Relationship between past academic performance and results of specialty in-training examinations

In this study, the authors review the records of 63 graduates of Northwestern University Medical School who were residents in its graduate medical education programs of anesthesia and orthopedic surgery. They examine the relationship among college grades, medical school performance, and the results of assessment by annual, nationwide, medical specialty in-training examinations. For the anesthesia group, the best predictors of in-training examination performance were the Medical College Admission Test (MCAT) Verbal Ability score, the college grade-point average for nonscience subjects, and the MCAT Science, General Information, and Quantitative Ability scores. For the orthopedic group, the best predictors were the MCAT Verbal Ability score, the college grade-point average in nonscience subjects, the MCAT Science score, and the National Board of Medical Examiners Part I and Part II examination scores. The previous academic records for the 63 residents contained little to presage results in the in-training examination. The correlation obtained between nonscience college subjects and the in-training examination results was negative.     

Traditional criteria as predictors of minority student success in medical school.

There have been many reports stating that the traditional criteria of the Medical College Admission Test (MCAT) and undergraduate grade-point average (GPA) have little, if any, value in predicting success in the preclinical years of medical school among students from underrepresented (racial and ethnic) minority groups. In contrast to previous articles this report emphasizes that traditional criteria and the quality of the undergraduate college attended are of some statistical value in predicting success in the preclinical years of medical school among accepted students from under represented minority groups. Of these criteria, the one with the greatest predictive value is the selectivity of the undergraduate college attended.     

Predictors of success in an anesthesiology residency

The selection of residents in medical specialty programs is a difficult task facing all selection committees. The present authors examined factors that contribute to successful residency performance by 26 anesthesiology residents in order to assist the program's selection committee in developing selection criteria. The best predictor of a resident's academic average in the anesthesiology program was the number of years the resident had spent in other specialties. Comparison of the residents' scores on the Anesthesiology In-Training Examination (ITE) and their scores on Part I of the examinations of the National Board of Medical Examiners (NBME) showed a statistically significant negative correlation. The higher that residents scored on the NBME Part I examination, the lower they scored on the ITE. No significant correlations were found between the residents' ITE scores and undergraduate grade-point average or nonacademic variables such as the residents' age or parents' level of education. The residents' composite grade-point average (GPA) for the residency and their interview scores had a strong positive relationship significant at the 0.06 level.     

Predictive validity of the MCAT for students with two sets of scores

Addressed in the study reported here is the question of which set of scores for those students who retake the Medical College Admission Test (MCAT) yields a better predictive validity. The sample was comprised of 304 students who retook the MCAT prior to entering Jefferson Medical College between 1978 and 1981. Five sets of MCAT scores were considered as predictors in the study: earlier, later, higher, and lower sets of MCAT scores and the average of the earlier and later scores for each MCAT subtest. Twenty-five criteria were used, including grades earned in the freshman and sophomore years and scores on the subtests of Part I and Part II of the examinations of the National Board of Medical Examiners. Correlational techniques, such as bivariate and multiple correlation analyses and canonical correlation followed by redundancy analysis were utilized. The magnitude of redundancy indices indicated that the set of MCAT scores in which the earlier and later scores were averaged was the best predictor, followed by the earlier, lower, higher, and later sets of MCAT scores. The implications of these findings for the admissions process and for validity studies are discussed.     

Using ratings of resident competence to evaluate NBME examination passing standards

The passing standards of the National Board of Medical Examiners (NBME) examinations were empirically evaluated by analyzing the distributions of examination scores received by 1,994 graduates of one medical school in relation to the clinical competence ratings given to the graduates by their first-year residency directors. A significant association was found between the NBME examination scores and ratings in the cognitive areas of clinical competence. The graduates who scored 420 or less on the NBME Part I or the Part II examination received significantly lower medical knowledge ratings than did the total group of graduates. A similar analysis of NBME Part III examination scores was less clear-cut but also suggested that a score of 420 or less could identify those graduates at significant risk of receiving low knowledge ratings. When low ratings were used as an outcome measure, analysis showed that the NBME Part II examination was not sensitive in detecting such graduates. Based on these data, the authors do not propose changes in the passing standards for the NBME examinations but recommend that these standards continue to be reassessed and further measures be taken to strengthen the internal evaluation methods in medical schools.     

Preclinical curriculum characteristics and institutional performance on NBME. Part I

A study was conducted that analyzed the relationship between various characteristics of the preclinical curriculum and institutional performance on the Part I examination of the National Board of Medical Examiners (NBME) at a sample of 85 U.S. medical schools. Total scheduled hours per week was the single curriculum characteristic having a positive and significant relationship with institutional NBME examination performance. However, when the data were controlled by medical school selectivity in admissions and institutional policy on the taking of the examination, total scheduled hours per week failed to make a significant contribution to the prediction of performance. The results were viewed as failing to provide support and justification for intense preclinical curriculum loads on the basis that this would enhance NBME examination performance. The authors conclude that simple comparisons of schools on the basis of mean NBME examination performance are meaningless unless the entering abilities of students and school policies on the examination are taken into account. The authors also suggest that those schools that prescribe heavy and intense preclinical curriculum loads should reexamine those policies in light of recommendations of the Project Panel on the General Professional Education of the Physician and College Preparation for Medicine.     

Student performance on the NBME Part II subtest and subject examination in obstetrics-gynecology

Clerkship directors in obstetrics-gynecology often use the National Board of Medical Examiners (NBME) norms to evaluate third-year medical students' performance on the NBME obstetrics-gynecology subject examination. A comparison of the scores of 342 students at the Medical College of Georgia School of Medicine showed that the students performed significantly better on the NBME subject examination than on the Part II subtest in obstetrics-gynecology. These results concur with the findings of the NBME, which advises directors wishing to adjust the criterion group norms to determine the average difference observed in a school's performance on the two examinations and to use that difference or some portion of it in their interpretation of percentile scores on the subject examination. Additional analyses revealed that a single, linear weight may inappropriately adjust these scores, that student performance on the Part II subtest depends on specialty choice (obstetrics-gynecology versus all others), and that time and sequence of the clerkship were unrelated to the students' performance on the two examinations.     

A model for predicting performance on the NBME Part I examination

Multiple regression analyses were employed to determine the relationships between achievement variables as predictors of the performance on Part I of the National Board of Medical Examiners examinations (NBME-I) of 366 medical students. The students' examination score averages in first- and second-year basic science courses were sequentially added to a composite Medical College Admission Test (MCAT) score to investigate the increases in prediction accuracy. Four prediction equations were cross validated on two subsequent medical school classes. While the multiple correlation between the predicted NBME-I score and the actual score significantly increased as first- and second-year course examination scores were added to the equations, the overall accuracy in predicting passing or failing did not substantially increase. The most useful equation employed a combination of the MCAT score and the first semester anatomy course examination score. The NBME-I predictions were given by letter to the 1984 and 1985 students approximately a year before they were to take the NBME-I examination so that students at risk of failure could undertake remedial study.     

Use of MCAT data in selecting students for admission to medical school

In the spring of 1986, medical school admissions personnel were surveyed on their institutions' admissions practices and the use of Medical College Admission Test (MCAT) data in student selection. The admissions officers listed sources of information considered in processing applications. The variables accorded high importance were: overall and science undergraduate grade-point averages, quality of degree-granting institutions, letters of evaluation, interview ratings, MCAT scores, extracurricular activities, work in areas related to health care, breadth and/or difficulty of course work, and state of legal residence. Variables judged of medium importance were: nonscience grade-point average, graduate study, narratives supplied through the American Medical College Application Service or supplemental narratives, demographic factors, and undergraduate research. The respondents divided the 15-point MCAT scale into exemplary, acceptable, and unacceptable ranges of performance. These responses varied widely among institutions. The mean response for the bottom of the acceptable range was a score of seven, and the mean for the beginning of the exemplary range was 11. In considering individual MCAT subtest scores, 34 percent of the admissions officers regarded the six scores individually and equally, 51 percent weighted the areas in ways related to their curricula, and 43 percent summed or averaged scores with equal weights at some point. Thirty-one percent used MCAT scores to adjust grade-point averages across undergraduate institutions.     

The effect of commercial coaching courses on performance on the MCAT

Two studies on the relationship between taking a commercial coaching course and performance on the Medical College Admission Test (MCAT) are reported. In both studies, one based on repeating examinees and another based on first-time examinees, differences in scores favoring coached examinees were found on the Science Knowledge and Science Problems subtests. The results on the Skills Analysis: Quantitative subtest were mixed, while neither study found differences between the two groups in scores on the Skills Analysis: Reading subtest. The analyses further revealed that the differences in scores on the Science Knowledge and Science Problems subtests were less pronounced for examinees with low scores on the Skills Analysis: Reading and Skills Analysis: Quantitative subtests, examinees with very low or very high undergraduate grade-point averages, and examinees from very selective undergraduate colleges. The substantive importance of the differences was also explored. The coaching effect was estimated to increase an average student's probability of acceptance by 5 percentage points. The results suggest that coaching for the MCAT has a limited effect that supports rather than undermines the validity of the test.     

Use of a medical reasoning aptitude test to help predict performance in medical school

A medical reasoning aptitude test (MRAT) was designed to assess aptitude for clinical problem-solving in medical school applicants. The purpose of the study reported here was to determine whether the information provided by this test, when used in conjunction with college grade-point averages (GPAs) and scores on the Medical College Admission Test (MCAT), would improve the prediction of medical school performance. Specifically, the authors investigated the incremental predictive value of the MRAT relative to students' overall performance in medical school and, more specifically, to their knowledge and clinical reasoning during preclinical years and clinical performance during the clerkship year. Overall, it was found that, with the exception of the first year, the addition of MRAT scores to the GPAs and MCAT scores increased precision in identifying students who performed poorly or exceptionally well in the second year and the clinical clerkship year. This last finding is especially useful, since few other tests have provided that information. On the basis of the preliminary findings, the authors propose further use and validation of the MRAT.     

Correlations between SAT scores and MCAT scores of black students in a summer program

The purpose of the study reported here was to determine the relationship between Scholastic Aptitude Test (SAT) scores and Medical College Admission Test (MCAT) scores of black students who had participated in a summer program for minority students interested in health careers held at the Medical College of Georgia. A significant correlation was established between these students' SAT and MCAT scores. These findings suggest that students who are likely to perform well in medical school can be selected for medical school while they are still underclassmen, accepted conditionally, and admitted after completing a prescribed curriculum. Early selection has implications for increasing the number of minority students in medicine and for permitting more liberal arts in the premedical education than the present selection process.     

Performance on NBME Part I examination in relation to policies regarding use of test

From 1971 to 1983, a number of administrative decisions were made at the University of Oklahoma College of Medicine regarding the National Board of Medical Examiners (NBME) Part I examination. Students' performance on this examination was found to be associated with administrative decisions that required (a) a passing score for promotion, (b) the student to take the examination, and (c) the student to take an integrated basic sciences review course. A modest improvement in student performance was noted when passage of the examination was made mandatory. The introduction of a review course into the curriculum effected a major upward change in scores that has persisted despite the removal of all requirements to pass, or even take, the Part I NBME examination.     

Undergraduate medical education

The number of applicants to US medical schools, which declined steadily between 1985 and 1988, increased slightly for the class entering in 1989. The profile of entering students showed a small decline from last year in the percentage of students with grade point averages categorized as "A" (3.5 or above on a 4-point scale) and slight declines in four of the six MCAT subtest scores. The percentage of both women and minority students in the entering class increased from the previous year. An interesting observation is the large percentage increase this year in students transferring to LCME medical schools from graduate and professional degree programs and from osteopathic medical schools. While the number of full-time faculty members in medical schools continues to increase, significant vacancy rates exist in some departments. More than 5% of full-time faculty positions are vacant in genetics, pathology, dermatology, family medicine, neurology, obstetrics-gynecology, orthopedics, otolaryngology, pediatrics, and surgery departments. Along with faculty vacancies, there has been a considerable turnover of medical school deans. The curriculum in most medical schools includes some innovative instructional formats, such as problem-based learning and computer-assisted instruction. However, current data do not allow a generalization about the extent to which these are being utilized. It seems that, at least in some institutions, multiple methods are being used to assess the clinical competence of medical students (observation by faculty members and residents, written and oral examinations, and multiple station examinations), including the use of standardized patients. About half of the medical schools require students to pass the NBME Part I examination and about one third require passage of Part II. The subject examinations provided by the NBME seem to be used widely, at least in the clinical disciplines. Within the past year, about 14% of medical schools have reported the presence of students or residents who have been diagnosed with human immunodeficiency virus infection, and 12% have had students or residents diagnosed with hepatitis B virus infection. It is critical that medical schools teach students how to prevent occupational exposure to these infections, as well as ensuring that adequate health insurance coverage be provided for these conditions.     

The use of Part I National Board scores in the selection of residents in ophthalmology and otolaryngology

A survey of ophthalmology and otolaryngology residency program directors was conducted to determine the extent to which National Board of Medical Examiners (NBME) Part I scores are used in selection of residents in these highly competitive specialty fields. Results from 218 completed questionnaires representing nearly 90% of all US ophthalmology and otolaryngology/head and neck surgery programs were analyzed. More than three fourths indicated that Part I NBME scores were used in selecting residents; more than half of the programs used these scores as a means to determine whom to interview. The direct use of NBME scores in the residency application process is widespread; however, such use of NBME scores is not consistent with the purposes of the National Board. The preeminent role of the faculty in the evaluation of medical students and in the assessment of their clinical competence needs emphasis.     

Effect of two preclinical curricula on NBME Part I examination performance

The Michigan State University College of Human Medicine conducts two preclinical medical education programs. In Track I (lecture-based), students attend classes 24 hours per week, and lecture time totals 908 hours over a 50-week period. In Track II (problem-based), students attend classes only 12 hours per week, and lecture time totals 112 hours over the same 50-week period. Institution of the Part I examination of the National Board of Medical Examiners (NBME) as a graduation requirement provided an opportunity to compare the performances of students in both tracks. When students from each track with similar Science Problems subtest scores on the Medical College Admission Test were compared, no significant differences were observed in the students' total scores or pass rate on the NBME examination. However, there were significant differences in scores on the microbiology subtest of the NBME examination, with the Track I students achieving higher scores. The 1984 report of the Project Panel on the General Professional Education of the Physician and College Preparation for Medicine of the Association of American Medical Colleges stressed the need to examine critically and consider reducing the scheduled instructional and lecture hours in preclinical medical education programs. In the study reported here, the authors demonstrated that reduction of scheduled instructional time, when replaced by a guided problem-solving program, is not detrimental to students' performance on the NBME Part I examination.     

Policies of U.S. medical schools on the use of the NBME Part I and Part II examinations

The study reported here provides information on National Board of Medical Examiners (NBME) examination policies from all U.S. medical schools (n = 67) that require students to pass the examination. Most of these schools have similar policies on use of a 380 total score as the passing level for the NBME Part I examination, allowing three attempts to pass the examination, interrupting a student's progress into the clinical curriculum upon failure of the examination, officially providing time for preparation for taking the test, and requiring students to pass the Part II examination.     

Undergraduate medical education

There were 35,944 applicants for the 1984-1985 year in US medical schools. Of this number, 17,194 were accepted by at least one school. First-year enrollment equaled 16,992. Be-accepted by at least one school. First-year enrollment equaled 16,992. Because some students were repeating the first year, the number of first-time enrolled students was 16,459. This represents a decrease of 113 from the previous year. The number of students enrolled in 127 US medical schools in 1984-1985 was 67,090; of this number, 21,287 (31.7%) were women. The estimated number of graduates in 1984-1985 was 16,347. The total enrollment of students of minority groups was 10,552 (15.7%), of which 3,663 (5.5%) were black not of Hispanic origin. The number of full-time medical school faculty was 58,767, another 123,002 were part-time or volunteer faculty. Medical school faculties also have teaching responsibility for a variety of other students, in addition to patient care and research responsibility. The average time needed to complete the curriculum requirements leading to the MD degree is 153 weeks. Twenty-one medical schools offered a combined college-medical school program. The length of these combined programs averaged 262 weeks. Ninety-six percent of students entering medical school in 1984-1985 had completed at least four years of college. More than two fifths of students had a premedical GPA of 3.6 or higher.     

Study of U.S. Medical School Applicants, 1977--78.

This annual report on applicants to the 1977--78 U.S. medical school freshman class describes the first substantial decline in applicants during the past quarter century. Since the decline was accompanied by the opening of several new medical schools, the chances of gaining admission improved to 39 percent, the highest in six years. For applicants making their first attempt the acceptance rate was 43 percent. There was no loss of qualified applicants, however; the 1977--78 pool presented record-high entrance test scores and undergraduate college grades. In the report comparisons are made with activity in 1974--75--the peak year for medical school spplicants--in order to measure the impact of the decline on individual medical schools and geographical regions and on various applicant subpopulations.     

Premedical and medical school performance in predicting first-year residency performance

In the study reported here the authors examined the relationships among 40 measures of undergraduate college and medical school performance and competence in 18 medical care tasks during the first year of residency. A rating form was developed for the study to assess residents' competency in the medical care tasks and was sent to the directors of the residency programs entered by the graduates of a medical school. Stepwise multiple regression procedures were used to analyze the relationship between these ratings of residency performance and the residents' premedical and medical school performance and to identify the best predictors of residency performance for the 1982 and 1984 classes. A Rasch model analysis of the residency performance ratings indicated the ease or difficulty of each of the 18 tasks. The results provide information that would allow medical educators to use premedical and medical school performance to predict residents' competencies. The task of "clinically evaluates research and clinical data" was the most difficult for the graduates; that is, they were rated lower on it than on any other task. Two groups of measures of undergraduate and medical school performance were significantly related to performance in the residency: the Part II examination of the National Board of Medical Examiners (particularly the scores on the obstetrics-gynecology, medicine, surgery, and pediatrics subtests and the overall score) and the clerkships (particularly the third-year medicine clerkship, the fourth-year medicine clerkship, and the surgery clerkship).