To examine the association of changes in pediatricians’ work characteristics with their satisfaction using longitudinal data.
Data from a cohort study, the American Academy of Pediatrics Pediatrician Life and Career Experience Study (PLACES), were used to examine self-reported work satisfaction from 2012 to 2020 among 2002–2004 and 2009–2011 residency graduates (N = 1794). Drawing from the Physician Worklife Study, work satisfaction was measured as a 4-item scale score and averaged [range, 1 (low)–5 (high)]. Mixed effects linear regression for longitudinal analysis examined work satisfaction with year as the lone explanatory variable and then with 11 variables that might change over time (time variant) to assess how changes in work might be linked to increased or decreased satisfaction.
In total, 85.9% of pediatricians in 2020 (September–December) thought their work was personally rewarding. Overall mean work satisfaction scale score displayed a small but significant (P < .001) decrease over time (3.80 in 2012 to 3.69 in 2020). Mixed effects modeling identified several changes associated with increasing work satisfaction over time: increased flexibility in work hours (B = 0.23; 95% confidence interval, 0.20 to 0.25) and personal support from physician colleagues (B = 0.18; 95% confidence interval, 0.15 to 0.21) had the largest associations. Pediatricians reporting increased stress balancing work and personal responsibilities and increased work hours had decreased satisfaction scores.
Early- to midcareer pediatricians expressed high levels of work satisfaction, though, satisfaction scale scores decreased slightly over time for the sample overall, including during 2020 (year 1 of the coronavirus disease 2019 pandemic). Pediatricians reporting increases in flexibility with work hours and colleague support showed the strongest increase in work satisfaction.
Recent studies examining satisfaction among pediatricians are scarce and tend to be cross-sectional. In studies of physicians, the majority reported satisfaction, with pediatricians reporting higher satisfaction than many specialties. Overall trends were mostly stable.
Nearly 9 in 10 pediatricians reported work was personally rewarding in 2020. Longitudinal analysis of 2012–2020 data showed work satisfaction decreased slightly over time. Pediatricians reporting increases in work hour flexibility and colleague support showed the strongest increase in satisfaction.
Physician careers can be both challenging and rewarding. Although many studies among physicians in the last decade have focused on burnout, few have examined work and career satisfaction. A 2009 systematic review of physician satisfaction found most studies to be cross-sectional, with only 12 repeated cross-sectional and 2 longitudinal studies meeting the review’s criteria.1 The majority of physicians reported being satisfied in these studies, with pediatricians reporting higher satisfaction than several other specialties. Overall trends for physician satisfaction were relatively stable, with small decreases among primary care physicians.
Recent US studies examining satisfaction among pediatricians are scarce and tend to be cross-sectional. Studies among early- and midcareer pediatricians found 8 in 10 reported they were satisfied with their career as a physician in 20132 and 2018.3 Nearly all (95%) of graduating pediatric residents surveyed in the spring and summer of 2020 reported they would choose pediatrics again, and this is consistent across the last 2 decades.4
Given the association of physician satisfaction with multiple aspects of patient care5–10 and the lack of national longitudinal data on satisfaction among pediatricians,1 we sought to better understand satisfaction trends through the first wave of the coronavirus disease 2019 (COVID-19) pandemic. The current study is a longitudinal analysis of pediatrician-reported work satisfaction. The same group of pediatricians answered questions on satisfaction across multiple years.
Methods
The American Academy of Pediatrics (AAP) Pediatrician Life and Career Experience Study (PLACES) launched in 2012 to track the career and life choices and experiences of pediatricians recruited early in their careers using a national, longitudinal design.11 In this study, we used data collected in 2012–2020 from 2 cohorts of pediatricians (2002–2004 and 2009–2011 residency graduates). In 2020, data were collected after the COVID-19 pandemic started (September–December). PLACES participants are surveyed twice annually, completing a primary, longer survey covering several domains measured each year (eg, work attributes, satisfaction, work–life balance, and life experiences) and a shorter survey on topics of importance to participants. The AAP institutional review board reviewed and approved the PLACES project.
Participants
Participants were recruited in 2012 using an AAP database that includes all pediatricians who completed a US residency program, comprising both AAP members and nonmembers. Four in 10 pediatricians (1925 of 4677) invited to participate in PLACES initially signed up for the longitudinal study. Additional detail regarding the PLACES methodology has been described previously.11 All of the current participants in the PLACES 2002–2004 and 2009–2011 residency graduates cohort (N = 1794) were included within the current analysis. Participation rates varied across survey years, from a high of 94% in 2012 to 79% in 2020.
Survey Content
The content domains and questions for the PLACES surveys were developed through (1) content prioritization by a project advisory committee, (2) literature review to identify related, existing questions, and (3) cognitive interviews and pilot tests to assess how questions were interpreted by participants. Where available, survey questions were adapted from other physician studies, including the Physician Worklife Study12,13 and Medicine in Australia: Balancing Employment and Life.14
Analytic Variables
All study data were self-reported. The main dependent variable was work satisfaction. We used a global work satisfaction measure adapted from the Physician Worklife Study.12,13 The measure included 4 items that used the following agreement scale: 1 = strongly disagree, 2 = disagree, 3 = neither agree nor disagree, 4 = agree, and 5 = strongly agree. The 4 items were the following: (1) I find my work personally rewarding; (2) overall, I am pleased with my work; (3) my work has not met my expectations; and (4) my current work situation is a major source of frustration. These 4 items were asked every year from 2012 to 2020. For each year, similar to the developers of the global work satisfaction scale and other researchers, we reverse coded the latter 2 items and then averaged scores from the 4 items to compute a scaled score to be used as a continuous variable.13,15,16
Time-Varying Explanatory Variables
We examined 11 time-varying explanatory variables focused on pediatrician-reported experiences that were selected based on previous research and the potential to be modifiable at the institution or individual level.1,2 Each variable measured a characteristic that could change from survey year to year. Thus, when each variable was examined across the 9-year study period, changes in the characteristics could be isolated and statistically examined. We dichotomized most variables at each time point to make the model results more comparable across variables. Several variables were dichotomized to strongly agree or agree versus neither agree nor disagree, disagree, or strongly disagree, including flexibility in pediatricians’ work schedule, adequate time to spend with patients, autonomy of clinical decisions, and personal support from physician colleagues. Other variables included total work hours in a typical week (continuous variable), changed jobs in the past year (yes or no), practice owner (yes: full or part owner, or no: employee, independent contractor, or other), academic setting (yes: medical school or university hospital or clinic, or no: other setting), stress balancing work and home responsibilities (very/moderately stressed or a little/not at all stressed) own income (continuous variable; 12 different income levels in $25 000 increments, eg, <$50 000, $50 000–$74 999, $75 000–$99 999, etc), and a COVID-19 variable (data collected during pandemic year: yes or no). We created deviation variables to capture the change effects for all variables.
Time-Invariant Demographic Characteristics
The demographic characteristics that were treated as time-invariant variables within the longitudinal analysis included PLACES age cohort (2002–2004 or 2009–2011 residency graduates), self-identified sex (male or female), underrepresented race and ethnicity in medicine (yes: Black or African American, Hispanic or Latino, American Indian or Alaska Native, and Native Hawaiian or Pacific Islander; or no: White, Asian, or other race), medical school location (United States/Canada or international), and subspecialty trained (subspecialty or no subspecialty). Respondents were included in the subspecialty group if they reported on any of the PLACES surveys that they completed or were in fellowship training.
Data Analysis
Using data from the years 2012 to 2020, we used descriptive statistics to examine work satisfaction. We also conducted mixed-effects linear regression for longitudinal analysis to examine work satisfaction over time. This approach uses all available data, regardless of whether the pediatrician did not respond in all years and had some missing values. It accounts for the fact that the same individual is responding on multiple occasions and provides results as coefficients and 95% confidence intervals (CI).17
We first conducted mixed-effects regression with work satisfaction as the dependent variable and with time (survey year) as the lone explanatory variable for the full sample. Next, we built a model with work satisfaction as the dependent variable and then simultaneously tested the 11 time-varying explanatory variables to assess how changes in work characteristics might be linked to increased or decreased satisfaction. This model also included the time-invariant variables.
For the time-varying variables, we tested change effects. The change effect focused on the deviation from the individual’s mean for the explanatory variable and examined its relationship with satisfaction. If someone had a flexible work schedule in 2012 but not in the other years, then the deviation would be large for 2012 and the model would test if the change toward a less flexible schedule across years was associated with reduced satisfaction.
For time-invariant variables, we tested both baseline and slope components. For example, the baseline value shows if there is a difference between men and women in year 1, and the slope shows if men and women change at the same rate across the duration of the study.
For analyses, we used P ≤ .05 as significant for all inferential tests and performed analyses in Stata version 16 (Stata Corp) and IBM SPSS Statistics version 25.
Results
Data from all current PLACES participants were included in the analysis (n = 1794). Half of the participants were in the 2009–2011 graduate cohort (n = 897) and half in the 2002–2004 graduate cohort (n = 897). Most participants were female (1309 of 1794, 73.0%). Fifteen percent self-identified as a race or ethnicity underrepresented in medicine (263 of 1794), 86.9% (1559 of 1794, 86.9%) graduated from US medical schools, and 41.6% (746 of 1793) had received subspecialty training.
Supplemental Table 2 shows aggregate responses to each of the 4 items that were used to compute the work satisfaction scale for each year. In 2012, 86.4% of pediatricians strongly agreed or agreed their work is personally rewarding and 25.5% strongly agreed or agreed their current work situation is a major source of frustration. In 2020, these percentages were 85.9% and 37.0%, respectively. We computed a scale score for each individual for each year by averaging the scores from the 4 items (2 of 4 items were reverse coded before averaging as described above).
Among pediatricians who responded to all 9 surveys (n = 1030), only 6 pediatricians had the same satisfaction score each year. Comparing scores of those responding in year 1 and year 9, 48.5% of scores decreased, 35.9% increased, and 15.5% remained the same. Among all PLACES respondents, the majority scored higher than 3.5 each year (76.8% in 2012 and 71.2% in 2020). Most scores on this point scale had an absolute change of less than 1.0 point (72.4%); 24.1% of the scores changed by 1.0 to 1.75 points, and 3.6% changed by 2.0 or more points. In 2012, 4 pediatricians had scores of 1.0 (very unsatisfied) and 5.9% (n = 104) had scores of 5.0 (very satisfied). In 2020, 2 different pediatricians had scores of 1.0 and 4.1% (n = 56) had scores of 5.0.
The mixed-effects model for the longitudinal analysis with the work satisfaction scale as the dependent variable and time as the lone explanatory variable found a small but significant decrease in the overall mean work satisfaction scale score over time (P < .001). Estimated means were 3.79 in 2012 [95% CI, 3.76 to 3.82], 3.78 in 2013 (95% CI, 3.75 to 3.80), 3.77 in 2014 (95% CI, 3.74 to 3.79), 3.75 in 2015 (95% CI, 3.73 to 3.78), 3.74 in 2016 (95% CI, 3.72 to 3.77), 3.73 in 2017 (95% CI, 3.70 to 3.75), 3.72 in 2018 (95% CI, 3.69 to 3.74), 3.70 in 2019 (95% CI, 3.68 to 3.73), and 3.69 in 2020 (95% CI, 3.66 to 3.72).
Table 1 presents findings from the mixed-effects model including the 11 time-varying explanatory variables and 5 time-invariant variables, displaying coefficients and 95% CIs for the change effects. The change effect assesses how changes in individual pediatricians’ work characteristics are associated with work satisfaction.
Pediatrician Experienced Changes Associated With Work Satisfactiona
| . | Change Effect . | |
|---|---|---|
| . | Coefficient . | 95% CI . |
| Pediatrician-experienced changes, time-varying | ||
| Increased flexibility in work schedule | 0.23 | 0.20 to 0.25* |
| Increased personal support from physician colleagues | 0.18 | 0.15 to 0.21* |
| Increased time with patients during visits | 0.15 | 0.12 to 0.18* |
| Increased autonomy of clinical decisions | 0.15 | 0.12 to 0.19* |
| Switched jobs | 0.12 | 0.09 to 0.14* |
| Practice owner status | 0.03 | −0.03 to 0.09 |
| Increased income | 0.01 | 0.001 to 0.01** |
| Academic setting | 0.01 | −0.03 to 0.04 |
| Increased work hours | −0.003 | −0.004 to −0.001* |
| COVID-19 year | −0.20 | −0.71 to −0.32 |
| Increased stress balancing work and personal responsibilities | −0.16 | −0.19 to −0.13* |
| Demographic characteristics, time-invariant, slope | ||
| Sex, male | 0.01 | −0.01 to 0.02 |
| 2009–2011 cohort | 0.01 | −0.003 to 0.02 |
| International medical school graduate | −0.003 | −0.02 to 0.01 |
| Underrepresented race and ethnicity in medicine | 0.01 | −0.01 to 0.02 |
| Subspecialty trained | −0.01 | −0.02 to −0.001** |
| . | Change Effect . | |
|---|---|---|
| . | Coefficient . | 95% CI . |
| Pediatrician-experienced changes, time-varying | ||
| Increased flexibility in work schedule | 0.23 | 0.20 to 0.25* |
| Increased personal support from physician colleagues | 0.18 | 0.15 to 0.21* |
| Increased time with patients during visits | 0.15 | 0.12 to 0.18* |
| Increased autonomy of clinical decisions | 0.15 | 0.12 to 0.19* |
| Switched jobs | 0.12 | 0.09 to 0.14* |
| Practice owner status | 0.03 | −0.03 to 0.09 |
| Increased income | 0.01 | 0.001 to 0.01** |
| Academic setting | 0.01 | −0.03 to 0.04 |
| Increased work hours | −0.003 | −0.004 to −0.001* |
| COVID-19 year | −0.20 | −0.71 to −0.32 |
| Increased stress balancing work and personal responsibilities | −0.16 | −0.19 to −0.13* |
| Demographic characteristics, time-invariant, slope | ||
| Sex, male | 0.01 | −0.01 to 0.02 |
| 2009–2011 cohort | 0.01 | −0.003 to 0.02 |
| International medical school graduate | −0.003 | −0.02 to 0.01 |
| Underrepresented race and ethnicity in medicine | 0.01 | −0.01 to 0.02 |
| Subspecialty trained | −0.01 | −0.02 to −0.001** |
Drawing from the Physician Worklife Study, work satisfaction was measured using a 4-item scale score (see Supplemental Table 2 for responses to individual items for each year); responses to the 4 items were averaged to create the scale.
P < .001;
P < .05.
Time-Varying Explanatory Variables
The variables in the mixed effects model with the strongest change effect with increased satisfaction are listed at the top of Table 1; variables associated with decreased satisfaction are at the bottom. For 8 of the 11 variables examined, statistically significant change effects were detected. The strongest associations with increased satisfaction were for increased flexibility with work hours (β = 0.23; 95% CI, 0.20 to 0.25), increased personal support from physician colleagues (β = 0.18; 95% CI, 0.15 to 0.21), and increased time with patients (β = 0.15; 95% CI, 0.12 to 0.18). Increased stress balancing work and personal responsibilities (β = −0.16; 95% CI, −0.19 to −0.13) and increased work hours (β = −0.003, 95% CI, −0.004 to −0.001) were associated with decreased satisfaction. The COVID-19 year variable was not significant.
Time-Invariant Demographic Characteristics
Higher levels of satisfaction were found at baseline for male pediatricians (β = 0.08; 95% CI, 0.02 to 0.15), but as shown in Table 1, the trajectory of male pediatricians was similar to female pediatricians (slope effect: β = 0.01; 95% CI, −0.01 to 0.02) in the full multivariable model. The trajectory of level of satisfaction for subspecialty pediatricians decreased slightly faster than those without such training (slope effect: β = −0.01; 95% CI, −0.02 to −0.001). The 2 age cohorts, underrepresented race and ethnicity in medicine, and medical school location did not differ in level of satisfaction at baseline or in the trajectory of satisfaction over time.
Discussion
In this national longitudinal study among early- to midcareer pediatricians, nearly 9 in 10 pediatricians in 2020 thought their work was personally rewarding. At the same time, almost 4 in 10 were frustrated with their work situation. Pediatrician-reported work satisfaction scale scores in 2012–2020 decreased slightly over time for the sample overall. Using longitudinal data analysis, we examined how changes in pediatricians’ work characteristics might be linked to increased or decreased satisfaction. Across the study years, pediatricians who reported increases in their flexibility with work hours, support from colleagues, and time with patients showed the strongest increase in work satisfaction. Pediatricians who reported increases in their stress balancing work and personal responsibilities and increases in work hours were associated with decreased satisfaction.
To our knowledge, no other longitudinal analyses of US pediatrician work satisfaction have been published. Although there is a longitudinal study of US emergency medicine physicians,18 studies identified in the literature on satisfaction using these data focused on data from 1994 to 201419,20 or on residents.21 Cross-sectional studies report some similar predictors of satisfaction that our longitudinal study found. A review of studies published between 2008 and 2013 that examined US physician work and/or career satisfaction found that autonomy, income, and perceptions of job demands were associated with physician satisfaction.22 Increased hours worked correlated with less job satisfaction in a cross-sectional survey of US surgeons.23 In a prospective study of physicians working for a single healthcare organization, physicians reporting less satisfaction were more likely to reduce their full-time equivalent units over the following 24 months.24 A 2008 study of female emergency physicians found that schedule flexibility, relationships with colleagues, and fair compensation was associated with career satisfaction.25
Physician studies on burnout and satisfaction with work–life balance have been more commonly published. A study comparing data collected in 2011 and 2014 on satisfaction with work–life balance and burnout reported that both measures worsened from 2011 to 2014.26 Another longitudinal study using PLACES data found significant increases in burnout over time, with 1 in 3 early- and midcareer pediatricians experiencing burnout.27 This study also found that increased flexibility in work schedules, change in jobs, increased time to spend with patients, increased autonomy at work, and increased colleague support were associated with decreased burnout.
Our 4-item measure of work satisfaction was adapted from the Physician Worklife Study. Studies in the 1990s of physicians using the 5-item Physician Worklife Study global job satisfaction measure reported a mean satisfaction score of 3.7016 and 79% percent satisfied with their jobs.28 We found similar or higher work satisfaction scores for the pediatricians in our study.
System approaches building from our results that might help address the small yet significant decline in work satisfaction include supporting flexible work hours, monitoring total work hours, increasing time with patients, increasing physician autonomy, and building in time and activities to foster relationships with colleagues. Studies suggest that physicians spend a large portion of their time on electronic health records (EHR) and administrative tasks during and after work hours.29–31 A study using data collected from PLACES pediatricians in 2018 found that three-fourths reported EHR documentation as a major or moderate burden, and such burden was associated with lower scores on work-life balance and career and life satisfaction measures.3 This study also reported that 9 in 10 pediatricians strongly agreed or agreed that improving the functionality of the EHR, providing protective time for administrative work, and obtaining physician input might help reduce administrative burdens. Finally, physician autonomy has been reported to be associated with satisfaction,1,2 including the current study, and burnout.27 It might be helpful for practices and institutions to ask pediatricians what interventions could be implemented to improve their satisfaction. This is particularly critical given the negative impacts of the current pandemic on families expressed in the lay literature, as both stress balancing work and personal responsibilities and increased work hours were associated with decreased satisfaction in this study.
Our study limitations included the fact that data are self-reported and limited to early- and midcareer pediatricians; therefore, the generalizability of the findings to other specialties or career stages remains unclear. Although the response rate among PLACES participants who completed the 2012–2020 surveys was high, the initial project sign-up rate was lower at 41%.11 Also, we were not able to investigate changes in patient satisfaction, administrative burden and documentation pressures, and payment and quality metric changes that might also be associated with work satisfaction. Finally, although this study used 2020 data that was collected during the COVID-19 pandemic (September–December), it does not include data collected during the heavy surges of the delta and especially omicron waves in 2021 and 2022. Many pediatricians have felt frustrated and exhausted by the pandemic and concerned about the toll COVID-19 has taken on children, and this may have impacted their perspectives. With PLACES, we will be able to compare work satisfaction before, during, and after the pandemic. Our currently available results, however, suggest that building in flexibility, collegial support, attention to total hours worked, and work–family balance are important levers of support to build on during the pandemic.
Conclusions
In summary, our study documented that most pediatricians during 2012-2020 thought their work was personally rewarding. We found a slight decrease in overall work satisfaction scores among early and midcareer pediatricians. Using longitudinal data analysis, we examined how changes in pediatricians’ work or personal experiences might be linked to increased or decreased satisfaction. Pediatricians who reported increases in their flexibility with work hours and support from colleagues showed the strongest increase in work satisfaction. Pediatricians who reported increases in their stress balancing work and personal responsibilities and increases in work hours were associated with decreased satisfaction.
ACKNOWLEDGMENTS
We acknowledge the contribution of the PLACES participants.
Ms. Frintner conceptualized and designed the study, carried out the analyses, supervised data collection, drafted the initial manuscript, and reviewed and revised the manuscript; Drs. Leslie and Starmer assisted with the conception and design of the study and critically reviewed the manuscript; Ms. Gottschlich coordinated data collection, reviewed all analyses, and reviewed the manuscript; Dr. Cull assisted with the conception and design of the study, reviewed and interpreted the analyses, and reviewed the manuscript; and all authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.
The research presented in this paper is that of the authors and does not reflect the official policy of the American Academy of Pediatrics.
FUNDING: This study was supported by the American Academy of Pediatrics.
CONFLICT OF INTEREST DISCLOSURES: The authors have indicated they have no conflicts of interest relevant to this article to disclose.
Comments