Electronic Health Record Use among Ophthalmology Residents while on Call

Abstract

Background As electronic health record (EHR) use becomes more widespread, detailed records of how users interact with the EHR, known as EHR audit logs, are being used to characterize the clinical workflows of physicians including residents. After-hours EHR use is of particular interest given its known association with physician burnout. Several studies have analyzed EHR audit logs for residents in other fields, such as internal medicine, but none thus far in ophthalmology. Here, we focused specifically on EHR use during on-call shifts outside of normal clinic hours. Methods In this retrospective study, we analyzed raw EHR audit log data from on-call shifts for 12 ophthalmology residents at a single institution over the course of a calendar year. Data were analyzed to characterize total time spent using the EHR, clinical volume, diagnoses of patients seen on call, and EHR tasks. Results Across all call shifts, the median and interquartile range (IQR) of the time spent logged into the EHR per shift were 88 and 131 minutes, respectively. The median (IQR) unique patient charts accessed per shift was 7 (9) patients. When standardized to per-hour measures, weekday evening shifts were the busiest call shifts with regard to both EHR use time and clinical volume. Total EHR use time and clinical volume were greatest in the summer months (July to September). Chart review comprised a majority (63.4%) of ophthalmology residents' on-call EHR activities. Conclusion In summary, EHR audit logs demonstrate substantial call burden for ophthalmology residents outside of regular clinic hours. These data and future studies can be used to further characterize the clinical exposure and call burden of ophthalmology residents and could potentially have broader implications in the fields of physician burnout and education policy.

Using event logs to observe interactions with electronic health records: an updated scoping review shows increasing use of vendor-derived measures

“…Reviewed articles mentioned several limitations of EHR log research ( Supplementary eTable 4 ). Three of the most frequently mentioned limitations echo those observed in the prior review of audit log research 11 : EHR logs do not provide a full view of clinical activity which can involve physical and digital interactions outside the EHR (22 articles), 3 , 26 , 34 , 43 , 55 , 57 , 58 , 61 , 65 , 68 , 73 , 74 , 84–86 , 89 , 91 , 92 , 95 , 103 , 109 , 112 qualitative methods are needed to better understand the context and motivation for observed work (15 articles), 32 , 46 , 53 , 62 , 64 , 68 , 72 , 80 , 85 , 91 , 96 , 97 , 101 , 115 , 118 and logs may not contain enough detail to observe complex workflows (13 articles). 4 , 26 , 32 , 43 , 51 , 56 , 59 , 69 , 75 , 81 , 89 , 90 ,…”

“…Looking at specific measures of EHR use reported in each study reveals additional differences, particularly regarding durations of EHR use ( Figure 4 ). Reported time-based measures included total time in the EHR (49 articles), 3 , 4 , 19 , 24–33 , 35–49 , 51–59 , 61 , 66 , 69–73 , 77 , 79 , 95 , 98 time in notes (32 articles), 3 , 4 , 7 , 19 , 24–32 , 35–38 , 40–46 , 48 , 51 , 55–58 , 79 , 109 time in inbox (23 articles), 2–4 , 19 , 24 , 25 , 27–29 , 35–38 , 42 , 43 , 45 , 47 , 48 , 51 , 61 , 62 , 70 , 71 time in chart review (23 articles), 3 , 4 , 24–27 , 29 , 31 , 32 , 35 , 36 , 40 , 42 , …”

“…While 70 articles measured a duration of time, 2–4 , 6 , 7 , 19 , 24–62 , 66–73 , 76 , 77 , 79–81 , 83 , 92–96 , 98 , 102 , 108 , 109 , 111 , 119 12 of these measured a duration between 2 specific points in time (e.g., duration of an appointment, or shift), 60 , 68 , 76 , 80 , 81 , 83 , 92–94 , 96 , 102 , 111 leaving 58 articles which measured a duration of active EHR use. 2–4 , 6 , 7 , 19 , 24–59 , 61 , 62 , 66 , 67 , 69–73 , 77 , 79 , 95 , 98 , 108 , 109 , 119 Of these 58 articles, 34 described (or referenced an article that described) how active use was defined. 3 , 4 , 19 , 24–29 , 31 , 32 , 35 , 39 , <...…”

“…Four of these limitations related to measure accuracy and granularity, each of which was mentioned in both investigator-measure and vendor-measure studies with no difference in reporting rates ( P > .05 in each case). Seventeen articles argued that current measures may systematically underestimate EHR use, 2 , 4 , 6 , 26 , 27 , 43 , 48 , 57 , 61 , 62 , 70 , 73 , 79 , 80 , 83 , 88 , 119 for example, undercounting time on activities such as chart review which may involve periods of reading without clicks or keystrokes. Alternatively, 6 articles suggested current methods may overestimate EHR use, often referring to being unable to distinguish ambulatory and inpatient EHR use or in-clinic and at-home use.…”