quantitative correlational study

OBJECTIVE: The aim of this study is to explore the relationship between reported sleep, perceived fatigue and sleepiness, and cognitive performance. BACKGROUND: Although evidence suggests that fatigue and sleepiness affect the provision of care in inpatient units, there is a lack of research on the sleep patterns of emergency nurses and the effects of dis- turbed sleep and fatigue on their cognitive abilities and susceptibility to medical errors. METHODS: A quantitative correlational design was used in this study; in each of 7 different statistical models, zero-order relationships between predictors and the dependent variable were examined with appropriate inferential tests. RESULTS: Participants reported high levels of sleepiness and chronic fatigue that impeded full functioning both at work and at home. CONCLUSIONS: Although high levels of self-reported fatigue did not show any effects on cognitive function, other factors in the environment may contribute to delayed, missed, or inappropriate care. Further research is indicated.

Studies of worker fatigue in the military and com- mercial trucking industries1,2 suggest that long hours, especially extending into overnight, can be hazardous in terms of the potential for errors due to sleepiness and reduced vigilance. Medical providers, both nurses and physicians, are at risk because of the need for around-the-clock coverage and the number and type of decisions they must make in a given time span; in 1989, the Bell Commission Report cited sleep depri- vation in medical house officers as a major contrib- utor to the 1984 death of Libby Zion, an 18-year-old who died at New York Hospital, leading to a restruc- turing of the hours involved in medical residencies.3

The long (up to 36 hour) shifts of the medical resi- dents in charge of her care and the consequences of their fatigue on the decisions that were made were cited as factors in her death. As a result of their inves- tigation of the case, the Bell commission recommen- ded limiting the work hours of medical residents to less than 80 hours a week and no more than 24 hours in a row, and subsequent research has led to similar recommendations for the nursing workforce.4,5

Emergency care settings are chaotic environ- ments, where there is high patient turnover, constantly changing priorities and frequent changes in patient condition. High demand work settings are associated with increased fatigue, which can impair nurses_ at- tentiveness and ability not only to recognize potential errors they might commit but also to recognize and mitigate the errors of others, including physicians.6,7

The work schedules of emergency nurses (ENs) are characterized by increasing demands, irregular shifts, and long hoursVall of which challenge their mental and physical stamina. Although 12-hour shifts are

JONA � Vol. 47, No. 1 � January 2017 41

Author Affiliations: Director, Institute for Emergency Nursing Research (Dr Wolf); Senior Research Associate (Mss Perhats and Delao), Emergency Nurses Association, Des Plaines; and Research Assistant Professor (Dr Martinovich), Department of Psychiatry, Northwestern University Feinberg School of Medicine, Evanston, Illinois.

The authors declare no conflicts of interest. Correspondence: Dr Wolf, Institute for Emergency Nursing

Research, Emergency Nurses Association, 915 Lee St, Des Plaines, IL 60016 (lisa.wolf@ena.org).

Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal_s Web site (www.jonajournal.com).

DOI: 10.1097/NNA.0000000000000435

Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved.

common, nurses frequently work additional hours; work days may last 16 hours or more, often spread- ing across evenings and nights with short intervals between shifts. More than a decade ago, Rogers et al8

presented compelling evidence of increasing incident/ error rates during extended periods of work: Rates rise after 9 hours, double after 12 consecutive hours, and triple by 16 hours. Further studies have demon- strated that insufficient sleep and inadequate recov- ery time from long hours and shift work affect a nurse_s ability to provide safe, effective patient care and increase risks to personal safety.9-11 The Joint Commission_s 2011 Sentinel Event Alert on health- care worker fatigue and patient safety states that BShift length and work schedules have a significant effect on healthcare providers” quantity and quality of sleep and, consequently, on their job performance, as well as on the safety of their patients and their individual safetyIStill, while the dangers of extended work hours (912 hours) are well known, the healthcare industry has been slow to adopt changes, particularly with regard to nursingI[12(p1) Multiple studies support that a significant number of inpatient nurses have im- paired sleep quality, excessive sleepiness, and abnormal fatigue, which are associated with a greater risk of making medical errors and potentially causing harm to patients.13-15 The research of Rogers and colleagues8

demonstrated that the 12-hour shift is past the point where safe decision making may be expected, and both the Agency for Health Research and Quality4

and the Institute of Medicine5 have since recommended limiting the length of nurses_ shifts. Although evidence suggests that both mental and physical fatigue and sleepiness affect the provision of care in inpatient units,16 there is a lack of research on the sleep patterns of ENs and the effects of disturbed sleep and fatigue on their cognitive abilities and susceptibility to making medical errors. The purpose of this study is to explore the relationship between reported sleep quality, per- ceived fatigue and sleepiness, and cognitive perfor- mance in a sample of ENs.

Methods

Before recruitment of subjects, institutional review board approval was obtained and the study was approved (Chesapeake Research Review, Columbia, Maryland). A Certificate of Confidentiality from the National Institutes of Health was obtained to further protect the respondents_ anonymity given the sensi- tive nature of the question. A quantitative correla- tional design was used to explore the relationship between nurses_ reported sleep patterns, perceived sleepiness and fatigue, and their ability to carry out timed cognitive tasks. As part of a 35-item online

questionnaire, participants performed 3 medication dosage calculations and correlations between fatigue and cognitive performance were measured in the same survey. Predictors were selected based on empirically based associations with work fatigue and its impact on task performance.17 In each of 7 different statistical models, zero-order relationships (ie, simple relationships that do not hold constant other predictors) between each of the predictors and the dependent variable were examined with appro- priate inferential tests.

Sample

Of the 2,419 ENs who were recruited via e-mail and social media, 2010 met eligibility criteria and consented to participate. Eligibility criteria required that study participants were English speaking, currently licensed ENs working at least 1 shift per week in a US emergency department (ED). Among those who consented, 504 survey respondents were excluded because they did not meet study criteria (eg, worked 1 shift or less in the past 30 days, did not complete survey in 1 sitting), resulting in a total sample of 1506 ENs.

Research Questions

On the basis of the existing body of evidence, we hy- pothesized that ENs_ patterns and quality of sleep, sleepiness, and fatigue would have an attenuating effect on their performance in timed cognitive skill tests. To investigate these effects, we posed the fol- lowing research questions:

Q1: Is there a relationship between ENs_ day- time sleepiness and performance in timed cognitive tests?

Q2: Is there a relationship between ENs_ patterns/quality of sleep and performance in timed cognitive tests?

Q3: Is there a relationship between ENs_ fatigue and performance in timed cognitive tests?

Data Collection

Using a Web platform (Qualtrics, Provo, Utah), study participants were asked to complete an online survey that included questions about participant and hospital demographics (Tables 1 and 2), nurses_ work sched- ules, reported sleep patterns and levels of fatigue, cog- nitive performance measures, and routine activities of daily living. Data on sleep patterns, sleep quality, and fatigue were collected using the following 3 validated and reliable instruments that were incorporated into the online survey.

Q1 Instrument: The Epworth Sleepiness Scale (ESS) is an 8-item self-rated questionnaire that measures the general level of daytime sleepiness. Each item is scored 0 to 3, giving a total score of 0 to 24, which is a measure

42 JONA � Vol. 47, No. 1 � January 2017

Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved.

of the subject_s average sleep propensity (ASP) in those 8 situations. An ESS score of 10 is most often consi- dered to be the upper limit of normal. There is good evidence for the validity of total ESS scores as a mea- sure of ASP and a high level of internal consistency within the ESS18 (Cronbach_s ! = .88-.74 in 4 different groups of subjects).

Q2 Instrument: We included 2 subscales (9-item subscale: Bhow often have you had trouble sleeping[ and 6-item subscale: Bhow often have you used the following substances to help you get to sleep[) from the Pittsburgh Sleep Quality Index (PSQI), which is a 19-item well-validated self-rated questionnaire for evaluating subjective sleep quality over the previous month19 (Cronbach_s ! = .83). We developed a third 8-item subscale to assess Bhow often has it been difficult to do the following things because you were fatigued[ (internal consistency and reliability coefficient = .85).

Q3 Instrument: The Occupational Fatigue Ex- haustion Recovery (OFER15) Scale is a self-report questionnaire of 15 questions, which form 3 sub- scales: chronic fatigue, acute fatigue, and intershift recovery, respectively. The OFER Scale produces

comparable scores between 0 and 100 for each subscale, with cut-points into levels of low, low/ moderate, moderate/high, and high on each subscale used to compute quartiles of scale score distribution. The subscales have high internal reliability (9.84) and face, construct, and discriminant validity.20

Timed cognitive performance was measured by the speed and accuracy for 3 weight-based medication calculations that are frequently performed in emer- gency nursing practice. The first 2 medication ques- tions required calculation of the initial bolus and the total dose for a thrombolytic medication order for an older adult patient with a diagnosis of ischemic stroke. The 3rd medication question required calculation of the 1st dose of an antibiotic for a pediatric patient with a diagnosis of strep throat. Survey participants were instructed to complete the survey in 1 sitting and without interruption. This was done so that we could determine the speed and accuracy of the medication calculations in relation to their perceived state of fatigue at that time. The online survey software recorded start and stop times for the medication calculations and for completion of the questionnaire. Seven people were eliminated from the study sample because their response times were extreme outliers. All results were scored and compiled in an SPSS database (version 22; Armonk, New York) for analysis.

Predictors

Predictors were selected because of a presumed impact of work fatigue on cognitive performance. Interval- scaled predictors included the standardized items (based on analytic sample norms) from each of the 3 instruments used in this study. We examined 29 categorical predictors including work schedules (total hours in the past 30 days), shift types (length

Table 1. Participants” Demographics (n = 1506)

Percentage of Survey Respondents

Gender Male 15.5 Female 84.5 Intersex 0.1 Missing 0

Age 18-24 2.0 25-34 23.0 35-44 24.0 45-54 24.9 55-64 25.0 Q65 1.1 Missing 0

Primary ED role Staff nurse 68.9 Charge nurse 16.3 Case manager 0.1 Clinical coordinator 1.8 Clinical educator 2.7 Clinical nurse specialist 0.6 Director 1.4 Manager 3.1 Nurse practitioner 1.6 Trauma coordinator 0.8 Other 2.7 Missing 0

Years of experience Mean (SD) In nursing 17.5 (12.5) In emergency nursing 12.8 (10.1) In current ED 8.2 (8.4) In all areas of emergency care,

excludingnursing (eg, tech, etc) 5.9 (9.3)

Table 2. Facility Characteristics

Percentage of Survey Respondents

ED patient population General ED 86.0 Adult only 9.4 Pediatric only 4.6 Missing 0

Facility type Nongovernment, not-for-profit 69.0 Investor owned, for-profit 18.8 State or local government 9.4 Federal government/VA/military 2.9 Missing 0

ED geographic location Urban 39.9 Suburban 34.9 Rural 25.2 Missing 0

JONA � Vol. 47, No. 1 � January 2017 43

Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved.

and time of day), sleep quality, and other life stressors that could potentially contribute to fatigue (eg, 2nd job, primary caregiver of children, and/or dependent adults). Interval-scaled predictors included 14 survey items regarding self-perceptions of fatigues, sleepiness, and demographic characteristics such as age and year of emergency nursing experience. Linear predictors were standardized so that all effect estimates refer to the impact of a 1 standard deviation change. Estimates for 2 interval-scaled predictors were also obtained for untransformed scores.

Data Analysis

Correlational analyses were performed using speed and accuracy of the medication dosage calculations as the dependent variable. Speed and accuracy data were examined via 7 statistical models: OLS regres- sion model predicting time in minutes on the page; Cox regression model predicting time to fully correct response (3/3 questions); Cox regression model predict- ing time to partially correct response (2/3 questions); 3 logistic regression models predicting the probability of a correct response for each individual medication dosage calculation; and a polytomous logistic regres- sion (PLUM) model predicting the number of correct responses (of 3).

In each of 7 different statistical models, zero- order relationships (ie, simple correlations or simple contrasts of group vs full sample) between each pre- dictor and the dependent variable (time and accuracy of response) were examined with appropriate inferen- tial tests (relative ratio, odds ratios, Wald z tests, P values). Probability estimates were calculated for each group and for the full sample. Inferential tests con- trasted each group_s estimated probability of a correct response to the medication calculation versus all other groups combined (ie, contrasting each group versus the full sample) and results were reported with associated Wald z tests and P values.

Results

Descriptive Statistics

In this sample of 1506 participants who provided valid responses on all items, it took an average of 21.6 minutes to complete the 34 items in the online questionnaire. Cog- nitive performance was measured by 3-timed medica- tion dosage calculations (see Table, Supplemental Digital Content 1, http://links.lww.com/JONA/A499). Overall, respondents averaged 3.78 minutes to com- plete the problem set and the average number of cor- rect answers was 1.95 of 3. More specifically, the percentages with 0, 1, 2, or 3 correct answers were 11.2% (n = 168), 17.4% (n = 262), 36.5% (n = 549), and 35.0% (n = 527), respectively.

Work Schedules

The primary work schedule for this group of nurses consisted of 12-hour shifts (74%), with a mean shift length of 11 T 2 hours and an average of 152 T 50 hours worked in the preceding 30 days. Survey data indicate that the mean number of shifts worked in that period was 13 T 4, with 30% of the nurses working all those shifts consecutively and 65% working at least some consecutive shifts. The re- ported numbers of extended work shifts ranged from 0 (39%), 1 to 3 (30%), 4 to 10 (25%), and 11 to 24 (5%). The primary reason given for working an extended shift was to meet patient care needs (47%).

Sleep, Fatigue, and Cognitive Performance

Study participants reported varying levels of sleepi- ness, sleep disturbance, and fatigue as measured by their responses on the 3 scaled instruments. Most nota- bly, 4 of every 10 nurses reported levels of daytime sleepiness above the upper limit of normal (an ESS score of Q10; sample mean, 9.5). In addition, most nurses reported moderately high (60-75) to high (975) levels of fatigue, with 50% of chronic and 75% of acute subscale scores falling between 60 and 100 (Figure 1). Roughly half (51%) of participants also had low to low/moderate scores on intershift recovery (score of 0-50), indicating that they are vulnerable to developing chronic or persistent fatigue.

Thirty-five percent of respondents rated their sleep quality as bad or fairly bad during the preceding 30 days; 26.8% reported that they slept an average of 5 hours or less per night and used various over-the- counter or prescribed medications, herbal remedies, and alcohol as sleep aids. A substantial number of nurses reported that their fatigue resulted in occasional or frequent difficulties with activities of daily living, such as driving (38.1%), eating regular meals (65.1%), exercising (66%), managing stress 76.3%), and attend- ing to personal (67%) and family (69.6%) needs (Table 3). Nearly half (46.8%) of study participants reported that their fatigue made it difficult to pro- vide safe patient care at least some of the time.

Figure 1. Reported scores for chronic and acute fatigue.

44 JONA � Vol. 47, No. 1 � January 2017

Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved.

Inferential Statistics

Seven different statistical models were used to examine zero-order relationships between each of the predictors and the dependent variable (speed and accuracy of response). Analysis of the interval-scaled predictors for sleep patterns, sleep quality and fatigue (ESS, modified PSQI, OFER15) did not detect any significant zero- order effects within any of the 7 models (Table 4). Nurses_ work schedules, caregiver status, and working a 2nd job did show a significant effect on cognitive performance. Nurses who worked the 7 AM to 7 PM shift (n = 511) showed significantly poorer response accuracy on several indicators (Table 5). These sub- jects had a 25% lower probability (time-adjusted) of a fully correct response set versus other nurses (mean, 1.87 vs 1.95; odds ratio [OR], 0.78; z = j2.45, P = .14). They were also more likely to give an incorrect answer for the

pediatric dose calculation (mean, 41.3% vs 47.3%; z = j3.36, P = .001). In contrast, nurses who worked the 7 PM to 7 AM shift (n = 418) had better time-adjusted accuracy estimates based on the 2 Cox regression models. Their probability of generating a fully correct response was 1.27 times greater (z = 2.48, P = .013). Similar effects were found for subjects with a 2nd part-time or full-time job (n = 426) and for those who were primary caregivers of children (n = 541), with both groups completing the problem set faster and more accurately than those with no additional job or caregiver responsibilities.

Discussion

Based on the existing body of evidence, we hypoth- esized that ENs_ sleep pattern, sleep quality, and

Table 3. Activities of Daily Living (ADL): Functional Impact of Fatigue (Modified PSQI Subscale)

In the past month, how often has it been difficult for you to do the following things because you were fatigued? (valid percent; internal reliability and consistency coefficient = .85)

Activity Never Difficult, % Sometimes Difficult, % Frequently Difficult, % Always Difficult, %

Engage in social activities 14.8 46.9 28.1 10.2 Manage stress 17.5 52.9 23.4 6.2 Attend to family needs 25.8 48.2 21.4 4.6 Eat regular meals 21.9 35.7 29.4 13.0 Provide safe patient care 52.7 43.0 3.8 0.5 Safely operate a motor vehicle 61.2 33.8 4.3 0.7 Exercise 10.8 32.4 34.0 22.8 Attend to personal needs 27.6 45.4 21.6 5.4

Table 4. Effect of Sleep and Fatigue on Responses to Medication Problem Set: 7 Statistical Models for Interval Scale Predictorsa (n = 1506)

DV1: Correlation With Time (Minutes

to Complete Problem Set)

DV2: RR of Fully Correct

Response (3/3)

DV3: RR of Partially

Correct Response (2/3)

DV4: % Correct

(Activase Initial)

DV5: % Correct

(Activase Total)

DV6: % Correct

(Amoxicillin 1st Dose)

DV7: Number Correct

(1-3)

r RR RR OR OR OR OR

Daytime sleepiness and fatigue Daytime sleepiness 0.00 1.02 1.00 1.06 0.96 1.01 1.02 Chronic fatigue 0.04 0.93 0.95 0.95 0.97 0.95 0.95 Acute fatigue j0.01 1.02 1.02 0.97 1.02 1.00 1.00 Persistent fatigue 0.01 1.01 0.99 1.02 0.98 1.03 1.01 ADL Functional

Impact Scale j0.03 1.05 1.04 0.99 0.95 1.06 1.01

Sleep patterns and quality Average hours of sleep j0.02 1.02 1.02 0.98 0.97 1.02 0.99 Average hours of

sleep (not standardized)

j0.02 1.02 1.02 0.98 0.97 1.02 0.99

Sleep quality j0.03 1.03 1.04 1.04 1.03 1.01 1.02 Sleep quality (not standardized)

j0.33 1.05 1.06 1.07 1.04 1.01 1.03

Abbreviations: OR, odds ratio; RR, relative risk. aNo significant deviations versus full sample for r, RR, or OR estimates were detected.

JONA � Vol. 47, No. 1 � January 2017 45

Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved.

fatigue would affect their performance on timed tests of cognitive skills that are commonly required in EN practice. Descriptive and inferential analyses performed on this data set describe frequent sleep loss, poor sleep quality, and difficulty with activities of daily living due to fatigue (Table 3). Twelve-hour work shifts were the norm among our study participants (74% of reported shifts), with an average shift length of 11 hours. Thus, any extension of work schedules places these nurses at increased risk of potential harm to patients and themselves.8-11 Of particular concern is the finding that nearly half (46.8%) of study participants reported that their fatigue made it difficult to provide safe patient care at least some of the time.

Previous researchers have reported higher error rates among nurses who worked rotating shifts21 and night shifts.13 In contrast, nurses in our study who worked the 7 AM to 7 PM day shift showed signifi- cantly poorer response accuracy when compared with nurses who worked any other shift (Table 5).

Participants who reported having specific stressors (ie, being a primary caregiver or working a 2nd full or part-time job in addition to their ED employment) answered the problem set with greater speed and accuracy than did those who did not have those res- ponsibilities. These unexpected findings stimulate a number of alternative hypotheses that have been raised in the literature and are worth further consideration.

Effects of Job Demands and High-Stress Environments

There is a substantial body of literature on the rela- tionship between job performance, fatigue, and ex- posure to repetitive, high levels of stress without adequate recovery between episodes.9,13-24 Our sample of ENs reported high levels of acute fatigue, low intershift recovery, and difficulties with manag- ing their daily routine (eg, eating regular meals, driv- ing safely, providing safe patient care), consistent with previous research that suggests that workers

Table 5. Effect of Shift Type on Responses to Medication Problem Set: 7 Statistical Models for Categorical Predictors

Survey question: During the past month, on which shift did you primarily work in your ED? (n = 1506)

DV1: Time in Minutes to Complete Problem Set

DV2: Relative RiskVFully Correct Response (3/3)

DV3: Relative RiskVPartially Correct (2/3)

n Mean t P RR Wald z P RR Wald z P

7 AM-3 PM 98 3.82 0.20 .842 1.07 0.38 .706 1.09 0.73 .463 7 AM-7 PM 511 3.90 1.36 .174 0.75 j3.04 .002 0.89 j1.85 .065 7 PM-7 AM 418 3.58 j1.94 .052 1.27 2.48 .013 1.15 2.06 .040 11 PM-7 AM 21 4.30 1.00 .316 0.45 j1.59 .112 0.71 j1.21 .226 11 AM-11 PM 136 3.63 j0.76 .446 1.08 0.54 .587 1.10 0.88 .380 3 PM-11 PM 31 3.61 j0.39 .694 1.81 2.40 .016 1.12 0.52 .606 3 PM-3 AM 51 3.67 j0.31 .758 0.96 j0.15 .883 0.93 j0.40 .687

DV4: % Correct (Activase Initial)

DV5: % Correct (Activase Total)

DV6: % Correct (Amoxicillin 1st Dose)

n % Wald z P % Wald z P % Wald z P

7 AM-3 PM 98 78.6 1.19 .236 79.6 1.19 .233 48.0 0.13 .900 7 AM-7 PM 511 72.8 j0.40 .686 73.2 j0.84 .403 41.3 j3.36 .001 7 PM-7 AM 418 74.2 0.39 .694 76.1 0.87 .385 49.5 1.05 .294 11 PM-7 AM 21 66.7 j0.70 .481 61.9 j1.32 .188 42.9 j0.41 .679 11 AM-11 PM 136 74.3 0.23 .819 72.8 j0.48 .632 52.9 1.37 .171 3 PM-11 PM 31 80.6 0.91 .362 74.2 j0.04 .968 64.5 1.90 .058 3 PM-3 AM 51 58.8 j2.36 .018 58.8 j2.56 .010 52.9 0.81 .416

DV7: Number of Correct Responses (Out of 3)

n Mean 0, % 1, % 2, % 3, % OR Wald z P

7 AM-3 PM 98 2.04 9.6 15.7 35.9 38.8 1.19 0.92 .355 7 AM-7 PM 511 1.87 12.8 19.1 36.7 31.4 0.78 j2.45 .014 7 PM-7 AM 418 2.01 10.2 16.3 36.2 37.3 1.15 1.31 .191 11 PM-7 AM 21 1.69 16.9 22.4 35.8 24.9 0.61 j1.24 .217 11 AM-11 PM 136 2.00 10.4 16.5 36.2 36.9 1.10 0.55 .581 3 PM-11 PM 31 2.24 6.6 11.7 32.7 49.0 1.81 1.73 .084 3 PM-3 AM 51 1.73 15.8 21.7 36.1 26.4 0.66 j1.63 .102

Data in bold indicate significant deviations versus the full sample.

46 JONA � Vol. 47, No. 1 � January 2017

Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved.

with similar levels of maladaptive chronic fatigue are more vulnerable to the detrimental consequences of work-related stress and fatigue, with adequate re- covery being a key mediating factor.24-26 Additional results from 2 Australian studies found that the psycho- logical strain that nurses experience in high-paced, demanding jobs and lower levels of fulfilling non- work activities were associated with poorer sleep quality and reduced ability to recover between shifts.27,28

Similarly, the unexpected finding that 12-hour night shift workers performed better than 12-hour day shift workers could be a function of differences in environmental stress both within and between hospitals. Possibly, the night shift manages fewer interruptions from other hospital departments and fewer distractions owing to comparatively fewer workplace and personal demands than are experienced on the day shift. Re- search suggests that fatigue, along with inadequate rest and recovery between ED shifts, may reduce ENs_ ability to be vigilant and recognize the potential for error.29 In his examination of the effects of sleep depri- vation on cognition, Killgore17 concluded that sleep loss has a differential effect on specific cognitive and emotional processes, with some indication that higher- level cognitive capacities (eg, memory, judgment, decision-making) might remain degraded despite resto- ration. The frequent extension of 12-hour shifts, inadequate time off between shifts, and constant exposure to demanding work may have contributed to difficulties with performing some nursing tasks (eg, medication calculations) within our sample of ENs, nearly half of whom reported that fatigue made it difficult for them to provide safe patient care.

Variations in nurses” vigilance between day and night shifts and/or over the duration of individual shifts may have contributed to performance differ- ences in our sample. Surani and colleagues14 found that vigilance improved over the duration of the shift; in particular, fast reaction times were significantly shorter in floor nurses after shift than before it. The authors concluded that Bthere may also be impair- ment related to the intensive care unit (ICU) setting, where higher patient acuity requires even more vigilance,[ suggesting that the high-acuity ED envi- ronment may have an attenuating effect on ENs” job performance. Because we did not collect data on these types of parameters, we were not able to assess similar effects that could have contributed to our findings. Given the considerable costs of work-related fatigue (eg, injuries), further research is warranted using chronobio- logical objective sleep measures to more effectively evaluate the effects of sleep deprivation30 as well as recommended strategies (eg, work hour limits)31

that could improve the health and safety of the EN workforce and the patients they care for.

Limitations

The sample was a self-selecting group of ENs who reported on their sleep patterns and levels of fatigue; no objective measures of fatigue were collected. We also did not collect data on environmental factors that may have affected results. In addition, although the study instruments have been widely used, to our knowledge, they have not been used to measure work-related fatigue among ENsVa subgroup that reports high levels of acute and chronic fatigue that may be above those found in the general nursing population. As an example, researchers found a greater discrepancy in test-retest ESS scores among patients with possible sleep-disordered breathing than has been previously reported in studies that examined reproducibility only in normal subjects.32

Similarly, high levels of acute and chronic fatigue within our study population could have resulted in limited ability to detect effects of poor sleep quality and fatigue on ENs” cognitive function. Moreover, because we did not collect data on variables such as work environment, vigilance, or the timing of survey completion (eg, whether study participants completed the medication problem set during midshift, postshift, or on a day off), we can neither support nor reject the null hypothesis that work-related fatigue has an attenuating effect on ENs_ cognitive function.

Conclusions

A significant percentage of our sample of 1,506 ENs reported high levels of sleepiness and chronic and acute fatigue that impeded full functioning both at work and at home. Although we could not determine from this study whether levels of self-reported fatigue affect cognitive function, participants did report difficulty with providing both self-care and patient care. Other factors in the environment may contrib- ute to the effects of EN fatigue and resultant delayed, missed, or inappropriate care. Further research on the effect of ENs_ work schedules, physical and emo- tional fatigue, patterns and quality of sleep, and work- place conditions on their ability to safely and effectively perform their jobs in the high-demand environment of the ED is required. There is compelling evidence to suggest that fatigue reduction interventions could contribute to overall improvements in the delivery of care, including reduction in medical errors.31 Nurse administrators may consider a fatigue management reduction system as described by Lerner et al in the guidance statement of the American College of Occu- pational and Environmental Medicine (ACOEM). The ACOEM, the Canadian Nurses Association, and the American Nurses Association33-35 offer specific sug- gestions to recognize and mitigate fatigue, including limiting weekly and consecutive work hours (eg, no

JONA � Vol. 47, No. 1 � January 2017 47

Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved.

more than 40 hours in a 7-day period and limit work shifts to 12 hours in a 24-hour period)35; develop- ing processes to document fatigue in the workplace and its relationship to overtime, maximum hours worked per day and per week, on-call hours, and data related to patient error, staff retention levels, and recruitment results; developing policies that provide time and space for rest periods, meals, and other health

promotion initiatives for sleep hygiene; educating nursing staff and management in recognizing and managing fatigue in self and others, to include understanding the science of sleep, the risks associated with fatigue, and approaches to circadian rhythm disturbances; and providing sleep facilities to enable nurses to minimize their circadian disruptions during evening and night shift work.

0 replies

Leave a Reply

Want to join the discussion?
Feel free to contribute!

Leave a Reply

Your email address will not be published. Required fields are marked *