Table 2.
Logistic Regression Models Testing the Association Between Race/Ethnicity (Non-URM vs URM) and Each Potential Categorical Mediator Among U.S. Medical School Graduates in 1997—2004 Who Planned Research-related Careers (N =27,521)a
| Potential categorical mediators | Adjusted odds ratio (95% confidence interval)b |
|---|---|
| Research-intensive medical school (yes vs. no) | 1.23 (1.14 – 1.32) |
| Medical school research elective (yes vs. no) | 1.12 (1.03 – 1.21) |
| Medical school authorship (yes vs. no) | 1.41 (1.31 – 1.52) |
| Degree program (each vs. MD degree) | |
| MD-other advanced degree | 0.85 (0.65 – 1.11) |
| MD-PhD | 1.55 (1.22 – 1.97) |
| Total debt (each vs. no debt) | |
| ≥ $100,000 | 0.42 (0.37 – 0.48) |
| $50,000 – $99,999 | 0.38 (0.34 – 0.44) |
| $1 – $49,999 | 0.47 (0.40 – 0.54) |
| Specialty choice (each vs. internal medicine) | |
| Family medicine | 0.49 (0.39 – 0.61) |
| Pediatrics | 0.94 (0.82 – 1.08) |
| Obstetrics-gynecology | 0.52 (0.44 – 0.60) |
| No/undecided about board certification | 0.82 (0.70 – 0.97) |
| Surgery specialties | 0.73 (0.66 – 0.82) |
| All other specialties | 0.88 (0.80 – 0.98) |
| ≥ 1 GME research year (yes vs. no) | 1.30 (1.19 – 1.43) |
| F32 award (yes vs. no) | 1.94 (1.23 – 3.06) |
Abbreviations: URM indicates underrepresented minorities in medicine; GME, graduate medical education.
This table provides the adjusted odds ratio data for the relationship described in Path A of Figure 1 between race/ethnicity and each potential mediator.
Each model is adjusted for sex and graduation year. An adjusted odds ratio (aOR) < 1.00 indicates, for example, that non-URM (vs. URM) graduates were less likely to report ≥ $100,000 total debt at graduation (vs. no debt). An aOR > 1.00 indicates, for example, that non-URM (vs. URM) graduates were more likely to have attended a research-intensive medical school. If 95% confidence intervals include the value 1.00, the associated aORs are not statistically significant.