Parenthood, Parental Benefits, and Career Goals Among Pediatric Residents: 2008 and 2019

Each year, the AAP surveys a national sample of 1000 third-year graduating residents selected randomly from an AAP database that includes categorical residents from all US pediatric programs. Participants are recruited to participate in the Annual Survey of Graduating Residents via e-mail and US Postal service. Survey requests alternate between mail and e-mail for a total of 8 requests, until the resident responds.

In 2008 and 2019, a set of focused questions on parental benefits during residency training and the importance of parental benefits and job characteristics in post-training positions were included on the survey, and we analyzed data from those years. Both years included the same questions on resident characteristics, availability of parental benefits, importance of job characteristics in a post-training position, limitations in selection of post-training positions, and career goals. The 2019 survey included questions on parental benefits during residency, including length of parental leave, payment, and time used for leave, American Board of Pediatrics waiver for missed time, satisfaction with leave, and perceived impact of having or adopting a child on training and career plans. All survey questions were developed by researchers experienced in survey development and reviewed by 5 pediatric residents from the AAP Section on Pediatric Trainees. Five randomly selected residents completed a pilot version of the survey as part of survey development.

Resident characteristics included age (>31 years, ≤31 years), self-identified gender (female, male, prefer to self-describe), medical school location (international, United States), race and ethnicity (White, non-Hispanic; Asian; underrepresented in medicine: Black or African American, Hispanic, American Indian or Alaska Native, Native Hawaiian or other Pacific Islander; other), partner status (nonphysician spouse or partner, physician spouse or partner, not married or partnered), parenthood (yes, no), program class size (small: <10 residents, medium: 10–19, large: ≥20), and total education debt including spouse’s debt (≥$200 000, <$200 000. Education debt was adjusted for inflation by using the yearly Consumer Price Index to convert 2008 dollars to 2019 dollars and dichotomized by the 2019 mean.

To assess availability of parental benefits, residents responded whether their program offered 5 benefits: maternity leave, partner leave, lactation room, on-site child care, and subsidized child care. In 2019, residents were asked if their program offered electives that allowed new parents to stay home longer with their new infant. To assess the importance of characteristics in selecting a post-training position, residents rated the importance of 7 characteristics: spouse or family considerations, control over work hours, number of overnight calls per month, option to work part-time, length of maternity or partner leave, on-site child care, and subsidized child care. Residents also responded whether their partners’ career plans or their family situation limited their selection of post-training positions. Finally, residents indicated their career goals (subspecialty, primary care, hospitalist).

We excluded respondents who were not postgraduate year 3. We included in our analytic sample residents who responded to gender, parental status, and at least 1 parental benefits question.

We assessed nonresponse bias for the 2008 and 2019 surveys on the basis of age and gender, which were available in the AAP administrative database, using a t test for age and χ² test for gender.

We used χ² tests to compare 2008 and 2019 resident responses to demographic characteristics, program size, and career goal. We used χ² tests to examine relationships between gender and parental status.

To examine whether resident report of available benefits differed in 2008 and 2019, we separately evaluated whether residency programs offered 5 parental benefits (yes versus no or unsure) using logistic regression models controlling for resident characteristics (described above). To provide a more intuitive alternative to odds ratios, we present predicted values (PVs) for each benefit for each survey year.¹⁵  We computed these values from logistic regression analyses holding all other independent variables at their respective sample means. The PVs or covariate-adjusted percentages provide the independent effect of survey year for each of the 5 benefits.

For limitations in selection of post-training positions and career goal, we used χ² tests to examine relationships with gender and parental status. Using logistic regression, we generated PVs as described above to test if (1) resident-rated importance of 7 characteristics in post-training position (essential or very important versus somewhat important or unimportant), (2) limited in selection of post-training positions because of partner career plans or family situation (yes versus no), and (3) career goal (subspecialty versus primary care or hospitalist) varied by year, gender, partner, and parental status. Logistic regression models included the resident characteristics described above and the PVs or covariate-adjusted percentages provided the independent effect of year, gender, partner, and parent status for each outcome, holding the other predictors at their sample means.

We used descriptive statistics to examine resident use of parental benefits during residency in 2019 (described above). We used χ² tests to examine relationships of gender with (1) satisfaction with leave (very satisfied or satisfied versus neutral, dissatisfied, or very dissatisfied), (2) perceived impact on training (strong positive effect or positive effect versus no effect, negative effect, or strong negative effect), and (3) perceived impact on career plans (strong positive effect or positive effect versus no effect, negative effect, strong negative effect).

The number of cases in each statistical analysis varied slightly because of missing values for specific questions. Data were analyzed with IBM SPSS Statistics 25 (IBM SPSS Statistics, IBM Corporation) and Stata 14 (Stata Corp, College Station, TX).

Combining 2008 and 2019, a total of 1049 graduating pediatric residents responded (54.7% overall, 58.8% in 2008, and 51.0% in 2019). In total, 28 respondents did not meet inclusion criteria, yielding an analytic sample of 1021 (2008: n = 518; 2019: n = 503). Compared with the random sample invited to participate, respondents were more likely to be female (75.1% vs 70.9%, P < .01), but did not differ in age (mean 31.3 vs 31.4 years, P = .31).

Demographic characteristics in 2008 and 2019 are presented in Table 1. Compared with 2008, the 2019 respondents were less likely to have children (24.5% vs 33.8%, P < .01) (Table 1). In 2019, a higher proportion of respondents were from large residency programs, had a subspecialty practice career goal, and had education debts ≥$200 000.

In both 2008 and 2019, a majority of respondents reported that maternity leave and lactation space were available at their program (Table 2). In 2019, 41.2% reported they were unsure if partner leave was available. Controlling for age, gender, medical school location, race and ethnicity, partner status, parenthood, education debt, goal, and program size, residents in 2019 were less likely than residents in 2008 to report their program offered maternity leave (PV = 78.5 [95% confidence interval (CI) = 74.5–82.4] and 89.5 [95% CI = 86.6–92.4], P < .001, respectively) and partner leave (PV =42.5 [95% CI = 37.9–47.2] and 59.2 [95% CI = 54.5–63.9], P < .001, respectively). Residents in 2019 were much more likely to report their program offered lactation rooms (PV = 77.8 [95% CI = 73.9–81.7] and 56.1 [95% CI = 51.3–60.9], P < .001).

In choosing their post-training position, the majority of residents in both 2008 and 2019 reported spouse or family considerations, control over work hours, and number of overnight calls per month as essential or very important characteristics (Table 3).

Controlling for resident characteristics, residents with children were more likely than residents without children to rate spouse or family considerations (PV = 94.3 [95% CI = 90.7–97.9] vs 85.3 [95% CI = 81.9–88.7], P < .01) and number of overnight calls (PV = 79.6 [95% CI = 74.1–85.1] vs 69.1 [95% CI = 65.2–73.1], P < .01) as important (Table 3). Female residents were more likely than male residents to rate as essential or very important 5 of 7 characteristics, including the option to work part-time (PV = 49.7 [95% CI = 45.8–53.6] vs 24.5 [95% CI = 18.8–30.3], P < .001) and number of overnight calls (PV = 74.9 [95% CI = 71.5–78.4] vs 64.2 [95% CI = 57.7–70.7], P < .01) (Table 3). Residents who were partnered were more likely than unpartnered residents to rate 6 of 7 characteristics as essential or very important, and a stronger effect was seen for those with a physician partner.

Half of residents (48.5%), combining 2008 and 2019 responses, reported their partner or family situation limited their selection of post-training position. In bivariate analysis, a greater proportion of female residents with children (64.7%) reported that their family situation limited their selection of position, as compared with female residents without children (47.4%, P < .001) and compared with male residents with children (42.7%, P < .01). In contrast, no statistically significant difference was seen for male residents with children (42.7%) compared with male residents without children (34.5%, P = .21). Female residents without children were also more likely to report this limitation as compared with male residents without children (47.4% vs 34.5%, P < .01).

Controlling for other resident characteristics, residents with children were not more likely to report that their family situation limited their selection of positions. However, 2019 graduates, female residents, and partnered residents were more likely to report that their family situation limited their selection of positions (Table 3).

Half of residents (49.2%), combining 2008 and 2019 responses, reported a subspecialty career goal. In bivariate analysis, 40.0% of female residents with children reported subspecialty career goals, as compared with 47.8% of female residents without children (P = .06) and 49.4% of male residents (P = .16). Female residents without children were less likely than male residents without children to report a subspecialty career goal (47.8% vs 65.3%, P < .001, respectively).

Controlling for resident characteristics, residents with children did not differ from residents without children in reporting a subspecialty career goal, but female residents were significantly less likely to report a subspecialty career goal than male residents (PV = 44.9 [95% CI = 41.1–48.7] vs 60.3 [95% CI = 53.8–66.8], P < .001) (Table 3).

Additional data on parent experiences were collected in 2019. Of 123 residents with children, 95 (77.2%) reported that they or their partner gave birth or adopted a child during residency. Of self-identified women who gave birth or adopted during residency, 82.1% reported using maternity leave and 80.6% reported using a lactation space. In comparison, 44.4% of men whose partners gave birth or who adopted reported using partner leave. Of those who became parents during residency, 60.0% used an elective to stay home longer with the new child, 11.8% used subsidized child care, and 10.6% used on-site child care. Mean length of parental leave was 7.2 weeks for women and 2.2 weeks for men, and 73.4% of women and 100% of men reported they were paid for their entire leave. Some or all of their vacation time was used during their leave by 89.1% of women and 23.1% of men. Only 21.0% reported petitioning the American Board of Pediatrics for a missed time waiver.

A minority of residents (39.5%) were satisfied or very satisfied with their parental leave, which did not differ by gender (P = .86). Among residents who became parents during residency, 50.0% reported a positive effect on their training and 44.7% reported a positive effect on their career plans; no differences were found for either measure by gender.

In a national survey of graduating pediatric residents conducted in 2008 and 2019, we found that 2019 graduates were less likely to be parents and less likely to report availability of maternity or parental leave than 2008 graduates. We hypothesized that having a child during training would impact future career goals, but when controlling for other resident characteristics, we did not find an association with parenthood. Instead, we found that gender correlated with future career goals and importance assigned to characteristics in post-training positions. These findings are important for informing policies to support trainees and the transition to post-training positions.

Similar to a previous study, we found that awareness of available parental leave during residency is not universal.¹⁶  A variability of parental leave policies at individual institutions, as has been shown in other studies, likely contributes to the decreased awareness.^17–19  A study from 1990 of female pediatric trainees who were pregnant during training reported that 37% would not choose to be pregnant during training if given a second opportunity; however, it is not known if that sentiment persists in the current training environment.²⁰  Transparency of leave policies has been proposed as an important part of creating a training environment supportive of parents.^14,21  We found that half of 2019 respondents who became parents during pediatric residency reported a positive effect on their training. Report of lactation space increased in 2019, which coincides with enactment of the Break Time for Nursing Mothers Law in 2010 as part of the Patient Protection and Affordable Care Act and increasing state laws supporting lactation for employees.²² 

Parental status did not correlate with future career goals after controlling for other resident characteristics, which likely reflects the many complex factors influencing an individual’s career path. Because a majority of early career physicians are parents, both trainees with children and trainees without children may be considering how their early career will integrate with family.²³  Therefore, those without children during training may be evaluating the impact of future career goals and positions on their anticipated family. Female trainees in our study were more likely to report certain characteristics in post-training position as important and that family situation limited their selection of post-training positions. Systemic barriers faced by women pursuing careers in pediatrics may be important, including those regarding work-life balance and household responsibilities.^6,7  A 2015 study of early career pediatricians found that female pediatricians spent more time on household responsibilities and the care of their own children than male pediatricians.^6,7 

Our findings mirror a similar study that showed that having a physician partner and being a woman correlated with seeking part-time positions after training.²  The characteristics women were more likely to rate as important (number of overnight calls per month, option to work part-time, length of parental leave, and availability of on-site child care in positions after residency) may be shaping the workforce of institutions and pediatric subspecialties. Institutions and programs hoping to recruit pediatric graduates should consider the benefits they offer.

Our study has several limitations, including limitation of data to the two years analyzed (2008 and 2019). The overall response rate was 54.7%, and respondents were more likely to be female, similar to other surveys of pediatricians.²⁴  Analysis by gender was limited to male and female on the basis of initial demographic survey data from 2008. The 2019 survey included a “prefer to self-describe” option, which no resident selected. Because all data are self-reported, including availability of parental benefits, bias may have been introduced if a resident had never investigated their program’s policies. We do not have data on whether trainees delayed becoming parents or were attempting to adopt, outcomes of pregnancies, or partner demographics. We did not differentiate between spouse and partner. Although we adjusted for race and ethnicity in our analytic models, researchers in future studies should investigate the impact being underrepresented in medicine has on parenthood. Finally, other sources of support for trainees who have children which we did not collect may have an impact on the need for parental benefits, the choice to have children, and future career goals.

In a national survey of graduating pediatric residents conducted in 2008 and 2019, we found that fewer residents in 2019 have children, and they were less likely to report availability of parental leave in 2019. Parenthood correlated with reported importance of 2 of 7 characteristics in post-training position. Female gender correlated with limitations in position selection and career goals, but parenthood did not.

AAP

American Academy of Pediatrics

PV

predicted value

CI

confidence interval