I recently took an in-depth look at match lists at a subset of top 10 schools because one disgruntled student from WashU complained on SDN about two things: 1) high inbreeding (where students stay at their home institution for residency) and 2) low numbers matching to “competitive” specialties.
After the cut/jump, I use the 2011 and 2010 match lists from WashU, Harvard, Hopkins, Stanford, Penn, and, for the last question, Columbia to address several questions: 1) what are the rates of inbreeding, 2) what proportion matches at the most competitive specialties, 3) what is the rate of matching to the most reputable internal medicine residencies, and 4) what is the rate of matching to the most reputable surgical residencies?
Question 1: inbreeding
What percentage of all categorical matches (i.e. excluding preliminary and transitional year matches) are to an affiliated hospital of the medical school?
WUMS 2011 - 26% WUMS 2010 - 29% HMS 2011 - 45% HMS 2010 - 37% JHU 2011 - 35% JHU 2010 - 34% Stanford 2011 - 33% Stanford 2010 - 26% Penn 2011 - 38% Penn 2010 - 38%
Conclusion: High end: HMS. Low end: WashU. “Inbreeding” is relatively common at top schools (average roughly one-third).
Question 2: competitive specialties
What percentage of matches were in a specialty considered “most competitive” (>81% of spots nationwide filled) – plastic and reconstructive surgery, orthopedic surgery, otolaryngology / ENT / head and neck surgery, neurological surgery, ophthalmology, diagnostic radiology, and radiation oncology? [Wow. Basically the surgical specialties plus rays, minus urology.] Data was unavailable for Hopkins’ 2011 match.
WUMS 2011 - 26% WUMS 2010 - 30% HMS 2011 - 26% HMS 2010 - 29% JHU 2010 - 21% Stanford 2011 - 31% Stanford 2010 - 33% Penn 2011 - 29% Penn 2010 - 17%
Conclusion: High end: Stanford. Everything else is a mixed bag. Hopkins’ single data point is low. A fifth to a third of graduates at top schools go into the seven most competitive specialties.
For other details and a breakdown by specialty on question 2, see my original post on SDN.
Question 3: internal medicine match
What proportion of students matched went to a top internal medicine program?
I made two lists based on USNWR reputation ratings: Primary (1) list (top 10 internal medicine AND hospital honor roll): Harvard/MGH, Harvard/BWH, Johns Hopkins/Johns Hopkins Hospital, UCSF/UCSF Medical Center, Duke, Penn/HUP, Columbia/NYP, WashU/Barnes-Jewish, UWashington, UMichigan.
Secondary (2) list (ranked [top 30] internal medicine AND hospital honor roll): Yale, Stanford, Cornell/NYP, Mayo, Cleveland Clinic, Mt Sinai, Vanderbilt, Pittsburgh/UPMC, UCLA/Ronald Reagan.
The third column below shows the percentage of matches that went to internal medicine (regardless of program reputation, excludes primary care, preliminary matches, preventive medicine, med-peds, etc.). Sorry for mixing up scales. To convert the third column into units of the left two columns, multiply by two (making it per 200 instead of percent).
1 1+2 % IM WUSM 2011 31.0 36.2 22% WUSM 2010 17.5 21.1 18% WUSM avg*** 24.3 28.7 20% HMS 2011 31.7 37.8 21% HMS 2010 32.4 38.0 21% HMS avg*** 32.0 37.9 21% JHU 2010 40.3 43.7 24% Stanford 2011 16.7 33.3 23% Stanford 2010 13.5 24.7 13% Stanford avg*** 15.0 28.9 18% Penn 2011 19.5 24.1 17% Penn 2010 28.8 37.4 21% Penn avg*** 24.3 30.9 19%
Conclusion: A mixed bag (or, if you’re an optimist looking at the big picture, they’re all damn good!). Strong: Hopkins 2010. Weak: WUMS 2010, Stanford 2010, Penn 2011. Year-to-year variability is high. A majority to vast majority of IM matches at these top schools go to the best (top 20?) IM programs.
For justification and other details on question 3, see my original post on SDN.
Question 4: surgery match
What proportion of students matched went to a top surgery or surgical specialty program?
This was tougher because numbers within each specialty is much smaller than for IM, so there is much more year-to-year variation. For that reason, I combine both years (2011+2010) when possible. “Surgical” includes general surgery, neurological surgery, orthopedic surgery, otolaryngology / ENT / head and neck surgery, plastic and reconstructive surgery, urological surgery. I added in Columbia to the mix because they’re known for producing large numbers of surgical matchers (“College of Surgeons & Surgeons”).
I calculated the number of surgical matches to hospitals that have “significant”, “strong”, or “very strong” reputations with specialists, and normalized them to match list/class size. The exception is plastics, which has no rankings that I know of, but I include all of plastics matches because they’re a small and highly competitive field. 1) Any plastics residency is probably a strong one. 2) The applicants who don’t match into plastics probably end up in general surgery. Out of any specialty, plastic surgery applicants have the highest average of distinct specialties ranked in the match (2.3 for matched applicants, 2.6 for unmatched applicants). In other words, the average plastics applicant also ranked a backup specialty.
School General Neuro Ortho ENT Plast Urology Total % surg WUMS 5.2 0.9 6.1 3.5 3.5 1.7 20.9 18.3% HMS 8.4 7.1 6.5 2.6 4.5 1.9 31.0 18.2% JHU 6.7 5.0 3.4 5.0 0.0 5.0 25.1 16.8% Stanfor 8.1 5.8 11.6 4.6 4.6 1.2 35.9 19.7% Penn 6.6 2.9 5.1 2.2 0.7 1.5 19.0 12.9% Columbi 6.6 3.3 13.2 3.9 2.0 4.6 33.6 24.0%
Conclusions: Another mixed bag (they’re all so good! on average three-fourths of surgery matches go to top programs!). In general surgery matches, WUSM is weakest compared to its peers, with HMS and Stanford dominating the group. In neurosurgery, WUSM is weakest compared to its peers, with HMS dominating the group, followed by Stanford. In orthopedics, the clear winners are Columbia and Stanford. In ENT, Hopkins and Stanford have small leads. In urology, Columbia and Hopkins take the lead. In plastics, WUSM falls in the upper half of its peers (in fact, one SEM above the mean), with HMS and Stanford taking a very slight edge.
The sum of all surgical specialties gives the following rank order: Stanford>Columbia>HMS>Hopkins>WashU>Penn.
For more justification and other details on question 4, see my original post on SDN.