Academic Performance of Clerkship Students Working on a Longitudinal Inpatient Pediatrics Team

The pediatric clerkship is a 6-week rotation consisting of a 3-week inpatient experience, a 1-week ambulatory clinic experience, a 1-week well-baby nursery experience, and a 1-week subspecialty experience. All services are completed consequently, but the order varies for each student. The clerkship was altered slightly during academic year (AY) 2020 to 2021 because of the coronavirus disease 2019 pandemic. The total duration was decreased to 4 weeks, 2 weeks of inpatient, 1 week each of ambulatory and nursery, and no subspecialty week. All students are required to achieve a passing grade in the clerkship.

During the inpatient portion of the clerkship, the majority of students work on the general wards of a 289-bed free-standing children’s hospital and care for patients with a variety of pathologies. A minority of students work on a single pediatric floor of a community hospital, consisting of 24 inpatient beds and a special care nursery. Students rotating through the free-standing children’s hospital work either on 1 of 3 traditional ward teams or on a single longitudinal team. Students rotating through the community hospital work directly with a pediatric hospitalist without a resident (Fig 1). All students remain on the same team throughout their inpatient experience. Team assignment is based on student preference, and ∼90% of students receive their first team choice. Students who do not receive their first choice are randomly assigned to any of the remaining teams. All students, regardless of service, have the same required clinical encounters to complete by the end of the clerkship, thus ensuring a comparable volume and mix of patients.

We defined the study population as all medical students who rotated on the pediatric clerkship from AY 2017 to 2018 through AY 2020 to 2021. The majority completed their pediatric clerkship during their third year of medical school, with a small minority deferring the rotation until their fourth year. To ensure consistency in the sample, we excluded data from MD/Ph.D. students, transfer students, and any second attempts by students after failing the first attempt of the pediatric clerkship. Additionally, students who opted out of having their academic data used for educational quality improvement were excluded from the analyses.

We defined the exposure group as those students who rotated on the longitudinal medical student team during the inpatient portion of the pediatric clerkship. The team consists of a hospitalist attending, a second-year pediatric resident, and 4 to 5 medical students. The students are listed as their patients’ primary contact in the electronic medical record, thus any correspondence from health care team members, including concerns related to a patient’s clinical deterioration, goes directly to them. They enter all patient orders which are cosigned by the resident and function without the supervision of a dedicated intern. In the majority of rotations, students work with the same resident and attending over their entire 3 weeks and are expected to care for 2 to 3 patients at a time. The acuity and complexity of the team’s census are intentionally lower than that of other teams to facilitate resident autonomy and keep the resident’s workload manageable.

We defined the comparison group as those students who rotated on any team other than the longitudinal team during the inpatient portion of the pediatric clerkship. We included in this group students who worked on traditional teams at the free-standing children’s hospital and students who worked directly with a hospitalist at the community hospital. Students who rotated on traditional teams tended to work with the same interns, resident, and attending for at least 1 week and occasionally longer. In contrast, students at the community hospital would work with a different attending every day. In addition to hospitalist attending physicians, students at the freestanding children’s hospital could also work with attending physicians from other subspecialties. Similar to the exposure group, students are expected to serve as the primary caregiver for 2 to 3 patients at a time. However, students are paired with an intern who is listed as the patient’s primary contact in the electronic medical record. Thus, the intern receives direct communication from other health care team members and is the first to be alerted if a patient is clinically deteriorating. Although students have the ability to enter orders for a resident to cosign, this is a task frequently handled by interns to improve team workflow efficiency.

We defined our primary outcome as students’ academic performance in the pediatric clerkship and our secondary outcome as students’ academic performance in the clerkship immediately after pediatrics. During each portion of the pediatric clerkship, a student’s performance is evaluated by attending and resident physicians by using a standardized rubric aligned with Entrustable Professional Activities (Supplemental Figure 3). Faculty and resident preceptors select a rating for each Entrustable Professional Activity utilizing word-anchor performance descriptors; no numbers are provided on the evaluation. To help improve objectivity, the clerkship director moderates a group evaluation in which agreement is reached between faculty and residents that worked alongside the student. The ratings are converted to a 0 to 4 scale, and students must achieve an average preceptor rating of 1.7 to pass the clerkship. The inpatient preceptor rating is worth 30% of the final clerkship grade and is combined with a student’s nursery preceptor rating, subspecialty preceptor rating, ambulatory preceptor rating, end of clerkship Objective Structured Clinical Examination (OSCE) score, and end of clerkship National Board of Medical Examiners shelf score to make up their final grade. The final grade is initially calculated on a continuous scale from 0 to 4 and then converted to “Fail,” “Pass,” “High Pass,” and “Honors” designation by using thresholds predetermined by the School of Medicine.

We defined students’ pediatric clerkship academic performance as their final grade, inpatient preceptor rating, nursery preceptor rating, subspecialty preceptor rating, ambulatory preceptor rating, mean preceptor rating across all 4 clinical portions of the clerkship (inpatient, nursery, ambulatory, and subspecialty), OSCE score, and shelf score. We defined students’ academic performance in the clerkship immediately after pediatrics as the final grade and average preceptor rating for that clerkship.

We described continuous student characteristic data with means and standard deviations, ordinal data with medians and interquartile ranges, and categorical data with frequencies and percentages. We stratified student characteristics by the inpatient team they worked on during the pediatric clerkship, either the longitudinal team or traditional team, and compared them using the χ² test, Wilcoxon rank test, or the student’s t test when appropriate. For continuous outcome data, we calculated unadjusted mean values and 95% confidence intervals (CIs) and tested for significance between groups using the student’s t test. For ordinal outcome data, we determined frequencies and percentages of students in each group and tested for significance between groups using the Wilcoxon rank test. Because of the unique educational experience of students rotating at the community hospital, who were included in the traditional team group, we performed a sensitivity analysis evaluating the impact of censoring their data on the primary outcomes.

For the adjusted analysis, we used multiple linear regression and multiple ordered logistic regression to compare the academic outcomes of students on the longitudinal team and students on traditional teams. The models were used to account for the potential confounders of sex, race/ethnicity, United States Medical Licensing Examination Step 1 score, average nonpediatrics clerkship final grade, average nonpediatrics clerkship preceptor rating, academic year, and the number rotation in the AY students rotated on pediatrics. Confounders were determined a priori on the basis of their relevance to the exposure and outcomes. We included race and ethnicity data because of previous studies documenting an association between these variables and clinical clerkship grades.^25,26  For continuous outcomes, we used the linear regression model to calculate adjusted means and 95% CIs. For ordinal outcomes, we used the ordered logistic regression model to calculate predicted probabilities and 95% CIs. We verified the ordered logistic regression model satisfied the proportional odds assumption using a likelihood ratio test.²⁷  We used 2-sided hypothesis tests and considered P values <.05 significant. We analyzed the data using Stata version 14 statistical software (Stata-Corp, College Station, TX). The study was deemed exempt by the study institution’s institutional review board.

An overview of student data included in the study analysis is shown in Fig 2. A total of 316 students were included in the longitudinal team group and 147 in the traditional team group. Students in the longitudinal group were more likely to be female and had significantly higher mean preceptor ratings and median final grades on all nonpediatrics clerkships (Table 1). We did not observe significant differences between groups regarding race and ethnicity, mean United States Medical Licensing Examination Step 1 score, AY, or rotation number.

Results of the regression analyses are shown in Table 2 with unadjusted results displayed in Table 2 and Supplemental Table 3. After adjusting for multiple confounders, students on the longitudinal team were found to have significantly higher mean preceptor ratings (mean 3, 95% CI 2.97 to 3.03 vs mean 2.85, 95% CI 2.81 to 2.90; P = .02; scale of 0 to 4), and mean inpatient preceptor ratings (mean 3.08, 95% CI 3.04 to 3.11 vs mean 2.88, 95% CI 2.82 to 2.93; P = .01) when compared with traditional team students. The adjusted analysis also revealed that working on the longitudinal team was associated with an increased probability of achieving a higher final grade in the pediatric clerkship (adjusted probability of achieving honors 22%, 95% CI 17% to 28% vs 11%, 95% CI 6% to 16%; P = .003). There was no significant difference between longitudinal and traditional team students’ adjusted pediatrics nursery, ambulatory, or subspecialty preceptor ratings. There was also no difference between groups in preceptor ratings (mean 3.28, 95% CI 3.26 to 3.31 vs mean 3.22, 95% CI 3.18 to 3.26; P = .27) or final grades (adjusted probability of achieving honors 24%, 95% CI 18% to 30% vs 27%, 95% CI 19% to 35%; P = .70) in the clerkship immediately after pediatrics. The mean shelf score, in which there was a significant difference between groups in the unadjusted analysis, was not significantly different after adjusting for confounders (longitudinal team mean adjusted shelf score 79, 95% CI 78 to 79 vs traditional team mean adjusted shelf score 77, 95% CI 77 to 78; P = .41).

Censoring data from students who rotated at the community hospital during their inpatient pediatrics weeks did not significantly impact the primary outcomes. The mean inpatient preceptor rating and median final pediatric clerkship grade of the traditional group did not change, whereas the mean preceptor rating of the traditional group increased slightly. The significant differences between groups observed in the unadjusted analysis persisted after the data were censored (Supplemental Table 4).

Our institution’s longitudinal medical student team was designed to enhance students’ inpatient pediatric experience by providing greater patient care responsibility and preceptor continuity, elements that are common to many LICs.^11,14  The research was conducted to determine if working on the longitudinal team was associated with improved academic performance in the pediatric clerkship and the clerkship immediately after pediatrics. Our results reveal that working on the longitudinal team was associated with a higher probability of receiving an honors or high pass final grade in the pediatric clerkship. After controlling for multiple confounders, the association persisted and was driven by students on the longitudinal team receiving higher inpatient preceptor ratings. We did not observe significant differences between groups in measures of academic performance unrelated to the inpatient portion of the clerkship and the differences did not persist in the clerkship immediately after pediatrics. This suggests the observed differences in academic performance were directly related to students’ inpatient pediatric experience.

Our results are consistent with previous studies that have revealed a positive association between enrollment in a LIC and academic achievement.^28,29,30  Given the longitudinal team in this study worked together for only 3 weeks, as opposed to months in LICs, our results suggest the benefits of a traditional LIC may take a shorter time to materialize. This could help justify the implementation of key elements of a traditional LIC within existing clinical education programs without the burden of a major curricular overhaul.

To our knowledge, there is limited literature describing medical students’ performance on inpatient teams without interns. Scheffer et al found that patients had positive perceptions of inpatient teams made up of final-year medical students and a supervising physician but did not investigate the impact of these teams on student performance.³¹  Antommaria et al investigated the impact of the same longitudinal team at our institution and identified that students in the intervention group received higher preceptor ratings in the domains of data interpretation and patient interaction.²⁴  We expanded on this previous study by including a larger number of students, and analyzing a broader range of student outcomes, including academic performance in the clerkship immediately after pediatrics, pediatrics OSCE score, preceptor ratings in the noninpatient portions of the clerkship, and final pediatric clerkship grade.

We hypothesize the improved academic performance seen in students on the longitudinal team is related to their more autonomous role and their longitudinal relationships with preceptors. These features directly relate to the SDT principles of autonomy and connectedness. They could foster students’ intrinsic motivation and thereby lead to improved performance.³²  An alternative explanation could be that the improved ratings are related to preceptors’ increased familiarity with students’ performance on the longitudinal team. The greater duration of interaction would allow for greater depth of connection, and possibly a less-critical view of the student, both of which could contribute to higher ratings independent of actual student performance.

Regardless of the explanation for the higher ratings, this finding has created 2 avenues for clerkship leadership. In the short term, the team is working to create equitable evaluations across the clerkship. In the longer term, the team plans to identify key elements of the longitudinal team that foster student success, and better incorporate these elements throughout the clerkship. More broadly, we suggest future research related to similar interventions focus on measuring higher-level learning outcomes, such as multisource student evaluations from patients, clinical and administrative staff, and peers, as well as patient-based outcomes.^33–36 

Our study has several important limitations. First, although our clerkship attempts to maximize preceptor objectivity through a rigorous evaluation process, the potential for preceptor subjectivity cannot be completely mitigated. Previous studies have revealed that observation of a learner by a preceptor over an extended period may be more prone to rater biases which could have directly impacted our results.^30,37,38  Second, students self-selected the team they worked on during the inpatient portion of their pediatric rotation. Thus, internal motivational factors, such as a desire to pursue a career in pediatrics, unrelated to the educational experience may have influenced academic performance and confounded the results. Third, although regression analyses were used to account for multiple confounders, it is possible the groups may have differed by other unmeasured factors. One such factor is patient complexity. At our institution, the longitudinal team tends to care for less medically complex patients, an aspect that may have allowed students on this team to appear more competent to their preceptors. Fourth, we did not evaluate patient-centered outcomes and thus cannot draw conclusions related to how the longitudinal team affected patient care.

In conclusion, for medical students in the pediatric clerkship, working on a longitudinal inpatient team was associated with a higher inpatient preceptor rating and final clerkship grade. The team serves as an example of how to increase medical students’ patient care responsibilities and preceptor continuity in an existing clinical education program. Our results suggest that implementing similar models within clinical clerkships may help foster optimal student performance.

The authors would like to thank Dr Eric Coon for his critical review and input during manuscript preparation.