Pediatric Resident Perspectives on the Impact of COVID-19 on Training

We performed a national, cross-sectional study of categorical pediatric residents in Association of Pediatric Program Director (APPD)–affiliated residency programs. The Institutional Review Board at Boston Children’s Hospital reviewed the protocol and determined that it qualified as exempt.

The study population included all US categorical pediatric residents in APPD-member pediatric residency programs at the time of survey distribution (academic year: 2019–2020), which includes the majority of Accreditation Council for Graduate Medical Education–accredited programs. We excluded residents in combined training programs because we anticipated those residents were uniquely impacted by their training experiences in their nonpediatric specialty.

After thorough literature review, the survey was designed by members of both APPD’s Associate Program Director (APD) executive committee and APPD Longitudinal Education Assessment Research Network with collective expertise in residency training and survey methodology, thereby achieving content validity evidence,¹⁸  and administered by using LimeSurvey (Hamburg, Germany). Before survey dissemination, we performed cognitive interviewing and pilot testing with a convenience sample of 5 residents and 7 chief residents, resulting in further refinement of the instrument and response process evidence.¹⁸  In addition to demographic questions, the survey (Supplemental Information) included questions focused on training experiences (rotations, continuity clinics, telemedicine, adult patients, exclusions from care, PPE, COVID-19–related missed work, and resident parents), postresidency employment plans, and attitudes and perceptions (about the impact on training, personal wellness, and their residency programs).

An e-mail describing the project with a survey link was distributed to APPD-member pediatric program directors and APDs. Program leadership was asked to forward the e-mail to their current residents. The survey was initially distributed on May 28, 2020, with 3 follow-up e-mails over 6 weeks.

To calculate the number of residents who received the e-mail inviting them to participate in the study at least once, we assumed that program leadership forwarded the survey to their residents if at least 1 resident from that program completed the survey. Statistical analyses were primarily descriptive. When estimating the effects of covariates on resident-level outcomes, we applied generalized mixed-effects models to control for clustering of residents within programs (linear models for continuous outcomes and logistic models for dichotomous outcomes).¹⁹  When comparing groups or estimating associations between variables at the program level, we used nonparametric tests for outcomes that did not appear normally distributed (χ² for categorical outcomes and Wilcoxon rank tests for continuous outcomes). As a sensitivity analysis, we repeated tests of selected research questions using the subsample of responses from programs with >40% response rates. The study team a priori identified those areas at high risk for within-program nonresponse bias, primarily related to the questions examining resident attitudes and perceptions. A total of 40% was intentionally chosen to balance reducing the risk of nonresponse bias with including enough programs to represent a broad sample. R 3.6 (R Core Team, Vienne, Austria) was used for all analyses. P values <.05 were considered statistically significant.

The survey response rate was 18.9% (1141 out of an estimated 6032 residents), which included at least 1 respondent from 63.2% (127 of 201) of APPD member residency programs. The median individual program response rate was 16% (interquartile range: 9% to 24%). Thirteen programs, representing 260 residents, had a >40% response rate. Respondent demographic data (Table 1) reveal that postgraduate year 3 (PGY-3) and above residents, male respondents and international medical graduates (IMGs) were underrepresented, in comparison with national data.²⁰  Demographic data of all respondents were comparable to that from programs with a >40% response rate except for male respondents, who were again underrepresented (Supplemental Table 4). Respondent program demographic data (Table 2) reveal that community and military sites were underrepresented in programs with responses, compared with those without. Program demographic data were not significantly different from programs with a >40% response rate (Supplemental Table 5).

Respondents reported frequent rotation adjustments during the pandemic (Supplemental Table 6). The 2 most frequently added rotations were jeopardy call (13.7%; 156 of 1141) and pandemic elective (10.3%; 118 of 1141). We also analyzed adjustments to scheduled rotations. Therefore, respondents that checked “N/A (never scheduled and have not completed)” were excluded from these analyses. Clinical electives (44.5%; 333 of 749) and advocacy and community health rotations (23.3%; 80 of 344) were the scheduled rotations that were most frequently canceled as a result of the pandemic. Respondents who were scheduled for specific rotations reported having fewer shifts than normal, most commonly on pediatric hospital medicine (29.0%; 260 of 896) and inpatient subspecialty rotations (22.1%; 142 of 643).

The vast majority of respondents had continuity clinics canceled, with 33.0% (374 of 1132) reporting all clinics canceled, 55.2% (625 of 1132) reporting some clinics canceled, and only 11.7% (133 of 1132) reporting no clinics canceled.

Resident participation in telemedicine increased significantly during the pandemic, compared to prior in both inpatient and outpatient settings (inpatient: 11.3% [129 of 1141] vs 1.0% [11 of 1141; P < .001]; shadowed outpatient: 20.3% [232 of 1141] vs 3.2% [37 of 1141; P < .001]; active role outpatient: 53.2% [607 of 1141] vs 2.5% [29 of 1141; P < .001]).

Approximately 17% (200 of 1141) of respondents reported caring in-person for adult patients who they would normally not have cared for, outside of the pandemic, either at their usual hospital or another health care setting. Of the 127 programs, 36.2% (46) had at least 1 respondent report caring for these adult patients. A total of 79% (158 of 200) of the respondents who reported caring for adult patients, from 39 programs, cared for adult patients in at least 1 pediatric practice setting; 32.5% (65 of 200), representing 20 programs, cared for adult patients in at least 1 adult practice setting; and 11.5% (23 of 200) cared for adult patients in both settings. Of the respondents who cared for adults in an adult practice setting, 55.4% (36 of 65) were mandated to do so, 42% (27 of 65) volunteered, and 3% (2 of 65) did not answer the question.

At least one-third of respondents reported being excluded from the care of patients with suspected or confirmed COVID-19 (36.5% [405 of 1109] and 36.9% [409 of 1109], respectively); 51.6% (572 of 1109) of respondents reported never being excluded from either patient type. In contrast, only 8.5% (94 of 1100) and 19.5% (215 of 1100) of respondents thought residents should be excluded from the care of patients with suspected or confirmed COVID-19, respectively. Additionally, 78.9% (868 of 1100) of respondents reported that residents should not be excluded from either patient type.

Among respondents who provided care during which the use of PPE was indicated, 29.4% (331 of 1124) reported a time when they did not have access to appropriate PPE, on the basis of guidelines specific to their institutions at the time. Of the 127 programs, 80.3% had at least 1 respondent report lack of access to appropriate PPE.

Although only 2.1% (24 of 1141) of respondents had confirmed COVID-19, 17.5% (200 of 1141) reported missing work because they developed symptoms consistent with COVID-19. Respondents who reported confirmed or possible COVID-19 infection reported missing several days of work (confirmed COVID-19: mean of 9.5 days [SD: 1.5]; possible COVID-19: mean of 4.0 days [SD: 0.3]).

A total of 15% of respondents reported being responsible for the care of a child or children living in their home. Of those respondents, 97% (166 of 171) reported experiencing challenges in that role, including worries about infecting their children (87.1%; 149 of 171), day care closure (44.4%; 76 of 171), school closure (27.5%; 47 of 171), concern about high-risk individuals providing child care (33.3%; 57 of 171), and nanny or child care provider unable to provide care (27.5%; 47 of 171).

Of the graduating resident respondents not entering fellowship or a chief residency year, 52% (64 of 123) reported that their postresidency employment was impacted by COVID-19 in ≥1 ways. Respondents reported hiring freezes (23.6%; 29 of 123), delayed employment start dates (13.0%; 16 of 123), and inability to travel for interviews (8.13%; 10 of 123) as the most common adversities.

For the following results, we performed a sensitivity analysis with only respondents from programs with a >40% response rate. This analysis reveals comparable results to all respondents (Supplemental Table 7; Supplemental Fig 2).

Overall, respondents reported a mostly negative impact of the pandemic on various components of their training (Fig 1), with the most negative impact on outpatient clinical education (70.9%; 809 of 1141 reported a negative impact) and educational conferences (63.0%; 719 of 1141 reported a negative impact). Interestingly, more respondents reported a positive (29.1%; 332 of 1141) rather than negative impact (5.8%; 66 of 1141) on duty hour compliance.

Personal and residency program impacts are described in Table 3. Most respondents reported no change in their desire to practice medicine as a result of the pandemic. However, respondents who were parents were more likely to report a decreased desire than residents without children (18.7% [32 of 171] vs 12.9% [125 of 970]; P = .042). Similarly, most respondents reported feeling equally prepared to advance to their next stage of training. However, more postgraduate year 1 (PGY-1) respondents reported feeling less prepared, as compared with postgraduate year 2 (PGY-2) and PGY-3 and greater respondents (42% vs 30.2% vs 25.4%, respectively; P < .001). Additionally, the majority (52.1%) reported that their personal wellness was worse as a result of the pandemic.

Respondents reported other impacts during the pandemic (Table 3). Respondents who identified their race or ethnicity as American Indian or Alaskan Native, Asian American, Black or African American, Hispanic, Native Hawaiian or other Pacific Islander, multiracial, or other were more likely to be impacted by personal experience with discrimination (odds ratio: 3.38; P < .001) and visa concerns (odds ratio: 3.48; P < .001) than respondents who identified as White.

Respondents reported a negative impact on morale but a more positive impact on camaraderie (Table 3). The top 3 most impactful wellness interventions offered by respondents’ programs or institutions were increased updates from residency leadership (67.8%; 774 of 1141), increased opportunities to socially connect virtually (42.8%; 488 of 1141) and the provision of more free meals (36.2%; 413 of 1141).

In the early months of the pandemic, respondents reported significant disruptions across several domains, including decreased in-person clinical care, increased use of telemedicine, and, in some cases, caring for adult patients. In general, respondents reported a negative impact on their training and personal wellness, but the majority had no change in their desire to practice medicine. To our knowledge, this is one of the first studies in which researchers examine the impact of COVID-19 on training from the perspectives of pediatric residents.²¹ 

At the onset of the pandemic and in the face of overwhelming uncertainty about the future, residency program leaders were forced to rapidly adjust resident schedules and training experiences.^11,12  This resulted in significant changes to respondents’ clinical experiences including canceled clinics and clinical electives, fewer inpatient shifts, and added jeopardy rotations. These adjustments may explain the positive impact on duty hour compliance. However, along with decreased inpatient census,⁹  these adjustments likely resulted in decreased patient exposure,^11,12  which may have contributed to some residents feeling less prepared for their next stage of training, a finding that was particularly noted among PGY-1 residents. Residency program and hospital leaders, along with clinical supervisors, should be aware of these decreased patient exposures and concerns about preparedness so that they can tailor supervision accordingly and offer support to these residents. Programs and hospitals have already started to innovate, supplementing missed experiences with simulation²²  and other similar experiences. Additionally, with more clear information about preventing COVID-19 transmission, increased supply and access to PPE, vaccination of trainees, and increased availability of COVID-19 testing, in-person clinical care of patients is likely safer. Therefore, program and hospital leaders should consider reprioritizing in-person clinical experiences to balance risk mitigation with the need for pediatric residents to gain clinical competence before graduation. Some changes to training may persist through the pandemic (eg, care for adult patients and virtual educational conferences) and, possibly, beyond (eg, participation in telemedicine). Program and hospital leaders will need to think creatively and carefully about how to prepare our trainees for these experiences.

Adequate PPE is vital to safely caring for patients with suspected or confirmed COVID-19. However, nearly 30% of respondents reported a time when they did not have access to appropriate PPE. Although these findings occurred during a national shortage of PPE,²  it is imperative that trainees be assured access to appropriate PPE at all times.²³  Although there was consensus that pediatric resident direct patient contact should not be suspended, no formal recommendations from national organizations were available regarding resident direct contact with patients with suspected or confirmed COVID-19. This left institutions and residency programs with the daunting task of developing their own policies, resulting in significant variation. One-half of respondents reported being excluded from the care of patients with suspected or confirmed COVID-19, but the vast majority did not believe they should be excluded from either, suggesting that residents without health concerns should be included in the care of patients with COVID-19 going forward. This is consistent with the findings of Kazmerski et al,²¹  who found that the majority of residents believed trainees were essential personnel and should participate in direct patient care during the pandemic.

Outside of disruptions to training experiences, specific populations of residents were affected in additional ways by the COVID-19 pandemic. For example, graduating resident respondents reported that their employment was impacted by COVID-19. Given concerns about the pediatric workforce,^24–26  it is critical that we monitor disruptions to postresidency employment. Additionally, attention should be paid to resident parents who face specific stressors related to parenting in a pandemic²⁷  and residents with concerns about taking care of patients with COVID-19 secondary to either a personal health or medical condition or that of someone living in their home. Residents of color reported experiencing discrimination and visa concerns. The COVID-19 pandemic and resultant stigmatization of Asian Americans,²⁸  global travel restrictions, policy changes that may jeopardize the IMG workforce, alarming racial and ethnic health disparities,²⁹  and widespread protests because of racial injustice ignited by the murders of George Floyd and Breonna Taylor, create a milieu of recurring stress and trauma for residents of color. It is critical that we recognize, address, and support our residents through these unprecedented times.³⁰  Lastly, nearly every resident was impacted by the pandemic in different ways. Some lost loved ones. Some missed important life events, such as weddings. It is critical to support all of our trainees as they navigate the different stressors of the pandemic.

Program and hospital leadership will also need to learn how to support trainee well-being and residency morale because the majority of respondents reported that their personal wellness and residency morale were negatively impacted by the pandemic. In this study, we suggest potential wellness interventions that trainees found helpful, but we acknowledge that it will likely require a multifaceted and adaptive approach to support trainees during this time. Shanafelt et al³¹  propose a framework of 5 different health care worker requests (hear me, protect me, prepare me, support me, and care for me) with suggestions on how to meet those needs that could also serve as a framework for program and hospital leadership.

Limitations of our study include the potential for nonresponse bias, across multiple domains. It is plausible that the training experiences were different in residency programs in which the PD forwarded the survey versus that of those that did not. Additionally, in our study, we did not include residency programs that were not APPD member programs at the time of survey distribution. Similarly, the characteristics of those who responded may be different from those that did not respond. For example, residents on clinically demanding and busy rotations may have been less likely to respond. Likewise, residents whose training was less impacted by the pandemic may also have been less likely to respond. Reassuringly, sampled programs were comparable to all programs nationally in terms of program size and location, and respondents from programs with a >40% response rate had similar responses to all respondents. Additionally, the survey was sent at the onset of the pandemic. Resident perspectives have likely shifted over the course of the pandemic and those shifting perspectives are not captured in this study. Other limitations include a lack of historical data to use for comparison and self-reported outcomes.

This article reveals the myriad of ways that pediatric residency training was impacted by the COVID-19 pandemic through the lens of respondents to this survey. Bringing awareness to these impacts may stimulate broader conversation about the critical concerns facing residency program and hospital leaders including the need to (1) provide enough patient care experiences to promote clinical competence while balancing the need to minimize resident exposure and preserve PPE, (2) supplement missed clinical experiences, (3) offer training to prepare residents for new experiences, and (4) support residents through the pandemic. These findings can also be used to inform the response to future pandemics. In future studies, researchers should explore objective outcomes resulting from the pandemic, including the impact of the pandemic on trainee clinical competence and certifying board examination pass rates as well as employment opportunities for graduating trainees.

We thank all the residents who completed the survey. Additionally, we thank Beth King and Dennis West at the Association of Pediatric Program Directors Longitudinal Education Assessment Research Network for administering the surveys and coordinating the data collection.