Exploring the Association Between Nurse Workload and Nurse-Sensitive Patient Safety Outcome Indicators

The Journal of Nursing Research h VOL. 20, NO. 4, DECEMBER 2012 Exploring the Association Between Nurse Workload and Nurse-Sensitive Patient Safety Outcome Indicators Li-Fang Liu 1 & Sheuan Lee 2 & Pei-Fang Chia 3 & Shu-Ching Chi 4 Yu-Chun Yin 5 * 1 MHA, RN, Advisor, Department of Administration, Kuang Tein General Hospital & 2 PhD, RN, Member, Examination Yuan, Taiwan, ROC, and Professor, College of Nursing, Chung-Shan Medical University & 3 MHA, RN, Vice Superintendant, Administration and Department of Nursing, Pingtung Christian Hospital & 4 MSN, RN, Director, Department of Nursing, E-Da Hospital & 5 MSN, RN, Course Associate Professor, Min-HweiCollegeofHealthCareManagement. ABSTRACT Background: Nurses affect patient safety. Although studies have associated patient safety with nurse staffing levels, Taiwan s Department of Health does not yet support changing nurse workforce standards for medical institutions. Purpose: This study was designed to gain insight into the workload of nurses employed at medical institutions and to determine the relationship between nurse workload and nurse-sensitive patient safety outcome indicators. Methods: This study adopted a cross-sectional quantitative method and collected data using a self-designed logbook. The study population comprised nurses from acute medical institutions, including medical centers and regional and district hospitals. One thousand five hundred logbooks were distributed to participants selected by random sampling from 21 city/county nursing associations across Taiwan. One thousand three hundred seventy-three questionnaires were retrieved; the 1,358 valid responses yielded a valid response rate of 90.5%. Nurses used the logbook to record individual working conditions for 2 weeks. Descriptive statistics included mean values, standard deviations, and percentages; inferential statistics included the Spearman rho correlation and odds ratios. Results: Nurse overtime working hours were positively associated with the following nurse-sensitive patient safety outcome indicators: patient falls, decubitus/pressure ulcers, near errors in medication, medication errors, unplanned extubation, hospitalacquired pneumonia, and hospital-acquired urinary tract infections; risks of patient falls, decubitus/pressure ulcers, unplanned extubation, hospital-acquired pneumonia, and hospital-acquired urinary tract infections significantly increased when the patient nurse ratio exceeded 7:1. Conclusion: Nurse workforce and nurse-sensitive patient outcome indicators are positively correlated. The results of this study will help professional nursing groups define suitable nursing workforce standards for medical institutions. KEY WORDS: nursing workload, patient safety, nurse-sensitive patient outcome indicators. Introduction Medical malpractice cases in Taiwan, such as a 2002 injection error incident at Bei Cheng Hospital and medication error incident at Chung Ai Hospital, highlight patient safety issues in this country (Taiwan Joint Commission on Hospital Accreditation [TJCHA], 2010). Patient safety is defined as the absence of preventable harm to a patient during the process of healthcare. Patient safety represents a coordinated effort to prevent harm attributable to the healthcare process (World Health Organization, 2010). Dr. Jesper Poulsen was estimated by the studies of different countries that the number of deaths caused by the unfavorable medical events in Taiwan was 6,000 20,000 per year (TJCHA, 2010). To minimize the occurrence of such incidents, the Department of Health established the Patient Safety Committee in February 2003 as a collaborative effort to establish patient safety goals. Since then, the JCHA has developed medical operation safety guidelines for hospitals. Unfortunately, these guidelines do not specify the level of workforce necessary for implementation. Many studies define nurse workload using working hours or patient nurse ratio (Baumann, 2007; Gaudine, 2000; Healthcare Financial Management Association [HFMA], Accepted for publication: August 3, 2012 *Address correspondence to: Yu-Chun Yin, No.17, Cuiping Lane, Zhennan Street, Qingshui District, Taichung City 43653, Taiwan, ROC. Tel: +886 (4) 2623-6106; E-mail: nelly922003@yahoo.com.tw doi:10.1097/jnr.0b013e3182736363 300

Nurses Workload and Patient Safety VOL. 20, NO. 4, DECEMBER 2012 2010; Needleman & Buerhaus, 2003). The American Nurses Association(ANA) s Donabedian structure, process, and outcomes model offers a comprehensive method of evaluating healthcare quality (ANA, 2007). Nurse-sensitive indicators reflect the structure, process, and outcomes of nursing care. Patient outcomes are considered nurse sensitive if they improve because of increased levels (quantity or quality) of nursing care (ANA, 2010b). Many researchers have reported higher levels of adverse patient events, for example, hospital-related mortality, medication errors, unplanned extubation, hospital-acquired pneumonia, and urinary tract infections (UTIs), during times when staff levels are low. Adverse events have been particularly associated with factors such as nurse overtime and patient nurse ratios (Agency for Healthcare Research and Quality, 2007; HFMA, 2010; Kane, Shamliyan, Mueller, Duval, & Wilt, 2007; Rogers, Hwang, Scott, Aiken, & Dinges, 2004; Stone et al., 2007). Rothberg, Abraham, Lindenauer, and Rose (2005) considered that, as a patient safety intervention, improved nurse staffing has a cost effectiveness that falls comfortably within the range of other widely accepted interventions. Although Taiwan s National Health Insurance is convenient for citizens, rising medical service costs have caused considerable financial burdens. Hospitals must reduce expenses. Personnel costs are the largest component of costcontrol models used by hospital management authorities. As nurses generally account for one third of total hospital staff, cost reduction strategies adopted by most hospitals have focused on reducing either nursing staff numbers or nurse salaries. Increased workloads and salary cuts have resulted in high turnover rates at Taiwanese hospitals, which ultimately contribute to greater frequencies of medical accidents. This poses a serious threat to nursing quality and overall patient safety. The Institute of Medicine in the United States confirmed that nursing practices make valuable contributions to patient health (Institute of Medicine, 1996). The policy statement principles on safe nurse staffing of the ANA (ANA, 2009) points out that determination of appropriate staffing must depend not only on the afford ability of the patient but also on factors including intensity of patient illness; contextual issues, including the environment and available technology; and the patient s physical and psychosocial condition. Because of extreme workplace pressure, medical institutions have experienced increased new employee turnover rates. For example, the turnover rate at the National Taiwan University Hospital in 2003 was 7.1% overall and 19.8% for newly employed (G1 year) nurses (Huang, 2004). The average turnover rate of nurses after 1 year of service nationwide was 28.02%; the Taipei Nurses Association reported a rate of 21.2% (Sun, 2005). For other hospitals outside Taipei City, turnover rates (for both public and private hospitals) were even higher. This study investigated the workload of nurses in medical institutions and assessed the association between workload and nurse-sensitive patient safety outcome indicators. These results may serve as a reference for health authorities working to establish guidelines for medical institutions and make nursing personnel policies. Figure 1 shows the research framework. Methods Participants This cross-sectional quantitative study targeted the 78,673 members of the Taiwan Province Union of Nurses Associations, power set at p =.05, the number of subjects required z in accordance with the formula N = 2 pð1 pþ d 2 ðn 1Þþz 2 pð1 pþ = 1,053 p =.05;N = 78,673;! =.025; d =.03; z.025 =1.96,and 1.91% members were randomly chosen from the nurses associations of 21 counties or cities across Taiwan. Participants entered information into the nurse workload and nurse-sensitive safety outcome indicators logbook over a 2-week period. One thousand five hundred logbooks were distributed, 1,373 were retrieved, and 1,358 were valid (valid response rate: 90.5%). Instruments The logbook was purposely designed for this study based on logbooks used in previous studies (Rogers et al., 2004). It gathered data on (a) respondent demographic data and workrelated variables including gender, marital status, education, position, nursing experience, ward or unit of service, and work shift type; (b) workload, including shift hours, actual working hours, whether overtime work was paid (with explanations), number of patients being cared for, number of patients in the unit, and number of nurses in each unit; and (c) eight nurse-sensitive patient safety outcome indicators compiled from previous studies including patient falls, decubitus/pressure Figure 1. Conceptual framework of the current study. 301

The Journal of Nursing Research Li-Fang Liu et al. ulcers, near errors in medication or missing medications, medication errors, failure to rescue, accidental tube removal or unplanned extubation, hospital-acquired pneumonia, and hospital-acquired UTI. Content validity was confirmed by six experts who were all presidents of county or city nurse associations. Each had more than 15 years of clinical experience; five held a master s degree and one held a doctoral degree. The average content validity index was.96. All items in the logbook and the logbook format were pilot-tested at a district hospital by 30 nurses before the study began. Although logbooks are not often used to collect information about medical errors, there is some evidence that daily, anonymous, end-of-shift reporting of events in a logbook is a valid approach to ascertain their nature and prevalence (Rogers et al., 2004). The data collection period began on December 1, 2008, and ended on February 28, 2009. Ethical Considerations The institutional review board at the Mennonite Christian Hospital approved this study. Approval was also granted by the nursing association presidents in the 21 counties or cities from which participants were randomly selected. In addition, all participants signed consent forms, and each received NTD-300 coupons redeemable at 7 11 for their participation. Data Analysis SPSS 10.0 for Windows (traditional Chinese software suite, IBM Taiwan Inc., Taipei, Taiwan) encoded, organized, and analyzed data. Durations of scheduled and actual work hours per shift were calculated and aggregated per nurse for the 2-week period. Shift duration was defined as 8 hours. The binary response for occurrence(s) of an event related to eight nurse-sensitive patient safety outcome indicators during a worked shift was used as the primary analysis outcome. The secondary outcome was near miss events, meaning that the respondent purposely stopped himself or herself before one of the safety outcome indicators occurred while on shift duty. Descriptive statistics, comprising mean values, standard deviations, and percentages described participant characteristics, shift hours, actual working hours, overtime hours, number of patients cared for, and number of events. Inferential statistics, comprising the Spearman correlation and odds ratio analysis, determined associations between nurse-sensitive patient outcome indicators and nurse workload. To increase sample size, we applied the exclude observations by analysis method rather than the exclude cases list-wise method to missing values. The odds ratio used in this experiment was odds ratio = [odds in favor of nurse-sensitive patient outcome indicators group] / [odds in favor of patient nurse ratio group]. Multiple logistic regression analysis wasusedtoinvestigatemultivariate adjusted odds ratios for nurse-sensitive patient safety outcome indicators. Significance was indicated by a p value of G.05intwo-sidedtests. Results Demographic Profiles of the Participants Table 1 shows patient demographics. Participants were mostly women (99.4%). Mean age was 29.91 years (SD = 5.76 years; range = 19 56 years), with 60.7% between 21 and 30 years; slightly more than half were single (54.9%), 48.6% had received a junior college education, and 48.2% had earned an undergraduate degree. In terms of clinical ladder, 40.5% possessed N2 and 23.3% possessed N1; 86.7% held a registered nursing license. About three quarters were staff nurses (76.3%), and mean nursing experience was 8.49 years (SD = 5.69 years). Average time spent at the current hospital was 5.99 years (SD = 5.02 years). Almost half (44.8%) worked at metropolitan teaching hospitals and nearly one-fifth (18%) worked at academic medical centers. In terms of nursing experience, 46.3% worked in medical and surgery wards and 22.7% worked in intensive care units. Analysis of Nurse Workload Shift hours, actual working hours, and overtime work Data on workload are provided in Tables 2 and 3. One thousand three hundred fifty-eight valid logbooks were retrieved. Statistically valid data contained 12,895 work shifts and 6,050 off days. Mean scheduled hours were 8.09 (SD = 0.99) per day, whereas actual working hours were 8.90 (SD = 1.38) per day; each shift worked overtime for an average of 0.86 (SD = 0.98) hours per day. Further analysis of scheduled and actual working hours revealed that 7,777 (60.4%) shifts worked overtime for at least 30 minutes every day. Among participants working overtime, compensation was offered to only 673 (8.7%); moreover, this compensation was provided only when overtime work was attributable to patient needs or was formally assigned. For the remaining 7,104 (91.3%) shifts, overtime work was unpaid. The reasons for this included hospital rules (1,807, 23.2% shifts), overtime unrelated to nursing work (1,222, 15.7% shifts), lack of willingness to apply for overtime hours (910, 11.7% shifts), or overtime as a consequence of inefficiency (959, 12.3% shifts). Other reasons included that nursing work was considered as a responsibility system, patient condition changes, patient transfers in or out of the hospital, meetings or in-service education, and assistance to colleagues and new employees. Patient nurse ratio In this study, the patient nurse ratio refers to the average number of patients formally assigned per nurse. The mean patient nurse ratio for all wards was 10.35 (SD = 6.92); psychiatric ward nurses had the highest ratio (17.83, SD = 12.66), and obstetrics and gynecology ward nurses had the lowest (7.21, SD = 5.79; Tables 4 and 5). The patient nurse ratio varied significantly among special unit nurses. The outpatient ratio was 29.27 (SD = 24.95) in outpatient departments, 302

Nurses Workload and Patient Safety VOL. 20, NO. 4, DECEMBER 2012 TABLE 1. Nurses Demographic Data and Work-Related Variables (N = 1,358) Variable M SD n % Gender Male 8 0.6 Female 1,350 99.4 Age 29.91 5.76 e20 9 0.7 21 30 824 60.7 31 40 440 32.4 41 50 81 6.0 Q51 4 0.3 Marital status Married 591 43.5 Single 746 54.9 Widowed/divorced 21 1.5 Education High school 29 2.1 Junior college 660 48.6 Undergraduate 654 48.2 Graduate school 15 1.1 License Intern a 3 0.2 LPN 168 12.4 RN 1,187 87.4 Institution Medical centers 245 18.0 Regional hospitals 685 50.5 Local hospitals 428 31.5 Clinical ladder (N = 1,357) N 216 15.9 N1 316 23.3 N2 549 40.5 N3 196 14.4 N4 80 5.9 Professional title Intern a 3 0.2 Nurse 168 12.4 Registered nurse 1,177 86.7 Advanced practice nurse 10 0.7 Position Nurse 1,036 76.3 Team leader 165 12.2 Vice head nurse 61 4.5 Head nurse 96 7.1 Nursing experience Total years (N = 1,354) 8.49 5.69 Current hospital 5.99 5.02 (N = 1,348) Ward/unit Medical and surgical 629 46.3 Intensive care unit 308 22.7 TABLE 1. Nurses Demographic Data and Work-Related Variables (N = 1,358), continued Variable M SD n % Obstetrics/Pediatrics 123 9.1 Outpatient department and emergency room 102 7.5 Other 181 12.3 Note. LPN = licensed practical nurse. a Graduate students who had not obtained a nurse/registered nurse license. 19.22 (SD = 17.54) in emergency rooms, and 2.81 (SD = 2.57) in intensive care units. The patient nurse ratios was 8.72 (SD = 10.24) during day shifts and 10.93 (SD = 9.04) during night shifts. When we divided patient nurse ratios into four patient number clusters (e7, 8 9, 10 11, and 911), the most prevalent was the 911 group (5,210 shifts, 40.4%; Tables 4 and 5). TABLE 2. Description of Shifts of Hospital Staff Nurses (N = 12,895) Variable Work Hour M SD n % Schedule shifts 8.09 0.99 Actual shifts 8.90 1.38 No overtime 5,118 39.6 Overtime (total) 7,777 60.4 0.5 1 hour 3,232 25.1 1 2 hours 2,585 20.1 2 3 hours 1,426 11.1 93 hours 534 4.1 TABLE 3. Analysis for the Reasons of Overtime Shifts (N = 7,777) Item n % Paid for overtime 673 8.7 Patient needs 525 6.8 Mandatory overtime 18 0.2 Other 130 1.7 Unpaid for overtime 7,104 91.3 Hospital rules 1,807 23.2 Unrelated to nursing 1,222 15.7 Not willing to apply 910 11.7 Inadequate efficiency 959 12.3 Other 2,206 28.4 303

The Journal of Nursing Research Li-Fang Liu et al. TABLE 4. Patient Nurse Ratio for Each Shift per Unit (N = 12,895) Variable Work Shift Patient Nurse Ratio M SD General wards 8,123 10.35 6.92 Medical 2,809 10.55 6.70 Surgical 1,725 9.89 5.95 Obstetrics/gynecology 636 7.21 5.79 Psychiatry 405 17.83 12.66 Pediatric 547 7.84 4.38 Medical/surgical 1,417 10.37 5.75 Respiratory center 468 11.60 7.06 Rehabilitation 116 9.63 3.77 Special units 4,772 8.00 13.38 Outpatient department 256 29.27 24.95 Emergency room 678 19.22 17.54 HC 38 5.71 3.13 Hospice 89 7.78 10.76 HD 157 5.67 2.83 Intensive care unit 2,962 2.81 2.57 Others 592 12.75 16.50 Work shift I. Day shift 7,405 8.72 10.24 II. Evening shift 2,972 10.22 9.97 III. Night shift 2,518 10.93 9.04 Note. HC = home care; HD = hemodialysis. Analysis of Nurse-Sensitive Patient Outcome Indicators Among the 12,895 valid shift records, there were 41 cases of patient falls including 21 mild injuries and one moderate injury (Table 6); 2,850 cases of decubitus/pressure ulcers including 346 cases that improved and 61 cases that worsened; 68 cases of medication near errors 11 cases of medication errors; 64 of 212 failed resuscitations, including two due to inadequate nurse numbers; 166 cases of unplanned extubation; 1,408 cases of hospital-acquired pneumonia including 828 attributed to tracheostomy suctioning; 1,240 cases of hospital-acquired UTI; and 781 cases of Foley catheter retention (Table 6). TABLE 5. Patient Nurse Ratio (N = 12,895) Group (patient per nurse per shift) Work Shift % e7 3,069 23.8 8 9 2,579 20.0 10 11 2,037 15.8 911 5,210 40.4 TABLE 6. Nursing Sensitive Outcome Indicators Analysis (N = 12,895) Nursing Sensitive Outcome Indicator Work Shift No. of Event Patient falls 41 41 No injury 19 Mild injury 21 Intermediate injury 1 Medication near errors 60 68 Medication errors 10 11 Decubitus 1,313 2,850 Improved 346 Worsened 61 No change 2,443 Hospital acquired UTIs 371 1,240 Retention Foley 781 Nonretention Foley 459 Rescue station 191 212 Rescued 148 Failed 64 Hospital-acquired pneumonia 339 1,408 No suctioning 169 Tracheostomy suctioning 828 Mouth and nose suctioning 182 Unplanned extubations 131 166 Not fixed appropriate 119 Nonphysical restraint 44 Other 3 Note. UTI = urinary tract infection. Analysis of Nurse Workload and Nurse-Sensitive Patient Outcome Indicators Nurse overtime and nurse-sensitive patient outcome indicators Using the Pearson product moment correlation coefficient to analyze the relationship between the overtime working hours and nurse-sensitive indicators, we found a statistically significant positive association between overtime working hours and the following six indicators: patient falls (r =.037, p =.000), decubitus/pressure ulcers (r =.062, p =.000), near errors in medication (r =.036, p =.000), rescue failures (r =.028, p =.001), unplanned extubations (r =.028, p =.002), and UTI (r =.022, p =.013). Overtime hours were delineated into five groups: e0.5 hr, 0.5 1 hr, 1 2 hrs, 2 3 hrs, and Q3 hrs. Using the Spearman rank order correlation coefficient to analyze the relationship between nurse overtime and nurse-sensitive indicators, we found a statistically significant positive association between the overtime working hours and the following six indicators: patient falls (r =.037, p =.000), decubitus/pressure ulcers (r = 304

Nurses Workload and Patient Safety VOL. 20, NO. 4, DECEMBER 2012 TABLE 7. Correlation of Nurse Overtime Work and Nurse-Sensitive Patient Outcome Indicators Variable Overtime Working Hour Overtime Work Group N r p N r S p Patient falls 12,675.037.000*** 12,936 0.030.000*** Decubitus/pressure ulcers 12,675.062.000*** 12,936 0.006.000*** Near errors in medication 12,675.036.000*** 12,937 0.034.000*** Medication errors 12,675 j.007.449 12,938 j0.003.659 Rescue failures 12,675.028.001** 739 j0.047.028* Unplanned extubations 12,675.028.002* 12,938 0.023.000*** Pneumonia 12,675.010.268 12,938 0.056.000*** Urinary tract infections 12,675.022.013* 12,937 0.039.000*** *p G.05. **p G.01. ***p G.001..062, p =.000), near errors in medication (r =.036, p =.000), rescue failures (r =.028, p =.001), unplanned extubations (r =.028, p =.002), and UTIs (r =.022, p =.013; Table 7). Patient nurse ratio and nurse-sensitive patient outcome indicators Logistic regression analysis of nurse-sensitive patient safety outcome indicators revealed that patient nurse ratios of 911, 10 11, and 8 9 resulted in patient fall OR values of 5.14 (p =.004),1.51(p =.004), and 3.59 (p =.007) compared with that when the patient number was G7. The OR for patients who developed decubitus/pressure ulcers increased with the patient nurse ratio; compared with the OR for patient nurse ratio of e7, the OR was 1.72 (p =.000) when patient nurse ratios were 8 9, 2.21 (p =.000) when 10 11, and 2.74 (p =.000) when 911.IncomparisonwiththeOR for unplanned extubation when the patient nurse ratio was e7, the OR was 2.40 (p =.000) when patient nurse ratios were 8 9, 3.57 (p =.000) when 10 11, and 5.58 (p =.000) when 911. In comparison with the OR for hospital-acquired pneumonia, when the patient nurse ratio was G7, the OR was 1.60 (p =.000),1.29(p =.000), and 0.86 (p =.000), respectively, when patient nurse ratios were 911, 10 11, and 8 9. For hospital-acquired UTI, the OR was 1.62 (p =.000), 1.38(p =.002), 1.05(p =.002) when patient nurse ratios were 911, 10 11, and 8 9 compared with when the patient nurse ratio was e7. The remaining indicators (near errors in medication, medication errors, and rescue failure) did not differ significantly between groups (Table 8). Results were adjusted for education, clinic ladder, and institution as follows. For education, the category high school was used as the reference, with junior college, and undergraduate school was used as the indicator variable. For clinical ladder, the category N was used as the reference, and N1, N2, N3, and N4 were used as indicator variables. For institution, the category medical centers was used as the reference, with regional hospitals, and local hospitals were used as the indicator variable. In addition to group, education, clinical ladder, and institution were forced into the model to perform multiple logistic regression analyses. Discussion In the current study, actual working hours of nurses was 8.90 (SD = 1.38) hours per day. Results showed that 50.4% of participants worked overtime. The average day shift patient nurse ratio was 8.72 (SD = 10.24), and the average patient nurse ratio was 10.93 (SD = 9.04) for the night shift. The most frequent patient nurse ratio group was 911 (5,210 shifts, 40.4%). These results show that Taiwanese nurses are overloaded in terms of both overtime and patient nurse ratios. Overtime affects physiological factors such as fatigue and system variables such as increased work intensity. Fatigue coupled with increased work intensity may contribute to an increased risk of errors (Rogers et al., 2004). The overtime and high patient nurse ratios documented here suggest a link between poor working conditions and threats to patient safety. DeLucia, Ott, and Palmieri (2009) proposed that nurse overloading is caused by a shortage of nurses. There is no nurse shortage in Taiwan; however, out of a total of 218,702 licensed nurses in May 2010, only 129,622 (59.27%) were actually practicing (National Union of Nurses Associations, 2010). This low practice rate supports the hypothesis that nurses avoid hospital work due to work overload and poor working conditions. Lin, Kao, and Cheng (2009) reported that the National Health Insurance implementation of its global budgeting system reduced government coverage of nurse salaries to 70%; individual hospitals must cover the remaining 30%. In response, medical institutions have cut nursing workforce to reduce costs, which has reduced nursing service quality. Medical institutions have continually reduced workforce, salaries, and benefits. This can only detract from the provision of excellent quality medical services, improved public health, and patient safety. Our analysis found a strong association between nurse overtime and the three patient outcomes of falls, decubitus/ 305

The Journal of Nursing Research Li-Fang Liu et al. TABLE 8. Association of the Patient Nurse Ratio With Nurse-Sensitive Patient Outcome Indicators (N = 12,895) Number of Patient Cared For Work Shift No. of Event % Odds Ratio p 95% Confidence Interval 1. Patient falls e7 3,069 3 0.1 1.00 8 9 2,579 9 0.3 3.59.007** 0.97, 13.28 10 11 2,037 3 0.1 1.51.004** 0.30, 7.50 911 5,210 26 0.5 5.14.004** 1.55, 17.02 2. Decubitus/pressure ulcers e7 3,069 176 5.7 1.00 8 9 2,579 234 9.1 1.72.000*** 1.39, 2.12 10 11 2,037 225 11.1 2.21.000*** 1.78, 2.74 911 5,210 678 13.0 2.74.000*** 2.28, 3.28 3. Near errors in medication e7 3,069 8 0.3 1.00 8 9 2,579 13 0.5 1.94.276 0.80, 4.70 10 11 2,037 12 0.6 2.28.228 0.93, 5.59 911 5,210 27 0.5 2.00.128 0.91, 4.41 4. Medication errors e7 3,069 1 0.0 1.00 8 9 2,579 1 0.0 1.19.298 0.07, 19.04 10 11 2,037 1 0.0 1.51.301 0.09, 24.15 911 5,210 7 0.1 4.13.081 0.51, 33.58 5. Rescue failure e7 3,069 11 0.4 1.00 8 9 2,579 6 0.2 0.36.027 0.12, 1.10 10 11 2,037 10 0.5 1.31.018 0.45, 3.80 911 5,210 37 0.7 1.46.046* 0.64, 3.36 6. Unplanned extubation e7 3,069 9 0.3 1.00 8 9 2,579 18 0.7 2.40.000*** 1.07, 5.36 10 11 2,037 21 1.0 3.57.000*** 1.63, 7.84 911 5,210 83 1.6 5.58.000*** 2.80, 11.13 7. Hospital-acquired pneumonia e7 3,069 65 2.1 1.00 8 9 2,579 47 1.8 0.86.000*** 0.58, 1.25 10 11 2,037 55 2.7 1.29.000*** 0.89, 1.87 911 5,210 172 3.3 1.60.000*** 1.19, 2.14 8. Hospital-acquired UTI e7 3,069 68 2.2 1.00 8 9 2,579 60 2.3 1.05.002** 0.74, 1.50 10 11 2,037 61 3.0 1.38.002** 0.97, 1.96 911 5,210 182 3.5 1.62.000*** 1.22, 2.16 Note. UTI = urinary tract infection. *p G.05. **p G.01. ***p G.001. pressure ulcers, and medication error near misses. Weaker associations were found with failure to rescue, unplanned extubation, and hospital-acquired UTI. Our results were similar to those of many other studies. Overtime work and insufficient break time directly affect the quality of nursing services and are the primary causes of dissatisfaction displayed toward nurses, increased medication errors, patient and family complaints, and patient falls (Agency for Healthcare Research and Quality, 2007; HFMA, 2007; Rogers et al., 2004; Stone et al., 2007; Zurn, Dolea, & Stilwell, 2005). The ANA (2010a) released a statement that, regardless of the number of hours worked, each registered nurse has an ethical responsibility to carefully consider his or her level of fatigue when deciding to accept any assignment extending beyond the regularly scheduled work day or week, including mandatory or voluntary overtime assignments. 306

Nurses Workload and Patient Safety VOL. 20, NO. 4, DECEMBER 2012 Taiwan extended the terms of the Labor Standards Act to include nurses in 1998. Items 24 and 30 stipulate working hours and overtime payments (Labor Standards Act, 2009). However, this study indicated that 91.3% nurses who worked overtime did not receive compensation due to hospital rules. Hospitals stated that nursing work is based on a responsibility system and that attending meetings, participating in in-service education, and assisting colleagues and new employees cannot be considered overtime work. Whether these requirements violate the Labor Standards Act is an issue meriting discussion. However, local nurses associations rarely receive complaints addressing this issue. We believe overtime work to be the primary cause of the high nurse turnover rate in Taiwan. Local nurses associations and the National Union of Nurses Associations should therefore work in conjunction to resolve issues related to nurse well-being. The relationship between patient nurse ratios and nursing outcomes is complex. The results of this study indicate patient falls, decubitus/pressure ulcers, unplanned extubation, and hospital-acquired pneumonia and UTI as positively correlated with patient nurse ratios. Weaker associations were found for medication errors, medication error near misses, and rescue failure. Findings regarding the relationship between patient nurse ratios and patient outcomes were consistent with those of previous studies, in which an increase of one registered nurse per patient per day had a significant and measurable effect on decreasing hospital-acquired pneumonia, unplanned extubation, respiratory failure, and cardiac arrest risks (Kane et al., 2007; McGillis et al., 2003; Needleman & Buerhaus, 2003; Needleman, Buerhaus, Mattke, Stewart, & Zelevinsky, 2001; Stanton, 2004; Stone et al., 2007). The patient nurse ratio and quality of care in acute care hospitals are strongly correlated. Rothberg et al. (2005) applied cost effect analysis using patient nurse ratios to compare expenses related to patient death and length of hospital stay. Their results showed that optimal cost effectiveness was achieved at a patient nurse ratio of 4:1. Therefore, the authors of the current article advocate increasing the patient nurse ratio as the most effective strategy to maintain patient safety in healthcare. Some patient outcomes were found more strongly related to other (nonnursing ratio) aspects of institutional care such as medical decision making and institutional policies. Additional research is thus recommended to better understand and control nurse-sensitive patient outcome indicators. Conclusion Our analysis provided evidence of a clear association between both working hours and patient nurse ratios and certain nurse-sensitive patient outcomes. We found nurses in Taiwan to be overloaded in terms of both overtime and patient nurse ratios. This has created our current situation in which Taiwan has an inadequate number of nurses working in hospitals despite having a sufficient number of licensed nursing professionals. In conclusion, the nursing practice environment has not improved. Nursing jobs increasingly fail to attract young and gifted students, whereas senior nurses tend to retire early or even take their services abroad. Over the next 10 years, the exacerbated nursing shortage can be expected to continue the deterioration of the quality of nursing in hospitals, resulting in compromised patient safety. Currently, government, institutions, and society focus largely on patient safety and nursing quality, thereby increasing workloads to unacceptable levels and diminishing the appeal of clinical jobs for young people. Creation of a positive working environment, deployment of sufficient personnel, and retention of young nurses will provide the best guarantee of patient safety and nursing quality. Future research and policy makers should focus on making practical, positive improvements in the overall working environment for nurses. Recommendations Professional nursing organizations and suggested government policies Study results indicate that the patient nurse ratio in hospital general acute care wards should not exceed 7:1. We strongly recommend the Department of Health to stipulate a maximum patient nurse ratio for each of hospitals three daily shifts as a basic requirement of medical institutions. Professional nursing organizations should educate the public regarding the strong correlation between patient nurse ratio and patient safety. The Taiwan Nurses Association and the National Union of Nurses Association should make higher patient nurse ratios at hospitals as a main thrust of their lobbying and education efforts. Nursing administrators 1. Nursing administrators should reduce the physical and spiritual loads on basic nursing staff. 2. Nurse working hours should strictly follow the Labor Standards Act requirements. 3. Nursing leaders should assist nurses to raise objections to unreasonable hospital demands. Primary nursing leaders and clinical nurses 1. Clinical nurses should possess sufficient competence. 2. Clinical nurses should report to and request assistance from the nursing administrator or local associations regarding unreasonable hospital demands that undermine their rights and interests. Nursing researchers Research on nursing workload and patient safety issues should be conducted in the various nursing domains to increase the scope and comprehensiveness of available data. Results may be referenced by professional nursing organizations seeking to plan and ensure adequate workforce. Study Limitations Owing to time limitations, we used a random sampling of Nurses Union members to recruit participants rather than a 307

The Journal of Nursing Research Li-Fang Liu et al. stratified sampling approach that considered patient illness severity. We were thus unable to explain correlations between the patient nurse ratio and certain nurse-sensitive patient outcome indicators. Future studies should focus on nurses in different specializations and investigate correlations between the patient nurse ratio and nurse-sensitive patient outcome indicators. Acknowledgments We are obliged to the Taiwan Provincial Nurses Association for providing the necessary facilities and resources for this study. We would also like to express our sincere gratitude to each of the 21 county and city nurses associations throughout Taiwan whose generous assistance helped complete this study. References Agency for Healthcare Research and Quality. (2007). Nurse staffing and quality of patient care. Evidence Report/Technology Assessment: Number 151. Retrieved from http://www.ahrq.gov/downloads/pub/evidence/pdf/nursestaff/nursestaff.pdf American Nurses Association. (2007). History of evaluating nursing care quality. Retrieved from http://www.medscape.com/viewarticle/569395_2 American Nurses Association. (2009). Safe staffing saves lives. Retrieved from http://www.safestaffingsaveslives.org/ WhatisSafeStaffing/SafeStaffingPrinciples.aspx American Nurses Association. (2010a). Mandatory overtime. Retrieved from http://www.nursingworld.org/mainmenu categories/anapoliticalpower/state/statelegislativeagenda/ MandatoryOvertime.aspx American Nurses Association. (2010b). Nursing-sensitive indicators. Retrieved from http://nursingworld.org/main MenuCategories/ThePracticeofProfessionalNursing/Patient SafetyQuality/Research-Measurement/The-National-Database/ Nursing-Sensitive-Indicators_1 Baumann, A. (2007). Positive practice environments: Quality workplaces = quality patient care. Information and action tool kit. Geneva, Switzerland: International Council of Nurses. DeLucia, P. R., Ott, T. E., & Palmieri, P. A. (2009). Performance in nursing. In F. T. Durso (Ed.), Reviews of human factors and ergonomics (Vol. 5, pp. 1 40). Santa Monica, CA: Human Factors and Ergonomics Society. Gaudine, A. P. (2000). What do nurses mean by workload and work overload? Canadian Journal of Nursing Leadership, 13(2), 22 27. Healthcare Financial Management Association. (2010). The true cost of overtime data trends overtime in nursing discussedv Industry overview. Retrieved from http://findarticles.com/p/ articles/mi_m3257/is_12_56/ai_95967107/ Huang, L. H. (2004). Transition from student nurse to staff nurse. The Journal of Nursing, 51(4), 33 36. (Original work published in Chinese) Institute of Medicine. (1996). Nursing staff in hospitals and nursing homes: Is it adequate? Washington, DC: The National Academies Press. Kane, R. L., Shamliyan, T. A., Mueller, C., Duval, S., & Wilt, T. J. (2007). The association of registered nurse staffing levels and patient outcomes: Systematic review and meta-analysis. Medical Care, 45(12), 1195 1204. Labor Standards Act. (2009). Modified 98 years Republic of China on April 22 Labor Disputes Law. Republic of China on 27 June 77 presidential decree announced amendments to text 45. (Original work published in Chinese) Lin, K. J., Kao, C. C., & Cheng, W. K. (2009). The relationship between nursing human capital and turnover rate. Soochow Journal of Accounting, 1(2), 93 127. (Original work published in Chinese) McGillis, L., Doran, D., Tregunno, D., McCutcheon, A., O Brien- Pallas, L., Tranmer, J., I Thompson, D. (2003). Indicators of nurse staffing and quality nursing work environments: A critical synthesis of the literature. Toronto, Canada: Faculty of Nursing, University of Toronto. National Union of Nurses Associations, ROC. (2010). Health information network medical personnel management systemv Taiwan and Fujian area nurses tables. Retrieved from http:// www.nurse.org.tw/userfiles/file/datasearch. (Original work published in Chinese) Needleman, J., & Buerhaus, P. I. (2003). Nurse staffing and patient safety: Current knowledge and implications for action. International Journal for Quality in Health Care, 15(4), 275 277. doi:10.1093/intqhc/mzg051 Needleman, J., Buerhaus, P. I., Mattke, S., Stewart, M., & Zelevinsky, K. (2001). Nurse staffing and patient outcomes in hospitals. (Health Resources Services Administration No. 230-99-0021). Boston, MA: Harvard School of Public Health. Rogers, A. E., Hwang, W. T., Scott, L. D., Aiken, L. H., & Dinges, D. F. (2004). The working hours of hospital staff nurses and patient safety. Health Affairs, 23(4), 202 212. doi:10.1377/hlthaff.23.4.202 Rothberg, M. B., Abraham, I., Lindenauer, P. K., & Rose, D. N. (2005). Improving nurse-to-patient staffing ratios as a costeffective safety intervention. Medical Care, 43(8), 785 791. doi:10.1097/01.mlr.0000170408.35854.fa Stanton, M. W. (2004). Hospital nurse staffing and quality of care. Retrieved from http://www.ahrq.gov/research/nursestaffing/ nursestaff.htm Stone, P. W., Mooney-Kane, C., Larson, E. L., Horan, T., Glance, L. G., Zwanziger, J., & Dick, A. W. (2007). Nurse working conditions and patient safety outcomes. Medical Care, 45(6), 571 578. doi:10.1097/mlr.0b013e3180383667 Sun, J. Z. (2005). National status of clinical practice survey of nursing practices. (DOH94-NH-36). Taipei City, Taiwan, ROC: National Union of Nurses Associations. (Original work published in Chinese) Taiwan Joint Commission on Hospital Accreditation. (2010). Patient safety in Taiwan. Retrieved from http://www.patientsafety.doh.gov.tw/en/content.asp?cid=2 World Health Organization. (2010). Patient safety. Retrieved from http://www.who.int/patientsafety/en/index.html Zurn, P., Dolea, C., & Stilwell, B. (2005). Issue paper 4-nurse retention and recruitment: Developing a motivated workforce, The Global Nursing Review Initiative. Retrieved from the ICN Web site: http://www.icn.ch/global/issue4retention.pdf 308

The Journal of Nursing Research VOL. 20, NO. 4, DECEMBER 2012 1 2 3 4 5 * 1 2 3 4 5 21 1,500 1,373 1,358 90.5% Spearman 7 101 8 3 * 43653 17 04 26236106 E-mail: nelly922003@yahoo.com.tw 309