Predicting performance using background characteristics of international medical graduates in an inner-city university-affiliated Internal Medicine residency training program

BACKGROUND IMGs constitute about a third of the United States (US) internal medicine graduates. US residency training programs face challenges in selection of IMGs with varied background features. However data on this topic is limited. We analyzed whether any pre-selection characteristics of IMG residents in our internal medicine program are associated with selected outcomes, namely competency based evaluation, examination performance and success in acquiring fellowship positions after graduation. METHODS We conducted a retrospective study of 51 IMGs at our ACGME accredited teaching institution between 2004 and 2007. Background resident features namely age, gender, self-reported ethnicity, time between medical school graduation to residency (pre-hire time), USMLE step I & II clinical skills scores, pre-GME clinical experience, US externship and interest in pursuing fellowship after graduation expressed in their personal statements were noted. Data on competency-based evaluations, in-service exam scores, research presentation and publications, fellowship pursuance were collected. There were no fellowships offered in our hospital in this study period. Background features were compared between resident groups according to following outcomes: (a) annual aggregate graduate PGY-level specific competency-based evaluation (CBE) score above versus below the median score within our program (scoring scale of 1 - 10), (b) US graduate PGY-level specific resident in-training exam (ITE) score higher versus lower than the median score, and (c) those who succeeded to secure a fellowship within the study period. Using appropriate statistical tests & adjusted regression analysis, odds ratio with 95% confidence intervals were calculated. RESULTS 94% of the study sample were IMGs; median age was 35 years (Inter-Quartile range 25th - 75th percentile (IQR): 33-37 years); 43% women and 59% were Asian physicians. The median pre-hire time was 5 years (IQR: 4-7 years) and USMLE step I & step II clinical skills scores were 85 (IQR: 80-88) & 82 (IQR: 79-87) respectively. The median aggregate CBE scores during training were: PG1 5.8 (IQR: 5.6-6.3); PG2 6.3 (IQR 6-6.8) & PG3 6.7 (IQR: 6.7 - 7.1). 25% of our residents scored consistently above US national median ITE scores in all 3 years of training and 16% pursued a fellowship.Younger residents had higher aggregate annual CBE score than the program median (p < 0.05). Higher USMLE scores were associated with higher than US median ITE scores, reflecting exam-taking skills. Success in acquiring a fellowship was associated with consistent fellowship interest (p < 0.05) and research publications or presentations (p <0.05). None of the other characteristics including visa status were associated with the outcomes. CONCLUSION Background IMG features namely, age and USMLE scores predict performance evaluation and in-training examination scores during residency training. In addition enhanced research activities during residency training could facilitate fellowship goals among interested IMGs.