|Home | About | Journals | Submit | Contact Us | Français|
Extended work shifts of twelve hours or longer are common and even popular with hospital staff nurses, but little is known about how such extended hours affect the care that patients receive or the well-being of nurses. Survey data from nurses in four states showed that more than 80 percent of the nurses were satisfied with scheduling practices at their hospital. However, as the proportion of hospital nurses working shifts of more than thirteen hours increased, patients’ dissatisfaction with care increased. Furthermore, nurses working shifts of ten hours or longer were up to two and a half times more likely than nurses working shorter shifts to experience burnout and job dissatisfaction and to intend to leave the job. Extended shifts undermine nurses’ well-being, may result in expensive job turnover, and can negatively affect patient care. Policies regulating work hours for nurses, similar to those set for resident physicians, may be warranted. Nursing leaders should also encourage workplace cultures that respect nurses’ days off and vacation time, promote nurses’ prompt departure at the end of a shift, and allow nurses to refuse to work overtime without retribution.
Traditional eight-hour shifts for hospital nurses are becoming a thing of the past:1,2 Nurses increasingly work twelve-hour shifts. This schedule gives nurses a three-day work week, potentially providing better work-life balance and flexibility.3,4 However, actual shift lengths are often unpredictable5 because of fluctuations in patient needs and unanticipated staffing changes. As a result, nurses often must put in unplanned overtime beyond the scheduled shift length. When long shifts are combined with overtime, shifts that rotate between day and night duty, and consecutive shifts, nurses are at risk for fatigue and burnout, which may compromise patient care.5
Despite regulations on shift length and cumulative working hours for resident physicians and people in other industries, there are no national work-hour policies for registered nurses. Several states, such as Maryland and California, have prohibited mandatory overtime for nurses, but there is no limit to nurses’ voluntary overtime hours. Furthermore, the distinction between voluntary and mandatory is often blurred: Nurses report feeling coerced into working “voluntary” overtime. Nurse shortages, coupled with a weak economy, have motivated nurses to work past the end of their scheduled shift or to work additional shifts.
There is limited research on the impact of long shifts on nurses6 or on the quality of the care they provide to patients.5,7–10 In addition, there is inadequate understanding of whether patients’ satisfaction with care is affected by the extended hours worked by nurses.
In this study we investigated the relationship between hospital nurses’ shift length and three nurse outcomes: burnout, job dissatisfaction, and intention to leave the job. Burnout is characterized by emotional exhaustion, depersonalization of patients, and feelings of lack of personal accomplishment by caregivers,11,12 which could negatively affect nurse job satisfaction and voluntary turnover.13,14
Our study also examined the relationship between nurses’ shift length and patient outcomes, using data from the Hospital Consumer Assessment of Healthcare Providers and Systems Survey, a national, standardized hospital-level data set that includes information about patients’ assessment of overall care and nursing care in acute care hospitals.15 Previous research has documented that hospitals in which a higher proportion of nurses experience burnout have lower patient satisfaction.16 However, to the best of our knowledge this is the first study to explore whether nurses’ shift length specifically is associated with patient satisfaction.
This study involved a secondary analysis of cross-sectional data from three sources linked by common hospital identifiers. The first source was the Multi-State Nursing Care and Patient Safety Study,17 which surveyed nurses in four states from 2005 to 2008 about their shift length, working conditions, burnout, job satisfaction, and intentions to leave their employer. The second was the Hospital Consumer Assessment of Healthcare Providers and Systems survey for 2006–07.15 This survey is a publicly available source of data on patients’ experiences during acute care hospitalizations. The third was the American Hospital Association Annual Survey of Hospitals for 2006, which we consulted to obtain information about additional characteristics of the study hospitals.
We used a sample of 22,275 registered nurses from the Multi-State Nursing Care and Patient Safety Study. The nurses in our study worked in 577 hospitals in California, New Jersey, Pennsylvania, and Florida. The sample included at least 10 nurses per hospital with an average of 39 nurses per hospital (the range was 10–205). A full description of the Multi-State Study is available elsewhere,17 and the online Technical Appendix includes explanations of the sampling methods we used.18
We included in our analysis nurse respondents in the Multi-State Study who reported working 1–24 hours on their last shift and providing care to at least one but fewer than twenty patients. These selection criteria were designed to eliminate nurses who were not working directly with patients, including those in supervisory or “on call” roles.
The analyses were further restricted to nurses on medical or surgical units and in intensive care units. We excluded nurses working in long-term care, outpatient services, and the operating room because they do not usually follow the same shift patterns as inpatient hospital units.
Nurses’ shift length was calculated as the difference between the start and end time of the most recent shift that the nurse worked. Shift length was grouped into four categories: 8–9 hours, 10–11 hours, 12–13 hours, and more than 13 hours. These ranges were used to account for changes in planned shifts. For example, shifts in the category of 10–11 hours might have involved overtime after an 8-hour shift. Scheduled shifts are typically eight or twelve hours in length.
We used the length of individual nurse shifts for our analyses of the nurse outcomes. For our analyses of patient outcomes, including patient satisfaction, we used the hospital-level proportion of nurses working each of the four shift-length categories. Patient outcomes data were available only at the hospital level. In our analyses, we estimated the effects of the four different shift category proportions—that is, the proportion of nurses working each shift category—on each patient outcome.
Job satisfaction was assessed using a four-point Likert scale–type question, “How satisfied are you with your job?” Responses, which ranged from very satisfied to very dissatisfied, were dichotomized to contrast satisfied and dissatisfied respondents.
Burnout was measured using the nine-item emotional exhaustion subscale of the Maslach Burnout Inventory, a reliable and valid instrument for assessing burnout in human service professions.11,12 Scores totaling twenty-seven or more are considered an indication of high emotional exhaustion,19 and our analysis contrasted nurses with burnout scores of twenty-seven and above and those with scores below that point.14,20–22
Nurses’ intent to leave their employer within the next year was assessed using a single yes/no item.
We used data on patients’ assessment of their care experience during short-term, acute care hospitalizations from the Hospital Consumer Assessment of Healthcare Providers and Systems survey that we found on the Hospital Compare website of the Centers for Medicare and Medicaid Services. The survey was conducted during 2006 and 2007, which was right in the middle of the period during which the Multi-State Nursing Care and Patient Safety Study data were collected (from late 2005 to early 2008). Thus patients and nurses were responding about hospital conditions at roughly the same time.
We analyzed all of the publicly reported items in the Hospital Consumer Assessment of Healthcare Providers and Systems survey, including two questions on global assessments of care; three questions on communication with doctors, nurses, and staff; three questions on nursing care; and two questions on hospital environment. All items were available only in aggregated form, and they were risk-adjusted for each hospital prior to release to the public.23
We used a variety of controls in our models to account for nurse characteristics, nursing organizational features, and hospital structural characteristics that have been shown to be related to our outcomes. For individual nurses, we included age and sex, characteristics that may influence intentions to leave.24 We controlled for the type of hospital unit to account for differences between intensive care and general care units in the acuity of patients’ conditions and nurse workloads. We controlled for two nursing organizational features—nurse staffing and the professional practice environment—that have been found to be associated with both nurse outcomes and patient satisfaction.20,25
Our measure of nurse staffing was calculated from the survey data as the average ratio of patients to nurses reported by the nurse respondents on their units during their last shift, aggregated to the hospital level. This method has been established in the literature.14,17
We also conducted parallel analyses using the American Hospital Association’s nurse staffing measure. We do not report these results because they were substantially similar to the results we obtained using the survey-based staffing measure. The nurses’ professional practice environment was measured using the Practice Environment Scale of the Nursing Work Index, endorsed by the National Quality Forum.26,27 The Technical Appendix18 describes how we categorized the environment based on the subscales of the Practice Environment Scale.
Finally, we used data from the American Hospital Association Annual Survey of Hospitals to control for hospital structural characteristics that have been associated with differences in patient outcomes.28–30 We defined high-technology hospitals as those that performed major organ transplant surgery, open heart surgery, or both. Teaching status was determined by the ratio of postgraduate medical residents to beds, and bed size was the number of licensed beds. Ownership, hospital state, and core-based statistical area—a measure of population density—were also included as hospital-level control variables in the patient satisfaction models.
Descriptive statistics were used to examine the major outcomes of interest and characteristics of nurses and hospitals by shift length category. First, we examined the bivariate relationship between nurses’ shift length and the nurse outcomes (burnout, job dissatisfaction, and intent to leave), using separate generalized estimating equation models. All comparisons between shift length categories were made using the category of 8–9 hours as the reference group.
Next, we added controls to the three nurse outcome models. All of the models used generalized estimating equations to take account of the clustering of nurses within hospitals.
Then we used the measures from the Hospital Consumer Assessment of Healthcare Providers and Systems survey to examine the relationship between the proportion of nurses in each hospital working each type of shift and patient satisfaction, using hospital-level data and estimating ordinary least squares regression models. After first estimating bivariate models, we added hospital-level controls to the models.
All analyses were completed using the statistical analysis software SAS, version 9.3. The significance level was p < 0:05 for a two-tailed test. This study was approved by the Institutional Review Board of the University of Pennsylvania.
Our study was cross-sectional, which limited our ability to make causal inferences about the effect of nurses’ shift length on nurse or patient outcomes. We attempted to include all known confounding factors in our models, but it is possible that some were not included.
In 2006–07, reporting on the Hospital Consumer Assessment of Healthcare Providers and Systems survey was voluntary. Therefore, the hospitals included in our analyses may not be representative of all hospitals. However, the controls in our models for hospital size, teaching status, and ownership adjust for some of the differences between the sampled and non-sampled hospitals that might otherwise represent a threat to external validity, or our ability to generalize our results.25
The response rate to the Multi-State Nursing Care and Patient Safety Study was 39 percent. In an intensive follow-up survey of a sample of 1,300 nonresponders, we achieved a 91 percent response rate. We found no major differences relevant to the variables of proximal interest in our study between the responses of the initial responders and those of the later responders.
Sixty-five percent of nurses worked shifts of 12–13 hours (Exhibit 1). Twenty-six percent worked shifts of 8–9 hours, and the remaining nurses worked shifts of either 10–11 hours or more than 13 hours. When we examined the hospital characteristics by shift length category, we found that nurses working shifts of twelve hours or longer were more numerous in teaching and high-technology hospitals.
Nurses working shifts that were twelve hours or longer were disproportionately male and nonwhite; their mean age was also lower than that of nurses in other categories of shift length (Exhibit 1). Nurses working in intensive care units were also more likely than medical or surgical nurses to work longer shifts.
The majority of nurses were non-Latino, white, and female. Most had less than a bachelor’s degree. In general, these demographics reflect the national profile of nurses found in the 2008 National Sample Survey of Registered Nurses, a survey conducted every four years to evaluate trends in the nursing profession.32
Most of the hospitals in the study had more than 100 beds; were high-technology institutions; and were minor teaching hospitals or non-teaching hospitals, rather than major teaching hospitals (Exhibit 1).
Across the four shift length categories, more than 80 percent of the nurses reported being satisfied with scheduling practices at their hospital (Exhibit 2). The percentages of nurses reporting burnout and an intention to leave the job increased incrementally as shift length increased. The percentage of nurses who were dissatisfied with the job was similar for nurses working the most common shift lengths, 8–9 hours and 12–13 hours, but it was higher for nurses working shifts of 10–11 hours and more than 13 hours.
We assessed the bivariate relationship between nurses’ shift length and nurse outcomes. Increases in shift length were associated with significant increases in the odds of burnout, job dissatisfaction, and intention to leave the job. The odds of burnout and job dissatisfaction were up to two and a half times higher for nurses who worked longer shifts than for nurses who worked shifts of 8–9 hours.
Even after we adjusted for potential confounding factors, the significant relationship between longer shift lengths and nurse reports of burnout and job dissatisfaction persisted. Compared to nurses who worked shifts of 8–9 hours, nurses who worked shifts of 10–11 hours had a greater likelihood of being burned out, being dissatisfied with the job, and intending to leave the job (Exhibit 3). So did nurses who worked shifts of 12–13 hours, although the impact of shift length was smaller for them than for nurses in the shift category of 10–11 hours. The odds of these unfavorable outcomes were highest for nurses who worked shifts of more than 13 hours.
We also found that nurses’ shift length was significantly associated with patient satisfaction, as measured by the Hospital Consumer Assessment of Healthcare Providers and Systems survey. In this analysis, we used a subset of nurses (n ¼ 16; 241) and hospitals (n ¼ 396), with an average of 41 nurses per hospital that participated in the survey.
In the fully adjusted linear regression models, we found that seven of the ten outcomes were significantly and adversely affected by the proportion of nurses in the hospital working shifts of more than thirteen hours, including both of the global assessments of care—patients’ rating of the hospital overall, and whether patients would recommend the hospital (Exhibit 4). That is, increases in the proportion of nurses working shifts of more than thirteen hours were associated with increases in patient dissatisfaction.
Although only 5 percent of the nurses in our sample worked shifts of more than thirteen hours, the hospitals in the sample varied considerably in terms of the percentages of their nurses working those shifts, from 0 percent to nearly 40 percent. A change from 0 percent to 40 percent of nurses working those shifts would result in nearly a five-percentage-point increase in the patients who gave the hospital a low rating.
We found a number of significant relationships between the other shift length categories and patient satisfaction. Notably, having higher proportions of nurses working shorter shifts—8–9 hours or 10–11 hours—resulted in decreases in patient dissatisfaction. Although not all of the coefficients in Exhibit 4 are significant, it is note-worthy that all of the significant ones associated with shifts of 12–13 hours and of more than 13 hours are positive (that is, they indicate more dissatisfaction), while all of the significant coefficients associated with shifts of 10–11 hours and 8–9 hours are negative (indicating less dissatisfaction).
This study provides new and valuable insights into the relationship between nurses’ shift length and patient and nurse outcomes. Specifically, we found that the longer the shift, the greater the likelihood of adverse nurse outcomes such as burnout. Patients were less satisfied with their care when there were higher proportions of nurses working shifts of thirteen or more hours and were more satisfied when there were higher proportions of nurses working eleven or fewer hours.
We also found that, despite poor nurse outcomes associated with longer shift lengths, nurses seem to be satisfied with their schedules. Most hospitals have transitioned to a regular use of the twelve-hour shift.
This study is the first to investigate the relationship between nurses’ shift length and patients’ assessment of care using the Hospital Consumer Assessment of Healthcare Providers and Systems survey. In hospitals in which a higher proportion of nurses reported working more than thirteen hours on their last shift, higher percentages of patients reported that they would not recommend the hospital to friends and family and gave the hospital lower overall ratings.
Furthermore, larger percentages of the patients in the hospitals with more nurses working the longest shifts reported that nurses sometimes or never communicated well; pain was sometimes or never well controlled; and they sometimes or never received help as soon as they wanted. For many patient outcomes, dissatisfaction decreased as the proportion of nurses working shifts of 8–9 hours or 10–11 hours increased. In sum, our results suggest that patients perceive worse care (overall and in nursing-specific domains) in hospitals that have high proportions of nurses working shifts of more than thirteen hours.
These findings are important given the changes being implemented in reimbursement by the Centers for Medicare and Medicaid. Beginning in fiscal year 2013, hospitals will face reductions in reimbursement if they do not meet national benchmarks on the two global measures of care on the Hospital Consumer Assessment of Healthcare Providers and Systems survey used in this study. Nurse working conditions, including shift length, is one area related to these benchmarks that we believe is readily amenable to change.
Nurses’ shifts of more than thirteen hours were also associated with greater likelihood of nurse burnout, job dissatisfaction, and intention to leave the job. The current literature cites twelve-hour shifts as a way to recruit and retain nurses because it is the preferred shift length among nurses.1,9We found that most nurses said they were satisfied with their schedule and that the majority of our sample worked shifts of at least twelve hours. But we also found that the nurses who worked shifts of 12–13 hours were more likely to intend to leave the job than nurses who worked shorter shifts, contrary to what the literature suggests.
One possible explanation for the findings is that nurses underestimate the impact of working long shifts because the idea of working three days a week instead of five seems appealing. Working longer but fewer shifts may also attract nurses who work a second job. However, the strain of those three long work days and the rest and recovery time needed may offset any perceived benefit, if our survey results are any indication. When a three-day week turns into more days or additional, unplanned-for overtime, nurses’ satisfaction appears to decrease.
Our findings contribute to a growing body of research associating nurses’ shift length with patient safety issues.5,7,8,10 The results also highlight an area of health care ripe for policy development at both national and institutional levels.
Policies by the Accreditation Council for Graduate Medical Education have resulted in mandatory restrictions in resident physicians’ duty hours to no more than eighty hours a week. Our results suggest that similar policies for nurses—perhaps restricting the number of consecutive hours worked—should be considered by accrediting bodies, such as the Joint Commission. Shift lengths could also become reportable evidence for magnet recognition by the American Nurses Credentialing Center, which denotes nursing excellence in hospitals.
In 2011 a “sentinel event” alert was issued by the Joint Commission,33 which called for hospitals to intensify their efforts to monitor and address health care workers’ risk for fatigue caused by extended shifts. A “sentinel event” indicates the risk for, or an actual unexpected occurrence of, death or serious injury in the health care setting. The Joint Commission recommended nine evidence-based actions, including assessment of off-shift hours and consecutive shifts worked, and the inclusion of staff in the design of work schedules to reduce risk for fatigue. Now a matter of hospitals’ policing themselves, these recommendations may become part of accrediting standards.
At the state level, boards of nursing should consider whether restrictions on nurse shift length and voluntary overtime are advisable. This idea has been raised in the past, but it has been met with emphatic pushback from nurses who wish to maintain the status quo.34
In addition, policies should be adopted by nursing management within hospitals to monitor nurses’ hours worked, including hours worked in second jobs. For example, chronic understaffing at a hospital may lead to nurses’ working longer shifts or overtime to make up for the staffing shortages. Hospitals that do not require overtime but are short staffed may also face quality-of-care issues. At a minimum, hospital administrators should establish practices designed to comply with the Institute of Medicine’s recommendations to limit nurses’ work hours to twelve hours in a twenty-four-hour period and to sixty hours in a week.35
Nursing leadership should also encourage a workplace culture that respects nurses’ days off and vacation time, promotes nurses’ prompt departure at the end of a scheduled shift, and allows nurses to refuse to work overtime without retribution.36 These types of policies that facilitate manageable work hours can contribute to the development of a healthier nursing workforce, prepared to manage the complex care needs of patients and their families.
This research was supported by the National Institute of Nursing Research (Grant Nos. T32-NR-007104 and R01-NR-004513) and the National Institute for Occupational Safety and Health (Grant No. T01-OH-008417-05). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institute of Nursing Research or the National Institute for Occupational Safety and Health.
Amy Witkoski Stimpfel is a research fellow at the University of Pennsylvania School of Nursing.
In this month’s Health Affairs, Amy Stimpfel and coauthors report on their study of the length of hospital nurses’ shifts, nurses’ job dissatisfaction and intention to leave their jobs, and patients’ dissatisfaction. Although survey data from four states showed that four out of five nurses were satisfied with scheduling practices at their hospital, nurses working shifts of ten hours or longer were up to two and a half times more likely than nurses working shorter shifts to experience burnout and job dissatisfaction. What’s more, as the proportion of hospital nurses working shifts of more than thirteen hours increased, patients’ dissatisfaction with care also increased. The authors observe that policies regulating work hours for nurses, similar to those set for resident physicians, may be warranted.
Stimpfel is a research fellow at the Center for Health Outcomes and Policy Research at the University of Pennsylvania School of Nursing. Her research is focused on nurses’ working conditions and patient and nurse outcomes. Stimpfel holds a doctorate in nursing from the University of Pennsylvania.
Douglas M. Sloane is an adjunct professor at the University of Pennsylvania School of Nursing.
Douglas Sloane is a sociologist and adjunct professor at the University of Pennsylvania School of Nursing. He is also assistant director, social science analyst, and supervisory statistician at the Government Accountability Office, in Washington, D.C. Sloane has a master’s degree and a doctorate, both in sociology, from the University of Arizona.
Linda H. Aiken is the Claire M. Fagin Leadership Professor of Nursing at the University of Pennsylvania School of Nursing.
Linda Aiken is the Claire M. Fagin Leadership Professor of Nursing, and a professor of sociology, at the University of Pennsylvania School of Nursing. She also directs the Center for Health Outcomes and Policy Research and is a senior fellow at the Leonard Davis Institute of Health Economics at the university’s Wharton School.
Aiken’s research focuses on the outcomes of nursing care in the United States and abroad. She has a doctorate in sociology from the University of Texas at Austin and is a member of the editorial board of Health Affairs.
Portions of this article were presented at the Academy Health Annual Research Meeting, Orlando, Florida, June 23–24, 2012.
Amy Witkoski Stimpfel, Research fellow at the Center for Health Outcomes and Policy Research at the University of Pennsylvania, School of Nursing, in Philadelphia.
Douglas M. Sloane, Adjunct professor at the University of Pennsylvania, School of Nursing.
Linda H. Aiken, Claire M. Fagin Leadership Professor of Nursing, a professor of sociology, and director of the Center for Health Outcomes and Policy Research at the University of Pennsylvania, School of Nursing.