Perceived adverse patient outcomes correlated to nurses workload in medical and surgical wards of selected hospitals in Kuwait

ISSUES AFFECTING PATIENT SAFETY AND WELL-BEING Perceived adverse patient outcomes correlated to nurses workload in medical and surgical wards of selected hospitals in Kuwait Fatimah Al-Kandari and Deepa Thomas Aim. This study was carried out to identify the perceived adverse patient outcomes as related to nurses workload. It also assessed nurses perception of variables contributing to the workload and adverse patient outcomes. Background. Several studies have been published on adverse patient outcomes in which a correlation was found between nurses workload and some adverse patient outcomes. Design. A cross-sectional survey was conducted between registered nurses (n = 780) working in medical and surgical wards of five general governmental hospitals in Kuwait. Data collection instruments. Data were collected using a self-administered questionnaire consisting of three sections to elicit information about the sample characteristics, perception of workload and perceived adverse patient outcomes during the last shift and last working week. Results. The three major perceived adverse outcomes reported by the nurses while on duty during their last shift were: complaints from patients and families (2%), patients received a late dose or missed a dose of medication (1Æ8%) and occurrences of pressure ulcer (1Æ5%). Similarly, the reported adverse outcomes over the past week were complaints from patients and families (5%), patients received a late dose or missed a dose of medication (5Æ3%) and discovery of a urinary tract infection (3Æ7%). Increases in nurse-patient load, bed occupancy rate, unstable patients condition, extra ordinary life support efforts and non-nursing tasks; all correlated positively with perceived adverse patient outcomes. Conclusion. This study sheds light on an important issue affecting patient safety and quality of care as perceived by the nurses themselves as caregivers. Relevance to clinical practice. Nurses perception of variables contributing to adverse patient outcomes and their workload could significantly affect the provided nursing care and nursing care recipients. The findings could help in policy formulation and planning strategies to decrease adverse patient outcomes in many countries with a health care structure similar to that of Kuwait. Key words: nurses, nursing, nurse-patient load, perceived adverse patient outcomes, workload Accepted for publication: 28 January 2008 Introduction Ensuring patient safety in the health care system is a major concern today. Without safety, there cannot be high quality of care. Improving patient safety is an ethical imperative for health care providers. Error prevention and safety promotion are part of everyone s work in the health care system. Patient safety is a priority of the World Health Organization and is on the policy agenda of many countries (WHO 2002). Authors: Al-Kandari Fatimah, DNS RN, Associate Professor, Dean, College of Nursing, Public Authority for Applied Education & Training (PAAET), Kuwait; Deepa Thomas, MSN RN, Nursing faculty, College of Nursing, Public Authority for Applied Education & Training (PAAET), Kuwait Correspondence: Fatimah Al-Khandari, Dean, College of Nursing, Public Authority for Applied Education & Training, PO Box 64923 A-Shuwaikh, Kuwait. Telephone: 00 965 4826798. E-mail: fatimah@paaet.edu.kw; f_alkandari85@hotmail.com Ó 2008 The Authors. Journal compilation Ó 2008 Blackwell Publishing Ltd, Journal of Clinical Nursing, 18, 581 590 581 doi: 10.1111/j.1365-2702.2008.02369.x

F Al-Kandari and D Thomas Adverse outcomes/events are a major challenge to the quality of care. These are unintended events caused by medical mismanagement rather than the underlying disease or condition of the patient (Kohn et al. 2000). Adverse events were placed between the fifth and eighth leading causes of death in the USA (Institute of Medicine 2000). A review of literature showed that most of the studies related to patient safety and adverse patient outcomes were conducted in European countries and the USA where the work environment is different from Kuwait and the Arabian Gulf region. The purpose of this study was to explore the situation in Kuwait with respect to nurses workload and adverse patient outcomes. The findings of this study will help in policy formulation and planning strategies to improve the nursing workforce with specific recommendations for measures to decrease adverse patient outcomes. Background Nurses are always available on the ward, monitoring patients conditions and delivering patient care. Many studies have found that nurses assume responsibility for first contact and ongoing care for all presenting patients (Laurant 2007). They are the key personnel in providing direct patient care. Anything that affects nurses work can directly affect the quality of nursing care and patient safety. Reports show that the most important issues affecting medical error rates are workload, stress or fatigue among health professionals, inadequate number of nurses and inadequate time spent with patients (National Survey on Consumers Experiences with Patient Safety and Quality Information, 2004). Hospitals with lower nurse staffing levels, nurses who spend less time with patients and practice settings with fewer registered nurses compared with licensed practical nurses or nurses aides tend to have higher rates of poor patient outcomes (Stanton & Rutherford 2004). Errors can occur in patient care settings at any time during a shift. The likelihood of error increases when wards are busy or the workload is heavy. Many health care professionals feel fatigued, stressed, overburdened, at risk and/or in pain and are not able to provide consistent quality care (Nicklin & Mc Veety 2002). Inadequate staffing results in adverse incidents and compromises patient safety (Beekmann et al. 1998). One study in Saudi Arabia reported that the average nurse to patient ratio in dialysis units was 1:5. A higher than average ratio of patients assigned to each nurse, language barriers and high turn over among nurses were challenges to the quality of care in Saudi Arabia (Souqiyyeh et al. 2001). Aiken et al. (2002) found that the risk of death following common surgical procedures increased by 7% for each patient added to the nurse s average workload and patient mortality was 30% higher in hospitals where nurses cared for an average of eight patients each than in those where nurses had an average load of 4 6 patients. Also failure to rescue rates for patients with complications were 30% higher in hospitals where nurses cared for eight or more patients (Aiken et al. 2002, Needleman et al. 2002). Yang (2003) studied the effects of nurse staffing variables on patient outcomes and found that nurses workload and daily average hours of nurses care were powerful predictors of adverse outcomes, such as patient falls, pressure sores, respiratory and urinary tract infections and patient/family complaints. In addition to patient load, studies have shown that nurses who are involved in non-nursing tasks such as clerical work, delivering and retrieving food trays and housekeeping and transporting patients have less time for tasks requiring their professional skills such as developing nursing care plans and client teaching (Aiken et al. 2002). Health care system in Kuwait The country is divided into five health regions with a general hospital in each region. Both the government and private health care system in Kuwait is monitored and regulated by the Ministry of Health (MOH). All the wards in the government hospitals are managed by registered nurses without any nurse aids or ward assistants. The Ministry of Health adopts three fixed shifts to meet the demand for nurses working round the clock a morning shift (7Æ00 AM 2Æ00 PM), an evening shift (2Æ00 PM 10Æ00 PM) and a night shift running from 10Æ00 PM 7Æ00 AM. Each nurse works six days and is entitled to one day off during the week. Nurses rotate between the morning, evening and night shifts during the month. There are three educational programmes in the country to train local registered nurses a nursing certificate, an associate degree in nursing and a Bachelor of Science degree programme. The minimum qualification to register as a nurse is a nursing certificate. The associate degree and the Bachelor of Science degree are similar to American and Canadian associate degrees and Bachelor of Science nursing degrees. In Kuwait, nationals represent only 9Æ4% of the nurse population (Ministry of Planning Statistics & Information Sector 2004). Nursing is not an attractive job for Kuwaiti nationals because of various social and cultural factors. Peer pressure, disapproval from family members and evening and night shift are some factors that have a negative impact on perceptions of nursing as a profession in Kuwait (Al-Kandari & Lew 2005). Therefore, foreign nurses are recruited from different Asian and African Arab countries. As the majority of the nurses are non-arabic speaking, sometimes language 582 Ó 2008 The Authors. Journal compilation Ó 2008 Blackwell Publishing Ltd, Journal of Clinical Nursing, 18, 581 590

Issues affecting patient safety and well-being problems exist between them and the Arabic speaking patients who do not speak English. The population of Kuwait was 2Æ48 million in 2003 and increased to 2Æ64 million in 2004. The annual population growth rate is 3Æ36%. A reduction in the number of beds per 1000 population from 1Æ9 in 2003 to 1Æ7 in 2004 has been noted. The bed occupancy rate increased to 68Æ7% in 2004. The nurse-population ratio was 3Æ8/1000 population in 2000, but decreased to 3Æ5 in 2004 (Ministry of Planning Statistics & Information Sector 2004), whereas higher ratios are reported from other countries such as UAE (4Æ18), Qatar (4Æ93), UK (4Æ97) and USA (7Æ73) (Al-Jarallah et al. 2000). The facilities and staffing are not increasing as fast as the population which could impose a burden of increased work on the shoulders of existing staff. Study aim: This study was conducted in an attempt to assess nurses workload and how it affects perceived adverse patient outcomes. Operational definitions Nurse-patient load The number of patients assigned to a nurse for complete nursing care in a shift. Nurses workload It is made up of nursing and non-nursing tasks/activities. Nursing tasks are the number of nursing tasks/activities carried out by a nurse during a shift. They include patient assessment, developing care plans and providing comprehensive nursing care. Non-nursing tasks/activities refers to the activities carried out by a nurse during a shift which do not require professional nursing skills and are not related to direct patient care. They include clerical work, house keeping, dietary services, coordinating ancillary services and transporting patients. Adverse patient outcome/event Any unexpected event that is perceived by the nurse during patient care caused by care processes (rather than the underlying condition) can and frequently does adversely affect the health of the patient. Research questions 1 What are nurses perceptions of their workload in the medical and surgical wards? 2 What are the adverse patient outcomes as perceived by the nurses in the medical and surgical wards? 3 What are the perceived variables contributing to adverse patient outcomes? 4 Is there a correlation between the perceived adverse patient outcomes and the nurses workload in the medical and surgical wards? Methods Design A cross-sectional survey was conducted in all the five regional hospitals in the State of Kuwait to collect information from registered nurses. The medical and surgical wards were chosen on the assumption that they were typical of the general wards in Kuwait hospitals. Data collection instrument A self-administered questionnaire was designed to collect data. It consists of three sections. Questions related to section 2 (workload) and 3 (adverse patient outcomes) were partially adopted from the study of the Nursing Personnel Survey Form which was developed by Dr Sochalski, Dr Aiken and colleagues for the International Hospital Outcomes Consortium and has been used in many studies in North Carolina and Pennsylvania. After getting permission from the authors, questions were modified and prepared at the level of eighth grade English language comprehension as the majority of the nurses are not native English speakers. It took 20 30 minutes for each nurse to complete the questionnaire. The three sections of the questionnaire are: 1 Section A: Contained questions relating to demographical data, place of work, specialty area, experience and variations in staffing of the wards for the past one year; 2 Section B: Related to perceptions of workload. Sixteen questions about the last shift worked by the nurse respondent, bed capacity of the unit, nurse-patient load, number of unstable patients assigned, emergencies and frequency of various nursing tasks/activities and nonnursing tasks were included; 3 Section C: Twelve questions relating to the perceived adverse patient outcomes were asked. Nurses were given a list of adverse events and were asked to indicate if any of the events had occurred with their patients and when they happened. Pilot study Nurses perception of adverse patient outcomes A pilot study was conducted first to evaluate nurses ability to understand the language and to test the reliability and Ó 2008 The Authors. Journal compilation Ó 2008 Blackwell Publishing Ltd, Journal of Clinical Nursing, 18, 581 590 583

F Al-Kandari and D Thomas content validity of the tool. The questionnaire was distributed to 54 nurses, 24 from a medical ward and 30 from a surgical ward from one general hospital. Both wards were later exempted from the actual study. Nurses were requested to give any additional information and any other nursing or non-nursing tasks performed which were not listed on the questionnaire. Some modifications were made to the questionnaire to ensure clarity of some words and the questionnaire was re-tested with the same groups after four weeks. A test-retest stability of 0Æ78 was recorded. Participants All registered nurses working in medical and surgical wards of the five general hospitals of MOH, Kuwait participated in the study. Nurses who were on leave at the time of the study were exempted. Nurse managers were also excluded as they are not always involved in direct patient care. Eight hundred and twenty copies of the questionnaire were distributed among nurses and 780 (95%) were returned fit for analysis. Data collection method Research assistants distributed the questionnaires to nurses working on each selected ward during a regular working day according to number of nurses working each shift present at the time of the study. Four working days were given to each nurse to complete and return the questionnaire. A copy of the questionnaire was left for nurses who were not on duty. Participants were instructed to place the completed tool in a sealed envelope in a special collection box. Follow up to collect the questionnaires was made by the same research assistants. Direct approach to nurses by the research assistants and the follow up could be reasons for the high return rate. Ethical considerations The proposal was reviewed by the Research Committee of the Public Authority for Applied Education and Training and permission was obtained from the MOH. Data were collected anonymously and the participants were informed and reassured about voluntary nature of participation and the confidentiality of the information disclosed. A cover letter included instructions for completing the questionnaire. The letter also explained that submitting a completed questionnaire would be considered as consent to participate in the study. Data analysis Collected data were analysed using the Statistical Package for Social Sciences, version 11 (SPSS-11) (SPSS Inc., Chicago, IL, USA). Missing values in the demographical data were replaced with dummy variables. Series mean was used for numbers and mode was used for string values. Both descriptive and inferential analyses were used. Nurses workload was calculated by obtaining the mean frequency of each nursing and non-nursing task performed and adding them together to find out on average how many times that procedure was performed by a nurse. The tasks were categorised as nursing tasks and non-nursing tasks depending on the nature of the task. The mean frequency of each adverse outcome was calculated to find out, on average, how many events were perceived by the nurses. These are reported as those that nurses perceived to have occurred during their previous shift and during the past week. Bed capacity of the unit, the nurse-patient load per shift, factors contributing to adverse outcomes and the nurses workload were crosstabulated and compared using t-tests and ANOVA. In addition, the workload was compared between the medical and surgical wards and t-test was performed. Pearson correlation was used with nursing tasks, non-nursing tasks and total workload to assess their relationship to perceived adverse outcomes. Two-tailed p-values at 0Æ05 or less were considered as a significant correlation. Workload and adverse outcomes in different hospitals were compared by ANOVA. Findings Sample characteristics Seven hundred and eighty nurses participated in the study of whom 417 (53Æ4%) were working in medical wards and 343 (44%) were in surgical wards. The remaining 20 (2Æ6%) were floaters who worked in other areas of specialisation and were asked to help in a medical or surgical ward for that shift to overcome the staff shortage. The demographical data are shown in Table 1. 72Æ7% of the sample were females while 27Æ3% were males. Almost 50% of the nurses belonged to the 21 30 years age group and 36Æ5% were in 31 40 years age group. The median age of the nurses was 29Æ9 years. The majority had an associate degree or equivalent (61Æ7%) in nursing education. Others had either a BSc degree (33Æ8%) or a nursing certificate (4Æ5%). Nurses mean years of experience was 5Æ9 years in Kuwait and 5Æ1 years in their current unit. 584 Ó 2008 The Authors. Journal compilation Ó 2008 Blackwell Publishing Ltd, Journal of Clinical Nursing, 18, 581 590

Issues affecting patient safety and well-being Table 1 Demographical data Variables n (780) % Gender Male 213 27Æ3 Female 567 72Æ7 Area of Work Medical 417 53Æ5 Surgical 343 44 Floaters 20 2Æ6 Nationality Kuwaiti 27 3Æ5 Arab 73 9Æ4 Indian 436 55Æ9 Filipino 129 16Æ5 Indonesian 83 10Æ6 Others 32 4Æ1 Professional qualification BSN 264 33Æ8 ADN 481 61Æ7 Nursing Certificate 35 4Æ5 Nurses perception of adverse patient outcomes significant difference in the number of nurses working in the two areas (p =0Æ002) was during the evening shift. The nurse-patient load in the medical wards during the morning, evening and night shifts was 4Æ3, 5Æ1 and 6Æ2 respectively; while in the surgical wards it was 3Æ6, 5Æ1 and 5Æ4 (Table 2). A statistically significant difference was noted in the nursepatient loads during night shifts between the medical and surgical wards (p =0Æ001). The workload according to different hospitals is shown in Table 2. The mean frequency of the nursing activities carried out by a nurse during a shift in the medical wards was 48Æ1 (SD = 38Æ5) and 41Æ7 (SD = 33Æ8) in the surgical wards. Mean frequency of non-nursing activities was 11Æ8 (SD = 14Æ2) and 11Æ7 (SD = 16Æ7) respectively in the medical and surgical wards. The non-nursing task with the highest mean frequency was clerical tasks for both areas. The t-test showed a statistical significant difference only in the nursing tasks (p = 0Æ02) between the medical and surgical wards. Nurses workload The bed capacity of the wards varied from 26 30 with a bed occupancy rate of 85%. The average number of patients in the medical wards during the studied shifts was 24Æ2 (SD = 7Æ4) and 21Æ2 (SD = 7) in the surgical wards. The t-test showed a statistically significant difference (p = 0Æ001) between the number of patients in the medical and surgical wards. The average number of nurses working in the medical wards was 6Æ8 (SD = 3Æ4) during the morning shift, 5Æ1 (SD = 3Æ0) during the evening shift and 3Æ9 (SD = 1Æ6) during the night shift. Similarly for the surgical wards the average number of nurses was 7Æ0 (SD = 3Æ8), 4Æ0 (SD = 1Æ4) and 3Æ6 (SD = 1Æ6) during the morning, evening and night shifts respectively. The t-test showed that the only statistical Perceived adverse patient outcomes The five major perceived adverse events by the nurses while on duty during their last shift were complaints from patients and families (2%), patient received a late dose or missed a dose of medication (1Æ8%) discovering pressure ulcer (1Æ5%), wound infection (1Æ2%) and infection at the site of IV cannula during the shift (1Æ2%) (Table 3). Perceived adverse outcomes reported over the past week (Table 3) were complaints from patients and families (5%), patient received late dose or missed a dose of medication (5Æ3%), discovering urinary tract infection (3Æ7%), wound infection (3Æ3%) and infection at the site of IV cannula (2Æ8%) during the shift. Table 3 shows the comparison of perceived adverse patient events between the medical and surgical wards during the last Table 2 Comparison of workload between hospitals Workload Hospital 1 Hospital 2 Hospital 3 Hospital 4 Hospital 5 F-value p-value Patients in the unit Mean 25Æ6 22Æ9 23Æ6 19Æ8 21Æ7 12Æ6 0Æ001 95% CI Lower 24Æ2 21Æ8 22Æ5 18Æ7 20Æ6 Upper 27Æ0 23Æ9 24Æ7 21Æ0 22Æ7 Nurses in the unit Mean 4Æ9 5Æ8 5Æ8 5Æ4 4Æ3 7Æ2 0Æ001 95% CI Lower 4Æ5 5Æ4 5Æ3 4Æ8 3Æ9 Upper 5Æ2 6Æ3 6Æ4 5Æ9 4Æ6 Patients assigned to a nurse Mean 5Æ4 4Æ7 4Æ9 4Æ0 5Æ5 6Æ1 0Æ001 95% CI Lower 5Æ1 4Æ5 3Æ9 3Æ7 5Æ1 Upper 5Æ7 4Æ9 5Æ8 4Æ1 5Æ9 CI, confidence interval. Ó 2008 The Authors. Journal compilation Ó 2008 Blackwell Publishing Ltd, Journal of Clinical Nursing, 18, 581 590 585

F Al-Kandari and D Thomas Table 3 Comparison of perceived adverse patient outcomes* between medical and surgical wards during the last shift and over the past week Total sample (n = 780 ) Medical (n = 417) Surgical (n = 343) Adverse outcomes Last shift Past week Last shift Past week Last shift Past week Complaints from patients and families 16 (2%) 39 (5%) 14 (3Æ4%) 30 (7Æ2%) 2 (0Æ6%) 7 (2%) Patient received late dose/missed a dose of medication 14 (1Æ8%) 41 (5Æ3%) 9 (2Æ2%) 22 (5Æ3%) 5 (1Æ5%) 19 (5Æ5%) Pressure ulcer discovered during the shift 12 (1Æ5%) 17 (2Æ2%) 8 (1Æ9%) 9 (2Æ2%) 4 (1Æ2%) 8 (2Æ3%) Wound infection discovered during the shift 9 (1Æ2%) 26 (3Æ3%) 4 (0Æ96%) 15 (3Æ6%) 5 (1Æ5%) 11 (3Æ2%) Infection at the site of IV cannula discovered during shift 9 (1Æ2%) 22 (2Æ8%) 5 (1Æ2%) 14 (3Æ4%) 4 (1Æ2%) 8 (2Æ3%) Urinary tract infection discovered during the shift 8 (1%) 29 (3Æ7%) 5 (1Æ2%) 19 (4Æ6%) 3 (0Æ87%) 10 (2Æ9%) Patient fall with injuries 6 (0Æ8%) 4 (0Æ5%) 1 (0Æ24%) 2 (0Æ48%) 5 (1Æ5%) 2 (0Æ6%) Errors in IV fluid administration 5 (0Æ6%) 6 (8%) 1 (0Æ24%) 4 (0Æ96%) 4 (1Æ2%) 1 (0Æ3%) Errors in sending laboratory specimen 1 (0Æ13%) 23 (2Æ9%) 1 (0Æ24%) 12 (2Æ9%) 0 10 (2Æ9%) Patient death 1 (0Æ13%) 2 (0Æ26%) 1 (0Æ24%) 1 (0Æ24%) 0 1 (0Æ3%) Deterioration of patient condition 0 5 (0Æ6%) 0 (0Æ24%) 5 (1Æ2%) 0 0 IV, intravenous. *Multiple responses were allowed; floaters are also included. shift and last week. The three most frequently perceived adverse outcomes reported by the medical nurses were complaints from patients and families, the patient received a late dose or missed a dose of medication and urinary tract infection. In comparison, the surgical nurses reported patient received late dose or missed a dose of medication, wound infection and urinary tract infections (Table 3). The t-tests showed significant differences in perceived adverse events of complaints from patients and families (p =0Æ001), patient received a late dose or missed a dose of medication (p = 0.05) discovering pressure ulcer (p = 0Æ001), infection at the site of IV cannula (p =0Æ02) and urinary tract infection (p =0Æ04) during the shift between the medical and surgical wards. Variables contributing to adverse patient outcomes Pearson correlations were used to determine the relationship between the reported perceived adverse outcomes and the studied variables. Significant positive correlations were found between the adverse patient events and nurse-patient load, bed occupancy of the unit, number of unstable patients in the unit and events of extraordinary life support efforts (Table 4). Perceived occurrences of adverse outcomes were least reported by associate degree nurses while they were the most reported by certificate nurses. No correlations were noted between the perceived adverse outcomes and other demographical variables such as nurses nationality and whether the wards were for male or female patients. Nurse-patient load A positive correlation was found between the number of patients assigned to a nurse and the reported adverse patient outcome of a patient fall with injury (r =0Æ1, p =0Æ01) and pressure ulcer (r = 0Æ08, p = 0Æ04). The average nurse-patient load in the medical wards was significantly more compared Table 4 Variables contributing to adverse patient outcomes Variable Adverse outcome r p (two-tailed) Patient load Patient fall with injuries 0Æ104* 0Æ005 Pressure ulcer discovered during the shift 0Æ075* 0Æ044 Bed occupancy Patient fall with injuries 0Æ096* 0Æ009 Pressure ulcer discovered during the shift 0Æ129** 0Æ001 Complaints from patient and families 0Æ081* 0Æ027 Urinary tract infection discovered during the shift 0Æ091* 0Æ014 Number of unstable patients in the unit Complaints from patient and families 0Æ08* 0Æ03 Extra ordinary life support Late/missed dose of medication 0Æ102* 0Æ05 Infection at the site of IV cannula discovered during the shift 0Æ073* 0Æ048 IV, intravenous. *Significant p-value 0Æ05. **Significant p-value 0Æ001. 586 Ó 2008 The Authors. Journal compilation Ó 2008 Blackwell Publishing Ltd, Journal of Clinical Nursing, 18, 581 590

Issues affecting patient safety and well-being Nurses perception of adverse patient outcomes with the nurse-patient load in the surgical wards. ANOVA showed a significant difference between the medical and surgical wards in perceived adverse events of pressure ulcer and complaints from patients and families. Bed occupancy A significant correlation was found between bed occupancy rate and reported adverse events of patient fall with injury (r =0Æ1, p =0Æ01), pressure ulcer (r =0Æ13, p =0Æ001), complaints from patients and families (r =0Æ08, p =0Æ03) and urinary tract infection (r = 0Æ09, p = 0Æ01). Number of unstable patients assigned/in the unit Pearson correlations showed statistically significant associations between the number of unstable patients in the unit and the perceived adverse event of complaints from patients and families (r =0Æ08, p =0Æ03). Events of extra ordinary life support efforts Events of extra ordinary life support efforts were associated with late or missed dose of medications (r =0Æ102, p =0Æ005) and infection at the site of IV cannula (r =0Æ07, p =0Æ04). Correlation between nurses workload and perceived adverse patient outcomes Correlation of nurses workload to perceived adverse patient outcomes is reported in Table 5. Increase in the nurses workload positively correlated with three perceived adverse patient outcomes late/missed dose of medications (r = 0Æ3, p = 0Æ001), errors in sending laboratory specimens (r = 0Æ15, p =0Æ001) and complaints from patients and families (r = 0Æ14, p = 0Æ001). Non-nursing activities, especially delivering and retrieving food trays, patient transport and clerical work also correlated significantly with perceived adverse outcomes of errors in sending lab specimen (r =0Æ35, p =0Æ001), patient fall with injury (r =0Æ3, p =0Æ001), errors in IV fluid administration (r =0Æ26, p =0Æ001) and late/ missed dose of medication (r = 0Æ14, p = 0Æ001). However, nursing activities did not correlate with any of the perceived adverse patient events. Similarly medical wards with higher workload than surgical wards reported higher numbers of perceived adverse outcomes. This will be discussed further with comparison between medical and surgical wards. Comparison between hospitals The mean number of patients in a ward varied between the five hospitals and ranged from 19Æ8 25Æ6 (Table 2). The mean number of nurses who worked during the last shift in a single unit of each hospital ranged from 4Æ3 5Æ8 (Table 2). The mean nurse-patient load varied from 1:4Æ0 1:5Æ5 (Table 2). ANOVAs showed that the differences in the mean numbers of in-patients, the number of nurses who worked in the unit and the nurse-patient load were statistically significant (Table 2) between the hospitals. Analysis of workload revealed that the highest nurse-patient load and nurses workload were in Hospital 5 (Table 2). When adverse patient events were compared, the highest number was also reported by Hospital 5 where the nursepatient load was the highest. ANOVA showed statistically significant differences in the perceived adverse patient outcomes of discovering pressure ulcer [F (4,775) =5Æ7, p =0Æ001], wound infection [F (4,775) =4Æ6, p =0Æ001], infection at the site of IV cannula [F (4,775) =4Æ6, p =0Æ001] and urinary tract infection during the shift [F (4,775) =4Æ4, p =0Æ002]. Comparison between medical and surgical wards The workload and adverse outcomes were compared between medical and surgical wards (Table 3). Average nurse-patient load in the medical wards was significantly more (p = 0Æ0098) compared with the nurse-patient load in the surgical wards. Medical nurses reported more nursing workload than Table 5 Correlation of nurses workload to adverse patient outcomes Workload Adverse outcome r p Increase in nurses workload Late/missed dose of medicine 5Æ3 0Æ022 Errors in sending laboratory specimen 7Æ16 0Æ008 Complaints from patient and families 5Æ95 0Æ015 Non-nursing workload Errors in sending laboratory specimen 5Æ73 0Æ017 Delivering and retrieving food trays Patient fall with injuries 4Æ59 0Æ032 Transporting of patients Errors in IV fluid administration 4Æ59 0Æ033 Clerical work Late/missed dose of medication 7Æ97 0Æ005 IV, intravenous. *Significant p-value 0Æ05. **Significant p-value 0Æ001. Ó 2008 The Authors. Journal compilation Ó 2008 Blackwell Publishing Ltd, Journal of Clinical Nursing, 18, 581 590 587

F Al-Kandari and D Thomas surgical nurses (p = 0Æ02). However, there was no significant difference in the nurses total workload (p = 0Æ142) and nonnursing workload (p =0Æ762) between the medical and surgical wards. Nevertheless, ANOVA showed a significant difference between the medical and surgical wards in the perceived occurrence of pressure ulcer (F = 2Æ72, p = 0Æ001) and complaints from patients and families (F =4Æ75, p = 0Æ001) as these adverse patient outcomes occurred more in the medical wards compared with surgical wards. Discussion Through this survey, the researchers attempted to explore nurses perception of adverse patient events as related to nurses workload and to identify factors contributing to the occurrence of adverse patient outcomes because of their impact on patient safety and quality of nursing care. Although the literature shows that several studies have been conducted on nurses workload and adverse patient outcomes, only a few studies have correlated the two variables together in the Middle East region and none in the State of Kuwait. Also, this study was conducted in different hospital settings with different cultural backgrounds and multicultural, multinational patients as well as nurse populations. As noted from the demographical data, the majority of nurses were in the age group of 21 40 and most of them had more than five years of experience in Kuwait and in their current unit. Thus, the nurses generally had enough work experience to deal efficiently with the demands of patient care in Kuwait hospitals. The findings revealed an agreement between the medical as well as the surgical nurses about the perceived adverse events. Complaints from patients and families and patient received late/missed dose of a medication were the first two adverse patient events perceived by the total sample and they were positively correlated to the nurses workload in Kuwait hospitals. The third and fourth ranked adverse events for the medical wards were urinary tract infection and pressure ulcers, while in the surgical wards wound infection and urinary tract infection were the third and fourth ranked. Patients age, general health conditions and medical diagnoses on the medical and surgical wards could be intervening factors influencing the third and fourth perceived events. The majority of patients in the medical wards are usually older and bedridden and in the surgical wards they were postoperative patients. Complaints from patients and families were that the patient received a late/missed dose of a medication and errors in sending laboratory specimen were three major adverse patient outcomes perceived to be associated with an increase in nurses workload. This is a similar finding to that cited in the literature (Stanton & Rutherford 2004, Tibby et al. 2004). Complaints from patients and families were greater when nurses were assigned patients who were more unstable, which further supports the above finding. In addition, perceived adverse patient outcomes were greater in the medical wards where the nurse-patient ratio was slightly higher. The reviewed literature also confirmed that the quality of patient care as well as adverse patient outcomes are directly affected by the nurses workload in the unit (Aiken et al. 2002, Cho et al. 2003). Late/missed dose of a medication was the most common form of perceived drug error by the nurses in this study. Patient falls, development of pressure ulcers and wound infections also were perceived to increase when the workload increased. Although patient acuity factors could influence the occurrence of these adverse outcomes, the reviewed literature reveals that previous studies have also reported a correlation between increased workload and adverse outcomes of patient falls and medication errors (Hickam et al. 2003). In the current study, the average nurse-patient load was 1:4Æ8, which is relatively low compared with other studies (Aiken et al. 2002, Lacey & Shaver 2002), yet nurses perceived an increase in their workload. Also a correlation was noted between extraordinary life support efforts which is directly related to patient acuity and late or missed dose of medication and infection at the site of cannula. Unfortunately, no official database is available in Kuwait in relation to patient acuity for benchmarking the average patient load according to patients conditions. The nursing division at the Ministry of Health uses a ratio of 0Æ7 nurses to one patient for the medical wards and 0Æ8 nurses to one patient for the surgical wards when staffing the wards. The health condition of the majority of patients in the medical and surgical wards was stable, but there were a few unstable patients in each ward who required close monitoring. These unstable patients were assigned to the nurses along with the stable ones. The high bed occupancy rate associated with performing nonnursing tasks in addition to the nurse-patient load explains the nurses perception of increase in their workload. Furthermore, the nurses had identified some adverse events being associated with an increase in workload which was more observable in the medical wards and in hospital 5 which had the highest workload; especially the perceived events of pressure ulcer, wound infection, infection at the IV site and urinary tract infection. Non-nursing tasks contributed heavily to nurses workload in Kuwait. The non-nursing task with the highest mean frequency was clerical tasks for both medical and surgical areas. Nurses involvement in non-nursing tasks limits their 588 Ó 2008 The Authors. Journal compilation Ó 2008 Blackwell Publishing Ltd, Journal of Clinical Nursing, 18, 581 590

Issues affecting patient safety and well-being time during the shift for direct patient care, adds to their workload and consequently contributes to nurses perception of adverse patient outcomes. Nurses had perceived four adverse patient events to be associated with the non-nursing tasks (Table 5). This finding is consistent with Aiken et al. (2002), where nurses were found to engage in non-nursing tasks such as clerical works, delivering and retrieving of food trays, transporting patients and housekeeping which decreased their time for tasks requiring their professional skills such as developing nursing care plans and client teaching. Nurses perception of adverse patient outcomes two levels of nursing education. Patient acuity, hours of care per patient day, nurse-patient load and acceptable average workload per shift are other aspects to be studied especially that patient acuity as a major factor influencing nurses workload and patient outcomes was not taken into consideration in this survey. Workload was only evaluated according to the existing ratio system in the country. This could be another limitation, yet the researchers focus was on perceived adverse patient outcomes as related to workload regardless of the workload measurement system. Relevance to clinical practice The clinical significance of this study is that nurses themselves perceive some adverse patient outcomes related to their workload. Nurses should explore the issue of adverse patient outcomes further and assess the contributing factors to the occurrence of adverse events and try to avoid or minimise further occurrence of such factors or events. Efforts should be exerted to use different approaches to enhance patient health and promote wellness. As the findings showed a positive correlation between the workload and adverse patient outcomes reducing the workload of nurses plays a vital role in promoting patient safety and quality of nursing care. Adequate numbers of clerical and other supportive staff are recommended to carry out non-nursing workload tasks. In light of the nursing shortage faced by many countries regionally and globally, introducing other levels of nursing personnel such as licensed practical nurses and assistant nurses is another option to be explored further to allow registered nurses more time to provide professional nursing care. Limitations and future recommendations Future research should take into consideration the limitations of the current study. One limitation was that workload and adverse patient outcomes were studied only as self-reported by the nurses. The researchers used a large sample size, yet the possibility of non-sampling errors cannot be ignored. Therefore, the reported perceived adverse patient outcomes do not reflect the actual hospital statistics or quantify the frequency of their occurrence. Nevertheless, the study helped to explore an area of investigation which has not been touched before in Kuwait. Review of patient files and hospital documents such as incident reports could provide further information on the actual number of adverse events reported. Recommendations for further studies include examining the influence of nurses level of educational preparation on perception of adverse patient outcomes, especially, that findings of this study revealed a significant difference between Conclusion In conclusion, this study addressed an important issue affecting patient safety and quality of care. It also shed light on an area that very often goes unexplored or is not openly discussed. Nurses play a vital role in the health care system. The cognitive aspect of nursing care should be taken into consideration when staffing units with adequate numbers of nurses. Employing supportive personnel could provide more time for nurses to carry out nursing tasks, especially with the universal nursing shortage threatening many health care sectors in different countries. It is also vital to monitor the common adverse outcomes where, why and how they happen and to critically evaluate errors such as medication errors which cannot be overlooked. Finally, recommendations for nurse managers are obvious in terms of setting quality care indicators to reduce adverse outcomes. Acknowledgements The researchers give due acknowledgement to Dr Julie Sochalski for giving permission to use and modify portions of the tool, as deemed necessary for the study. The tool was originally developed by Dr Sochalski, Dr Aiken and their colleagues. This study was funded by The Public Authority for Applied Education and Training (2005). Contributions Study design: F A-K and D T; data collection and analysis: F A-K and D T; manuscript preparation: F A-K and D T. References Aiken LH, Clarke SP, Sloane DM, Sochalski J & Silber JH (2002) Hospital nurse staffing and patient mortality, nurse burn out and job dissatisfaction. Journal of the American Medical Association 288, 1987 1993. Ó 2008 The Authors. Journal compilation Ó 2008 Blackwell Publishing Ltd, Journal of Clinical Nursing, 18, 581 590 589

F Al-Kandari and D Thomas Al-Jarallah KF, Moussa MAA, Al-Duwairy Y, Zaatar E, Hakeem SK, Al-Khanfar KF & Mathews AT (2000) Health Vision 2020 Workforce Needs of Health Professionals in Kuwait. Kuwait Institute of Medical Specialization, Kuwait. Al-Kandari FH & Lew I (2005) Kuwaiti high school students perception of nursing as a profession: implications for nursing education and practice. The Journal of Nursing Education 44, 533 540. Beekmann U, Baldwin M, Durie M & Morrison LS (1998) Problems associated with nursing staff shortage: an analysis of the first 3600 incident reports submitted to the Australian incident monitoring study. Anaesthesia and Intensive Care 26, 396 400. Cho SH, Ketefian S, Barkauskas VH & Smith DG (2003) The effects of nurse staffing on adverse events, morbidity, mortality and medical costs. Nursing Research 52, 71 79. Hickam DH, Severance S, Feldstein A, Ray L, Gorman P, Schuldheis S, Hersh WR, Krages KP & Helfand M (2003) The Effect of Health Care Working Conditions on Patient Safety. Available at: http://www.ncbi.nlm.nih.gov/books/bv.fcgi?rid=hstat1a.chapter. 9710 (accessed 31 July 2006). Institute of Medicine (2000) In To Err is Human: Building a Safer Health System. Chapter 2, Errors in health care: a leading cause of death and injury (Kohn LT, Corrigan JM & Donaldson MS eds). National Academy Press, Washington DC, USA, pp. 26 48. Kohn LT, Corrigan JM & Donaldson MS (2000) To Err is Human: Building a Safer Health System. National Academy Press, Washington DC, USA. Lacey LM & Shaver K (2002) Staff Nurse Satisfaction, Patient Loads and Short Staffing Effects in North Carolina. Findings from the 2001 Survey of Staff Nurses in North Carolina. Available at: http:// www.nursenc.org/research/staff_demos_work.pdf (accessed 14 October 2004). Laurant M (2007) Substitution of Doctors by Nurses in Primary Care. Available at: http://www.medscape.com/viewarticle/511405 (accessed 20 September 2007). Ministry of Planning (2004) Annual Statistical Abstract. Statistics & Information Sector, Ministry of Planning, State of Kuwait. National Survey on Consumers Experiences with Patient Safety and Quality Information (2004). Available at: http://www.kff.org/ kaiserpolls/pomr111704pkg.cfm (accessed 8 October 2006). Needleman J, Buerhaus P, Mattke S, Stewart M & Zelevinski K (2002) Nurse staffing levels and the quality of care in hospitals. The New England Journal of Medicine 346, 1715 1722. Nicklin W & Mc Veety JE (2002) Canadian nurses perceptions on patient safety in hospitals. Canadian Journal of Nursing Leadership 15, 11 21. Souqiyyeh MZ, Al-Attar MBA, Zakariah H & Shaheen FAM (2001) Dialysis centers in the Kingdom of Saudi Arabia. Saudi Journal of Kidney Diseases and Transplantation 12, 293 304. Stanton MW & Rutherford MK (2004) Hospital Nurse Staffing and Quality of Care. Research in Action. Available at: http:// www.ahrq.gov/research/nursestaffing/nursestaft.pdf (accessed 31 July 2006). Tibby SM, Correa-West J, Durward A, Ferguson L & Murdoch LA (2004) Adverse events in a paediatric intensive care unit: relationship to workload, skill mix and staff supervision. Intensive Care Medicine 30, 1160 1166. WHO (2002) Patient Safety: Rapid Assessment Methods for Estimating Hazards. Report of the WHO Working group meeting, Geneva, 17 19 December 2002. Yang KP (2003) Relationship between nurse staffing and patient outcomes. The Journal of Nursing Research 11, 149 158. 590 Ó 2008 The Authors. Journal compilation Ó 2008 Blackwell Publishing Ltd, Journal of Clinical Nursing, 18, 581 590