Dark Clouds With Silver Linings: Resident Anxieties About COVID-19 Coupled With Program Innovations and Increased Resident Well-Being

Burnout among health care workers has been associated with health effects including depression, substance abuse, and suicidal thoughts, as well as work effects including medical errors, lower productivity, and greater job turnover.¹  Although an estimated 46% of health care workers experience burnout at any given time, residents display greater proportions of burnout than medical students, fellows (subspecialty trainees), and attendings.^2–4 

The COVID-19 pandemic may have increased the risk of burnout by straining residency training programs and residents. Residents in multiple specialties experienced many of the same COVID-19 stressors as other health care workers, including lack of personal protective equipment, inadequate training to care for COVID-19 patients, and the associated feeling of helplessness, childcare concerns, and the risk of contracting or passing the virus on to family and patients.^5–15  Indeed, these concerns are not unlike those which have been reported for other outbreaks such as SARS and Ebola.^16–18  Residents experienced disruption of their training programs, uncertainty about day-to-day assignments, and concerns about future fellowships or jobs.⁹  Residency program directors reconfigured their programs abruptly, some relying on the Extraordinary Circumstances policy of the Accreditation Council for Graduate Medical Education.¹⁸  The sudden suspension of program requirements and the focus on the singular challenge of caring for COVID-19 patients led many programs to innovate quickly and implement new approaches to manage the clinical learning environment and resident well-being.

The iCOMPARE study of resident work hours in internal medicine revealed that a substantial proportion of residents experienced burnout and dissatisfaction.¹⁹  This follow-up study was designed to solicit a prospective evaluation of internal medicine residents through biweekly surveys during the 2019–2020 academic year, during which participating residents would respond to short questions on different aspects of their well-being. When the COVID-19 pandemic spread across the United States, we redirected our prompts to focus on the pandemic's effect on resident well-being. Here, we report on those COVID-19-specific responses, highlighting common themes that drove local innovation in the early days of the pandemic. These findings may inform residency programs more generally.

All 394 residents in 4 residency programs that participated in the original iCOMPARE trial were invited to participate in a longitudinal study over academic year 2019–2020. The 4 programs comprised 2 large university programs, one small university program, and one small community-based university-affiliated program. The 88 participating residents were asked to respond to an open-ended prompt on 18 occasions roughly 2 weeks apart.

Of the residents expressing interest in participation, we randomly selected 22 to 24 per program to participate, based on a goal of 15 to 18 residents per site and an expected 20% to 30% dropout rate. Surveys responses were tracked with a blinded study ID. Analytic data sets were anonymized. Participants received a $100 Amazon gift card each time they responded to 3 prompts. The study ran from October 2019 to May 2020. Prompts related to COVID-19 were disseminated to participants in March and April 2020.

Prompts were designed by the study team of current and former program directors, health services researchers, and qualitative researchers, many of whom had been studying resident well-being for years. Data were collected and analyzed by a team of research coordinators who were trained in qualitative methods but were not directly involved in graduate medical education. In general, ideas for the prompts were consistent with the biopsychosocial model of health that acknowledges the multiple layers of influence on individual actions and reactions and also with the conceptualization of burnout from Shanafelt and Noseworthy.^20,21  The 4 COVID-19-related prompts are listed in the Figure. Prompts and responses were distributed by RedCap email invitations. It is worth noting that in these early weeks of the pandemic there were substantial changes to operations for the residents in all 4 participating programs. Many fewer patients overall were admitted, and fewer caretakers of all types were in the hospital. But patients admitted with COVID-19 were primarily cared for by internal medicine residents and faculty. Specifically, the rapidly increased volumes of patients with COVID-19 with concomitant increases in critical care volumes resulted in redeployment of physicians to COVID-19-specific wards and ICUs.

The data were coded by 4 trained coders who met at least biweekly with a subset of the larger team. After all data were collected, the 4 coders met in alternating teams of 2 to initiate a formal content analysis of the responses, utilizing NVivo 12.0 (QSR International). Codes were developed within each prompt and iteratively through weekly meetings, resulting in a master codebook that captured the frequently occurring themes across prompts but also retained those unique to some prompts. Within a prompt, coding proceeded once coders achieved an average reliability rating of 0.80 on a sample of at least 20% of the data within a prompt and interrater agreement was assessed repeatedly.

The study was reviewed and deemed exempt by the University of Pennsylvania Institutional Review Board.

A total of 186 residents expressed interest, and 88 residents were enrolled (47%). Sociodemographic data for the residents are shown in Table 1. Of the 88 residents, 59% (n = 52) were female, and 68% (n = 60) identified as White. Site and postgraduate year (PGY) were nearly evenly distributed across participants. The average response rate was 89%. Participants responded to a mean of 16.8 of the 18 prompts overall, and a mean of 3.6 of the 4 prompts related to the pandemic.

Residents' early responses to the pandemic largely fell into 2 main themes: emotional impact and impact on day-to-day life. After a few weeks, some routine was established, and observations largely reflected the training program with themes of communication and innovations to program structure. Each of these themes are described below with exemplary quotations within the text; additional quotations characterizing these themes are included in Table 2.

Residents offered many reflections that highlighted their anxiety and fear about the pandemic, the new uncertainties in daily life, and shifting senses of stability. Specific worries were about lack of personal protective equipment and contracting the disease—mostly based on a fear of transmission to their family members and patients.

Early on, residents in all programs expressed frustration with communication from program and hospital leaders. Information was constantly provided, but often with conflicting or uncertain meaning.

Residents expressed disappointment and some anger with forced schedule changes. Bans on travel, requirements for social distancing, and the canceling of long-awaited vacations or celebrations were a hard adjustment. As the pandemic progressed there was worry about their education and learning. The COVID-19 pandemic had dominated residents' clinical experiences, and they worried about holes in their learning of “bread and butter” internal medicine.

Amid all the anxiety and worry, several residents expressed excitement. Being a physician on the ground in the wake of a pandemic reinforced primary motivations for entering the field.

Thanks to innovative schedule changes described below, residents also expressed appreciation for unstructured time in their days. While the pandemic unsettled residents, many embraced the schedule changes and work from home shifts that allowed for healthy lifestyle changes, related to sleep, exercise, meal planning, and reflection.

Residents missed interacting in-person with their friends and family but were engaged in frequent virtual socialization. Many mentioned reaching out to old friends and engaging in virtual dining, movie watching, and game nights.

Taking advantage of their time at home, some residents aimed to use the time to read and study.

As residents settled into new routines, many complimented their program leadership. They gave numerous examples of messaging platforms and improved communication that helped residents feel valued and heard. For example, some of the programs implemented weekly updates and forums.

The frequent forums and program director availability were widely appreciated. Residents emphasized the focus on wellness within their programs. There was a sense that it was genuine (sometimes contrasted to past perceptions of lip service), and there was a shared hope that the messaging and activities could be retained.

When prompted to comment on interprofessional teamwork, nearly all responses were positive. Residents wrote of more collaboration and shoulder-to-shoulder work with nurses, fellows, and attendings, sometimes attributing it to the lower census in the hospitals. They also commented that more teaching was occurring and that they were all in this together.

The COVID-19 pandemic required unprecedented changes to the training environment and schedules. Several residents complimented the agility of their program leadership, contrasting prior notions of stagnant training models. The innovation most often lauded was the removal of pre-rounding (daily check-ins with patients that commonly involve a physical examination, interview, and discussion of the patient's course and diagnostic plan).

Also appreciated was the general move to shorter rotations and shorter shifts. Sometimes this extended to schedule changes that included canceling some previously required rotations and providing more flexibility in selecting learning opportunities. Residents in all programs noted that program directors intentionally followed “hard” rotations with “easier” rotations.

Residents appreciated the opportunity for more self-directed learning in choice of conferences and some subspecialty rotations.

Responses were varied but generally positive regarding conference modality and schedules. Fewer conferences were scheduled and required. The move to virtual conferences was going well, although residents did miss getting together with their peers.

There were some mentions that a virtual format opened the doors for a chance to have more external speakers and perhaps even shared national curricula. In some cases, residents appreciated the opportunity for more teaching and preparation.

The move to telemedicine was largely, though not unanimously, well-received. Some residents noted that telemedicine provides more opportunity to get to know their patients.

Residents hoped that telemedicine would persist beyond the pandemic because of its ease of use for patients and physicians alike, and because it would allow residents to work from home.

Many residents reflected on how the structural changes to training “proved” that fewer people needed to be in-person, on a team, and around the bedside. Members of the team at home were still effective and engaged and clearly helped with the workload. Overall, the schedule changes enhanced efficiency.

This study has 3 main findings. First, the COVID-19 pandemic added stress to residency, but residents' own messages focused on resiliency and adaptation. Early on they expressed a cluster of negative sentiments (eg, anxiety, disappointment, and fear). Sudden and frequent schedule changes and mixed signals from their program leaders unsettled residents. But residents adjusted relatively quickly; as a whole, residents took advantage of more unscheduled time and more time at home for wellness activities, distanced social connections, and reading. Our findings of resilience in internal medicine residents resemble the adaptability reported in other specialties in response to the COVID-19 pandemic^9–14  and other outbreaks such as Ebola and SARS.^16–18 

Second, the historically slow pace of change that characterized graduate medical education pre-COVID-19 was dramatically accelerated as programs quickly developed new solutions to long-standing challenges. Residents were surprised by their programs' agility in creating many structural alterations in schedules, education, and patient interactions. Many of the program changes reported by our participants mirror what was happening in multiple specialties.^6,7,9–14  The majority agreed that many changes should become permanent; for example, elimination of pre-rounding, shorter blocks and shifts, more attention to the sequence of rotations, virtual rounds, and telemedicine. Equally appreciated were the program directors' attention to well-being and opportunities for more individualized learning. However, positive sentiments were not unanimous. In each program some residents wished for (almost) everything to go back to how it was. Some residents missed past approaches to learning, especially missing “bread and butter” internal medicine in exchange for the COVID-19-focused education.

Third, residents themselves identified tired conventions to abandon forever. Just as the pandemic has led people in other professions to question whether they should return to commuting to an office worksite, business travel, or to formal clothes and uncomfortable shoes,²²  resident participants now question all the conventions of residency that couldn't get a fresh look until now: medical hierarchy, long shifts, and infrequent messaging from administration.

We aimed to engage internal medicine residents for a year, to gain insight into how residency affects their well-being. Unexpectedly, the COVID-19 pandemic became a central part of those experiences^23,24  providing us with a unique resident view we report here. The story's arc takes us from personal and professional concerns through empowerment and purpose toward innovative ways to restructure residency training in the post-pandemic era. We could not have anticipated the pandemic when we designed and launched this study, but it supported our goal to understand how to help residency programs support resident well-being.

This study has limitations. Residents came from programs in the northeastern United States who were volunteers in a study about well-being. They had already received 12 rounds of prompts encouraging them to think about well-being before the pandemic, possibly skewing their responses to highlight well-being. The results we report cannot be disentangled from the prompts we asked. While we did collect data on residents' sociodemographics, we chose not to dive deeply into subanalyses given few observed global differences. As in any study, there is a risk of desirability bias. Finally, we collected these data in the early weeks of COVID-19 in the United States. Residents' perspectives on the impact of the pandemic, and the durability of early program changes, may be different a year later.

Repeating this study in other specialties, as COVID-19 cases continue and programs recalibrate, would help to generalize our findings. More importantly, these data were collected at a unique point in time. Residents' internal responses to external events and the many program adaptations should be reevaluated to learn what was sustained and why. Future efforts should examine the feasibility of maintaining the discussed program changes made in response to the COVID-19 pandemic and their effects on resident job performance, patient outcomes, and program director satisfaction.

The COVID-19 pandemic added stress to the lives of internal medicine residents, but has also revealed human resilience, program agility, and what may be new and enduring approaches to internal medicine residency.

The authors would like to thank Andrea Bilger from the University of Pennsylvania Mixed Methods Research Lab and Claire Bocage for their assistance.