. Eval Health Prof, 30(2), Retrieved from ISI: Michel, P., Quenon, J. L., de Sarasqueta, A. M., & Scemama, O. (2004). Comparison of three methods for estimating rates of adverse events and rates of preventable adverse events in acute care hospitals. Brit Med J, 328(7433), A. Retrieved from ISI: Morey, J. C., & Salisbury, M. (2002). Introducing teamwork training into healthcare organizations: Implementation issues and solutions. Proceedings of the human factors and ergonomics society 46th annual meeting (pp ) Santa Monica, CA: Human Factors and Ergonomics Society. Morey, J. C., Simon, R., Jay, G. D., & Rice, M. M. (2003). A transition from aviation crew resource management to hospital emergency departments: The MedTeams story. Proceedings of the Twelfth International Symposium on Aviation Psychology. The Aviation Psychology Laboratory of the Ohio State University (pp ) Columbus,OH. Morey, J. C., Simon, R., Jay, G. D., Wears, R. L., Salisbury, M., Dukes, K. A. et al. (2002). Error reduction and performance improvement in the emergency department through formal teamwork training: evaluation results of the MedTeams project. Health Serv.Res., 37(6), Retrieved from PM: Nap, R. E. (2009). Work compliance of hospital personnel during pandemic influenza. Nichol, K. L., & Hauge, M. (1997). Influenza vaccination of healthcare workers. Infect Control Hosp Epidemiol, 18(3), Retrieved from ISI:A1997WP Nieva, V. F., & Sorra, J. (2003). Safety culture assessment: A tool for improving patient safety in healthcare organizations. Qual Saf Health Care, 12, Retrieved from ISI: O Connor, M. F. (2007). Safety culture: Easy to advocate, difficult to create. Crit Care Med, 35(5), Retrieved from ISI: Paling, J. (2003). Strategies to help patients understand risks. Br Med J, 327(7417), Retrieved from ISI: Parker, L. E., Kirchner, J. A. E., Bonner, L. M., Fickel, J. J., Ritchie, M. J., Simons, C. E. et al. (2009). Creating a qualityimprovement dialogue: utilizing knowledge from frontline staff, managers, and experts to foster health care quality improvement. Qual Health Res, 19(2), 229. Retrieved from Sage Publications. Pearson, M. J. (1999). The clinical role of clinical biochemists. Ann Clin Biochem, 36, Retrieved from ISI: Presidential Commission on the Space Shuttle Challenger Accident. (1986). Report of the presidential commission on the space shuttle Challenger accident. table-of-contents.html. Pronovost, P., Berenholtz, S., Dorman, T., Lipsett, P. A., Simmonds, T., & Haraden, C. (2003a). Improving communication in the ICU using daily goals. Am J Crit Care, 18(2), Retrieved from PM: Pronovost, P., Needham, D., Berenholtz, S., Sinopoli, D., Chu, H., Cosgrove, S. et al. (2006). An intervention to decrease catheter-related bloodstream infections in the ICU. N.Engl.J.Med., 355(26), Retrieved from PM: Pronovost, P., & Sexton, B. (2005). Assessing safety culture: Guidelines and recommendations. Qual Saf Health Care, 14(4), Retrieved from ISI: Reference List

126 Pronovost, P. J., Weast, B., Holzmueller, C. G., Rosenstein, B. J., Kidwell, R. P., Haller, K. B. et al. (2003b). Evaluation of the culture of safety: Survey of clinicians and managers in an academic medical center. Qual Saf Health Care, 12(6), Retrieved from ISI: Pronovost, P., Weast, B., Rosenstein, B., Sexton, J. B., Holzmueller, C. G., Paine, L. et al. (2005). Implementing and Validating a Comprehensive Unit-Based Safety Program. J Patient Saf, 1(1). Retrieved from lww.com/journalpatientsafety/fulltext/2005/03000/implementing_and_validating_a_comprehensive.8.aspx Reader, T. W., Flin, R., Mearns, K., & Cuthbertson, B. H. (2007). Interdisciplinary communication in the intensive care unit. Br J Anaesth, 98(3), Retrieved from ISI: Reason, J. (2000). Human error: Models and management. Brit Med J, 320(7237), Retrieved from ISI: Reffay, C., & Chanier, T. (2003). How social network analysis can help to measure cohesion in collaborative distancelearning. In B. Wasson, S. Ludvigsen, & U. Hoppe (Eds.). Designing for change in networked learning environments. Proceedings of the International Conference on Computer Support for Collaborative Learning (pp ) Dordrect: Kluwer. Risser, D. T., Rice, M. M., Salisbury, M. L., Simon, R., Jay, G. D., & Berns, S. D. (1999). The potential for improved teamwork to reduce medical errors in the emergency department. The MedTeams Research Consortium. Ann Emerg Med, 34(3), Retrieved from PM: Rivers, R. M., Swain, D., & Nixon, W. R. (2003). Using aviation safety measures to enhance patient outcomes. AORN Journal, 77(1), Retrieved from PM: Rogers, B. (1997). Health hazards in nursing and health care: an overview. Am.J.Infect.Control, 25(3), Retrieved from PM: Saari, L. M., & Judge, T. A. (2004, WIN). Employee attitudes and job satisfaction. Hum Resource Manage, 43(4), Retrieved from ISI: Salancik, G. R., & Pfeffer, J. (1978). Social Information-Processing Approach to Job Attitudes and Task Design. Admin Sci Quart, 23(2), Retrieved from ISI:A1978FA Salas, E., Prince, C., Bowers, C. A., Stout, R. J., Oser, R. L., & Cannon-Powers, J. A. (1999). A methodology for enhancing crew resource management training. Hum Factors, 41(1), Retrieved from ISI: Schiøler T, Lipczak H, Pedersen BL, Mogensen TS, Bech KB, Stockmarr A et al. (2002). [Incidence of adverse events in hospitals. A retrospective study of medical records]. Ugeskr Laeger, 164(38), Schoorman, F. D., Mayer, R. C., & Davis, J. H. (2007). An integrative model of organizational trust: Past, present, and future. Acad Manage Rev, 32(2), Retrieved from ISI: Schwappach, D. L. B., & Boluarte, T. A. (2009). The emotional impact of medical error involvement on physicians: a call for leadership and organisational accountability. Swiss Med Wkly, 139(1-2), Retrieved from ISI: Scott, J., Tallia, A., Crosson, J. C., Orzano, A. J., Stroebel, C., Cicco-Bloom, B. et al. (2005). Social network analysis as an analytic tool for interaction patterns in primary care practices. Ann Fam Med, 3(5), Retrieved from PM: Scott, J. (1991). Social network analysis: A handbook. Thousand Oaks, CA: Sage. Scott, S. D., Hirschinger, L. E., Cox, K. R., McCoig, M., Brandt, J., & Hall, L. W. (2009). The natural history of recovery for the healthcare provider second victim after adverse patient events. Qual Saf Health Care, 18(5), Retrieved from ISI: Sepkowitz, K. A. (1996a). Occupationally acquired infections in health care workers. Part I. Ann.Intern.Med., 125(10), Retrieved from PM: Sepkowitz, K. A. (1996b). Occupationally acquired infections in health care workers. Part II. Ann.Intern.Med., 125(11), Retrieved from PM: Sexton, J. B., Helmreich, R. L., Neilands, T. B., Rowan, K., Vella, K., Boyden, J. et al. (2006a). The safety attitudes questionnaire: Psychometric properties, benchmarking data, and emerging research. BMC Health Serv Res, 6. Retrieved from ISI: Reference List 125

127 Sexton, J. B., Holzmueller, C. G., Pronovost, P. J., Thomas, E. J., McFerran, S., Nunes, J. et al. (2006b). Variation in caregiver perceptions of teamwork climate in labor and delivery units. J Perinatol, 26(8), Retrieved from ISI: Sexton, J. B., & Klinect, J. R. (2001). The link between safety attitudes and observed performance in flight operations. Proceedings of the Eleventh International Symposium on Aviation Psycholog (pp. 7-13) Columbus OH: The Ohio State University. Shapiro, M. J., Morey, J. C., Small, S. D., Langford, V., Kaylor, C. J., Jagminas, L. et al. (2004). Simulation based teamwork training for emergency department staff: does it improve clinical team performance when added to an existing didactic teamwork curriculum? Qual Saf Health Care, 13(6), Retrieved from ISI: Shendell-Falik, N., Feinson, M., & Mohr, B. J. (2007). Enhancing patient safety: improving the patient handoff process through appreciative inquiry. J Nurse Admin, 37(2), Retrieved from PM: Singer, S., Meterko, M., Baker, L., Gaba, D., Falwell, A., & Rosen, A. (2007). Workforce perceptions of hospital safety culture: Development and validation of the patient safety climate in healthcare organizations survey. Health Serv Res, 42(5), Retrieved from ISI: Singer, S. J., Falwell, A., Gaba, D. M., & Baker, L. C. (2008). Patient safety climate in US hospitals: variation by management level. Med Care, 46(11), Singer, S. J., Falwell, A., Gaba, D. M., Meterko, M., Rosen, A., Hartmann, C. W. et al. (2009a). Identifying organizational cultures that promote patient safety. Health Care Manage Rev, 34(4), Retrieved from ISI: Singer, S. J., Gaba, D. M., Falwell, A., Lin, S., Hayes, J., & Baker, L. (2009b). Patient safety climate in 92 US hospitals differences by work area and discipline. Med Care, 47(1), Retrieved from ISI: Singer, S. J., Gaba, D. M., Geppert, J. J., Sinaiko, A. D., Howard, S. K., & Park, K. C. (2003). The culture of safety: Results of an organization-wide survey in 15 California hospitals. Qual Saf Health Care, 12(2), Retrieved from ISI: Sjoberg, L. (1999). Consequences of perceived risk: demand for mitigation. J Risk Res, 2, Sjöberg, L. (2003). Neglecting the risks: The irrationality of health behavior and the quest for La Dolce Vita. Eur Psychol, 8(4), Retrieved from ISI: Sjoberg, L. (2003). The different dynamics of personal and general risk. Risk Manage. Sjoberg, L., & Drottz-Sjoberg, B. M. (2008). Attitudes towards nuclear waste and siting policy: experts and the public. In A.P. Lattefer (Ed.), Nuclear Waste Research: Siting, Technology, and Treatment (pp ). New York: Nova Science Publisher. Smedley, J., Palmer, C., Baird, J., & Barker, M. (2002). A survey of the delivery and uptake of influenza vaccine among health care workers. Occup Med (Lond), 52(5), Retrieved from ISI: Smith, R. (2003). Communicating risk: the main work of doctors. Br Med J, 327(7417). Solomon, R. L. (1980). The Opponent-Process Theory of Acquired Motivation - the Costs of Pleasure and the Benefits of Pain. Am Psychol, 35(8), Retrieved from ISI:A1980KD Soop, M., Fryksmark, U., Koster, M., & Haglund, B. (2009). The incidence of adverse events in Swedish hospitals: a retrospective medical record review study.int J Qual Health Care, 21(4), Retrieved from ISI: Sudman, S. (1985). Mail surveys of reluctant professionals. Eval Rev, 9(3), Retrieved from ISI:A1985AJZ Templeton, L., Deehan, A., Taylor, C., Drummond, C., & Strang, J. (1997). Surveying general practitioners: does a low response rate matter? Br J Gen Pract, 47(415), Retrieved from ISI:A1997WG Thomas, E. J., Sexton, J. B., & Helmreich, R. L. (2003). Discrepant attitudes about teamwork among critical care nurses and physicians. Lippincott Williams & Wilkins. Thomas, E. J., Sexton, J. B., Neilands, T. B., Frankel, A., & Helmreich, R. L. (2005). The effect of executive walk rounds on nurse safety climate attitudes: A randomized trial of clinical units. BMC Health Serv Res, 5. Retrieved from ISI: Reference List

128 Thomas, E. J., Studdert, D. M., Burstin, H. R., Orav, E. J., Zeena, T., Williams, E. J. et al. (2000). Incidence and types of adverse events and negligent care in Utah and Colorado. Med Care, 38(3), Retrieved from ISI: Tucker, A. L., & Edmondson, A. C. (2003). Why hospitals don t learn from failures: Organizational and psychological dynamics that inhibit system change University of California, Berkeley, Haas School of Business. Turner, B. A., & Pidgeon, N. F. (1997). Man-made disasters (second ed.). Boston: Butterworth-Heinemann Ltd. Tzeng, H. M. (2002). Satisfying nurses on job factors they care about - a Taiwanese perspective. J Nurs Adm, 32(6), Retrieved from ISI: VanGeest, J. B., Johnson, T. P., & Welch, V. L. (2007). Methodologies for improving response rates in surveys of physicians - a systematic review. Eval Health Prof, 30(4), Retrieved from ISI: Vaughan, D. (1997). The Challenger launch decision: Risky technology, culture, and deviance at NASA University of Chicago Press. Vincent, C. (2003). Understanding and responding to adverse events. New Engl J Med, 348(11), Retrieved from ISI: Vincent, C., Neale, G., & Woloshynowych, M. (2001). Adverse events in British hospitals: preliminary retrospective record review. Br Med J, 322(7285), Retrieved from ISI: Vogelsmeier, A., Scott-Cawiezell, J., Miller, B., & Griffith, S. (2010). Influencing Leadership Perceptions of Patient Safety Through Just Culture Training. J Nurs Care Qual, 25(4), 288. Walker, P. H., Baldwin, D., Fitzpatrick, J. J., Ryan, S., Bulger, R., DeBasio, N. et al. (1998). Building community: Developing skills for interprofessional health. Nurs Outlook, 46(2), Retrieved from ISI: Ward, J. (1994). General-practitioners experience of research. Fam Pract, 11(4), Retrieved from ISI:A1994QE Wassermann, S., & Faust, K. (1994). Social network analysis: methods and applications. New York: Cambridge University Press. Weinstein, N. D. (1987). Unrealistic optimism about susceptibility to health-problems - conclusions from a communitywide sample. J Behav Med, 10(5), Retrieved from ISI:A1987L Wicker, S., Rabenau, H. F., Doerr, H. W., & Allwinn, R. (2009). Influenza vaccination compliance among health care workers in a German university hospital. Infection, 37(3), Retrieved from ISI: Wilson, J. M., Straus, S. G., & McEvily, B. (2006). All in due time: The development of trust in computer-mediated and face-to-face teams. Organ Behav Hum Decis Process, 99(1), Retrieved from ISI: Wilson, R. M., Runciman, W. B., Gibberd, R. W., Harrison, B. T., Newby, L., & Hamilton, J. D. (1995). The Quality in Australian health care study. Med J Australia, 163(9), Retrieved from ISI:A1995TM Winsor, D. A. (1988). Communication failures contributing to the Challenger accident: an example for technical communicators. Professional Communication, IEEE Transactions on, 31(3), Retrieved from IEEE. Wolf, K. N. (1999). Allied health professionals and attitudes toward teamwork. J.Allied Health, 28(1), Retrieved from PM: Xiao, Y., Mackenzie, C. F., Patey, R., & LOTAS Group. (1998). Team coordination and breakdowns in a real-life stressful environment. Proceedings of the Human Factors and Ergonomics Society 42nd Annual Meeting (pp ). Zegers, M., de Bruijne, M. C., Wagner, C., Groenewegen, P. P., Waaijman, R., & van der Wal, G. (2007). Design of a retrospective patient record study on the occurrence of adverse events among patients in Dutch hospitals. BMC Health Services Research, 7. Retrieved from ISI: Zegers, M., de Bruijne, M. C., Wagner, C., Hoonhout, L. H. F., Waaijman, R., Smits, M. et al. (2009). Adverse events and potentially preventable deaths in Dutch hospitals: results of a retrospective patient record review study. Qual Saf Health Care, 18(4), 297-U79. Retrieved from ISI: Zohar, D., Livne, Y., Tenne-Gazit, O., Admi, H., & Donchin, Y. (2007). Healthcare climate: A framework for measuring and improving patient safety. Crit Care Med, 35(5), Retrieved from ISI: Reference List 127

129

130 Acknowledgement Having a passion for research is not enough to make one finish a doctorate thesis, let alone enjoying the journey of becoming an independent scientist. One should also be lucky enough to be surrounded with the right people in the right environment at the right time. I have been among the lucky ones. The visible product of my luckiness is definitely the thesis itself. But, let us not overlook other invisible products such as my enhanced writing skills, broadened scientific knowledge and maturity. If I had not met the following people, I may not have achieved all that and may not have acquired many valuable experiences and lessons. This section is dedicated to these people who have made me lucky to have them crossing my life path. I am thankful to have met Prof. dr. Addie Johnson who offered me the PhD position. Her phone call brightened my winter in Finally, I had the chance to start working on a research topic that had fascinated me. Under her supervision, my passion in research about patient safety has been polished into skills to do science. My love for personal writing has also been polished into skills to write scientific papers. She has taught me not to fear obstacles in reaching goals and how to be adaptive in dealing with many difficulties in research. She has especially taught me to always try to reach the excellence. I am thanking her for taking part in shaping the scientist in me. In the beginning of my PhD study, Addie introduced me to Dr. Stephen Peuchen who then helped me in building networks within the UMCG to do studies. Thank you for your willingness to dress up in the fascinating surgical gowns and accompany me to many meetings in the operation center. We should have taken pictures! Also, thank you for sharing your knowledge, and giving me encouragement in dealing with meetings in Dutch and explaining many important points after the meetings. The meetings have been fruitful because we were finally introduced to Dr. Raoul E. Nap who then became my Copromotor. Raoul has given me not only precious data, but also much knowledge in health care industry. He has taught me about the organizational structure of health care and how theories are put into practiced. Our conversations have been ones of my sources to create research ideas and to understand study results. I would also like to thank my research group members for their academic inputs, supports and insights: Dr. Candice Morey, Dr. Hedderik van Rijn, Dr. Jacob Jolij, Dr. Dina Burkolter, Tadeusz Kononowicz, Rasa Gulbinaite, Ari Widyanti, Dr. Paolo Toffanin, Jonathan Mall, Dr, Anastasios Sarampalis, and Edyta Sasin. Acknowledgement 129

131 Apart from academic helps, having a personal support system during your PhD time is crucial because your hypotheses of studies may not come true. You may end up with zero results and you can forget about getting beautiful data such as those published in high impact journals. In addition to that, the slow process of research may dominate your self-esteem and make you question yourself whether you are eligible as a scientist. There are times when you may feel that you have to run uphill with many kilos of weights to carry, with rain pouring your ways and there is no one there to give you an umbrella. In times like this, quitting the PhD was never the answer for me. Thinking low about me also never lasted too long. This is because I am blessed with many, many support systems. Because of them, I dared to make an umbrella from whatever was available while still carrying those weights up to the highest hill, and was still be able to smile and wholeheartedly enjoyed the journey. Cinthya Sopaheluwakan, Aree Witoelar, Imeza Saraswaty, Erwin Smit, Nadia Gozali, Johann Eddy Theodore, Shinta Tan, Ridwan Maulana and Christin are among those of my blessings. Thank you for loving me as if I were your own family. I really cannot ask for more. To Rasa Gulbinaite, Tadeusz Kononowicz, Edyta Sasin, Paolo Toffanin, Marleen Schippers, Jonathan Mall, Juanita Moreno, Elizabeth Teracino, Dejan Petrovic, Gaston Sendin, Primoz Pirih and Aysa Arylova. Thank you for showing me the artistic part of life. Each of you colors my days in a very different way. A common thing you all share is the fact that if I had not interacted with you, I may have not enjoyed the bitterness of life and laughed at my unfortunates so much. Thank you for transforming my dark sides into dark humors and my melancholic sides into violence. I still greed my teeth during sleeps, though. There is more aggression to be put up to surface, people! Although Groningen is beautiful, sometimes I needed to get a weekend getaway to Randstad, where I used to study. Here, I usually recharged my battery by making visits to all-you-can-eat sushi and cheap but delicious Chinese/Indonesian restaurants. What is it with great food without great companionships? Luckily, I always have both and there goes my thanks to Farabi Fakih, Tri Astraadmadja, Enira Ella, Louis and other Randstad inhabitants who are too many to be mentioned here. You make my food tastes much better! Living in a rather small country such as the Netherlands means that you will bump into the same people multiple times. Some of these people have occupied my daily life and left footprints on my heart, even after some of them left the Netherlands. Martin William, Mackenzie Hadi, Agnes Ardiyanti, Asteria Yasunari, Oktavia Hendrawati, Laura Junistia, Jerome Wirawan, Rosalina Wisastra, Dani Setiawan, Lydia Metasari, Stephanie Parasibu, Jardi Huzen, Tejas Gandhi and Diana Dimitrova, thank you for enlightening my dark days. Sri 130 Acknowledgement

132 Margana, Shiskha Prabawaningtyas, Nadia Koesnanto, Angie Theonis, Kristano Reinhart and Lola Devung, I absolutely miss your presence. Being far away from my immediate family has never been easy. Fortunately, we have never made the distance and time differences barriers in keeping our unconditional love a circle. There are not enough words to express how grateful I am to have been given such open and devoted family. To my dad in heaven, you are the farthest person from me (geographically), but you are the first and main person who had introduced me to the beauty of learning, so that, now, I am always hungry of learning things. I know that you would be proud of me. To my mother in Indonesia, you are always the first person to whom I want to call up to during my rainy and sunny days. Just because I know that you are always there for me, no matter what. You have taught me to become this strong woman who can enjoy whatever life is throwing onto her way. I am so proud to have you as my mom (Untuk mama di Indonesia, mama akan selalu jadi orang pertama yang aku akan telepon di saat aku sedang sedih dan gembira. Hanya karena aku tahu bahwa mama akan ada selalu untukku, untuk terus mendukungku, apapun keadaanku. Mama adalah orang yang mengajarkanku untuk menjadi wanita yang kuat, yang bisa menikmati hidup apapun yang hidup berikan, baik dan buruk. Aku selalu, selalu bangga akan mama sebagai mamaku.) To my older sister, Arivia Listyowardojo, in the United States. I have learned so much from you in both personal and professional contexts. Thank you for always trying to take the bullets first. To my younger sister in South Korea: Teresa Listyowardojo. Your spirits, passions, ambitions and ways of thinking are what have inspired me in many ways. You are a source of my insights. To my Anspach family in the United States, especially Brian, Ginny and Bob, thank you for taking care and loving me unconditionally. You make my heart full of joy every time I get off the plane in Philadelphia. Finally, it is worth acknowledging that I have learned many things from many different people during my PhD time. These learning experiences have enriched myself and helped me winning many battles. My sincere thank you naturally goes to every person who has come into my PhD life, even those who came only for a second, and given me a new perspective about life. Acknowledgement 131

133