Skip to main content

The Fellowship Match Game: Gearing Residents and Programs up for the NRMP Match



Now is the time of year that residents are starting to interview at programs across the U.S. in preparation for the NRMP Specialty Match and the hopeful placement in their top fellowship program (here at GeriPal, we are still hopeful that Palliative Medicine will soon enter the match, but the good news is that geriatrics is now fully in). One of the most important aspects of the match is that everyone understands the ground rules, the most important of which is to always submit your true preferences for programs/applicants when you fill out your Rank Order List.  Residents applying to fellowships should be told this golden rule, as breaking it will NEVER be in their best interest.  To learn why this is true, try playing the following game to learn how NRPM matches residents to programs.

RULES OF THE GAME

A.  Each resident can only be accepted into one program.
B.  For the purposes of simplicity, every program in this game has only 1 fellowship position available (in real life you may have more than one fellowship position)
C. Your job is to act as the NRMP matching algorithm by following these 4 steps:
  1. Start off by proposing a Resident to her top-ranked program based on her rank order list (ROL).
  2. The program accepts the proposal if it has an open spot, retaining the right to reject it later in the algorithm if a better suitor arrives.
  3. If the proposal is rejected, propose the Resident to the next program on her list. 
  4. If the proposal is accepted, repeat the process for the next Resident B.

There is some more complexity in the system (mainly from couples matching), but this is basically what the Matching algorithm does in several milliseconds.  So, lets play the matching game.  Can you get all of the questions right?


QUESTION 1

There are two residents, two programs, and all put their true preferences in the Rank Order List (ROL) of the Match as outlined below:


What is the final match result?

Answer:
  1. Start with Resident 1 by proposing to program A, which is tentatively accepted as a match
  2. Go to Resident 2. Propose to program A, which is rejected, as Resident 1 is ranked higher
  3. Resident 2 Proposes to Program B, which is tentatively accepted as a match
  4. Final Match Result: Resident 1 – Program A; Resident 2 – Program B


QUESTION 2

Same example as question 1, with two residents and two programs. However, Resident 1 likes program A better but doesn't think she can get in, so ranks B higher (so she doesn't follow her true preferences, but believes it better serves her to rank the less preferred program higher)


What is the final match result?

Answer:
  1. Start with Resident 1 by proposing to program B, which is tentatively accepted 
  2. Go to Resident 2. Propose to program A, which is tentatively accepted
  3. Final Match Result: Resident 1 – Program B; Resident 2 – Program A

Take Home Point: Residents should always order their rank list based ONLY on their own preferences.  Resident 1 would have been accepted to program A if she ranked it higher than program B


QUESTION 3 (getting tougher)

There are three residents, three programs, and all put their true rank order list (ROL) preferences as stated in the following picture:


What is the final match result?

Answer:
  1. Start with Resident 1 by proposing to program A, which is tentatively accepted
  2. Go to Resident 2. Propose to program B, which is tentatively accepted
  3. Go to Resident 3. Propose to Program A. As Resident 3 is ranked higher than Resident 1, Resident 3 is tentatively accepted by Program A, and Resident 1 is displaced
  4. Go to Resident 1. Propose to Program B. As Resident 1 is ranked higher than Resident 2, Resident 1 is tentatively accepted by Program B, and Resident 2 is displaced
  5. Go to Resident 2. Propose to Program C, which is tentatively accepted
  6. Final Match Result: Resident 1 – Program B; Resident 2 – Program C; Resident 3 - Program A


QUESTION 4

Somewhat similar to example 3. Three residents, three programs, Resident 3 really would like A, but feels she has a better chance of getting into C. So, Resident 3 ranks C higher than A, going against the true preference strategy.


What is the final match result?

Answer:
  1. Start with Resident 1 by proposing to program A, which is tentatively accepted 
  2. Go to Resident 2. Propose to program B, which is tentatively accepted 
  3. Go to Resident 3. Propose to Program C, which is tentatively accepted
  4. Final Match Result: Resident 1 – Program A; Resident 2 – Program B; Resident 3 - Program C

Learning point: If Resident 3 would have stuck to her true preferences, she would have been accepted by A.  Residents should always order their rank list based ONLY on their own preferences.



QUESTION 5 

Three residents, three programs, program C really doesn't want to take Resident 3, but decides to rank her anyway just in case.


What is the final match result?

Answer:
  1. Start with Resident 1 by proposing to program A, which is tentatively accepted 
  2. Go to Resident 2. Propose to program B, which is tentatively accepted 
  3. Go to Resident 3. Propose to Program C, which is tentatively accepted 
  4. Final Match Result: Resident 1 – Program A; Resident 2 – Program B; Resident 3 - Program C


QUESTION 5 Extra Credit

Extra Credit: What would happen if program C didn’t rank Resident 3


What is the final match result?

Answer:
  1. Start with Resident 1 by proposing to program A, which is tentatively accepted 
  2. Go to Resident 2. Propose to program B, which is tentatively accepted 
  3. Go to Resident 3. Propose to Program C, which is rejected. Next, propose to program B. As Resident 3 is ranked higher than Resident 2, Resident 3 is tentatively accepted by Program B, and Resident 2 is displaced 
  4. Go to Resident 2. Propose to Program A. As Resident 2 is ranked higher than Resident 1, Resident 2 is tentatively accepted by Program A, and Resident 1 is displaced 
  5. Go to Resident 1. Propose to Program C, which is tentatively accepted
  6. Final Match Result: Resident 1 – Program C; Resident 2 – Program A; Resident 3 - Program B

Learning Point: If the stars align perfectly, there is a very small benefit of a truncated list for programs (in real life, this happens very rarely). However, programs don't have enough information prior to submitting their match results to take advantage of this in a systematic way.  This though is never true for residents - it is always best to put your true preferences in as your rank order lists.


QUESTION 6

Three residents, four programs (getting a little closer to geriatrics now). However, Program A likes Resident 1 but would like to truncate their list for the very, very small chance of getting better results (see example 5). They decide to remove Resident 1 from their rank.


What is the final match result?

Answer:
  1. Start with Resident 1 by proposing to program A, which is rejected as a didn't rank Resident 1 
  2. Propose Resident 1 to Program C, which is tentatively accepted 
  3. Go to Resident 2. Propose to program B, which is tentatively accepted  
  4. Go to Resident 3. Propose to Program C, which is rejected. Next, propose to program B. As Resident 3 is ranked higher than Resident 2, Resident 3 is tentatively accepted by Program B, and Resident 2 is displaced 
  5. Go to Resident 2. Propose to Program C. As Resident 2 is ranked higher than Resident 1, Resident 2 is tentatively accepted by Program A, and Resident 1 is displaced 
  6. Go to Resident 1. Propose to Program D, which is tentatively accepted
  7. Final Match Result: Resident 1 – Program D; Resident 2 – Program C; Resident 3 - Program B

Learning Point:  If program A listed Resident 1, they would have matched together.  Truncated Rank Order Lists for programs (not for applicants) may slightly improve final assignments but also significantly increases chance of not matching, and the only way to guess right is know exactly what everyone else is putting in as their rank order lists.   

That's all, but read the following for more on the Match:

by: Eric Widera (@ewidera)

Comments

Nice post Eric. thanks for doing this! nancy
Alex Smith said…
5 residents leave UCSF for the match traveling at 20mph. There is a headwind of 5mph. How long does it take them to reach Mt Sinai Medical Center (aka UCSF East)?
Samir Desai said…
Eric has done a great job of emphasizing an important point in the development of the fellowship rank-order list. I have had numerous conversations over the years about this with residents, and I've found that trainees can be tempted to rank based on interest expressed by programs. It's always best, as Eric has indicated above, to rank programs based on your interest. It's sad for me to see applicants go unmatched because they truncated their rank-order list based on interest expressed by programs. This unfortunate scenario can be avoided if you rank all programs you would be willing to attend in order of desirability to you.

Samir Desai, M.D.
Assistant Professor of Medicine
Baylor College of Medicine

Popular posts from this blog

Geroscience and it's Impact on the Human Healthspan: A podcast with John Newman

Ok, I'll admit it. When I hear the phrase "the biology of aging" I'm mentally preparing myself to only understand about 5% of what the presenter is going to talk about (that's on a good day).  While I have words like telomeres, sirtuins, or senolytics memorized for the boards, I've never been able to see how this applies to my clinical practice as it always feels so theoretical.  Well, today that changed for me thanks to our podcast interview with John Newman, a "geroscientist" and geriatrician here at UCSF and at the Buck Institute for Research on Aging.

In this podcast, John breaks down what geroscience is and how it impacts how we think about many age-related conditions and diseases. For example, rather than thinking about multimorbidity as the random collection of multiple different clinical problems, we can see it as an expression of the fundamental mechanisms of aging. This means, that rather than treating individuals diseases, targeting …

The Dangers of Fleet Enemas

The dangers of oral sodium phosphate preparations are fairly well known in the medical community. In 2006 the FDA issued it’s first warning that patients taking oral sodium phosphate preparations are at risk for potential for acute kidney injury. Two years later, over-the-counter preparations of these drugs were voluntarily withdrawn by the manufacturers.  Those agents still available by prescription were given black box warnings mainly due to acute phosphate nephropathy that can result in renal failure, especially in older adults. Despite all this talk of oral preparations, little was mentioned about a sodium phosphate preparation that is still available over-the-counter – the Fleet enema.

Why Oral Sodium Phosphate Preparations Are Dangerous 

Before we go into the risks of Fleet enemas, lets spend just a couple sentences on why oral sodium phosphate preparations carry significant risks. First, oral sodium phosphate preparations can cause significant fluid shifts within the colon …

Length of Stay in Nursing Homes at the End of Life

One out of every four of us will die while residing in a nursing home. For most of us, that stay in a nursing home will be brief, although this may depend upon social and demographic variables like our gender, net worth, and marital status. These are the conclusions of an important new study published in JAGS by Kelly and colleagues (many of whom are geripal contributors, including Alex Smith and Ken Covinsky).

The study authors used data from the Health and Retirement Study (HRS) to describe the lengths of stay of older adults who resided in nursing homes at the end of life. What they found was that out of the 8,433 study participants who died between 1992 and 2006, 27.3% of resided in a nursing home prior to their death. Most of these patients (70%) actually died in the nursing home without being transferred to another setting like a hospital.

 The length of stay data were striking:

the median length of stay in a nursing home before death was 5 months the average length of stay was l…