The Etiology of Application Fever

I saw my first case of Application Fever three years ago.

A medical student asked me to write a letter of recommendation for her internal medicine residency applications. She was a great student – easily in the top 10% from my work with her on the wards – and I was happy to do it.

I wanted to be sure my letter would was hit the high points of her application, so I met with her to review her CV. As we chatted, I casually asked where she was applying, and she pulled out a printed list.

The list was two pages long.

She explained that her primary advisor recommended applying to 60-75 programs.

Dumbfounded, I actually asked her if she had a criminal record or some other huge black mark in her background.

She didn’t. She was every bit the outstanding student I thought she was. She’d just been infected with Application Fever.

See, I graduated from medical school in 2007. And back in them days, folks just didn’t apply to so many programs without a darn good reason.

I do remember one of my classmates – who wanted to be a dermatologist – applying to 50+ programs. But whenever that statistic came up in conversation, it elicited a gasp, as everyone else applied to far fewer. Personally, I carefully chose a group of 10 programs in three disciplines (pediatrics, med-peds, and internal medicine). To me, the idea that a strong applicant in a historically non-competitive discipline would be need to apply to 60-75 programs to ensure matching was just astounding.

But then again, like so many other physicians, I hadn’t spent much time following trends in Match statistics after my own Match Day.

Well, this is what happened while I wasn’t paying attention.

Applicants today apply to many more programs than those in the past. As shown above, in 2007, the average U.S. medical graduate applied to 32 programs. By 2018, that rose to 60.

International medical graduates (IMGs) apply to even more programs: 79 in 2007, increasing to 136 by 2018.

There is a nationwide epidemic of Application Fever. The question is, why? Why do candidates apply to so many programs now? Who benefits from this?

Application Fever is not driven by what you think it might be

By far, the most common explanation for the figure above is that it’s just harder to get into residency programs nowadays. There are more medical schools and more applicants, but residency spots haven’t kept up. Competition for residency positions is tighter than it’s ever been – so students have to apply to more programs. Or so the story goes.

Do the data support that?

Not really.

Here’s a graph of Match rates over time.

As you can see, Match rates for U.S. graduates are flat; rates for IMGs have actually been increasing slightly for the past few years.

Similarly, there’s been essentially no change in the number of residency positions available per applicant.

In 2007, there were 0.78 residency spots for every applicant in the Match (1.44 positions for every U.S. graduate).

In 2018, there were 0.81 residency positions for every applicant (1.61 for every U.S. graduate).

In other words, Application Fever is not driven by a change in the basic number of residency positions and applicants.

So why is this myth so pervasive? Why do students feel like the residency marketplace is tightening, even in the absence of objective data to suggest that it is?

The self-perpetuating psychology of Application Fever

Let’s take a moment and think about what rising application numbers means – in both practical and psychological terms.

Let’s start with some real numbers.

In 2018, there were 171 orthopedic surgery residency programs that participated in the Match (offering a total of 742 individual positions). Altogether, 1017 candidates applied for these positions.

If you’re a student, and you want to be an orthopedic surgeon, you can apply to all 171 programs if you wish. (Some people do.)

But if you’re an orthopedic surgery program director, and all 1017 candidates apply to your program, there’s no way you can interview them all. As applications rise, you might try to squeeze in an extra interview days here or there, or cram in couple of additional candidates each interview day.

And although some programs are interviewing more significantly more candidates than they were 10 years ago, the increase definitely doesn’t mirror the rise in applications per student. The point is, applicants can apply to every program – but programs can’t interview every applicant.

What does that mean?

It means that, even without a change in the number of candidates or positions, when students apply to more programs, a smaller percentage of candidates get interviewed – which contributes to students’ perception that they need to apply to more and more programs.

Even worse, as programs receive more applications, the decisions about who to interview become less predictable.

Example #1

A total rockstar of an applicant applies to a not-so-competitive program, and is not offered an interview.

From the program’s standpoint: “She’s overqualified. She’s just gonna take an interview spot and then cancel at the last minute. Why waste our time?”

From the student’s standpoint: “That was my safety school! This whole process is so random! I need to apply to more programs!”

Example #2

A solid candidate applies to 20 programs where his advisor assures him he’ll be competitive. He gets interview offers from only a fraction – even though the programs are similarly competitive by any objective measure. There is no rhyme or reason that he can discern for why one program offered him an interview and another did not.

There’s an important relationship between the predictability of the Match process and the number of applications that students submit. And that relationship is bidirectional. So the more applications students submit, the more capricious the whole process seems – and in response, students submit even more applications to ensure that they don’t go unmatched.

Repeat this process iteratively over a decade, and you get a graph that looks like the application trends above.

So why don’t students just stop overapplying? I mean, if match rates are exactly the same now as they were in 2007, when students applied to so many fewer programs, why can’t we just turn back the clock? If all students agreed to just apply to fewer programs, everyone could save a lot of cash and stress, and the Match would shake out essentially the same way. Everybody wins, right? So why doesn’t that just happen naturally?

This is where game theory can help us.

Game Theory and Application Fever

To explain why rational individuals make decisions that result in a suboptimal outcome for everyone, we can use the Prisoner’s Dilemma. This is probably the most famous model in game theory, and has been applied to all manner of real-world conflicts.

The relevance of the Prisoner’s Dilemma to Application Fever was beautifully analyzed in this paper, from which the key graphic is shown below.

If you haven’t looked at this type of matrix before, let’s unpack things a bit.

The game involves two players, Applicant A and Applicant B.
The goal of the game is to avoid not matching.
Going unmatched is a potentially devastating outcome. For a student who has invested $200,000+ in her education, the chance that it could all go for naught is a powerful incentive.

Each applicant can choose one of two strategies – to apply to the ideal number of programs, or to overapply.

However, Applicant A can’t control Applicant B’s decision about how many applications to submit. Applicant A can only choose to overapply or not.

Which strategy should each applicant choose?

If Applicant A chooses to apply to an ideal number of programs, he’ll have a 5% unmatched rate only if Applicant B chooses the same strategy. But if Applicant B chooses to overapply, his chance of going unmatched will double.

However, by choosing to overapply, Applicant A guarantees that he’ll avoid the worst outcome (a 10% unmatched rate), and gives himself a chance of getting the best possible outcome (a 2% unmatched rate). In other words, choosing to overapply is a dominant strategy: on average, it will work out better for Applicant A to overapply, regardless of the strategy Applicant B chooses.

Application Fever is rational – but harmful

The Prisoner’s Dilemma shows us that Application Fever is not the students’ fault. It’s a natural consequence of the game we’ve forced them to play. We shouldn’t expect them to choose any differently.

But just because overapplying is a rational choice in the game we’ve forced them to play doesn’t mean it’s without harm.

First and foremost, it’s expensive. Overapplying may be a dominant strategy, but it ain’t a cheap one.

So the first thing we should consider is if this really how we want to allocate residency positions. Do we really want to advantage wealthier applicants, or reward students who go into greater debt? Probably not.

Just how expensive is Application Fever?

Well, here’s the current fee schedule for the Electronic Residency Application Service (ERAS):

Interestingly, each of the hypothetical examples on from the ERAS website were for students applying a number of programs well below the national average. So let’s quickly do the math for an “average” U.S. graduate applying to 60 programs:

$99 + (10 x $14) + (10 x $18) + (30 x $26) = $1199

(I’m not even going to calculate the cost for the average IMG applying to 136 programs, because I know it will make me sad.)

Who benefits from Application Fever?

When you consider that there were 37,103 participants in the Match in 2018… it’s clear that we’re talking about a lot of money.

And where do all those ERAS fees go? Why, to its sponsor, the Association of American Medical Colleges (AAMC).

As I have done for certain other organizations in the past, I evaluated AAMC revenues using their publicly available tax documents.

Turns out that AAMC has benefitted greatly from the epidemic of Application Fever, with revenue growth that has dramatically outpaced that seen in their other programs.

And where does this revenue go? After all, the AAMC is a non-profit corporation, so it can’t distribute extra profit to shareholders.

Well, around 5-6% of the AAMC’s budget goes toward executive salaries. That proportion has been constant over time – so when revenues climb, so does everyone’s salary.

Top executives at the AAMC do quite well. AAMC President and CEO Darrell Kirch received total compensation of $2,196,377 in 2016.

As you can see from the table above, their other executives do pretty well, too. Overall, the AAMC has 247 executives and employees who receive >$100,000 in reportable compensation annually – a figure that has more than doubled from 119 in 2007.

Do you think the AAMC has a financial conflict of interest in pursuing policies that might decrease the number of applications per student? I do.

Even worse, AAMC’s financial interests are closely tied to another powerful organization in medical education, which I’ve discussed elsewhere. So unless we all speak up and demand change, the AAMC is not going to prescribe the antipyretic we need for Application Fever.

NOTES: Financial data for the AAMC were obtained from the organization’s IRS Form 990 tax returns, obtained online via ProPublica’s Nonprofit Explorer. Want to check the figures? Please do. You don’t have to agree with my conclusions – but it’s important to me that I get the data right.