The Electronic Medical Record and Nephrology Fellowship Education in the United States

Discussion

Our survey elicited opinions and observations from program directors, clinical faculty, and fellows at United States nephrology training programs regarding the perceived effect of the EMR on the general and nephrology-specific clinical learning experience, system performance, and clinician time commitment to data entry. There were potentially multiple respondents at each training site, thus reducing bias when comparing fellows, faculty, and program directors. Programs surveyed were similar in size to the average (six fellows per program), and proportions of first-year and second-/third-year fellows were approximately those reported for training year 2019 (24,25).

Limitations include potential nonresponse bias, which may be (but not necessarily) associated with a lower response rate (26,27). Response rates were as expected for an online physician survey, considering delivery complexity and because reminders were delivered by program directors rather than investigators (28–⇓30). Wave analysis indicated that late responders (surrogate “nonresponders”) were similar to earlier responders. We attempted to reduce nonresponse and social desirability bias using anonymity, few demographic questions, and no questions regarding training site or EMR vendor. We did not ask how long the EMR had been in use; unless a program had a new EMR, first-year fellows would have been using it for at least 6 months.

Nephrology patients are among the most medically complex in terms of medications prescribed, number of comorbidities, and mortality risk (31). The EMR, which facilitates access to laboratory and imaging data, updated medication lists, past medical history and diagnoses, and standardized dialysis order sets, should improve nephrology patient care and the clinical learning environment.

About 50% of fellows indicated that the EMR contributed positively to their education. Faculty and program directors did not differ significantly from nephrology fellows. This is consistent with previous surveys demonstrating positive perceptions of EMR effect on graduate medical education. At a large academic health center, >50% of faculty and trainees (residents and fellows) had a positive overall impression of the EMR during the initial implementation and 2 years afterward (17). Another physician survey (including residents and fellows) at an academic hospital showed that 58% were satisfied with their EMR, and 61% felt that it had a positive effect on patient care (32). In a 2007 survey of a large, representative sample of United States physicians, the majority reported being satisfied overall with their outpatient EMR (33). However, in our survey many respondents described unintended and deleterious consequences for fellows’ clinical experience. Although 50%–60% reported that patient medications could be accurately determined, for others medication lists are often incorrect and not updated—a focus issue for the 2019 Centers for Medicare & Medicaid Services (CMS) Promoting Interoperability Program (PIP; i.e., “Meaningful Use”) (6). Few EMRs seem to assist with drug dosing on the basis of eGFR. Problem lists and past medical history may be a series of uncurated “billing diagnoses.” Drop-down picklists and prompts for dialysis and continuous RRT orders assist fellows in entering correct and complete orders, but program directors were concerned that these aids may prevent fellows from learning to write orders without prompts. Nephrology-specific EMR tracking of CKD and dialysis quality indicators, patient mix, and post-transplant protocol adherence are underused and in some cases, prohibited.

Many respondents reported serious technical, interoperability, and usability problems. Over 40% reported that function was “slowed, disrupted, or completely lost” monthly or more. Not surprisingly, these were significantly less likely to agree that their EMR contributed positively to education. EMR downtime is associated with adverse patient safety events (34), and frequent disruptions likely adversely affect the clinical learning environment. Over 30% of respondents reported that outpatient, inpatient, and dialysis unit EMRs did not intercommunicate. Interoperability, another focus of the 2019 CMS PIP, is impeded by proprietary issues (1). Usability problems were common. Poor usability may lead to increased cognitive load and clinician error, directing attention away from data synthesis and clinical decision making; improving usability has been shown to associate with improved cognitive performance (3,4,9,35).

The major negative reported by all respondent categories was excessive EMR documentation time associated with fellow reluctance to prolong direct patient interactions. Reported documentation time for fellow clinic encounters would, on average, leave little time for interacting directly with the patient. Outpatient practice physicians may spend up to twice the time in EMR documentation as in patient face-to-face time, and outpatient data entry time is inversely associated with patient satisfaction (13,36). The majority of fellow respondents also felt reluctance to perform nephrology procedures, perform case-directed reading, and attend didactic conferences—shifting emphasis from education (the purpose of fellowship) to data entry. Procedural competency in kidney biopsy and temporary dialysis catheter insertion is required by the Accreditation Council for Graduate Medical Education, and the survey suggests that EMR time burden may affect procedural proficiency. It may also affect attendance and engagement at academic conferences, the predominant educational modalities in nephrology fellowship curriculum (30).

Bloated, inaccurate progress notes (due to copy forward, prepopulation, and irrelevant data entry) lead to data obfuscation, and the majority of faculty surveyed had observed copy-forward errors in nephrology fellow notes. Copy forward and prepopulation decrease EMR documentation time—creating a vicious cycle (3,2,9,13,18). EMR entry time burden is associated with frustration and burnout (6,37). For fellows, who must comply with work-hours limitations, EMR documentation demands may lead to severe work compression, with associated irreconcilable and opposing professionalism requirements. This is reflected by the majority of fellows reporting sometimes or often exceeding work-hours limitations—choosing medical record completion and patient care over work-hours adherence (38).

What can be done? Hospital administration should focus on minimizing EMR dysfunction/downtime, which in the interest of patient safety, provider wellness, and the clinical learning environment, should be exceedingly rare. Documentation policies addressing copy forward and other practices that lead to data obfuscation should be considered (3,13). The problem of work compression must be faced—there are limits to clinical efficiency, especially in graduate medical education. Usability, interfaces, and interoperability must be improved, with reduction of data entry time (2,39). New technologies are needed, including digital scribes, continuously updated smart links to published guidelines (stimulating evidence-based decision making), and multidisciplinary, clinically driven, pertinent, fluid patient problem lists.

In summary, this survey of nephrology program directors, faculty, and fellows indicates that the EMR is perceived to enhance educational experience by promoting efficient and flexible access to laboratory/imaging results and to a lesser extent, medication reconciliation and standardized dialysis order sets. EMR data entry time is excessive due to poor functional performance/usability and medical record obfuscation. Time pressures of EMR data entry and resulting work compression may increase fellow reluctance to engage in educational opportunities, encourage work-hours noncompliance, and stifle direct patient interactions. This may apply to graduate medical education generally, perhaps more significantly in those specialties, like nephrology, with highly complex patients. Faculty should be aware of trainee documentation time, advocate for improved EMR functionality/usability, and promote practices that improve efficiency and reduce work compression.