Parenthood and Parental Leave Decisions in Pediatric Residency

We invited residents (N = 845) enrolled in pediatric residency programs at 13 institutions to participate in a voluntary, one-time, confidential online survey distributed through the Association of Pediatric Program Directors Longitudinal Educational Assessment Research Network (APPD LEARN).¹⁷  An e-mail invitation was sent to residents by a faculty member at their institution. Sites solicited responses on a rolling basis over a 6 to 8 week period between October 2019 and May 2020. Three reminders (2, 4, and 6 weeks) were sent to nonresponders. To maintain confidentiality of responses, sites held a list of unique research identifiers corresponding to participants but had no access to responses. The data coordinating center held research identifiers and responses but no participant list. The data coordinating center informed sites of research identifiers returning a survey, allowing sites to send reminders to nonresponders, without links to responses.

We adapted the resident survey from a previously validated cross-specialty tool on the basis of feedback from physicians with personal experience in pregnancy and parenthood during medical training and literature reviews of pregnancy and parenthood in graduate medical education and pediatric training.^6,9,18,19  Outcome measures included satisfaction with breastfeeding duration, perceptions of support, satisfaction with parental leave length, postpartum depression screening results, and burnout. Postpartum depression screening was presented to respondents who had given birth or whose partners had given birth in the previous 12 months by using the Edinburgh Postnatal Depression Scale (EPDS).²⁰  A single-item measure shown to be a reliable substitute for the emotional exhaustion domain of the Maslach Burnout Inventory was used to assess burnout in all respondents, regardless of parental status.^21,22 

Respondents who took parental leave during residency were asked to identify the top 3 factors determining leave length from a list of options, based on the initially described validated tool, including a free-text response (“Other”).^6,9,18,19  Respondents with children were asked if they wanted additional children in the future, and respondents without children were asked if they wanted children in the future. For respondents desiring future children, we assessed whether childbearing was intentionally delayed during residency, reasons for doing so, and satisfaction with that decision. We present the complete survey in the Supplemental Information.

Each site’s institutional review board deemed this study exempt (11 programs) or approved (2 programs).

Descriptive statistics were used to characterize study participants and responses. Based on previous work, we dichotomized parental leave length as ≤6 weeks or >6 weeks.⁶  To assess the association between parental leave length and other factors, linear or logistic random-effects models were fitted. In addition to adjusting for clustering within programs, we also examined models with the program response rate as a program-level covariate. To examine the impact of program leave norms, we fitted a random-effects logistic regression model to our dichotomous burnout measure, with the program mean leave, an individual resident’s deviation from their program mean leave, and a random intercept for program as predictors, using a generalized additive model framework. We also repeated the analysis using the ordinal burnout measure as the outcome.

Of 845 eligible participants, 639 (76%) residents responded. Program response rates ranged from 37% to 100%; the response rate was not associated with any outcome. Respondent and program characteristics are listed in Table 1. Among respondents, 70% identified as female and 16% (n = 100) were parents (64 mothers and 36 fathers). Four percent (n = 27) were pregnant or had pregnant partners at the time of response. Thirty-three percent (n = 213) were in postgraduate year 1. Of respondents who were parents, 45% had their first child before residency and 55% had a child born during residency. Thirty-seven percent of parents had >1 child.

Among all respondents, 85% endorsed wanting children or additional children in the future and 52% (n = 330) reported delaying having children during residency. Wanting children was significantly associated with higher age (P < .001), but decision to delay was not associated with age (P = .79). Sixty-four percent (248 of 386) of female respondents who wanted children reported delaying childbearing, compared with 53% (82 of 155) of male respondents who wanted children (P = .02). The most common reasons for delaying having children included “busy work schedule” (89.7%), “finances” (50.9%), and “desire not to extend residency training” (41.2%) (Fig 1). “Single/lack of partner” was prominently represented among free-text responses (including 4 of 7 men). Female respondents were more likely than male respondents to cite “desire not to extend residency training” (46% vs 28%; P = .005) and “residency training might increase pregnancy complications” (31% vs 0%; P < .001) among top reasons for delaying. Female respondents were less likely than male respondents to select “finances” (47% vs 63%; P = .01) and “my partner is not ready” (20% vs 37%; P = .003). Other factors cited by >1 respondent in free-text comments included “desire not to burden colleagues,” “demands of current child/ren,” and “lack of support system.”

Among respondents delaying childbearing during residency, 29 were parents and 300 were not. Parents were more likely than nonparents to report “desire to not delay taking my board examination” (24% vs 6%; P < .001) and “desire to not extend my residency training” (66% vs 39%; P = .007) as reasons for delay. Parents were less likely than nonparents (single and partnered) to select “finances” as a reason (28% vs 53%; P = .01). Of respondents delaying having children, 29% (n = 97) were dissatisfied or very dissatisfied with the decision.

Sixty-one previous or planned parental leaves during residency were described, including 42 maternity leaves and 19 partner leaves, with leave length ranging from 0 to 30 weeks. Sixty-four percent of partner leaves were ≤2 weeks in length, whereas 27% of maternity leaves were ≤6 weeks in length (Fig 2). Among reported leaves, 77% (n = 47) included vacation time; 34% (n = 21), sick time; and 16% (n = 10), neither. Excluding vacation and sick leave, other paid leave was included in 72% (n = 44) of reported parental leaves. The median unpaid leave was 3 weeks (interquartile range 0–5.75), with longer leave lengths due predominately to unpaid leave (Fig 3). We did not specifically inquire whether programs offered discrete return-to-work rotations for new parents, but 77% (n = 47) of residents reported returning to full-time work after parental leave, without schedule adjustments. Twenty-three percent (n = 14) of residents needed to “pay back” call not worked during parental leave; another 26% (n = 16) were unsure whether payback was required.

The most common self-reported determinants of parental leave length among female and male respondents were “desire not to extend residency training further” and “newborn bonding” (Fig 4). Free-text responses included “amount of nonclinical time” (female respondent), “schedule flexibility and elective time,” and “residency policies” (male respondents). Female respondents were more likely than male respondents to report “desire not to extend residency training further” (88% vs 42%; P < .001), or “desire not to delay taking my board examination” (29% vs 5%; P = .04) as determining factors. Male respondents whose partners had delivered or planned to deliver a child were more likely than female respondents to report “partner’s leave” (26% vs 5%; P = .04). Pediatric residents with leaves >6 weeks were significantly more likely to have a spouse employed full-time (versus a spouse not employed full-time) than residents with leaves ≤6 weeks (82% vs 52%; P = .04).

Forty-two female residents rated satisfaction associated with parental leave duration. The majority (69%) reported leave duration as “less than I would like,” whereas 31% reported leave duration as “about right” (Table 2). Among 19 male residents rating satisfaction associated with leave duration, 63% described it as “less than I would like.”

Among all respondents, 38% (244 of 639) reported ≥1 symptom of burnout, with no difference between parents and nonparents (34% vs 39%; P = .45). Among 400 female nonparents completing burnout items, 173 (43%), reported ≥1 symptom. These rates did not significantly differ from those of female parents with ≤6 weeks of leave (P = .30) or female parents with >6 weeks of leave (P = .08). Resident-reported measures of well-being for female parents are shown in Table 2, stratified by leaves of ≤6 weeks and >6 weeks. Among female parents with leaves ≤6 weeks, 58% (7 of 12) reported ≥1 symptom of burnout. Among female parents with leaves >6 weeks, 27% (8 of 30) reported ≥1 symptom of burnout (P = .06). Positive postpartum depression screen results (EPDS score >9) were identified in 50% (3 of 6) of women who delivered in the previous 12 months with ≤6 weeks of leave and in 17% (2 of 12) of women who delivered in the previous 12 months with >6 weeks of leave (P = .15). Residents in both groups had similar perceptions of support from co-residents, attending physicians, program directors, and administration.

The program mean weeks of parental leave, but not individual resident deviations from the program mean, were significantly negatively associated with burnout, whether measured as a dichotomous outcome (odds ratio = 0.81; 95% confidence interval [CI] 0.68–0.98; P = .02) or an ordinal measure (cumulative threshold odds ratio = 0.83; 95% CI 0.71–0.97; P = .02). At the median program mean leave length (8 weeks), predicted probability of burnout was 36%; in programs in which mean leave was 4 or 12 weeks, predicted probability of burnout was 56% or 20%, respectively (Fig 5).

Of female residents who had children or were pregnant, 91% (n = 71) reported breastfeeding or plans to breastfeed; 30% (n = 23) were breastfeeding at the time of response. Of mothers who delivered during training who were no longer breastfeeding (n = 11), 73% (n = 8) were not satisfied with the length of time they provided breast milk (Table 2). There was no significant difference in satisfaction between mothers with ≤6 weeks of leave and those with >6 weeks of leave. Among mothers who breastfed or were breastfeeding after leave, 72% (n = 56) selected “time to pump at work” and 56% (n = 44) selected “place to pump at work” as factors most influencing duration of breastfeeding after return to work. Other determinants reported as free text included other work responsibilities, culture of program support, and wearable pump technology.

This large multicenter study reveals that pediatric residents delay having children during residency, and many are dissatisfied with this decision. Resident parental leaves remain short, and a desire to avoid extending training is the most common factor determining leave length, particularly for female residents. A longer program mean leave length is negatively associated with burnout. Breastfeeding is common among female residents with children, but time and space to pump are barriers to continued breastfeeding after return to work.

Timing of childbearing is a complex and multifactorial decision for residents and their families. Previous research has shown that pediatric residents may be unaware of options for parental leave during training.²³  This lack of awareness about program leave policies may contextualize our finding that a majority of residents desire children but often delay having them during training. Additionally, previous work has shown that residents perceive a potential negative career impact associated with parenting.^13,24  Our study supports these findings, with respondents reporting concern for extending residency or fellowship or job start dates as reasons to delay childbearing. However, we also found that a majority of respondents reported busy work schedules and finances as reasons for delay. This suggests that adequate leave is necessary but likely insufficient for supporting trainee parents. Schedule modifications, removal of call payback requirements, and facilitating access to affordable child care are additional program actions that may support residents desiring children. A transition to competency-based (rather than time-based) residencies may allow more flexibility to accommodate parental leave and improve the experience of trainee parents; however, research in another specialty has shown that milestone attainment may differ by gender and further burden female trainees.^25,26  Ultimately, the decision to become a parent is impacted by many factors not fully captured in this study.

After survey data collection, the American Board of Medical Specialties (ABMS) announced a new parental, caregiver, and medical leave policy for residents and fellows. Effective July 2021, ABMS member boards, including the ABP, will allow for 6 weeks away once during training, without exhausting vacation or sick leave and without requiring training extension.²⁷  At the time of this study, however, the ABP permitted a maximum of 3 months of leave over 36 months of pediatric residency training, inclusive of vacation, sick leave, or family leave.¹¹  Historically, a training extension was required for a total leave of >3 months, unless the program director requested an ABP waiver for up to 2 months of elective training for family or parental leave.¹¹  In this study, we did not assess frequency of use of the waiver on behalf of respondents, but many residents cited the desire to avoid extending residency as a reason for delaying parenthood and limiting leave, suggesting that the waiver may be underused.

We found that nearly half of parental leaves were ≤6 weeks in length and that a majority of resident parents were dissatisfied with their leave length. A shorter program mean leave length was significantly associated with a higher likelihood of resident burnout. Analysis of postpartum depression screen results by leave length did not reach statistical significance because of the small sample size; in future work, researchers should investigate whether leave length for pediatric residents correlates with outcomes such as postpartum depression screening results. Although the new ABMS policy will allow for 6 weeks of leave, this may be inadequate to mitigate mental health impacts. We also found trainee dissatisfaction with breastfeeding support after return to work, including reports of inadequate time and space for lactation. Although legislation mandating breastfeeding support has improved after the Affordable Care Act of 2010, our findings suggest that further lactation support is required for trainees, as previously proposed.^8,28  In addition to extending leave beyond the ABMS’s recommended 6 weeks, work modifications, access to high-quality lactation facilities, and using the ABP’s existing leave waivers could effectively improve the experience and mental health of residents who have children during training. Future studies should assess the impact of ABMS policy on residents’ parental and leave decisions.

Although improving the experience of trainee parents is important for their well-being, it may also influence patient-facing outcomes. In other personal domains, such as breastfeeding and personal vaccination status, the lived experiences of physicians and medical trainees are strongly associated with the clinical care they offer to patients.^29,30  Adequate leave and support for pediatric trainees may impact the advocacy, clinical care, and counseling they provide to patients and families regarding parental leave, although this has not yet been studied.

Although this study has numerous strengths, including the high response rate, use of a validated survey measure, its multicenter design, the broad geographic catchment, and inclusion of both female and male residents, there are several limitations. As an observational study, associations noted may reflect the influence of confounding variables. Our findings regarding the prevalence of residents who delay having children may reflect selection bias, if residents who were most seriously considering parenthood were more likely to respond than those who were not. Although we did not observe statistically significant differences related to postpartum depression screening results, leave length satisfaction, breastfeeding duration satisfaction, and perceptions of support when resident leaves of ≤6 weeks were compared with those >6 weeks, there may be another leave duration that represents an inflection point for these measures; we had insufficient data to assess the impact of longer leave lengths. We did not investigate the reasoning behind certain respondent answers, such as why individuals were dissatisfied with their decision to defer childbearing. Furthermore, although individual leave length was not correlated with burnout, the mean program leave length did reach statistical significance, suggesting that the program mean leave length may be a surrogate for unmeasured factors in program environment and culture impacting burnout. All responses were self-reported and subject to recall bias. Finally, our survey primarily collected responses before the start of the coronavirus disease 2019 pandemic, characterized by significant disruptions in child care, schooling, and the experience of parents. As a result, we are unable to assess how pandemic-related changes in pediatric training may impact parenthood and parental leave decisions.³¹ 

This study provides insight on trainee decisions regarding parenthood and characteristics associated with parental leave during pediatric graduate medical education. A majority of trainees delay having children during residency, and many are dissatisfied with their decision to delay. Although experiences vary across a wide range of parental leaves, a majority of parents are not satisfied with leave length or duration of breastfeeding. Future studies should assess the impact of ABMS policy on pediatric resident decision-making surrounding parenthood, average length of resident parental leave, breastfeeding support, and well-being.

Members of the APPD LEARN Parenthood in Pediatric Residency Study Consortium include the following: Vasudha Bhavaraju, Sharon Dabrow, Hillary Franke, Adam Frischknecht, Sharon Istfan, Erin King, Elizabeth Rodriguez Lien, Lanier Lopez, Meredith Plant, Caroline Rassbach, Kira Sieplinga, Dave Swaby, and Stephen Thacker.

AAP

American Academy of Pediatrics

ABMS

American Board of Medical Specialties

ABP

American Board of Pediatrics

APPD LEARN

Association of Pediatric Program Directors Longitudinal Educational Assessment Research Network

CI

confidence interval

EPDS

Edinburgh Postnatal Depression Scale