Robert C. Bowman, M.D.
Correlations in US medical schools compare the characteristics most likely to graduate more physician researchers. Be cautioned that research graduates are skewed toward 15 medical schools in the nation where half of the nation's medical researchers hail from. Attempts to graduate physician researchers are driven by NIH grants, which continue to gain strength in all medical schools as a source of funding. NIH grants are extremely competitive and schools seem to be willing to risk more and more to "move up" in the funding. Some colleges have been disciplined because they have diverted student tuition and other college funding inappropriately to research.
The characteristics related to graduation of more physician researchers include NIH grant dollars, higher MCAT score at a school, longitude (eastern vs western), and more urban born students at a school. States with better education resources and distribution also tend to graduate more physician researchers (east and west). The Midwest and the south do not tend to be sources of physician researchers.
When assessing medical schools research and younger students go together and are opposite from choice of family medicine and older student characteristics. However in private allopathic schools, those choosing physician research career paths are also older. Family medicine and research also share a common base with education. When this area is thoroughly addressed, states graduate more family physicians and more researchers.
Feeder programs attempting to recruit students to medical school commonly involve research at all levels from middle school to admission. Medical schools are often frustrated with the lack of graduation of physician researchers, but this may be because they are admitting a narrower group of students. Elite allopathic private medical schools are a uniform spike of 25-27 year olds with similar interests, backgrounds, schools, scores, and grades. They are the ones who have been working all of their lives in the academic arena with a focus on intellectual skills in a narrow area. Good research flows out of the ability to recognize patterns, similarities, and differences. Good research requires borrowing from other disciplines, sometimes quite unrelated to your own.
Our training process is unlikely to involve any sort of "broadening." The major risk of the research today is that it will be more and more irrelevant, expensive, and unlikely to benefit the nation. Each new drug costs $60,000 to $80,000 per capita for each person who will use that drug. Unless you are a very rich American, most would be more than happy to take that $60,000 check and use it and die a few months sooner.
Weighted linear regression involving allopathic medical schools
Dependent is % of medical school graduates from a school who are researchers in AMA Masterfile
You can even predict research physician outcomes Before Admissions
For more researchers in the nation:
Just need higher MCAT scores, more eastern medical schools, more private medical schools, and younger schools
| Descriptive Statistics | ||||||
| Mean | Std. Deviation | N | ||||
| RESEARCH | 0.434682267 | 2.298045469 | 109 | |||
| MCATALL | 9.381263946 | 3.580524779 | 109 | |||
| LONGIT | -88.929439 | 61.69824605 | 109 | |||
| NIHAMT | 67124564.24 | 331600060.9 | 109 | |||
| PUBLIC | 0.719477208 | 2.165623582 | 109 | |||
| AGESCHOO | 110.2334237 | 243.5675559 | 109 | |||
| Correlations | RESEARCH | MCATALL | LONGIT | NIHAMT | PUBLIC | AGESCHOO |
| RESEARCH | 1 | 0.667630287 | 0.202381332 | 0.666280027 | -0.52648564 | 0.050501859 |
| MCATALL | 0.667630287 | 1 | 0.01926247 | 0.662381543 | -0.48110963 | 0.180240883 |
| LONGIT | 0.202381332 | 0.01926247 | 1 | 0.098501307 | -0.22423538 | 0.38474912 |
| NIHAMT | 0.666280027 | 0.662381543 | 0.098501307 | 1 | -0.31128174 | 0.280103977 |
| PUBLIC | -0.52648564 | -0.48110963 | -0.22423538 | -0.31128174 | 1 | -0.20041595 |
| AGESCHOO | 0.050501859 | 0.180240883 | 0.38474912 | 0.280103977 | -0.20041595 | 1 |
| Model Summary | ||||||
| Model | ||||||
| R | R Square | Adjusted R Square | Std. Error of the Estimate | Change Statistics | ||
| R Square Change | F Change | df1 | ||||
| 0.807421 | 0.651928493 | 0.635031818 | 1.388308937 | 0.651928493 | 38.5832414 | 5 |
| ANOVA | ||||||
| Model | Sum of Squares | df | Mean Square | F | Sig. | |
| Regression | 371.827026 | 5 | 74.36540519 | 38.5832414 | 3.83282E-22 | |
| Residual | 198.5223755 | 103 | 1.927401703 | |||
| Total | 570.3494014 | 108 | ||||
| Unstandardized Coefficients | Standardized Coefficients | t | Sig. | 95% Confidence Interval for B | ||
| B | Std. Error | Beta | Lower Bound | |||
| (Constant) | -0.43446752 | 0.539456507 | -0.80538007 | 0.422455725 | -1.50435227 | |
| MCATALL | 0.187296312 | 0.05465865 | 0.291821505 | 3.426654535 | 0.000879162 | 0.078893771 |
| LONGIT | 0.007178996 | 0.002406496 | 0.192742683 | 2.983174066 | 0.003561971 | 0.002406279 |
| NIHAMT | 3.08594E-09 | 5.52E-10 | 0.445289903 | 5.590465628 | 1.86744E-07 | 1.99117E-09 |
| PUBLIC | -0.27034877 | 0.072733351 | -0.25477027 | -3.71698488 | 0.00032813 | -0.41459821 |
| AGESCHOO | -0.002378 | 0.000616872 | -0.25204138 | -3.85492787 | 0.000201709 | -0.00360141 |
Dependent Variable: RESEARCH
Weighted Least Squares Regression - Weighted by RURAL92