Myles, Thomas D. MD; Henderson, Robert C. II MD
Department of Obstetrics and Gynecology, Texas Tech Health Sciences Center, Amarillo, Texas.
Address reprint requests to: Thomas D. Myles, MD, St. Louis University, 6420 Clayton Road, Suite 559, St. Louis, MO 63117; E‐mail: email@example.com.
Received November 15, 2001. Received in revised form April 10, 2002. Accepted April 18, 2002.
OBJECTIVE: We studied the correlations between both Step 1 and Step 2 of the United States Medical Licensure Examination (USMLE) and the National Board of Medical Examiners Obstetrics and Gynecology Examination (NBME‐OB/GYN).
METHODS: From July 1994 until June 2001, all third‐year medical students at Texas Tech University Health Center at Amarillo were studied. The scores from the first attempts for the three examinations were obtained. We investigated for correlations between the examination scores and the score extremes.
RESULTS: A total of 258 students were evaluated. The mean USMLE Step 1 score was 203.7; mean score for Step 2 was 205.0. The NBME‐OB/GYN final examination was 86.0. We found linear correlations between the NBME‐OB/GYN and the USMLE Step 1 (r = .463, P < .001) and USMLE Step 2 (r = .595, P < .001), as well as between the USMLE Steps 1 and 2 (r = .666 P < .001). Students failing the USMLE Step 1 were more likely to fail their USMLE Step 2 (relative risk [RR] 9.3 [95% confidence interval (CI) 2.1, 41.0]) and/or the NBME‐OB/GYN (RR 3.3 [95% CI 1.03, 10.8]). Students scoring in the lowest 25th percentile of the NBME‐OB/GYN were more likely to fail the USMLE Step 2.
CONCLUSION: Both USMLE Steps 1 and 2 correlated with the NBME‐OB/GYN scores, as well as with each other. Students failing either the USMLE Step 1 or NBME‐OB/ GYN were more likely to fail the USMLE Step 2. The use of this information could predict students at risk for low scores on future examinations.
The objective grade for most obstetric and gynecology clerkships is determined most commonly by a written examination. In spite of the efforts of clinical faculty, some third‐year medical students continue to fail this examination. Limited data is published that can assist in the early detection of the at‐risk student. With respect to the clerkship itself, Riggs and Blanco1 reported that poor lecture attendance is associated with lower test scores. Armstrong and colleagues2 and Myles3 each noted relationships between poor United States Medical Licensure Examination (USMLE) Step 1 performance and poor performance on the National Board of Medical Examiners Obstetrics and Gynecology Examination (NBME‐OB/GYN). Each also noted a linear correlation between the USMLE Step 1 score and the NBME‐OB/GYN score. Armstrong's group also noted that a low grade point average from the first 2 years of medical school could predict poor performances on the NBME‐OB/ GYN.
Conversely, the ability to predict the student at risk of failure of the USMLE Step 2 is also of importance. Spellacy and Dockery4 reported a correlation between NBME Part 2 and NBME‐OB/GYN. Campos‐Outcalt and colleagues5 reported correlations between the Family Medicine Final Examination and both Part 1 and Part 2 of the NBME.
The primary objective of this study was to confirm the association between USMLE Step 1 and the NBME‐OB/ GYN for our facility, as well as to determine if a similar relationship existed between the USMLE Step 2 and the NBME‐OB/GYN. To determine the students at risk of failure, further analysis was also completed. We also evaluated for possible relationships between the USMLE Step 1 and Step 2 scores.
A MEDLINE search from 1966 to March 2002 using the key words of “National Board Examination,” “United States Medical Licensure Examination,” “OB/GYN Final Examination,” and “Clerkship Examination” was executed. Only the studies previously cited were identified. No studies evaluating the USMLE Step 2 and any clerkship examination were found, and no comparisons between Step 1 and Step 2 of the USMLE were found.
MATERIALS AND METHODS
The NBME‐OB/GYN scores and the USMLE Step 1 and Step 2 scores of 258 Texas Tech University School of Medicine at Amarillo students (1994–2001) were obtained from the Registrar's Office of Texas Tech University School of Medicine. Only first attempts at each of the three examinations were included in the analysis. Step 1 of the USMLE was taken upon completion of the second year. As with the rest of the country, these examinations were taken using a computer beginning in June 1999. At Texas Tech, students must pass Step 1 to advance to third‐year clinical rotations. Although passage of Step 2 is not a requirement for advancement or graduating, all students must take the examination. Step 2 is taken after completion of the third‐year requirements. All students take the NBME‐OB/GYN on the final Friday of a 6‐week clerkship.
Texas Tech University School of Medicine assigns students to three campuses for third‐ or fourth‐year clinical rotations. Only the Amarillo campus was studied. Lectures and case discussions are given by the faculty over the 6‐week rotation to cover the educational objectives outlined by the Association for Professor of Gynecology and Obstetrics.6 The campus in Amarillo includes two hospitals. The obstetric services are at one hospital, whereas gynecologic services are at both hospitals through which the students rotate. All students participate in activities at both hospitals.
Associations between test scores were tested using regression analysis. Evaluations of students (on the Amarillo campus) who either failed the first attempt, scored in the lowest tenth percentile, the lowest quartile, upper quartile, or finished in the 90th percentile on their first attempt of each of the examinations were also derived using correlation coefficients (Pearson), odds ratios, Student t test, and χ2 analysis. Sequentially pertinent data will be presented. Data was analyzed using SPSS 9.0 (SPSS Inc., Chicago, IL) Significance was set at P < .05.
The mean score (± standard deviation) for the USMLE Step 1 was 203.7 ± 19.1. The mean score for the NBME‐OB/GYN was 86.6 ± 7.1. The mean score for the USMLE Step 2 was 205.0 ± 20.7 (Table 1). Linear fits were observed between the NBME‐OB/GYN and the USMLE Step 1 score (r = .463, P < .001) and the USMLE Step 2 score (r = .595, P < .001). A linear correlation was also seen between the USMLE Step 1 and Step 2 (r = .666, P < .001).
Eighteen students (7.0%) failed their first attempt at the USMLE Step 1. Eleven (5.0%) of the students failed the USMLE Step 2, and 15 (5.5%) of the students failed the NBME‐OB/GYN. One student did not take either the NBME‐OB/GYN or the USMLE Step 2 because of academic difficulties.
Students failing the USMLE Step 1 were at an increased risk of failing the USMLE Step 2 (relative risk [RR] 9.3, [95% confidence interval (CI) 2.1, 41.0]) (Table 2). These students were also more likely both to have scores in the lowest tenth percentile of the USMLE Step 2 and to score in the lowest 25th percentile (Table 2). When evaluating the students failing the USMLE Step 1, we found an increased risk for failing the NBME‐OB/ GYN. These students were also more likely to score in the lowest 25th percentile on that examination. Students scoring either in the lowest tenth or lowest quartile on USMLE Step 1 were also more likely to score similarly on both the NBME‐OB/GYN and USMLE Step 2. Several of the comparisons of the test scores (USMLE Step 2 and NBME‐OB/GYN) at far extremes from the USMLE Step 1 score extremes showed significant differences (all P < .01) (Table 3). As with the USMLE Step 1, comparisons of the USMLE Step 2 scores at extreme variance with the USMLE Step 1 and NBME‐OB/GYN scores showed significant differences (Table 3).
Equally important, students who failed their NBME‐OB/GYN were at a increased risk of scoring below the 25th percentile on their first attempt of the USMLE Step 2. Students scoring in the lowest quartile on the NBME‐OB/GYN either failed, scored below the tenth percentile, or finished in the lowest quartile on the USMLE Step 2 (Table 4).
With respect to the students who obtained a Step 1 score in the 90th percentile, none finished in the lowest quartile of (or failed) Step 2. Nor did any fail or finish in the lowest quartile on the NBME‐OB/GYN. None of the students who had a NBME‐OB/GYN score in the top 10% failed or finished in the lowest quartile on USMLE Step 2. Students scoring in the 90th percentile on the USMLE Step 1 were more likely to achieve similarly high scores on both the NBME‐OB/GYN (RR 4.4, [95% CI 1.7, 11.4]) and the USMLE Step 2 (RR 12.3, [95% CI 4.7, 36.6]). A student scoring in the 90th percentile on the NBME‐OB/GYN was also likely to score well on the USMLE Step 2 (RR 15.0 [95% CI 5.8, 38.8]). In fact, none of the students who finish in the highest 25th percentile (217) on the USMLE Step 1 finished in the lowest 25th percentile on Step 2. Five students finished in the lowest quartile on the NBME‐OB/GYN, but none finished below the tenth percentile. Additionally, none of the students achieving a NBME‐OB/GYN score in the highest 25th percentile (91) failed or obtained a USMLE Step 2 score in the lowest tenth percentile (two were in the lowest quartile).
The findings of this study confirm those of both Armstrong and colleagues2 and Myles,3 that a linear relationship between the USMLE Step 1 and the NBME‐OB/ GYN exists. Our finding of a relationship between the NBME‐OB/GYN and the USMLE Step 2 had not been reported previously, though the findings are similar to those of Spellacy and Dockery.4 They noted correlations between the NBME‐OB/GYN and the National Board of Medical Examiners, Part II.
To the best of our knowledge, this is the first study to directly compare USMLE Step 1 and Step 2. A MEDLINE review failed to reveal any additional studies making such comparisons. Our finding of a strong linear relationship between the two examinations was not a complete surprise, because Spellacy reported a correlation between both Part I and Part II of the National Board of Medical Examiners and the Council on Residents Education in Obstetrics and Gynecology (CREOG) examination scores of obstetrics and gynecology residents.7 The nonsecure nature of the CREOG examination during the period studied by Spellacy probably does not significantly affect the results. Similarly, Sosenko and colleagues8 noted a correlation between scores on the National Board of Medical Examiners and the American Board of Internal Medicine Certification Examination.
Our findings validate these older studies as to the continued usefulness of the licensure examination scores in assessing student performance and academic potential. The correlations observed in this study, as well as by others,2–4 indicate their importance for identifying the student at risk. Early intervention through mentoring, tutoring, and additional assistance could be implemented as a result of lower test scores on both the USMLE Step 1 and the NBME‐OB/GYN. Additional support and tutoring for the at‐risk student on all third‐year clerkships seem warranted. Any student finishing in the lowest quartile of either the USMLE Step 1 or the NBME‐OB/GYN should be considered to be at risk for poor performance on future examinations. In our study, 88.9% of the students who failed the USMLE Step 2 fell into one or both of these categories. Additional education interventions, such as special tutoring, small group review, and prescribed self‐study could be of benefit to these students. Specific studies of various interventions are warranted in view of these findings.
1. Riggs JW, Blanco JD. Is there a relation between student lecture attendance and clinical science subject examination score? Obstet Gynecol 1994;84:311–3.
2. Armstrong A, Dahl C, Haffner W. Predictors of performance on the national board of medical examiners obstetrics and gynecology subject examination. Obstet Gynecol 1998;91:1021–2.
3. Myles TD. United States Medical Licensure Examination step 1 scores and obstetrics-gynecology clerkship final examination. Obstet Gynecol 1999;94:1049–51.
4. Spellacy W, Dockery JL. A comparison of medical student performance on the obstetrics and gynecology National Board Part II Examination and a comparable examination given during the clerkship. J Reprod Med 1980;24:76–8.
5. Campos-Outcalt D, Witzke DB, Fulginiti JV. Correlations of family medicine clerkship evaluations with scores on standard measures of academic achievement. Fam Med 1994;26:85–8.
6. Association for Professors of Gynecology and Obstetrics. Instructional objective for a curriculum in obstetrics and gynecology. 7th ed. Washington, DC: Association for Professors of Gynecology and Obstetrics, 1997.
7. Spellacy WN. The use of national board scores in selecting residents for obstetrics and gynecology. Am J Obstet Gynecol 1985;153:605–7.
8. Sosenko J, Stekel KW, Soto R, Gelbard M. NBME Examination Part I as a predictor of clinical and ABIM certifying examination performances. J Gen Intern Med 1993;8:86–8.