Measuring Documentation Burden in Healthcare

Abstract

Background

The enactment of the Health Information Technology for Economic and Clinical Health Act and the wide adoption of electronic health record (EHR) systems have ushered in increasing documentation burden, frequently cited as a key factor affecting the work experience of healthcare professionals and a contributor to burnout. This systematic review aims to identify and characterize measures of documentation burden.

Methods

We integrated discussions with Key Informants and a comprehensive search of the literature, including MEDLINE, Embase, Scopus, and gray literature published between 2010 and 2023. Data were narratively and thematically synthesized.

Results

We identified 135 articles about measuring documentation burden. We classified measures into 11 categories: overall time spent in EHR, activities related to clinical documentation, inbox management, time spent in clinical review, time spent in orders, work outside work/after hours, administrative tasks (billing and insurance related), fragmentation of workflow, measures of efficiency, EHR activity rate, and usability. The most common source of data for most measures was EHR usage logs. Direct tracking such as through time–motion analysis was fairly uncommon. Measures were developed and applied across various settings and populations, with physicians and nurses in the USA being the most frequently represented healthcare professionals. Evidence of validity of these measures was limited and incomplete. Data on the appropriateness of measures in terms of scalability, feasibility, or equity across various contexts were limited. The physician perspective was the most robustly captured and prominently focused on increased stress and burnout.

Discussion

Numerous measures for documentation burden are available and have been tested in a variety of settings and contexts. However, most are one-dimensional, do not capture various domains of this construct, and lack robust validity evidence. This report serves as a call to action highlighting an urgent need for measure development that represents diverse clinical contexts and support future interventions.

Supplementary Information

The online version contains supplementary material available at 10.1007/s11606-024-08956-8.

INTRODUCTION

Satisfaction and burnout of healthcare professionals are urgent challenges facing the healthcare system.¹ There are multiple contributors driving these issues, but documentation burden is commonly cited as a key factor.^2–4 Many clinicians report that electronic health records (EHRs), electronic prescribing, electronic patient portals, and computerized physician order entry (CPOE) lead to information overload, frequent interruptions and distractions, and a change in the content of professional work to tasks less connected to meaning and purpose.^2,5,6 Greater time spent on administrative tasks is associated with decreased career satisfaction and increased burnout,⁵ and greater use of EHRs and CPOE is associated with increased burnout.⁵ EHR usability is generally described as poor, and physician assessment of poor EHR usability is strongly associated with burnout.⁷ In addition, clinical documentation requirements often cannot be completed during the work day,^8–10 and “work outside of work” is a strong driver of burnout.¹¹ Aside from burnout, documentation burden may affect patient outcomes. One study has shown that the total in-basket notifications and delivery of alerts over the weekend can impact the opening of time-sensitive EHR alerts, and another has suggested a role of health information technology in diagnostic delays.^12,13 Furthermore, documentation burden on clinicians may be associated with reduced patient satisfaction.¹⁴

Numerous factors contribute to documentation burden, including regulatory demands, payor needs, organizational structure and needs,¹⁵ and possibly fear of litigation.^16–18 The 2009 enactment of the Health Information Technology for Economic and Clinical Health (HITECH) Act is often identified as the beginning of the modern era of clinical burdens associated with health information technology (IT), including documentation burden.¹⁹ Although the adoption of EHRs in the USA has increased since the mid-2000s and burdens have been a concern for many years, clear definitions and measures for documentation burden are lacking.^6,20,21

Measures of documentation burden have been described largely in the context of time and effort associated with specific clinical documentation tasks.⁶ Some are too simplistic and unidimensional, such as the number of messages or time spent on documentation.⁶ Expanded and more granular measures have been introduced through the American Medical Association’s Joy in Medicine Health System Recognition Program.²² Considering the lack of clarity about the available measures, their validity, and applicability, the Agency for Healthcare Research and Quality (AHRQ) commissioned this systematic review as a Technical Brief to identify and characterize measures of documentation burden in healthcare and was complemented with interviews with relevant stakeholders identified as key informants of this work.

METHODS

We followed the established methodologies of Technical Briefs as outlined in the Agency for Healthcare Research and Quality (AHRQ) Content and Procedures Guide for the Evidence-based Practice Center (EPC). The study protocol was published on the AHRQ Effective Healthcare website.²³ This manuscript reports measures of documentation burden that have been developed or are under development, characterizes the populations in which they have been used, and summarizes available validity evidence for these measures.

Inclusion Criteria and Literature Search

We applied the following inclusion and exclusion criteria for the studies identified in the literature search (supplemental Table 1).

We searched Embase, Epub Ahead of Print, In-Process & Other Non-Indexed Citations, MEDLINE Daily, MEDLINE, Cochrane Central Register of Controlled Trials, Ovid Cochrane Database of Systematic Reviews, and Scopus from January 1, 2010, to December 7, 2023. Reference mining of relevant systematic reviews and eligible primary studies was conducted to identify additional literature. The search strategy was developed by a medical librarian and peer-reviewed by an independent information specialist. The detailed search strategy is listed in Appendix. Each abstract was screened by two reviewers and conflicts were advanced for full-text screening. Independent pairs of reviewers screened the full-text version of eligible references and discrepancies were resolved through consensus.

Engagement of Key Informants and Gray Literature Search

To supplement findings from the literature search and gain additional contextual information, we recruited nine Key Informants (KIs) including an internist, cardiologist, surgeon, nurse, researchers, policymakers, and an EHR trade association representative. We conducted two group conferences to collect input on the review questions, experiences, opinions, and challenges related to documentation burden. We searched various websites of regulatory agencies, patient advocate groups, EHR vendors, and professional societies. A Supplemental Evidence and Data for Systematic Reviews portal was posted to collect additional information from interested parties. A Federal Register Notice was posted for this review. Public comments were solicited by posting the draft report on the AHRQ Effective Health Care website.

Data Extraction and Synthesis

We developed a standardized data extraction form to extract study characteristics, settings, measures of documentation burden, and validity evidence. We summarized measures of documentation burden in evidence tables and visual depictions. Documentation burden measures were narratively synthesized into distinctive categorizes based on similarity of the identified measures through consensus among the team. We built on an existing framework from Moy et al.⁶ and added additional measures identified in the current search. Evidence supporting validity of measures was evaluated following the model developed by Messick and adapted by Cook and Beckman.^24,25 The model identifies five sources of evidence that can support construct validity, which are content, response process, internal structure, relations to other variables, and consequences.

RESULTS

Literature Search

The literature search strategy identified 5653 citations, of which 135 articles were included. Description of the included studies is provided in supplemental Table 2. The process of study selection is depicted in Fig. 1.

Figure 1.

The process of study selection.

Settings

Measures were used across a diverse range of settings, including ambulatory and inpatient practices, primary care and specialty care, surgical and nonsurgical settings, rural and urban sites, academic and community-based medical centers, and private practices. The most commonly studied group of healthcare professionals was physicians in the USA, but there were studies from Canada, the UK, Germany, the Netherlands, Switzerland, Saudi Arabia, China, Taiwan, and Australia. Clinician populations included faculty and resident physicians, nurse practitioners, physician assistants, clinical psychologists, registered and licensed nurses, social workers, occupational and physical therapists, dietitians, and speech pathologists.

Available Measures

Narrative synthesis identified 11 types of measures of documentation burden, summarized in Fig. 2.

Figure 2.

Categories of measures of documentation burden with examples.

Overall Time Spent in EHR

Total time spent in the EHR was tracked most commonly through EHR usage logs.^8,14,26–56 This was tracked through time-motion analysis in five studies.^8,57–60 Two studies tracked time remotely through a smartphone or tablet applications.^61,62 Five studies reported subjective EHR use.^63–67

Activities Related to Clinical Documentation

Time specifically spent in clinical documentation activities was most commonly tracked through EHR usage logs, as both total time spent on clinical documentation specifically and the proportion of total EHR time spent in clinical documentation.^{8,30,32–36,40,41,43–45,47,48,51–55,62,66,68–86} Additional time measurement of clinical documentation activities occurred via smartphone or tablet applications^{39,62,77,87–98} or video time-motion recording.^58,60 Measures related to clinical documentation other than time spent in clinical documentation activities included the number of flowsheet-related entries such as vital sign entry,^59,60,99,100 documentation length,^{30,42,44,62,70,73,85–87,94,101–105} number of notes,⁷⁵ and number of actions taken to complete each note.^74,84,85,106 Some studies provided subjective assessments of clinical documentation burden, including perceived sufficiency of time for documentation and estimated time spent on clinical documentation.^{10,65,107–118}

Inbox Management

Measures of inbox management-related documentation burden included tracking of time spent specifically in this activity,^{8,30,32,33,35,37,39,40,46,47,51–55,71,78,79,81,83,87,101,103,119,120} and volume/number of messages.^{44,46,49,52,55,119–122} Self-reported assessment of inbox burden was described in several studies, including time elapsed before responding to inbox messages,¹²³ time spent on inbox management,¹²⁴ and estimated number of messages received.⁶⁴

Time Spent in Clinical Review

Time spent in clinical review activities was tracked most commonly through EHR usage logs,^{8,32,35–37,40,42,46,47,52–54,61,62,71,78,79,81,83,87,101,120} followed by time-motion analysis.^{39,58,61,91,93,97,98} One study tracked time spent in clinical review remotely through a smartphone or tablet application.⁶¹ Subjective reporting of time spent in clinical review was described in three studies.^34,117,124

Time Spent in Orders

Documentation burden associated with time spent in clinical orders was assessed through EHR usage logs,^{30,32,33,37,39,45,53,61,71,81,83,87,125} in time-motion studies monitored via iPad,⁹⁶ or in person.^58,94 One study examined orders placed remotely through a smartphone or tablet application.⁶¹ Subjective reporting of time spent in orders was described in one study using a 5-point scale.¹²⁴ Other measures measuring the burden of orders included an assessment of the number of orders entered.^39,101

Work Outside Work/After Hours

Work activities occurring outside of usual clinical time were most often tracked through the EHR, with varying definitions. Sample definitions included total time active after hours (7:00 P.M.–7:00 A.M.) on scheduled clinic days and time active anytime on unscheduled days,¹²¹ time spent in the EHR on unscheduled days, and “pajama time” (5:00 P.M.–7:00 A.M.).¹²⁶ Several studies reported subjective measures of work outside of regular work hours (self-reported), such as with a survey.^123,124

Administrative Tasks

Three studies objectively evaluated documentation burden associated with billing- and insurance-related administrative tasks using EHR automated tracking features.^8,32,53 One study used video time-motion analysis to evaluate the burden of administrative tasks.⁵⁸ Five studies subjectively examined the burden of administrative tasks.^{116,124,127–129}

Fragmentation of Workflow or Multitasking

Task switching has been evaluated in one study that measured switching between different EHR tasks⁷¹ and another that assessed time spent on uninterrupted EHR documentation.⁵⁹ Moy et al.^81,130 measured time spent on single tasks and the number of task switches per minute to measure workflow fragmentation in different clinical settings (e.g., intensive care unit (ICU), emergency department (ED), acute care, and ambulatory clinic). Additional measures of fragmentation included number of weekly interruptive alerts¹³¹ and number of interruptions per hour.⁸⁹ One study showed that the majority of attendings physician multitasked while residents were actively staffing cases.¹³² A time-motion study demonstrated a high frequency of multitasking and task transitions in clinical documentation, with resulting fragmentation in clinical work.^95,96

Measures of Efficiency

Several studies reported objectively assessed efficiency measures which can be further subcategorized as follows: timely completion of documentation,^34,66,72,133 time to chart/encounter closure,^{55,72,105,132,134,135} visits closed the same day,^{35,43,49,52,69,87,126,136} clinical encounters that were closed on the same day as the visit,^{30,42,55,56,62,104} charts closed within 72 h,⁷⁹ time spent in the EHR relative to expected time based on clinical workload,¹²¹ timely inbox completion,^43,46,49,52 documentation compliance,⁹⁷ number of steps in healthcare staff workflow,²⁸ mean nursing admission database dataset completion rate,⁷⁴ and time to completion of results, prescription requests, and patient messages.¹³⁷ Two studies subjectively assessed measures of efficiency via surveys asking about the frequency of closing encounters on the same workday and the number of notes closed within a 72-h period.^117,123

EHR Activity Rate

The number of actions required to complete clinical notes (defined as clicks, keystrokes, transitions, and mouse-keyboard switches) was a measure of burden in one study.¹⁰⁶ A Cerner Advance definition of three or more mouse clicks per minute, 15 or more keystrokes per minute, or 1700 or more mouse miles (pixels) per minute of mouse movement were applied in other studies.^26,27,78,83 Other measures included number of mouse clicks required to complete a nursing admission patient history, number of chart clicks in a 1-month period, chart clicks per minute, words per minute, number of logins per shift, number of charts reviewed per shift, and number of patient charts documented per shift.^{37,62,74,77,138}

Usability

Various aspects of EHR usability were subjectively assessed from the perspective of the clinician in several studies. Four of these studies used the system usability scale (SUS), which is a validated survey instrument.^7,91,110,139 The rest of the studies used nonvalidated surveys. Kadish et al.³⁵ assessed confidence in the EHR overall and in five key activities: placement of orders, documentation, chemotherapy ordering, clinical review, and inbox message management, using a 5-point scale. Tell et al.¹⁴⁰ evaluated “technostress” (stress experienced by end users in organizations attributed to EHR) on a 5-point scale.

Validity Evidence

The majority of the included studies did not provide sufficient evidence to establish validity of documentation burden measures. Figure 3 depicts the proportion of studies that adequately provided each of the five validity evidence types and supplemental Table 3 provides details about the individual studies:

1. Content evidence was judged to be inadequate in studies that simply measured time, number of clicks, or completion of tasks without clear linkage to documentation burden, whereas studies that associated time with specific documentation tasks and linked it to burden (approximately 11% of the studies) were considered to provide adequate evidence. For example, Gardner et al. associated the perception of insufficient time for documentation and excessive use of EHR at home with odds ratios for burnout of 2.8 and 1.9, respectively.¹⁰⁹

2. Response process evidence was considered adequate when the actions and thoughts of researchers and respondents in the studies intended to measure the burden associated with documentation (approximately 12% of the studies). For example, Kroth et al. conducted focus groups in which researchers and participants purposefully targeted documentation burden and its correlation with technology stress, ergonomic problems, poor interoperability between systems, EHR use at home, and excessive data entry requirements.¹¹²

3. Internal structure evidence was considered adequate when the measure was judged to be reliable and reproducible (approximately 8% of the studies). While many studies used EHR logs to capture documentation time, which is a reliable and reproducible method, we judged these studies to “partially” fulfill this criterion because time itself was a surrogate for burden and because EHR logs may miss inactive users and those who are multitasking on their screens. Time-motion studies and studies with observers may capture true documentation time better. For example, Arndt et al.⁸ and Karp et al.⁷⁷ used parallel time-motion studies to validate EHR measures.

4. Relation to other variables evidence was considered adequate when studies correlated the time measure with other measures of burden or clinician stress or satisfaction (approximately 27% of the studies).

5. Consequences evidence was considered adequate when documentation time was associated in a study with burnout, patient outcomes or satisfaction, or teaching time (approximately 31% of the studies). For example, Baugh et al. demonstrated that every minute spent on documentation in the emergency department was associated with 0.48 fewer minutes spent on teaching (p < 0.05).¹⁴¹

Figure 3.

Percentage of studies providing validity evidence.

Several studies applied instruments to assess factors related to documentation burden such as work stress and well-being but did not report on the validity of documentation burden measures themselves.^{42,113,114,129,142} Some used the SUS, a validated survey, although SUS captures only limited aspects of documentation burden.^7,91,110,139 Benson et al.¹²² applied a validated survey instrument to evaluate the impact of health IT, including the EHR, on clinician job satisfaction. Gesner et al.¹¹⁰ applied the Burden of Documentation for Nurses and Mid-wives which had a previous content validity analysis.¹⁴³

DISCUSSION

Documentation burden in healthcare continues to increase. A recent study comparing the period from 2019–2020 to 2022–2023 showed a significant increase in the time primary care physicians spent in the EHR across most tasks.¹⁴⁴ A strong association between documentation burden and burnout has been described in many studies, adversely impacting personal and professional job satisfaction, work-life balance, and experiences of burnout.^42,128 Studies report decreased satisfaction with the EHR across clinical roles, specialties, and geographic locations.^76,112 Total time spent in the EHR is associated with decreased patient satisfaction, lower communication ratings, and lower likelihood of recommending the physician.¹⁴ Time spent documenting is significantly associated with less time spent on teaching.¹⁴¹

While many researchers and some EHR vendors are trying to develop documentation measures, these measures are in preliminary stages, do not appear to be consistent with proposals of multidisciplinary informatics workgroups, and are not multi-dimensional.^145,146 No single established definition for documentation burden exists, which contributes to the difficulty in creating valid and scalable measures. From a total of 135 studies, we identified 11 categories of measures for documentation burden, from overall time spent in the EHR to usability. The most common source for most measures was EHR usage logs. Specific work functions contributing to documentation burden were represented in the set of measures, including clinical notes, flowsheet entry, inbox management, clinical review, order entry and review, and administrative clinical support tasks, such as billing and insurance-related documentation. The capture of the latter administrative tasks is likely underestimated in existing measures. For example, medication-related pre-authorizations can involve lengthy phone calls with pharmacy staff and insurers, as well as providing information through websites outside of the EHR. Measures included both time during scheduled work and time required outside of work and after hours, although definitions of these parameters varied. Efficiency was also an important category of measures addressing workflow fragmentation, multitasking, time to complete documentation requirements, and EHR activities down to the level of individual keystrokes and mouse clicks.

Strengths and Limitations

A key strength of this review is the systematic search of the literature, applying methodologically rigorous techniques including a medical librarian-developed search strategy of multiple databases and duplicate assessment of study selection. The timeline selected for this report aligns with a major shift in documentation burden occurring with enactment of the Health Information Technology for Economic and Clinical Health Act in 2009; hence, the conclusions reflect modern practice experiences.

In terms of limitations, the literature offers the greatest relevance to the US healthcare system and may not represent documentation burden in other systems. Thematic synthesis can be conducted in multiple ways that can all be reasonable. For example, clinician satisfaction and system usability are interrelated concepts, as are scalability and feasibility. Thus, a different categorization of the measures or their validity evidence types is possible. Heterogeneity in how different organizations implemented and configured their EHR limits comparative inferences across studies. There was limited knowledge of validation of vendor-derived measures and their reproducibility. The literature also contains very little discussion about the appropriateness of measures according to considerations of scalability, feasibility, or equity across clinical settings, clinician and patient/caregiver populations, and geographic locations. Although clinician perspectives were generally consistent about the impact of the burden, certain groups were poorly represented (e.g., pharmacists, medical assistants). Patient perspectives were also not commonly reported. Literature evaluating the role of patients in documentation burden beyond the volume of communications and associated time demands for associated paperwork, chart review, and responses is extremely limited.

CONCLUSIONS

While numerous measures of documentation burden exist and have been tested in a variety of settings and contexts, renewed attention to fundamental instrument development procedures such as establishing validity and reliability is needed. Measures should explicitly map to specific burden domains and address multiple domains of this construct. Parts of the documentation process that do not serve patient needs may need to be eliminated and subsequently not measured. Future research should explore the impact of patient portals on documentation burden. This report serves as a call to action and emphasizes the urgency of the problem. Identifying measurement gaps is just the first step that is required as a basis for developing interventions and solutions.

Supplementary Information

Below is the link to the electronic supplementary material.

Declarations:

Conflict of Interest:

The authors declare that they do not have a conflict of interest.

Disclaimer:

The findings and conclusions in this document are those of the authors, who are responsible for its contents; the findings and conclusions do not necessarily represent the views of AHRQ or the U.S. Department of Health and Human Services. Therefore, no statement in this report should be construed as an official position of AHRQ or of the U.S. Department of Health and Human Services.