Improving USMLE Step 1 outcomes in academically vulnerable students through a targeted, competency-based curriculum Authors: Meghan M. Kiefer, Edith Wang, Kellie Engle, Maya Sardesai, Heather McPhillips, Matthew J. Cunningham Published: 2025 ABSTRACT IntroductionFailing the United States Medical Licensing Exam (USMLE) Step 1 can have significant and lasting negative consequences on medical student well-being, academic performance and time to graduation. Although academic predictors can identify students at higher risk for Step 1 failure, there is limited evidence on how to best support these students in achieving first-attempt success and maintaining on-time graduation. Time-variable progression is one feature of competency-based medical education, but simply extending the dedicated study period is not associated with improved Step 1 outcomes.MethodsBeginning in 2019, the University of Washington School of Medicine offered a 10-week Comprehensive Basic Sciences Review (CBSR) course to students identified as having an elevated risk for Step 1 failure. This optional curricular intervention precedes the dedicated study period for these students. Program effectiveness was evaluated by comparing Step 1 scores and on-time graduation rates of 50 students who took CBSR in 2019–2021 with 50 academically propensity-matched controls.ResultsCBSR students scored significantly higher on Step 1 (mean of 212.7 vs. 202.1, P < 0.001) than their academically-matched controls. Rates of on-time graduation were similar between the two groups (65.3% [32/49]) for CBSR students, compared to (60.0% [30/50] for the control group).DiscussionThis proactive, optional curriculum is associated with significantly higher Step 1 scores for at-risk students than their propensity-matched controls, without negatively affecting on-time graduation. Background Supporting students in achieving success in their licensing exams is a key component of undergraduate medical education, and one aligned with a mission of developing excellent physicians [1]. 94% of U.S. medical schools require passing the United States Medical Licensing Exam (USMLE) Step 1 for graduation. To support students in this endeavor, most institutions offer additional Step 1 resources beyond the core curriculum, including access to tutors, learning specialists, practice question banks, and review sessions, as well as medical or mental health support when indicated. Failing Step 1 can have a lasting negative impact for students [2, 3] and can lead to increased administrative burden from a scheduling, support, and curricular standpoint. Preventing Step 1 failures benefits all involved. The literature has consistently demonstrated associations between Step 1 outcomes and earlier academic variables, including Medical College Admissions Test [MCAT] scores, pre-clerkship academic performance, and National Board of Medical Examiners’ Comprehensive Basic Science Exam [CBSE] scores [4–8]. Therefore, we can with some accuracy identify students at high risk of Step 1 failure in advance of the dedicated Step 1 study period (“dedicated period.”) However, targeted, evidence-based interventions for these students are lacking. In many medical schools, additional time to prepare for Step 1 is only offered to struggling students at the end of the dedicated period, when it is apparent that the student is not ready to sit for the exam. These students must decide between two unpalatable choices: either (1) drop their clerkships and continue high-intensity studying as their classmates move on, or (2) sit for the exam and hope for the best. Both options carry disruption and risk. As most students who fail Step 1 on initial attempt later succeed in subsequent attempts [9]; additional study time appears to support competency development for these students. However, more time spent in the dedicated period has not been associated with better outcomes [8]. We also know that students experience high rates of burnout and strain during the dedicated period [10]. Offering flexible timing around Step 1 has been proposed as part of a competency-based approach to medical education [11, 12]. Some schools offer flexible Step 1 timing, and others offer it after core clerkships. There is some evidence for higher pass rates when schools placed Step 1 after clerkships, albeit with lower clerkship subject exam scores [13, 14]. Thus, although academic predictors make it possible to identify students at high risk of negative Step 1 outcomes, the optimal approach to lower their risk is unclear. They may benefit from more time, although perhaps not in continuous dedicated study, and not necessarily by delaying study until a later date. Delaying Step 1 may also have unintended consequences: it can reduce opportunities for timely remediation, prolong academic progression, and, in some cases, delay the recognition that a student is unlikely to complete medical school—resulting in greater personal, financial, and emotional costs [13]. At our institution, students complete an 18-month pre-clerkship phase in December. Beginning in mid-January, they enter the Consolidation course, an 8-week dedicated period with curricular supports (review sessions, practice exams, dedicated time to study for and take the United States Medical Licensing Exam [USMLE] Step 1 exam). Clerkships then begin in mid-March. Our school’s passing rate for first-time test-takers in 2017 and 2018 had fallen to 93% (mean score, 224.4) and 94% (mean score, 226.4), respectively, below the national average of 96% (mean, 229–230). To address this problem, we conducted a targeted needs assessment through conversation with students who either failed Step 1 or delayed the exam, learning specialists, Student Affairs leadership, and review of our internal academic performance indicators. Through this it was apparent that many of these students lacked sufficient (1) understanding of some core physiological principles, (2) skill with interpreting and answering USMLE-style test questions, (3) test-taking confidence and self-efficacy, or some combination of the three. We posited that for students at high risk of failure, building these skills via a targeted curriculum prior to initiating a dedicated study period would allow students to be more productive once they began. We set a primary goal of improving Step 1 performance among our students at increased risk of Step 1 failure, with a secondary aim to avoid increased risk of delayed graduation. To do this, we expanded our curricular offerings to account for different timeframes for Step 1 competency development. Starting in 2019, we began to offer a course titled Comprehensive Basic Sciences Review (CBSR) that provided additional review of basic sciences and test-taking skill development prior to beginning their dedicated Step 1 study period. After this course, when other students begin their clerkships, students in this track begin their dedicated study period. To offset the implications of starting clerkships later, the CBSR course is credited towards an elective graduation requirement typically completed in the fourth year, as seen in Fig. 1. Fig. 1Comparison of the schedules of students following a standard schedule, the old extended Step 1 study schedule, and the new CBSR study schedule. Prior to implementation of CBSR, students who required additional time to prepare had additional unstructured study time after the standard dedicated period. CBSR offers a proactive readiness phase prior to initiation of dedicated studying. *Since students receive clinical elective credit for CBSR, the time needed to complete the required fourth-year curriculum is shortened, allowing them the same flexibility in their fourth year for away rotations, Step 2 studying, and residency interviews as other students Comparison of the schedules of students following a standard schedule, the old extended Step 1 study schedule, and the new CBSR study schedule. Prior to implementation of CBSR, students who required additional time to prepare had additional unstructured study time after the standard dedicated period. CBSR offers a proactive readiness phase prior to initiation of dedicated studying. *Since students receive clinical elective credit for CBSR, the time needed to complete the required fourth-year curriculum is shortened, allowing them the same flexibility in their fourth year for away rotations, Step 2 studying, and residency interviews as other students Course design CBSR is a 10-week, faculty-led review course offered to students identified as being at elevated risk of Step 1 failure. Delivered fully online to accommodate the school’s regional campus model, the course combines foundational science review with structured support for test-taking strategies and academic self-efficacy. Although not a formal competency-based curriculum, CBSR was designed using principles aligned with competency-based education as outlined by Van Melle et al. [15], including individualized pacing, targeted instruction, and performance-based feedback (Fig. 2). Weeks are` organized by subject area and target topics where prior cohorts had performed less well. Each week includes curated materials from a commercial resource not otherwise provided by the institution (ScholarRx: specifically, Rx Bricks modules, instructional videos, and quizzes), a set of internally developed boards-style practice questions, and a faculty-led small-group seminar. During the seminar sessions, faculty facilitate in-depth discussion of question logic, clinical reasoning, and physiological concepts. Students are encouraged to share their thought processes, compare problem-solving strategies, and learn from one another’s reasoning, creating a peer-supported learning environment within a structured, low-stakes setting. To allow students to individualize their study program, students must complete the work for any 8 of the 10 weeks for credit, based on their areas of greatest content review; for the other two weeks, they may attend sessions or focus on other areas where they need extra review. The course is graded pass/fail, with passing determined by completion of the required weekly assignments and active participation in sessions. After CBSR, students enter their dedicated study period, with continued academic oversight and check-ins from the course team. CBSR is designed as a proactive intervention that helps students build readiness and confidence before transitioning to independent Step 1 study. Study aims Given the primary goal of improving Step 1 performance among academically vulnerable students, with a secondary goal to avoid increased risk of delayed graduation, we sought to evaluate the effectiveness of the program by comparing Step 1 outcomes and on-time graduation of students who enrolled in CBSR with academically-matched controls who did not participate in CBSR. Fig. 2CBSR Design Alignment with Principles of Competency-Based Education. The principles of Competency Based Curriculum as outlined by Van Melle [15] are listed in the left column, and the corresponding features of CBSR are in the right column CBSR Design Alignment with Principles of Competency-Based Education. The principles of Competency Based Curriculum as outlined by Van Melle [15] are listed in the left column, and the corresponding features of CBSR are in the right column Methods Course eligibility Students are considered for enrollment if they meet any of the following criteria, set after review of our internal data of Step 1 performance of prior cohorts: pre-clinical course grade average 1.75 standard deviations below the cohort mean, history of one or more pre-clinical course failures, USMLE Comprehensive Basic Sciences Exam (CBSE) score 1.75 standard deviations below the cohort mean, and/or a recommendation from an academic dean with in-depth knowledge of the student’s individual situation and performance. We extend invitations to CBSR to students meeting these criteria in December of Year 2 of the pre-clerkship phase. Data and study design This project and analysis were submitted to the University of Washington Institutional Review Board and deemed exempt from their review (STUDY00015607). We analyzed data from second-year medical students (M2s) at University of Washington from 2017 to 2021, including Step 1 scores, average preclerkship academic performance, National Board of Medical Examiner’s Comprehensive Basic Sciences Examination (CBSE scores, total Medical College Admission Test (MCAT) scores, and demographic data from admissions. Total MCAT scores were converted to percentiles using the published AAMC percentile ranks for both the current and old MCAT exams [16]. Pre-clerkship course average was converted to Z-scoring to better analyze across multiple M2 cohorts. Statistical analysis To understand the impact of CBSR enrollment on Step 1 scores, passing rates, and graduation delays, we used propensity score matching to compare the outcomes of CBSR students with those of academically matched controls who did not take the CBSR course. SPSS Statistics v30 was used for analysis. Logistic regression to predict participation in CBSR was performed on the 2019–2021 M2 cohorts with the following predictor variables: pre-clerkship grade average (Z-score by cohort), total MCAT score (percentile rank), and Comprehensive Basic Science Exam performance (percentile rank). The parameters from the logistic regression were used to create a propensity score equation. Propensity scores from CBSR students were hand-matched to non-CBSR students by an author (MJC) who was blinded to Step 1 outcomes. Results CBSR student population From 2019 to 2021, 74 students participated in CBSR out of the 101 that were invited. Students who accepted CBSR invitations, when compared to those who declined to participate, had no significant differences in their academic metrics or demographic characteristics. As expected, the CBSR students had poorer academic performance when compared to the rest of their cohort who were not invited: they had a lower mean Z-score for their pre-clerkship academic performance (−1.57 vs. 0.22), a lower CBSE percentage rank (14.1 vs. 48.0%), and lower MCAT percentage rank (63.7 vs. 79.7%); all p-values by t-test were < 0.0001. (Table 1). CBSR students were more likely to have a racial/ethnic identity that is underrepresented in medicine (29.7% vs. 10.5%, p-value < 0.0001.) Table 1Academic and demographic characteristics of 2019–2021 M2s at university of Washington by (Comprehensive basic science Review) CBSR enrollment statusVariableCBSR participants(n = 74)Declined CBSRN = 27Non-CBSR eligible students(n = 703)Block average79.4 (2.58)79.9 (3.41)88.1 (4.04)Block average Z-score−1.57 (0.534)−1.49 (0.715)0.223 (0.837)CBSE percent14.0 (8.75)15.5 (8.23)48.5 (24.2)Average MCAT total percentile (mean, SD)65.0 (16.3)68.8 (16.7)79.8 (15.5)Average age during dedicated period (mean, SD)28.5 (4.89)27.9 (4.98)26.4 (2.97)Gender (% female)43 (58.1%)18 (66.7%)390 (55.5%)Underrepresented in medicine racial/ethnic identity (%)22 (29.7%)6 (22.2%)74 (10.5%) Table 1Academic and demographic characteristics of 2019–2021 M2s at university of Washington by (Comprehensive basic science Review) CBSR enrollment statusVariableCBSR participants(n = 74)Declined CBSRN = 27Non-CBSR eligible students(n = 703)Block average79.4 (2.58)79.9 (3.41)88.1 (4.04)Block average Z-score−1.57 (0.534)−1.49 (0.715)0.223 (0.837)CBSE percent14.0 (8.75)15.5 (8.23)48.5 (24.2)Average MCAT total percentile (mean, SD)65.0 (16.3)68.8 (16.7)79.8 (15.5)Average age during dedicated period (mean, SD)28.5 (4.89)27.9 (4.98)26.4 (2.97)Gender (% female)43 (58.1%)18 (66.7%)390 (55.5%)Underrepresented in medicine racial/ethnic identity (%)22 (29.7%)6 (22.2%)74 (10.5%) Academic and demographic characteristics of 2019–2021 M2s at university of Washington by (Comprehensive basic science Review) CBSR enrollment status Propensity matching Of the 74 students enrolled in CBSR from 2019 to 2021, twelve were excluded from analysis because they were missing either predictor or outcome data. Of the remaining 62 CBSR students, we were able to match 50 to non-CBSR students (Table 2). Given that a majority of the academically struggling students during those years chose to enroll in CBSR, there were an insufficient number of matches from the 2019-21 cohorts, and we expanded our search for non-CBSR controls to the two preceding cohorts (2017 and 2018) as well. For the 50 matched pairs, the standardized mean difference between the two groups in the pertinent variables was as follows: propensity score (0.01), grade average (0.057), CBSE (0.07), MCAT (0.156) (Fig. 3). Fig. 3Comparison of the academic predictors of students enrolled in CBSR (blue, left side) compared to academically matched students not enrolled in CBSR (green, right side). Panel A shows MCAT percentile rank, Panel B shows the Z-score of student’s preclerkship GPA, and Panel C shows student CBSE percentile rank. Figure created using SPSS Statistics software Comparison of the academic predictors of students enrolled in CBSR (blue, left side) compared to academically matched students not enrolled in CBSR (green, right side). Panel A shows MCAT percentile rank, Panel B shows the Z-score of student’s preclerkship GPA, and Panel C shows student CBSE percentile rank. Figure created using SPSS Statistics software Table 2Academic and demographic characteristics of propensity-matched comprehensive basic science review (CBSR) and non-CBSR studentsVariableCBSR matched cases(n = 50)Non-CBSR matched controls (n = 50)Block average (SD)80.0 (2.48)80.1 (2.26)Block average Z-score−1.48 (0.52)−1.51 (0.52)Average MCAT percentile (SD)65.2 (16.6)67.58 (12.9)CBSE percentile (SD)15.22(9.24)15.86 (8.73)Average age during dedicated period (SD)29.0 (5.08)28.0 (4.15)Gender (% female)28/5031/50Underrepresented in medicine racial/ethnic identity (%)15/508/50 Table 2Academic and demographic characteristics of propensity-matched comprehensive basic science review (CBSR) and non-CBSR studentsVariableCBSR matched cases(n = 50)Non-CBSR matched controls (n = 50)Block average (SD)80.0 (2.48)80.1 (2.26)Block average Z-score−1.48 (0.52)−1.51 (0.52)Average MCAT percentile (SD)65.2 (16.6)67.58 (12.9)CBSE percentile (SD)15.22(9.24)15.86 (8.73)Average age during dedicated period (SD)29.0 (5.08)28.0 (4.15)Gender (% female)28/5031/50Underrepresented in medicine racial/ethnic identity (%)15/508/50 Academic and demographic characteristics of propensity-matched comprehensive basic science review (CBSR) and non-CBSR students Overall outcomes From 2019 to 2021, 101 students were invited to participate in CBSR; 74 (73.3%) accepted and 27 (26.7%) declined. Among those invited to participate, 14.7% (14/95) (12 who enrolled, 2 who declined) failed Step 1, compared to 1.3% (9/703) of students who were not invited. On-time graduation rates were lower in those who enrolled in CBSR than those who declined (Table 3). Table 3Step 1 outcomes and on-time graduation rates of 2019–2021 M2s at university of Washington by CBSR statusVariableCBSR participants(n = 74)*Declined CBSR(n = 27)**Non-CBSR eligible students(n = 703)Total students(n = 804)Mean score on Step 1 first attempt (SD)212.1 (16.7)213.0 (12.3)234.0 (15.7)232.4 (16.1)% passing on first attempt82.9% (58/70)92% (23/25)98.7% (694/703)97.1%(774/797)% On-time graduation49.3% (35/71)68.0% (17/25)90.3% (604/669)85.5%(656/767)*Of the 74 CBSR participants, 3 students withdrew from the medical school; 1 is still enrolled but has been on a leave of absence and has not yet taken Step 1**Of the 27 students who declined CBSR, 2 students withdrew from the medical school. The remaining 25 took Step 1 Table 3Step 1 outcomes and on-time graduation rates of 2019–2021 M2s at university of Washington by CBSR statusVariableCBSR participants(n = 74)*Declined CBSR(n = 27)**Non-CBSR eligible students(n = 703)Total students(n = 804)Mean score on Step 1 first attempt (SD)212.1 (16.7)213.0 (12.3)234.0 (15.7)232.4 (16.1)% passing on first attempt82.9% (58/70)92% (23/25)98.7% (694/703)97.1%(774/797)% On-time graduation49.3% (35/71)68.0% (17/25)90.3% (604/669)85.5%(656/767)*Of the 74 CBSR participants, 3 students withdrew from the medical school; 1 is still enrolled but has been on a leave of absence and has not yet taken Step 1**Of the 27 students who declined CBSR, 2 students withdrew from the medical school. The remaining 25 took Step 1 Step 1 outcomes and on-time graduation rates of 2019–2021 M2s at university of Washington by CBSR status 97.1% (774/797) 85.5% (656/767) *Of the 74 CBSR participants, 3 students withdrew from the medical school; 1 is still enrolled but has been on a leave of absence and has not yet taken Step 1 **Of the 27 students who declined CBSR, 2 students withdrew from the medical school. The remaining 25 took Step 1 Matched outcomes Among the matched pairs, students who enrolled in CBSR (n = 50) had significantly higher mean scores on Step 1 (212.7 vs. 202.1, P < 0.001 by ANCOVA) than academically matched students. Forty-one out of fifty (82.0%) of matched CBSR students passed Step 1 on the first attempt, compared to 37/50 (74.0%) of controls (P = 0.33 by Chi-Square). To determine the impact of CBSR on graduation timing when compared to similar students who did not enroll in CBSR, we compared time to graduation for the matched pairs. One student was excluded from analysis due to enrollment in a concurrent degree program. Within the matched group, students who enrolled in CBSR had a 65.3% (n = 32/49) graduation rate within 4 years, compared to 60.0% (n = 30/50) for those who did not enroll in CBSR (P = 0.74 by Chi-square) (Table 4).Table 4Comparison of step 1 outcomes and on-time graduation rates of propensity-matched comprehensive basic science review (CBSR) and non-CBSR studentsVariableCBSR students(n = 50)Matched non-CBSR students (n = 50)P-valuesStep 1 score212.7202.1P < 0.001 by ANCOVAStep 1 pass rate first attempt41/50 (82.0%)37/50 (74.0%)P = 0.33 by chi-squareOn-time graduation rate65.3% (n = 32/49)60.0% (30/50)P = 0.74 by chi-square Table 4Comparison of step 1 outcomes and on-time graduation rates of propensity-matched comprehensive basic science review (CBSR) and non-CBSR studentsVariableCBSR students(n = 50)Matched non-CBSR students (n = 50)P-valuesStep 1 score212.7202.1P < 0.001 by ANCOVAStep 1 pass rate first attempt41/50 (82.0%)37/50 (74.0%)P = 0.33 by chi-squareOn-time graduation rate65.3% (n = 32/49)60.0% (30/50)P = 0.74 by chi-square Comparison of step 1 outcomes and on-time graduation rates of propensity-matched comprehensive basic science review (CBSR) and non-CBSR students Discussion Our findings demonstrate that enrollment in a proactive, targeted, and voluntary curricular intervention for students at risk of Step 1 failure is associated with significantly improved Step 1 performance, indicated by a 10-point higher mean Step 1 score, when compared to academically matched students. Additionally, this intervention was not associated with delayed graduation. To our knowledge, this is the first description of a formalized curricular intervention targeted to high-risk students which demonstrates a measurable improvement in Step 1 outcomes. This work aligns with broader trends towards competency-based, time-variable medical education and precision education [17]. Our risk identification strategy, based on internal and external academic predictors, appears effective: only 1.3% of non-invited students failed Step 1 on their first attempt. This suggests the criteria were appropriately calibrated to identify students most at risk. Among those who were invited, the minority of students who declined CBSR participation had lower rates of Step 1 failure than those who enrolled. This group had some slight trends towards stronger academic indicators, but it is most likely that those who chose to decline CBSR also differed in unmeasured ways that influenced both their risk of failure and their willingness to participate in an optional academic support program. These potential confounders – such as confidence and perceived need – were a key reason we decided a priori to use propensity score matching to evaluate CBSR effectiveness. It is also possible that the act of being identified as high-risk carries an intrinsic benefit. That CBSR participants had significantly higher Step 1 scores than their academically matched controls suggests that the program is effective for those who choose to participate. As expected, the higher average Step 1 scores among CBSR students compared to their matched controls were also associated with a higher passing rate. However, due to sample size limitations, the study was not powered to detect statistically significant differences in pass rates; detecting a true 5% difference would require hundreds of students in each group. Still, the combination of significantly higher mean scores and a trend toward an 8% higher pass rate suggests that the intervention may continue to offer meaningful benefit even as Step 1 transitions to pass/fail scoring. Limitations Although our study has several strengths, it is not without important limitations. While propensity-matching is a robust comparative technique, it cannot account for potential motivational and other non-academic differences between the students who voluntarily chose to participate in the CBSR program and their matched counterparts who declined or were not invited. A randomized trial would be challenging to conduct but would provide stronger evidence of the positive effect of CBSR itself on Step 1 outcomes and could be a valuable target for future study. Additionally, while students were matched academically using Z-scores and percentiles to ensure cohort compatibility and the curriculum remained stable during this timeframe, students could not be matched within cohort due to limited sample sizes. This analysis also predates the transition to pass/fail Step 1 scoring and the increased passing threshold. However, the substance of the exam has not changed, only the granularity of the reported outcome, and we anticipate that we will see this success continued once we have accrued sufficient data for analysis in the pass/fail era. Future directions The Step 1 literature is poised to evolve from risk identification to risk modification. There remains much to learn about optimal intervention timing and content for students at risk of Step 1 failure, including how to intervene early without being overly deterministic. Expanding CBSR-style competencies to all students, potentially as early as the end of the first year of medical school, could provide additional benefit. Moreover, as some students who fail Step 1 were not identified by pre-dedicated period metrics, future studies should explore targeted interventions within the dedicated study period to better support these students. Conclusions Academically vulnerable students who enrolled in a proactive, extended, remote, and peer-focused dedicated Step 1 study program had significantly higher Step 1 scores compared to academically matched peers. Notably, participation in this extended study program was not associated with delayed graduation, suggesting that targeted academic support can enhance exam performance without negatively impacting time to graduation.