Factors associated with nurse well-being in relation to electronic health record use: A systematic review

Abstract

Objective

Although nurses comprise the largest group of health professionals and electronic health record (EHR) user base, it is unclear how EHR use has affected nurse well-being. This systematic review assesses the multivariable (ie, organizational, nurse, and health information technology [IT]) factors associated with EHR-related nurse well-being and identifies potential improvements recommended by frontline nurses.

Materials and Methods

We searched MEDLINE, Embase, CINAHL, PsycINFO, ProQuest, and Web of Science for literature reporting on EHR use, nurses, and well-being. A quality appraisal was conducted using a previously developed tool.

Results

Of 4583 articles, 12 met inclusion criteria. Two-thirds of the studies were deemed to have a moderate or low risk of bias. Overall, the studies primarily focused on nurse- and IT-level factors, with 1 study examining organizational characteristics. That study found worse nurse well-being was associated with EHRs compared with paper charts. Studies on nurse-level factors suggest that personal digital literacy is one modifiable factor to improving well-being. Additionally, EHRs with integrated displays were associated with improved well-being. Recommendations for improving EHRs suggested IT-, organization-, and policy-level solutions to address the complex nature of EHR-related nurse well-being.

Conclusions

The overarching finding from this synthesis reveals a critical need for multifaceted interventions that better organize, manage, and display information for clinicians to facilitate decision making. Our study also suggests that nurses have valuable insight into ways to reduce EHR-related burden. Future research is needed to test multicomponent interventions that address these complex factors and use participatory approaches to engage nurses in intervention development.

INTRODUCTION

The Agency for Healthcare Research and Quality defines burnout as “a long-term stress reaction marked by emotional exhaustion, depersonalization, and a lack of sense of personal accomplishment.”¹ It is often discussed alongside professional well-being as the 2 constructs have many overlapping factors.² Research on these topics has recently grown, especially as leaders studying clinician well-being advocate for reforming the “triple aim” model to include a fourth pillar focused on clinician well-being.³ Burnout is observed across clinical specialties, including physicians,^4,5 nurses,^6–9 and physician assistants.^10,11 Although some attributing factors may be shared among the various healthcare professionals, systematic differences may arise as a result of different roles in the U.S. healthcare system. For example, staffing ratios and job autonomy reportedly affect nurse burnout more so than physician burnout.²

Even though nurses outnumber physicians by 4 to 1,¹² studies in MEDLINE on “physician burnout” outnumber those on “nurse burnout” by 4 to 1. Additional attention on nurse burnout is needed. Among nurses, the estimated prevalence rate of burnout ranges from 15% to 45%.^6–9 Because burnout is associated with turnover and the U.S. reports a nursing shortage,^13–15 failure to address nurse burnout will likely have multilevel consequences on patient care. For instance, high nurse burnout is associated with higher healthcare-associated infection rates,¹⁶ and worse perceived quality of care provided.¹⁷ It is indirectly related to higher mortality rates by way of higher patient-to-nurse ratios and higher turnover.^17–20 The average turnover cost to the U.S. healthcare system is estimated to be between $37,700 and $58,400 per registered nurse (RN) in 2015, or between $5.1 million and $8.1 million in aggregate over the same time period.²¹ Taken together, these data present a compelling case for the importance of addressing nurse well-being as a mechanism to avoid eventual burnout.

One potential contributor to nurse well-being is the electronic health record (EHR) system.^22–25 EHRs have been adopted by most U.S. hospitals²⁶; however, concerns have been raised about how well EHRs meet nurses’ needs. Best practices in quality improvement entail consulting with the frontline personnel (ie, clinical nurses) to glean insight into problematic issues encountered and ideas regarding meaningful solutions.^27,28 To this end, nurses’ input will be necessary to help healthcare administrators and EHR vendors address EHR-related burden that negatively affects nurse well-being. Although a recent review studied burnout among nurse practitioners and physician assistants,¹³ to our knowledge, there is no review that focuses exclusively on the broader nursing populations.

This systematic review has 2 objectives: (1) to assess multifactorial and multilevel (organizational, nurse, and health information technology[IT]) factors associated with EHR-related indicators of nurse well-being and (2) to summarize potential EHR improvements recommended by frontline nurses.

MATERIALS AND METHODS

This is a systematic review that follows the design recommended by the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analysis) guidelines.²⁹

Search strategy

The MEDLINE database was searched for all relevant works (eg, peer-reviewed studies, dissertations, think tank reports, conference abstracts) that focus on EHR systems, nurses, and indicators of well-being. Other academic databases (ie, Embase, CINAHL, PsycInfo, ProQuest, Web of Science) were also queried. All searches occurred on April 7-8, 2020, through controlled vocabulary and subject headings and truncated keywords in the title or abstract fields as permitted by the databases. For instance, our nurse construct included terms such as advanced practice nurses and critical care nursing. Supplementary Appendix S1 provides a listing of all search terms. Studies’ reference lists were manually searched to identify other potentially relevant articles. Duplicate studies were removed before the screening.

To gain a comprehensive understanding of how the EHR relates to nurse well-being, we included studies that examined the link between EHR use and multiple indicators of well-being (or lack thereof). Included studies examined broad EHR use or focused on specific features (eg, interface displays). The variables of interest and associated terms were identified a priori through the collaboration of a health sciences librarian, a health IT manager (O.T.N.), and a clinical psychologist that specializes in clinician well-being (L.J.M.). We refer to the collection of measures studied as “EHR-related indicators of nurse well-being” in our included studies throughout this review. Supplementary Appendix S2 displays each variable that falls under this construct, along with its definition.

Study selection

Our final inclusion criteria were articles with the following characteristics: (1) written in English, (2) used empirical study designs (ie, reports original findings), (3) U.S. study setting, and (4) examined EHR-related indicators of well-being and burnout among nurses. Analyses were limited to studies conducted in the United States because systemic differences may be present in other healthcare systems, such as the documented differences in length of clinic notes.³⁰ Studies examining nonnurse clinicians were excluded because each clinical specialty has distinct needs, expectations, and functionalities despite using the same EHR system.^31,32 For this review, nurses were broadly defined to include licensed practical nurses (LPNs) (or licensed vocational nurses), RNs, and nurse practitioners (NPs) (or advanced practice registered nurses). Studies that combined data for nurses and nonnurses were excluded.

Two reviewers (O.T.N., S.S.) independently screened all studies’ titles and abstracts against the inclusion criteria discussed previously. Articles flagged for full-text review were independently assessed against the inclusion criteria. If the reviewers disagreed on an article, a third reviewer (A.P.) reviewed the article to provide input and the reviewers reached a final decision through consensus.

Data extraction and synthesis

For each included study, study design, type(s) of nurses studied, sample size, and relevant findings were recorded in Microsoft Excel. We adapted a previously developed risk-of-bias assessment tool to assess study quality.³³

For all studies, we evaluated for (1) thorough description of the inclusion criteria of participants, (2) bias from funding source, (3) indirectness bias from utilizing samples differing from the target population, and (4) indirectness bias from eliciting participants’ comments on EHRs that they never personally used. All qualitative studies were assessed for (1) reporting of data collection methodologies, (2) description of data analysis methodologies, (3) evaluation of interrater reliability during data coding, and (4) sufficient sample sizes to conclude saturation. All quantitative studies were assessed for (1) flawed measurements of variables (eg, asking participants to estimate their EHR use duration instead of examining EHR activity logs), (2) response rate exceeding 50% for surveys, (3) description of significant and nonsignificant findings, (4) combined sample size of 385 nurses or higher (a threshold commonly used for sampling larger populations at the 95% confidence level),^34–37 and (5) use of validated instruments to assess outcome measures. For mixed-methods studies, both quantitative and qualitative components were assessed individually using the previously mentioned criteria.

For each study, 1 point was assigned for each criterion met. Zero points were assigned when the study lacked a particular criterion, if it was not applicable, or if the authors were unable to make a definitive decision. Studies utilizing objective measurement components could score up to 9 points, whereas those with survey designs could score up to 10 points. Qualitative components could score a maximum of 8 points. Total quality scores determined the risk of bias: 7-10 points were deemed low risk of bias, 5-6 points were deemed moderate risk of bias, and 1-4 points were deemed high risk of bias. Notably, the total score for a mixed-methods study was determined by evaluating both the quantitative and qualitative components, and assigning the lower score. This was added to the subscore on criteria applicable to either design type (eg, description of inclusion criteria).

We grouped studies’ empirical findings on nurses’ recommendations using in vivo coding (ie, using keywords from article).³⁸ The conventional standard of P < .05 was used to indicate the significance of findings.

RESULTS

In total, 4583 unique articles were included in the initial screening. Based on review of the title and abstract, 26 of the 4583 articles met our initial inclusion criteria. After full-text review, 12 articles ultimately met the inclusion criteria. Figure 1 details our study screening process.

Figure 1.

PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analysis) flowchart detailing the study selection process. Adapted with permission from Moher et al.²⁹ EHR: electronic health record.

Study characteristics

Six quantitative studies, 1 qualitative study, and 5 mixed-methods studies were found. Of studies with quantitative components, 4 used observational designs and 7 used quasi-experimental designs. Three of the quantitative studies used a cross-sectional design and 3 had a cohort design for repeated measurements. Of studies with qualitative components, surveys were the most common design followed by interviews. Two studies examined only 1 type of nurse (eg, only RNs, only NPs) while the remainder studied multiple types of nurses or did not report this information at all. Of the studies that reported the type of nurses represented, acute care and intensive care unit (ICU) nurses were overrepresented, with 1 study examining nurses in ambulatory settings. A summary table of study characteristics is included in Supplementary Appendix S3.

Study quality assessment results

As detailed in the Materials and Methods, points were attributed to each study based on whether it met the specific criterion. Across the 12 studies, 4 were deemed to have a high risk of bias,^39–42 4 had a moderate risk of bias,^23,43–45 and 4 had a low risk of bias.^22,25,46,47 Detailed quality assessment results are in Table 1. Each column represents 1 of the 14 quality assessments, which are provided in detail under the table. The right column represents the total amount of criteria that were met.

Table 1.

Study quality assessment results

Citation	1	2	3	4	5	6	7	8	9	10	11	12	13	14	Total
Staggers et al, 200739a,b	N	Y	N	NA	N	Y	N	N	Y	N	Y	UC	Y	Y	2 (All) + 2 (Quant) + 2 (Qual) = 4
Sockalow et al, 200941b,c	N	Y	UC	UC	UC	UC	UC	UC	UC	N	Y	UC	UC	UC	0 (All) + 1 (Quant) + 1 (Qual) = 1
Harshberger et al, 201142a,d	N	Y	N	UC	Y	NA	NA	NA	Y	N	NA	UC	Y	N	4
Anders et al, 201244b,c	N	Y	Y	NA	Y	NA	NA	NA	Y	N	NA	UC	Y	Y	6
Koch et al, 201240b,c	N	Y	N	UC	Y	Y	UC	N	UC	N	N	UC	Y	Y	1 (All) +1 (Qual) + 3 (Quan) = 4
Laramee et al, 201246b,c	Y	Y	N	N	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	4 (All) + 4 (Quan) + 4 (Qual) = 8
Moreland et al, 201247a,b	Y	Y	Y	UC	Y	NA	NA	NA	Y	Y	NA	Y	Y	N	8
Schenk et al, 201645a,d	Y	Y	N	UC	Y	Y	Y	UC	Y	N	Y	Y	Y	N	4 (All) +2 (Quan) + 3 (Qual) = 6
Bristol et al, 201843e	N	NA	NA	NA	NA	Y	Y	Y	Y	NA	Y	UC	Y	NA	6
Harris et al, 201822a,d	Y	Y	Y	N	Y	NA	NA	NA	Y	N	NA	Y	Y	Y	8
Summers et al, 201923a,b	N	Y	N	UC	Y	NA	NA	NA	Y	N	NA	UC	Y	Y	5
Yen et al, 201925a,d	Y	Y	Y	NA	Y	NA	NA	NA	Y	N	NA	Y	Y	Y	8

The numbers in the heading columns indicate (1) All studies: Is eligibility criteria (ie, the population to be included in study) described thoroughly?; (2) Quantitative: Is the measurement of outcome variables free from flaws? (eg, asking participants to self-report EHR use as opposed to examining EHR data logs); (3) Quantitative: Does the study control for confounders in analyses? (eg, linear or logistic regression models; (4) Quantitative: If a survey was involved, was there at least a 50% response rate?; (5) Quantitative: Were p-values reported for both significant and non-significant findings?; (6) Qualitative: Was there a structured format to collect the data (eg, standardized form) or were researchers recording information they personally deemed important?; (7) Qualitative: Was there a formal qualitative analysis approach done? (eg, grounded theory approach); (8)Qualitative: Did the study report any interrater reliability assessments during data coding?; (9) All studies: Did the funding source play no role in the study design and results reporting (eg, company that provides services on nurse well-being)?; (10) Quantitative: Was the combined sample size 385 or more?; (11) Qualitative: Is the sample size sufficient to achieve saturation (ie, interviewing more individuals would not lead to additional/different responses)?; (12) All studies: Is the desired population of study the same population that participated in study?; (13) All studies: Were the nurses asked about the EHR they actually use (compared with discussing an EHR they have no direct experience with)?; (14) Quantitative: Did the study measure the outcome variable using recognized scales or instruments (eg, Well-Being Index, System Usability Scales)?

EHR: electronic health record; N: no; NA: not applicable; UC: unclear; Y: yes.

^aCross-sectional design.

^bQuasi-experimental design.

^cCohort design.

^dObservational design.

^eQualitative study.

We excluded the results of the 4 studies that were deemed to have had a high risk of bias from the data review and synthesis that follows.

Organizational characteristics

The number of clinicians in the organization,²² practice type,²² and use of scribes had no association with nurse well-being.²² However, those who reported use of an EHR as opposed to paper charts were more likely to report burnout.²² It is worth noting that these findings were from the same study. Nurses’ recommendations focused on EHR training and insufficient computer workstations.^43,45,46

Table 2 summarizes the individual results per study. Each row represents the study and its outcome measures, as well as characteristics of the study site (eg, location, practice type, number of beds reported). Most studies provided the total number of beds in the hospital, while others reported beds in the ICU or a minimum number of beds (eg, 1200+).

Table 2.

Organizational characteristics of study sites and organization-level findings

Citation	Outcome Measures	Location of Study Site	Practice Type	Number of Beds in Study Site	Relevant Findings	Organization-Level Recommendations
Anders et al, 201244	NASA-TLX, usability	Tennessee, Utah	Teaching hospitals	Tennessee: at least 150 in ICU Utah: at least 62 in ICU	None	None
Laramee et al, 201246	EHR attitudes, belief that documentation improved	Vermont	Teaching hospital	500 total	None	Invest in additional computer workstations
Moreland et al, 201247	eMAR satisfaction	Ohio	Tertiary hospital	At least 1200 total	None	None
Schenk et al, 201645	EHR perceptions	Montana	Tertiary hospital	250 total	None	EHR training needs to have tailor content specifically to nurses
Bristol et al, 201843	EHR usage experiences	NR	NR	NR	None	EHR training needs to be offered when implementing an EHR Invest in additional computers
Harris et al, 201822	Burnout	Rhode Island	67.6% of respondents worked in hospital settings, 32.4% worked in office settings	NR	Burnout was associated with EHR use compared with paper charts. It was unrelated to the number of clinicians in the organization, practice type, and use of scribes.	None
Summers et al, 201923	Burnout, EHR satisfaction	Michigan	NR	NR	None	None
Yen et al, 201925	Stress, perceived workload	Missouri	Teaching hospital	NR	None	None

EHR: electronic health record; eMAR: electronic medication administration record; ICU: intensive care unit; NASA-TLX: National Aeronautics and Space Administration Task Load Index; NR: not reported.

Nurse characteristics

Sex,^22,25 education level,²³ years of prior work experience with EHRs,²³ amount of hours worked weekly,²³ shift type (day vs night),²⁵ time spent on the EHR,^22,23,25 personal usage of remote EHR access,²² perceived EHR mental demand,²⁵ and perceived effort were unrelated to EHR-related well-being.²⁵

Older studies reported that age influenced EHR-related well-being,^46,47 but several recent studies found no association.^22,23,25 Most studies were in agreement that more experienced nurses reported worse EHR-related well-being,^23,46,47 but 1 study found no differences.²⁵

Improved EHR-related well-being was associated with LPNs compared with RNs.⁴⁷ Other factors associated with better EHR-related well-being included personal comfort with computer use,⁴⁷ caring for adult patients compared with pediatric populations,⁴⁷ and working on surgical floors compared with ICUs or medical floors.⁴⁷ When assessing attitudes, improved EHR-related well-being was associated with higher EHR satisfaction,²³ lower EHR frustration,²⁵ perception that the time allotted for documentation is sufficient,²² and lack of the belief that the EHR adds to daily frustration.²²

Table 3 summarizes each study’s results. Each row represents 1 study and includes the outcome measures and sample characteristics.

Table 3.

Sample characteristics on nurses studied and nurse-level findings

Citation	Outcome Measures	Sample Size	Age	Reported Sex	Race/Ethnicity	Nurse Type	Relevant Findings
Anders et al, 201244	NASA-TLX, usability	32	29 y (median)	87% female, 13% male	NR	NR	None
Laramee et al, 201246	EHR Attitudes, belief that documentation improved	984	57% were younger than 50 y	Majority reported female, but exact proportion not reported.	NR	LPNs, RNs, NPs	Nurses older than 50 y reported worse EHR attitude scores (P < .001). Less experienced nurses (<5 y) generally reported better EHR attitude scores (P < .001).
Moreland et al, 201247	eMAR satisfaction	722	38.5 y (mean)	90.9% female, 9.1% male	NR	RNs, LPNs	Younger nurses, those with more personal comfort with computers, and those with less nursing experience reported lower eMAR satisfaction (P < .05). Compared with RNs, LPNs reported higher eMAR satisfaction (P < .05). Nurses taking care of pediatric patients reported lower eMAR satisfaction than did those taking care of adult patients (P < .001). Nurses on ICU or cardio/medical floors reported lower eMAR satisfaction than did nurses on surgical floors (P < .001). No differences in scores were seen between ICU and cardio/medical nurses.
Schenk et al, 201645	EHR perceptions	285	NR	NR	NR	RNs	None
Bristol et al, 201843	EHR usage experiences	144	46-55 y (median age group)	92.36% female, 7.64% male	NR	NR	None
Harris et al, 2018 22 b	Burnout	371	41-60 y (median age group)	87.7% female, 12.3% male	NR	NPs	Burnout was associated with beliefs that the EHR adds to daily frustration and insufficient time allotted for documentation. It was not affected by age, sex, total time spent on the EHR at home, or use of remote EHR access.
Summers et al, 201923	Burnout, EHR satisfaction	162	35-44 y (median age group)	NR	NR	NR	Higher EHR satisfaction was associated with less nursing experience. It was unrelated to age, education level, prior years of experience working with EHRs, amount of hours worked, or total time spent on the EHR. Burnout was associated with lower EHR satisfaction. It was not associated with age, education level, prior years of experience working with EHRs, or total time spent on the EHR.
Yen et al, 201925	Stress, perceived workload	7	30 y (mean)	NR	NR	RNs	Higher EHR stress was associated with perceived EHR frustration. It was not associated with age, sex, years of nursing experience, shift type (day vs night), total time spent on the EHR, perceived EHR mental demand, or perceived EHR effort.

EHR: electronic health record; eMAR: electronic medication administration record; LPN: licensed practical nurse; NASA-TLX: National Aeronautics and Space Administration Task Load Index; NP: nurse practitioner; NR: not reported; RN: registered nurse.

^aThis study stratified the sample based on those with and without an EHR. We report only on the subgroup who self-reported EHR use.

EHR characteristics

Studies assessing the association of EHR characteristics and indicators of well-being focused exclusively on interface displays. One study reported that integrated displays were associated with higher usability scores when compared with nonintegrated displays, despite no differences in perceived workload scores.⁴⁴ Three studies reported nurse recommendations to improve the EHR. Identified themes included EHR usability issues, misaligned EHR workflows with nursing tasks, and the desire to participate in improving EHR designs.^43,45,46

Results are shown in Table 4 by outcome measures, EHR vendor used, and individual study findings. Qualitative EHR recommendations are separated from quantitative findings.

Table 4.

EHR-level findings and EHR recommendations from frontline nurses

Citation	Outcome Measures	EHR Vendor Used	Relevant Findings	EHR-Level Recommendations Reported by Nurses
Anders et al, 201244	NASA-TLX, usability	NR	Integrated displays were associated with higher usability scores than non-integrated displays (P < .05). No differences were found.	None
Laramee et al, 201246	EHR attitudes, belief that documentation improved	Epic	None	Improve EHR usability through intuitive interfaces and better placement of information on the screen Redesign flowsheets with more flexibility for cases (ie, specific patients need additional data tracked, some data are not applicable to certain patient types) Optimize login processes, as logging in is cumbersome Directly involve nurses when designing nurse-facing systems
Moreland et al, 201247	eMAR satisfaction	Epic	None	None
Schenk et al, 201645	EHR perceptions	NR	None	Reduce the number of clicks needed to complete EHR tasks Align EHR workflows to nursing workflows
Bristol et al, 201843	EHR usage experiences	NR	None	Location of information needs to be more intuitive Nursing input is needed in EHR design Standardize EHR design to streamline nursing workflows
Harris et al, 201822	Burnout	NR	None	None
Summers et al, 201923	Burnout, EHR satisfaction	Epic	None	None
Yen et al, 201925	Stress, perceived workload	NR	None	None

NASA-TLX: EHR: electronic health record; eMAR: electronic medication administration record; National Aeronautics and Space Administration Task Load Index

DISCUSSION

To our knowledge, this is the first systematic review that summarizes the current literature on how the EHR affects the broader domain of nurse (LPN, RN, NP) well-being, including nurses’ suggestions for EHR improvement. Although current evidence suggests that time spent in the EHR is associated with physician well-being,^48–53 time spent using the EHR was not associated with nurse well-being,^22,23,25 suggesting that the quantity of effort may be less important than the quality. For example, RNs, LPNs, and nursing assistants have reported that frequent EHR downtime forces them to document on paper charts.⁵⁴ As a result, reports of total EHR time may underestimate the documentation burden that nurses experience. Although we would expect NPs to be more prone to after-hours EHR use due to increased documentation responsibilities, only 1 study examined NPs’ time spent in the EHR. Thus, it is difficult to discern the effect of total EHR time on well-being. In addition, nurses working in different units or settings (eg, ICU, general hospital floor, outpatient clinic) are likely to vary significantly in terms of their reliance on the EHR to perform job tasks, as well as the objective stress of their work environment due to the nature of the population they serve (eg, age, acuity, mortality rates of their patients). Indeed, previous research has documented varying burnout rates for nurses in different settings.^6–9,55 We also found that interface displays may reduce cognitive load,⁴⁴ leading to decreased burnout and medical errors. Consistent with literature on physician well-being,^50,56–63 recommendations offered by nurses suggested a significant desire for improvements to reduce cognitive load. Notable interventions to highlight in this discussion include (1) digital literacy’s role in preparing nurses for the workforce, (2) the use of integrated displays to consolidate meaningful information necessary for decision making, and (3) organizational policies for hospitals.

Our review suggests that the development of functional digital literacy among nurses may serve as a protective factor against EHR-related well-being issues and is deserving of further, more deliberate study.⁴⁷ A recent study reported that, although the proportion of clinicians reporting high digital literacy has risen over the years, about 20%-30% still report that they have not adequately developed this skill.⁶⁴ This rate is still unacceptably high, given that between 84% and 96% of acute care hospitals utilize an EHR system.²⁶ Although not explicitly discussed among our included studies, we argue that nursing school curricula must be intentional about preparing future generations of nursing students for the increasing and expanding role that health IT plays in health care delivery, both currently and in the future.^65,66 One recent review recommended standardizing the expected nursing informatics competencies for graduates and providing nursing faculty additional support and training to enable them to teach these skills in the classroom setting.⁶⁷ Continuing education programs for nursing faculty could focus on methods and platforms to utilize when integrating these skills into their curricula. We posit that adding a segment on digital literacy to licensure exams, such as the National Council Licensure Examination, would create additional incentive for nursing schools to adapt their curricula in order to best prepare students to sit for these exams. Notably, the American Association of Colleges of Nursing has already assembled the Essentials Task Force, who are updating nursing curricula and competencies used across United States nursing schools for training RNs and NPs. Of their drafted 10-domain model in October 2020, the eighth domain focuses on digital health competencies, such as delivering care through telemedicine and using information from health information exchanges (HIEs).⁶⁸ Should this model be implemented into nursing schools, this will create an area ripe for further evaluations on the model’s impact on nurses’ comfort with using health IT and EHR-related well-being.

Policy changes and technological advances will likely increase the amount of information available in EHRs for clinical decision making. Therefore, it is reasonable to assume that the typical nurse may see increases in the amount of data, patient contributed or otherwise, in the EHR. Skills to turn those data into actionable information to make clinical decisions could be augmented with increased digital literacy and simulations in nursing programs. For instance, one study on HIE usage reports that nurses were more likely to access this system to view or pull information.⁶⁹ However, some studies report that the EHR is already perceived as inducing information overload for clinicians.^70,71 This poses a problem, as the coming deluge of additional information may exacerbate the existing problem of information overload.

Our review found that the use of integrated displays may improve issues related to cognitive load for nurses. Other mechanisms to consolidate meaningful information on one screen, such as patient dashboards,⁷² may also help to reduce cognitive burden for nurses and cushion against the potential impact of additional information entering the EHR via HIEs and other external health system EHRs. Although additional study is needed to identify the best EHR solutions to improve cognitive processing, the Office of the National Coordinator for Health Information Technology Health IT Certification program may become an important facilitator in incentivizing EHR vendors to offer their customers effective ways to process incoming information.

At a more proximal level, our findings suggest that hospitals can implement helpful policies. Increasing computing resources by way of providing additional workstations to improve accessibility to the EHR and enhancing infrastructure to improve network speed would improve nurses’ EHR task-related efficiency. Other works suggest that nurses report overly frequent encounters with system downtime.⁵⁴ This suggests that additional attention is needed to identify root causes of unplanned system downtime. Furthermore, the cadence of planned system downtime should be reconsidered to bundle updates and minimize outages and interruptions in patient care.^73,74 Implementing protected time policies, particularly for NPs with significant documentation responsibility, would allow time to manage documentation and administrative burden outside the direct patient care setting where interruptions are frequent.⁷⁵ Last, on-site EHR training provided should incorporate principles of adult learning to deliver EHR training on an ongoing basis.^76,77

Nurse well-being in association to EHR use is an understudied area. From this systematic review, further areas of research emerged. For example, more work is needed to examine how differences in EHR vendor, autonomy, clinical decision making, and documentation requirements impact EHR-related nurse well-being. Additionally, more focused and rigorous studies are needed that examine the relation of digital literacy and nurse well-being as a modifiable factor in managing EHR-related well-being. We also observed a need for research into other organizational characteristics that can influence EHR-related well-being among other clinician types, such as the use of scribes, availability of EHR training, and the number of years an organization has had its current EHR.⁷⁸

Despite illuminating several areas of future research that would increase the knowledge around nurse well-being, this study has several limitations. First, there were methodological challenges that limited our ability to draw firm conclusions from the current evidence base. Earlier studies generally had high risk of bias, so their results were excluded from the synthesis. Second, we chose to include articles that either assessed burnout directly through a validated tool or that assessed a proxy measure for burnout. We acknowledge that the state of burnout is rarely an immediate reaction within an individual; rather, a progressive series of events typically contribute to the development of burnout. As a result, we chose to include indirect measures that could provide us insight regarding nurses who have not yet developed burnout but are displaying intermediary symptoms that could lead to burnout if left unchecked. Third, it was infeasible to pool estimates across the quantitative studies due to the heterogeneity in types of measures used. Fourth, because this area of research is relatively nascent and growing, we included 2 quantitative works from the gray literature in our analysis. Had they been excluded, our review may have produced different results. Fifth, we are unable to comment on whether our results are more applicable to certain nurse types due to the lack of this information in more than half of the studies. Last, our results may not be generalizable outside the United States and may not generalize across U.S. states, as differences in licensure and scope of practice may exist.

CONCLUSION

The overarching finding from this synthesis reveals a critical need for multifaceted, multifactorial interventions that better organize, manage, and display information for clinicians to facilitate the clinical decision-making process. Poor nurse well-being has a cascade of negative consequences for healthcare systems, including reduced quality of care, increased healthcare costs, and system inefficiency. Future research is needed to test multicomponent interventions that address these complex factors and use participatory approaches to engage nurses in intervention development.

AUTHOR CONTRIBUTIONS

This work represents the original research of the authors. This work has not been previously published. OTN, KT, and LJM conceptualized the study. OTN, KT, SSF, and LJM drafted the manuscript. All authors participated in the analyses and interpretation of data. SS, KT, SSF, and LJM provided critical revisions to the manuscript. All authors approved the submission.

SUPPLEMENTARY MATERIAL

Supplementary material is available at Journal of the American Medical Informatics Association online.