Health information technology and clinician burnout: Current understanding, emerging solutions, and future directions

Burnout among healthcare providers has been increasingly recognized as a significant problem.¹ The National Academy of Medicine has defined burnout as “a syndrome characterized by high emotional exhaustion, high depersonalization (ie, cynicism), and a low sense of personal accomplishment from work.”² The Agency for Healthcare Research and Quality similarly defines Burnout as a long-term stress reaction marked by emotional exhaustion, depersonalization, and a lack of sense of personal accomplishment.³ Clinician burnout is both costly and has been associated with reduced job satisfaction, quality and safety of care, and patient health outcomes.⁴ Burnout is common—affecting between 35% and 54% of U.S. nurses and physicians and between 45% and 60% for medical students and residents.²

Research to date has identified a number of contributing factors as associated with burnout.⁵ Among these, health information technologies (HITs) are often implicated. Electronic health record (EHR) systems, for example, are often seen as cumbersome to use, failing to fulfill the promise of improved healthcare delivery, and little more than a means of meeting regulatory and billing requirements.⁶ However, there remains considerable debate in the informatics community as to the actual role health information technologies play in the problem of clinician burnout.⁷ Existing research suggests that technologies may be confounded with other important causes, including regulatory mandates, clinical volumes, increasing hyperspecialization among healthcare providers, and a mismatch between the incentives driving system designers and purchasers and those driving providers. Regardless of the role information technology plays in clinician burnout, innovative solutions to prevent or mitigate burnout are urgently needed.

In this special focus issue of Journal of the American Medical Informatics Association, we target articles evaluating the role that health information technologies have in causing and mitigating burnout, identify confounding factors, and consider informatics and policy-based solutions. This special focus issue is an outgrowth of the 2020 American College of Medical Informatics Symposium, “Clinician Burnout: Is it Informatics’ Fault, and What Can We Do About It?” In parallel with this special issue, the American College of Medical Informatics (ACMI) Symposium also led to the 25x5 initiative,⁸ to reduce the burden imposed by clinical documentation on healthcare providers in the United States to 25% of its current level within 5 years. The 25x5 initiative in turn resulted in a National Library of Medicine–funded 6-week symposium that concluded in February 2021. Follow-on work will lay out concrete steps that can be taken to reduce burden, create a community of like-minded stakeholders, and will work with key organizations and associations to guide this change. Taken together, the 2020 ACMI symposium, the 25x5 initiative, and this special issue comprise concrete steps the informatics community is taking to address the problem of clinician burnout.

This special issue includes 24 articles across the variety of JAMIA formats: Research and Applications (n = 6),^9–14 Brief Communications (n = 4),^7,15–17 Reviews (n = 5),^18–22 and Perspectives (n = 9).^23–31 In the following paragraphs, we summarize selected papers reflecting 3 key themes: (1) understanding the relationship between HIT and clinician burnout, (2) emerging HIT approaches to mitigate clinician burnout, and (3) future directions.

Understanding the relationship between HIT and clinician burnout

Several articles in this special focus issue anchor our understanding of the relationship between HIT and clinician burnout. Two review articles, one by Yan et al²¹ and another by Nguyen et al¹⁹ both identified documentation burden, high inbox message volumes, and negative perceptions of EHR functionality and usability as key EHR-related factors most consistently associated with objective measures of provider burnout in the extant literature. Both review articles also identified time spent on EHR after work hours—often called “pajama time”—but not total time spent on EHR, as being associated with burnout.

Findings from these review articles point to opportunities for HIT systems to identify clinicians at increased risk of burnout. Baxter et al¹⁶ demonstrated that 3 leading EHR vendors currently provide “off-the-shelf” metrics to measure provider activity on the EHR through log data. While further work is needed to harmonize the metrics’ definitions so that meaningful cross-vendor comparisons can be made, these measures are now routinely available to healthcare organizations and informatics researchers. Two articles in this special issue offer practical insights on how these metrics could be used to identify the subset of clinicians at elevated risk of burnout and to target burnout mitigation interventions. Eschenroeder et al¹⁵ analyzed data from the KLAS Arch Collaborative and found that physicians who spend 6 or more hours per week performing after-hours charting were more likely to report burnout. Similarly, Peccoralo et al¹⁴ found using survey data from a single institution that faculty members who used EHR for more than 90 minutes a day after hours or who spent more than 60 minutes a day performing clerical tasks were more likely to report burnout. Taken together, these findings suggest that risk of burnout for full-time clinicians may rise significantly if they spend more than 60 to 90 minutes per day on the EHR after hours.

Emerging HIT approaches to mitigate clinician burnout

Articles in this special issue also highlight opportunities to leverage HIT to address the pervasive problem of clinician burnout. Several contributions build on the evidence base for approaches that healthcare organizations could adopt. Lourie et al¹² reported that personalized customization and training sessions across 14 specialty and 31 primary care ambulatory care practices led to improved self-reported efficiency and burnout perception. Simpson et al¹⁷ found that a 2-week EHR optimization sprint consisting of EHR changes and one-on-one training sessions led to an improvement in clinicians’ satisfaction toward the EHR in a single-specialty practice but did not impact measures of emotional burnout. A qualitative study conducted by Tran et al¹³ found that medical scribes are commonly used to offload 7 categories of clinical or clerical tasks as a way to a alleviate burnout attributable to the use of HIT. While these approaches require dedicated resources, these articles should help healthcare organizations build the case for these investments.

As suboptimal EHR usability has often been cited as a significant contributor to clinician burnout in the United States, the editors of this special focus issue invited key EHR vendors and usability experts to elucidate current approaches to and opportunities for vendors to improve EHR usability. Leading EHR vendors were invited to respond to a semi-structured written survey on how they meet or exceed the 2015 EHR usability (or user-centered design) requirements issued by the ONC (https://www.healthit.gov/test-method/safety-enhanced-design#ccg). Anonymized responses from 4 major vendors (Supplementary Appendix) were sent to usability experts to comment on the strengths and weaknesses adopted by the EHR industry, and to suggest improvement opportunities. When compared to research from 2015 on the usability of EHR products, Hettinger et al²³ noted that vendors have increased their adoption and maturity of user-centered design practices. This observation highlighted the ongoing efforts from U.S. federal policy makers, as summarized by Gettinger et al²⁵ to promote HIT usability by implementing usability standards and funding research to examine the efficacy of these policies. However, much work remains. Hettinger et al highlighted the usability reality gap between EHR as designed by vendors and EHR as implemented by each healthcare organization, citing the paucity of the workforce trained to optimally configure and usability or safety test local configurations as a key driver of this gap. Carayon and Salwei²⁴ further pointed out that the path for reducing clinician burnout through improving EHR usability requires a continuous approach, as vendors and their clients need to work together to turn their focus away from technology embedded in work-as-imagined toward sociotechnical systems supporting work-as-done.

EHR vendors should also recognize that they can and should partner with informatics innovators to advance EHR usability and mitigate clinician burnout. In a block-randomized study, Semanik et al¹¹ found that problem-oriented summaries of clinical data, built directly into a vendor EHR, allowed clinicians across three academic medical centers to retrieve data faster and with fewer errors. With the use of this tool, clinicians also reported a reduced cognitive load and increased satisfaction. This article by Semanik et al demonstrates how EHR vendors could support efforts to mitigate burnout by spreading and sustaining usability innovations coming from an individual customer across their customer base.

Future directions

So where does the topic of informatics and clinician burnout go from here? Perspectives articles from several ACMI members offer new lenses through which to view, understand, and address HIT-associated clinician burnout. Williams³⁰ contended that moral injury associated with EHR, as defined by EHR use that leads clinicians to transgress deeply held moral beliefs and expectations, may be a hidden contributor to clinician burnout. Weir et al³¹ further postulated that burnout may be linked to drivers of intrinsic motivation, and that goal-based decision making, sense making, and agency or autonomy should be considered in the design of future technological interventions to mitigate clinician burnout.

From a methodological perspective, significant opportunities remain. Moy et al²⁰ pointed out in their scoping review that standard and validated measures of documentation burden are still lacking, which in turn forms a barrier to the rigorous study of documentation burden. Moy et al further called for efforts to operationalize the concept of documentation burden and develop best practices for measurement. Kannampallil et al²⁶ proposed a conceptual framework that would allow the informatics community to build on EHR activity measures evaluated by Baxter et al¹⁶ and use technology to assess holistically clinicians’ workload, cognitive burden, and well-being.

The editors of this JAMIA special issue recognize that this body of work is but a snapshot of a rapidly growing and evolving topic. New technologies such as ambient voice speech to text, internet of things, natural language processing and machine learning–driven data visualization, and Fast Healthcare Interoperability Resources, as highlighted by Dymek et al²⁸ and Gettinger and Zayas-Cabán,²⁵ may yet open up opportunities to support more meaningful clinician-patient interactions and more efficient workflows. The policy landscape is also constantly changing, as evidenced by the recent simplification in documentation requirements initiated by the Centers for Medicare and Medicaid Services Burden Reduction efforts intended to place “Patients Over Paperwork.”³² As a target for multidisciplinary scientific inquiry, the subject of HIT-associated clinician burnout must continue to evolve through future empirical studies. Its current evidence base remains modest, at best. We therefore encourage readers of this special issue to participate in and accelerate the ongoing work in this area.

AUTHOR CONTRIBUTIONS

STR, KZ, and EGP presented the special issue proposal to JAMIA; STR, KZ, and EGP fulfilled Associate Editor duties; EGP fulfilled EIC duties; EGP contributed to instrument design, data collection from vendors and commissioning of invited perspectives articles; and EGP, STR, and KZ contributed to drafting and finalization of the editorial.

SUPPLEMENTARY MATERIAL

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

CONFLICT OF INTEREST STATEMENT

The authors have no relevant conflicts of interest to declare.