A Changing Landscape: Exploring Resident Perspectives on Pursuing Pediatric Hospital Medicine Fellowships

A 19-question anonymous online Qualtrics survey (included in Supplemental Information) was developed by 2 of the authors (H.C., Y.N.W.) and modeled after a similar survey published by Goodman and Januska,¹⁴  which asked employers, hospitalists, and internal medicine residents about their perspectives on clinical hospital medicine fellowships. Our survey was subsequently tailored to PHM-specific content. The survey began with 2 screening questions asking if (1) respondents were currently considering practicing PHM or (2) had ever considered practicing PHM. If respondents answered “definitely no” to question 1 and “no” to question 2, they were directed to complete only the 8 demographic questions. The remainder of the respondents were directed to complete 8 questions regarding likelihood of pursuing PHM given introduction of the fellowship requirement, factors influencing reported perspectives, perceived value of additional training in specific areas, and other careers being considered, before being directed to the demographic section. At the end of the survey, respondents were asked if they would like to include a free-text response. Finally, they were given the option to provide their e-mail addresses to be entered into a drawing to win an Amazon gift card. These were kept separate from the remainder of the data to ensure deidentification. The survey was pilot tested with 2 chief residents, and modifications were iteratively reviewed by authors (H.C., Y.N.W., C.P.L., C.H.M.) to develop the final survey.

In August 2018, the categorical pediatric and Med-Peds program directors (PDs) for the 205 residency programs listed in the Association of Pediatric Program Directors directory were emailed and asked to distribute the survey to their residents or to provide us with the residents’ electronic mailing lists. Two reminder emails were sent 1 week apart to the PDs (or directly to the residents in the cases in which the PDs provided us resident emails). Fifteen-dollar Amazon e-gift cards were given to 25 randomly selected respondents.

We collected demographic data, including sex, postgraduate year (PGY) status, type of residency program (eg, categorical, Med-Peds, other combined pathways), student loan burden, and additional degrees held. We tabulated desired future career characteristics, including goal of practicing as a temporary or career hospitalist and desired practice setting (ie, community or major teaching hospital). For the question “does (or did) the introduction of a PHM fellowship affect your likelihood of pursuing a job as a pediatric hospitalist,” responses of “much less likely” and “somewhat less likely” were grouped into “less likely to pursue PHM” and compared between demographic groups with the second grouping of “neutral,” “somewhat more likely,” and “much more likely to pursue PHM.” We analyzed categorical variables using χ² tests in GraphPad software (GraphPad Holdings, LLC, San Diego, CA).¹⁵ 

Responses to questions asking residents to rate the importance of various factors in encouraging or discouraging them from pursuing PHM on a 1 to 5 scale, with 1 representing “not at all important,” 3 representing “moderately important,” and 5 representing “extremely important” were also tabulated and grouped into “at least moderately important” (3–5 on scale) for discussion.

The Boston Children’s Hospital Institutional Review Board determined the protocol to be exempt.

PDs at 48 programs (23.4%) agreed to participate in our study, of whom 41 agreed to forward the survey to their residents and 7 provided their residents’ e-mail addresses to contact residents directly (Fig 1). In one additional institution’s case, the Med-Peds PD forwarded the survey to Med-Peds residents, but the pediatric PD at the same institution did not. This program was counted as a nonresponding program for purposes of the following size and geographic comparison, given that the majority of the residents did not receive the survey. On the basis of the number of positions each program was approved for by the ACGME, we estimated that the survey was distributed to ∼2657 residents, which includes the ∼14 Med-Peds residents in the single program where only Med-Peds residents received the survey. Participating program size ranged from 14 to 164 residents (median: 41), and there was a geographic distribution of 27.1% from the Northeast, 31.3% from the Midwest, 29.2% from the South, and 12.5% from the West. In comparison, nonparticipating programs ranged in size from 12 to 140 residents (median: 38), and there was a geographic distribution of 29.2% from the Northeast, 19.5% from the Midwest, 37.7% from the South, and 13.6% from the West. PDs at 2 institutions in Puerto Rico and 1 in Canada were emailed and did not respond; these are not reflected in the geographic spread. In addition, program size information was not available for the 1 Canadian institution included.

Eight hundred fifty-five residents completed the survey (32.2% response rate). Of these, 94 (11.0%) residents answered “definitely no” or “no” to the questions “are you considering” or “have you ever considered,” respectively, “practicing pediatric hospital medicine after completing residency” and were directed only to the demographic questions. The remaining 761 (89.0%) residents were directed to the entirety of the survey. These 761 respondents are analyzed in more detail. Of these 761 respondents, 66.9% were female. The PGY level composition included 31.8% PGY-1 residents, 33.1% PGY-2 residents, 28.8% PGY-3 residents, and 5.9% PGY-4 or higher. Of the 761 respondents, 123 (16.2%) were Med-Peds residents, and 15 (2%) were part of another combined pathway (9 child neurology residents, 3 combined pediatrics-medical genetics residents, 2 combined pediatrics-anesthesiology residents, and 1 who did not specify).

Residents were asked to rate the impact of the introduction of a PHM fellowship requirement for board certification eligibility on their likelihood of pursuing hospital medicine (Fig 2). In total, 48.9% of residents (46.4% of non–Med-Peds residents [ie, categorial pediatric and other combined pathway residents], 62.6% of Med-Peds residents) reported that they were “much less likely to pursue hospital medicine,” and an additional 30.5% (30.7% of non–Med-Peds residents, 29.2% of Med-Peds residents) reported being “somewhat less likely to pursue hospital medicine.” These 2 groups of respondents were combined as “less likely to pursue PHM” for further analysis. The remaining 20.6% of residents (23% of non-Med-Peds residents, 8.1% of Med-Peds residents) reported being either “neutral,” “somewhat more likely to pursue hospital medicine,” or “much more likely to pursue hospital medicine” and were grouped together for further analysis.

The impact of the introduction of a PHM fellowship requirement on likelihood of pursuing PHM was compared among demographic groups (Table 1). Responses did not differ significantly by sex, PGY status, or additional degrees held. Residents planning on practicing at community hospitals were more likely to be deterred as compared with residents planning on practicing at major teaching hospitals (85.0% vs 76.3%, P = .005). Similarly, residents planning on being temporary hospitalists were more likely to be deterred compared with peers planning to be career hospitalists (89.5% vs 75.0%, P < .001). Residents with higher student loan burdens reported being less likely to pursue PHM (81.5% in highest tertile of student loan burden vs 73.3% in lowest tertile, P = .045). Med-Peds residents were also less likely to pursue PHM as compared with non–Med-Peds respondents (92.3% vs 76.9%, P < .001).

In the next part of the survey, residents were asked to rate the importance of various factors in encouraging or discouraging them from pursuing a PHM fellowship. A majority of residents (53.0%) rated “opportunity to gain new skills” as important (4 or 5 on 5-point scale) (Fig 3A). Slightly fewer residents (40.2%) valued potential increase in salary. Opportunity for a master’s degree and increased prestige were only considered important by 24.3% and 20.0% of residents, respectively. Residents were also asked about factors negatively affecting their decision to pursue a PHM fellowship (Fig 3B). More than two-thirds of residents considered forfeited earnings during fellowship (72.9%), family obligations (75.7%), student loan obligations (68.8%), and residency training sufficient (86.5%) as important (4 or 5 on 5-point scale). A minority of residents (41.2%) reported intent to practice in a community hospital as a reason discouraging them from pursuing a PHM fellowship.

Residents were then asked to rate the value of additional training in 5 areas: medical education, QI, hospital administration, clinical training, and research training. Distribution of responses was similar for each subject area, with ∼50% of respondents rating each as at least moderately valuable (≥3 on 5-point scale).

Finally, residents were asked to select which career they were now considering if they were less likely to do PHM given the introduction of the fellowship requirement. Three hundred forty-one (44.8%) respondents reported considering another subspecialty career, 248 (32.6%) reported primary care, 59 (7.8%) reported other, and 113 (14.8%) did not respond.

In this study of pediatric and combined pediatric (eg, Med-Peds) residents’ perspectives regarding PHM fellowships, many residents saw value in pursuing a fellowship but also reported concerns. Over half of residents expressed interest in the opportunity to acquire new skills during a PHM fellowship. However, 79% of respondents reported that they were “somewhat” or “much” less likely to pursue PHM after the introduction of a fellowship requirement for board certification, with the concerns centering around opportunity cost, both financial and personal.

Respondents perceived value in a PHM fellowship. A majority of respondents rated “opportunity to gain new skills” as an important factor encouraging them to pursue a PHM fellowship. Half of respondents reported that additional training in medical education, QI, hospital administration, clinical training, and research training would be valuable. This perceived value in new skills was despite nearly 87% of respondents reporting that they felt residency training was sufficient. Contrary to our respondents’ perceived adequacy of training, in a previous survey of pediatric hospitalists, nonfellowship-trained hospitalists felt less comfortable than fellowship-trained hospitalists in intravenous access and phlebotomy, emergencies in technology-dependent children, QI methods, statistics, and identifying resources for research.¹⁶  Interestingly, authors in a 2016 study surveying PHM fellowship graduates suggested that procedures, QI, and research were areas in which additional training would be beneficial, suggesting that even early PHM fellowships were not meeting all training needs.¹⁷  Subsequently, through a rigorous, multiyear process, many stakeholders considered the gaps in both residency training and in existing PHM fellowship training to develop the PHM Curricular Framework, which has informed the ACGME PHM Core Program requirements.^18,19 

The factors discouraging respondents from pursuing a PHM fellowship centered around financial and personal opportunity costs. A majority of respondents reported both forfeited earnings and student loan obligations as important factors discouraging them from pursuing a PHM fellowship. The debt reported by respondents was consistent with previously reported average pediatric resident debt of >$150 000 per resident.²⁰  Previous research indicates that residents with higher debt are less likely to pursue any type of fellowship, opting to pursue primary care or hospitalist positions, which would allow them to pay off their debt faster.²¹  Additionally, family factors significantly influence the decision to pursue a PHM fellowship, as 75% of respondents in our study reported that family obligations deterred them from pursuing a PHM fellowship. This, however, is in contrast to the result that nearly 45% of respondents reported now considering another subspecialty career that may have similar opportunity costs to a PHM fellowship. Authors of another study analyzing resident survey data from 2006 to 2015 showed that residents who reported that family factors strongly influenced their postresidency plans were more like to pursue PHM (before fellowship training requirement for certification) and less likely to pursue a subspecialty fellowship than their peers for whom family factors were not as strong of an influence.²⁰  This raises concern that these individuals may now feel that PHM is no longer an option. It is possible that this concern about opportunity cost could be decreased by shortening the duration of a PHM fellowship, although it may be challenging to fit the skills that applicants desire into a shorter time period. Alternatively, it may be worth considering competency-based evaluation and promotion, where trainees may take variable amounts of time to acquire certain skills on the basis of their individual needs.^22,23 

In our study, 3 specific populations were significantly more deterred from PHM by introduction of fellowship requirement for board certification eligibility: those intending to temporarily pursue PHM before moving on to a different area of practice, Med-Peds residents, and those intending to practice in a community hospital. For those intending to temporarily pursue PHM before moving on to a different area of practice, the time cost of a multiyear fellowship is likely hard to justify, because they are uninterested in PHM as a long-term career. In a survey published in 2015, less than half of pediatric hospitalists (43%) reported that this was their long-term career plan.²⁴  In contrast, a 2018 study that analyzed the 2006 to 2015 data from the AAP Annual Survey of Graduating Residents found that 71% of residents entering hospitalist positions planned for PHM as their long-term career.²⁰  Regardless, these studies suggest that the introduction of PHM fellowships may affect the size of the “temporary hospitalist” workforce if employers require board certification for these positions.

The second group that was more significantly deterred from pursuing PHM fellowship was Med-Peds residents. A survey of Med-Peds PDs reporting on the experiences of recent graduates found that 26.4% of Med-Peds residents practiced as hospitalists, of whom 65% (ie, 17.2% of the cohort surveyed) provided care to both hospitalized children and adults.¹¹  Med-Peds practitioners are estimated to currently compose at least 10% of the pediatric hospitalist workforce, suggesting that changes in the practice patterns of Med-Peds graduates could impact the PHM workforce.¹¹  Of particular note, the field of internal medicine does not offer board certification in the subspecialty of hospital medicine, raising a discrepancy between the fields of pediatrics and internal medicine that uniquely affects Med-Peds practitioners. Going forward, it will be important to see whether Med-Peds graduates who plan to treat both adult and pediatric patients pursue PHM fellowship training and, if so, how they can maximize their fellowship training to best prepare them for this combined career.

Finally, residents intending to practice in a community setting were significantly more deterred by a PHM fellowship. Although this survey did not collect in-depth data about the reasons that these residents were deterred from PHM by the introduction of a fellowship, we hypothesize that their perception of their need for training in areas such as research and care of technology-dependent children may be less as compared with residents who intend to work at university-based programs. However, that is not to say that community pediatric hospitalists would not benefit from additional training. Compared with their counterparts at university-based hospitalist programs, community-based pediatric hospitalists are more likely to be involved in delivery room management, newborn nursery care, and emergency department pediatrics consultations.²⁵  Many of these needs could be met through the individualized curriculum portion of the PHM fellowship curriculum.^18,19  A targeted survey of perceived competencies and training needs among hospitalists practicing in community settings may provide additional helpful information about this population.

There are several limitations of this study. Although our survey was modeled after a similar survey published by Goodman and Januska,¹⁴  the majority of the final survey was developed de novo and has not been validated. Our survey did not include questions on training site (eg, freestanding children’s hospital, general medical center, etc) or questions regarding exposure to PHM, such as a PHM track, PHM fellows, or PHM-trained faculty. Moreover, we had a low response rate, with only 23.4% of programs participating and only 32.2% of residents in these programs responding. We speculate that the PDs who did not agree to distribute the survey have policies about external surveys or try to limit the number of surveys their residents receive, rather than concerns about the topic. It is somewhat reassuring that our responding and nonresponding programs are similar in terms of program size and geographic distribution, but it is possible that they may differ from one another in other significant ways that we have not captured. Moreover, we do not know the factors that led some residents to respond to the survey and others not, therefore raising the possibility of nonresponse bias. Of note, Med-Peds residents were slightly overrepresented, given that 16% of our respondents were Med-Peds, but Med-Peds residents compose an estimated 13% of categorical pediatric positions.²⁶  Despite our low response rate, however, a notable 855 residents nationwide did complete the survey, and we believe that the data presented here reflect the perspectives of at least a subset of residents. The vast majority of the respondents (89%) had at least considered a career in hospital medicine, which is the population we were hoping to target because this is the group whose perspective on PHM fellowships matters most. However, it is worth noting that we left the term “considered” open to interpretation. In addition, we did not formally analyze respondents’ free-text comments. Another limitation is that this survey was administered during 2018, which was a time of change, so the responding residents may have been particularly influenced negatively by change, whereas current and future residents who only know the new fellowship and certification system may not have the same perspectives.

The majority of pediatric and combined pediatric residents who responded to our survey and who had considered a career in PHM reported that the requirement of a PHM fellowship for board certification made them less likely to pursue this path. Respondents intending to temporarily pursue PHM, Med-Peds residents, those intending to practice in a community hospital, and those with high student loan burden were more likely to be deterred. Deterring factors included financial opportunity cost, family obligations, and a feeling that residency training was sufficient. However, respondents also cited an opportunity to gain additional skills as a benefit of a PHM fellowship, which will be supported by the PHM Curricular Framework and resulting ACGME PHM Program Requirements. As the rapidly growing field of PHM moves forward, an ongoing understanding of residents’ needed skill sets and concerns about the opportunity costs of fellowships will be important to monitor as well as monitoring the trends in residents’ ultimate PHM career decisions.

We thank the PDs who helped distribute this survey as well as the residents who took the time to share their opinions. In addition, we thank the Fred Lovejoy Housestaff Research and Education fund in supporting this project.