The Single Student Loan Servicer That Wasn’t

InsideHigherEd on the rapidly abandoned plan to consolidate all student loan servicing into a single monolithic “too-big-too-fail” mega contract:

DeVos has taken heat since May from members of Congress and representatives from the loan-servicing sector over the plan to pick a single servicer that would hire subcontractors to collect loan payments. Department officials at the time argued that the plan would make oversight of servicers by the government more efficient.

But the proposal found critics among both Republicans like Senator Roy Blunt of Missouri, who argued that the system would remove choice and competition, and Democrats like Massachusetts Senator Elizabeth Warren, who warned against creating a federal contractor “too big to fail.”

Blunt and Warren were part of a bipartisan group of senators who introduced legislation ahead of the department’s announcement Tuesday to block the single-servicer plan. Their bill would instead require the participation of multiple loan servicers.

Instead, 2019 will still see the consolidation of the different servicer websites to a single portal that interacts with the borrowers (which was the original Obama-era plan already in motion). Everything else will remain separate. How that will happen basically remains a mystery, but the site itself will be developed and managed by the Office of Federal Student Aid.

Hopefully, they don’t take any tips from Navient, because their website is singularly terrible.

Medscape resident survey, embarrassingly interpreted, again

Medscape’s newest resident compensations survey is out and discussed in “Most Residents Say They Deserve Big Raise, Survey Shows.”

The main thrust is fine, discussing that today’s residents feel more underpaid than generations past, which is no surprise given the proliferation of mid-levels who work alongside them making considerably more (and likely combined with the envy caused by their better-off friends parading happily on social media [when #YOLO, #FOMO can be devastating]).

But then this:

Resident salaries in 2017 vary considerably by specialty. Trainees in hematology lead the pack, at $69,000, while family medicine residents bring up the rear, at $54,000.

The gender gap in resident pay is negligible. Men averaged $57,400 or 1.2% more than women, who received $56,700.

Ugh. Who writes up these Medscape survey articles? I even wrote about the same misleading fake resident gender pay gap back in 2014.

To summarize:

All trainee salaries are based on PGY year and location. There are absolutely no differences between specialties or genders of trainees of the same seniority. Any differences are related to the differing duration of training between specialties as well as the geographic spread of the relatively small sample.

Ultimately, any attempt to differentiate annual salaries by specialty is intrinsically misleading. Any differences that can be created between genders or specialties are simply reflective of different numbers of respondents at different levels of seniority within the PGY scale. The difference between a family practice resident and a “hematology resident” is that almost every family medicine resident finishes in three years while any hematology fellow will be at least a PGY4 or higher. The fact that hematology “led the pack” and not—let’s say—cardiology or gastroenterology just means of the respondents of the survey, slightly more senior hematology fellows answered compared to their other IM-fellowship peers.

There is a real gender wage gap in medicine, but it does not apply to residency. As I discussed almost exactly three years ago, any differences in gender pay during training are related to the known disparities in gender representation among certain fields, particularly surgical specialties (which have longer training lengths and thus get “paid more”). Now, if we want to talk about the “gender surgeon gap,” that would be a different and worthy story. Because there are fields in which women are underrepresented—that’s the story when it comes to residency. Not a misinterpretation of the statistics.

This sort of willfully misleading interpretation has no place on a website that caters to physicians. Medscape should know better. And, reading some of the comments suggests that some readers (primarily the nonphysicians) do latch on to these “differences” despite simply being a distraction from the real issues at play.

Mammogeddon 2017: The Conclusion?

From the ABR’s July 19 email:

Some of you are wondering why it has taken so long for the ABR to provide a solution. We apologize for the delay as we know this has been stressful for you. More than 450 candidates were affected by this situation. The cause of the problem was not initially apparent, and it was important for us to have time to investigate, review preliminary scores of all candidates, obtain direction from our board members and some program directors, and devise a solution that was most appropriate for all stakeholders, including you, your program, and your patients.

The ABR board also received input from the breast imaging community, which feels it is imperative for residents to be tested on breast content at some point in the certification process. The board members considered requiring a breast module on the Certifying Exam for those who did not receive the module on the Core Exam. However, all were concerned that more than two years of delay would require you to study again for the breast module.

The board feels strongly that we must administer the content as soon as possible, and that we should not require travel, other expenses, or additional resources, which is aligned with what we have heard from the breast imaging community. Therefore, we decided that we should trust you to take the online module in a setting of your choice. In addition, the breast module has been carefully edited to ensure that all findings are visible without the need for a high resolution monitor.

[…] We will schedule residents who need to take the breast imaging module at specific times on two dates: September 7 and September 18, 2017. You will select your desired start time when you register.

Still missing: what actually happened in Chicago, what the technical glitch was, how they’ve taken steps to prevent this from happening again, how this module is graded, how “hard” it will be to pass, if it’s actually possible to fail, and a finally—what happens if someone actually manages to fail.

It is interesting that you can take it anywhere you want but that you still must take it at specific times—presumably a compromise to prevent cheating/sharing of the exam content without resorting to using an official testing center. The real exam is proctored with a bathroom monitor, but the fabled mammo content is on the honor code. To me, this is highly suggestive of lip service to an apparently deeply hurt mammography community.

And, speaking of testing centers, the ABR recently released the following narrative about why they haven’t been able to disseminate the exam:

These delivery requirements have proven to be insurmountable obstacles for the numerous commercial testing vendors that we’ve engaged over the years. It’s important to remember that the vast majority of these vendors’ clients deliver text-based question exams with little or no multimedia content.


Just last year, we engaged two prominent commercial testing vendors to explore our goal of delivering the diagnostic radiology initial certification exams at local testing centers. Both vendors were given in-depth details of our exam delivery needs and asked to provide a proposal for our consideration […]

…but neither was interested.

I like that they’ve finally publically responded to these perennial requests.

I imagine these two were Prometric and Pearson VUE, because (despite the claim of “numerous” vendors) there are only a handful of large commercial testing centers around that could possibly furnish the exam. I suppose it’s possible the big two passed in years past. I have no doubt that the ABR’s demands for administration are not worth the time and expense for most vendors to meet given the low exam volume. The follow-up question, however, is whether or not it’s possible to write a Core exam that can be disseminated.

For example, the video portions of the exam are small in number and generally useless outside of cardiac MR (which, if we’re being honest, plays a comically outsized role on the test). The multi-slice scrolling capacity is rarely used and usually only a handful of images anyway. Mammo and radiographs could be selected that do not require high-resolution high-filesize images. The ACR in-service exam, of note, was able to snag a contract and is also image-based.

We are committed to making the initial certification process as facile as possible. While our past efforts have not been successful, we will continue to pursue our goal (and your wish) of delivering diagnostic radiology exams in local commercial testing centers. As we all know, technology is constantly evolving, and perhaps local exam delivery will become more feasible in the future.

I don’t doubt that the exam the ABR created couldn’t be ported to Prometric as is. Shucks, it didn’t even work in Chicago. But couldn’t we have a Core Exam that was functionally equivalent but wasn’t so off-putting? Exams need to be written with the administration in mind from the onset, not just as an afterthought.

Perhaps putting our hopes in the possibility that bandwidth and memory will be so cheap one day that testing companies won’t find our poorly written and conceived exam so unpalatable isn’t the best plan.

Switching from REPAYE to PAYE after residency

From a reader:

Incoming PGY1. Your posts are tremendously helpful! My question is, why aren’t all residents (or at least the majority) doing REPAYE and then switching to PAYE at the end of their training? I feel like I am missing a key pitfall or something. Is it a pain in the ass to switch? Are new residents scared they will not be able to switch? Or is this just information not everyone has? I understand some people are not eligible, have a large spousal income, have private loans, etc etc…But just wondering why this isn’t a more ‘popular’ way to go about it?

Both REPAYE and PAYE calculate payments based on 10% of your discretionary income. But because PAYE monthly payments are capped at the amount that would be due for the standard 10-year repayment, PAYE payments can be lower than REPAYE payments for high earners. PAYE also allows for the married-filing-separately loophole that REPAYE closes.

The first thing we do as human beings when considering any big important action is to look around and see what everyone else is doing. (One imagines that, overall, this is a helpful survival mechanism.)

The downside to this approach is when the crowd is wrong. Or, when the crowd is right, but you’re different in some critical way.

So, the first/main reason people aren’t all switching from REPAYE to PAYE? They haven’t had a chance to yet.

Reason 1: Too Fresh

The switch probably will be popular over the next few years, but REPAYE is still pretty much brand new and a lot of current residents who should be aren’t on it. It was released at the end of 2015 (which is mid-cycle for almost everyone’s annual recertification), so really only students graduating in 2016 (i.e. new PGY2s) even had the option to pick it when they first selected a repayment plan.

Reason 2: Too Mislead

Another reason is that even for the more industrious residents who considered switching when it came time for their annual recertification, it seems that a lot of servicers have been misleading borrowers about the ability to switch out. For example: Yes, you can switch back from REPAYE to IBR or PAYE or even Navient is still lying to borrowers despite lawsuit.

Reason 3:  Too Ignorant or Lazy

Most borrowers choose and set-it and forget-it strategy to student loans, which means that they don’t critically re-evaluate their decisions or maximize their strategies. I’d like to think most people fall into the reason #1 camp, but the reason #3 group is one of the reasons why I wrote a book about it. A lot of folks are just lost.

Reason 4: They Actually Don’t Need To

A final big reason is that many borrowers won’t benefit from switching to PAYE: it depends on what happens after training. Switching only makes sense if you’re trying to minimize payments for PSLF. Otherwise, having smaller payments just means paying more over the life of the loan. Additionally, the accrued interest will capitalize, which is not relevant for PSLF but is for everyone else. For PSLF purposes:

And even for those PSLF-bound:
– A lot of people don’t need to. It isn’t that easy to break past the pay cap for a lot of docs. Thus, if you’re single, have a non-working spouse, or have a spouse with a similar debt to income ratio, PAYE isn’t going to make a big difference unless you are in a high paying specialty. There are definitely attendings who will continue to earn a REPAYE interest subsidy throughout their 10 years of qualifying payments, even with spousal income (particularly heavy borrowers).
– Similarly, depending on their spouse, many won’t gain enough in lower payments by filing separately to offset the tax penalty of switching to PAYE in order to file taxes separately worth it.

The real take home

As you approach the end of training, it’s time to sit down and make your real repayment plan. You may have been in REPAYE because it was the no-brainer choice while in training, but now—with a new job and a salary increase on the horizone—you’ll have the information you need to figure out if you should stay the course, switch to PAYE or IBR, or prepare to refinance privately.

How I wrote my second book

I published my second book a little over a month ago. It took about 11 months from conception to release, clocks in at around 45,000 words, and the bulk of the first draft was written (actually dictated) on my iPhone using Siri and an app called WorkFlowy. A significant fraction of that was “penned” walking down a quarter-mile long sky-bridge that attaches the decrepit parking garage I park in to the hospital I work in, which I typically traversed a few times a day as a resident.



(Mmhmm, dictation errata.)

The workflow basically went like this: dictate fragments and ideas while walking. Lots of them.

Some of these started off as headings and things to cover later in further detail ( especially those parts that require crunching numbers). Other parts were fully fleshed out (sometimes with placeholders for the data I didn’t have offhand). Back at my computer, I intermittently organized these entries into categories using the browser/web-version of Workflowy.

Workflowy—which is a note-taking/outlining/list-making platform—is perfectly suited to this because it allows for an endlessly large and endlessly nested hierarchal outline. (It’s a freemium app; it’s free but you can pay for upgraded features. That link is a referral that doubles the number of list items you can add per month. Using that would add to my maximum as well, but I already have more than I need). I use Workflowy for basically everything I write: post ideas, drafts, quotations, rapid to-do lists, etc. You can expand and collapse different levels of the outline with the click of a bottom and drag/drop to reorganize elements on the fly. It’s basically frictionless.

So, I spitballed the first half of the book based on a latent outline I had in my head and the topics I knew I would need to cover. Once I had some volume on paper [sic], I went back in and plotted out the chapters I would include and then nested everything I had already written in their proper locations. I had approximately 10,000 words dictated before I organized all these fragments into chapters. The outline format makes it easy to generate new content and then filter and clump it together in batches, so then I knew where the gaps were.

As I got the core content done and it was time to synthesize the disparate elements, flesh out certain paragraphs and arguments, etc, I then transferred the whole outline into Ulysses: the only distraction free writing environment I’ve ever used meaningfully (I also wrote my Texas JP exam book in it). Ulysses allows you to create a smart-folder with multiple separate text documents (“sheets”) in it while displaying them all in a sidebar on the left. So I transferred all the primary chapters into their own sheets within this larger group. Then I begin the process of fleshing out the writing, fixing innumerable dictation errors, adding examples, figuring out all the data I needed, crunching the numbers, and trying to limit repeating myself too much (it’s easy to have an epiphany and dictate what you think is an amazing well-argued point only to realize you’ve already said it twice).



The Ulysses’ sheet system allowed me to keep all my chapters in separate places to move effortlessly back and forth between them, keeping the format simple with Markdown instead of the usual poorly-implemented Word styles (this is particularly helpful when trying to format e-books, as well as preventing me from wasting my time fiddling with formatting when I should be writing). Ulysses isn’t great for complex data like tables, so as a consequence, those had to wait until the final push. When I was done writing, I exported the whole thing into Microsoft Word for the final additions, table of contents, etc.

The folks who make Ulysses finally released a fantastic iOS version during this process, which I could theoretically use for my next project from the very beginning. The mobile app is a separate purchase but well worth it. Once I move a project into Ulysses, I can now edit it on my phone or on my computer and the iCloud sync works perfectly every time. But I’ll probably still use Workflowy for the initial draft; it’s just so flexible that nothing I have beats it. On the other hand, I’m slowly revamping my “Guide to Fourth Year” and doing it on straight on Ulysses (because I had the initial drafts already written from the old blog posts).

Ultimately, Medical Student Loans: A Comprehensive Guide was a bit scarier to write in some ways than my first book. The JP book was in some ways straightforward: distil a large amount of known boring material into a reasonable amount of condensed material with the hopefully right balance of precision, clarity, concision, and humor. I knew there was a small but underserved market of people (i.e. all physicians who need a Texas license), and I felt that there were several workable but no good options.

No one needs a student loan book. There is no mandatory test. I’m competing with a bunch of free websites and a few mostly crappy books that I am confident no one is actually buying or reading. The vast majority of these don’t tailor well to doctors, but enough cover the issues well enough such that a dedicated person can learn enough in a couple hours to feel like they can (in many cases correctly) make reasonable decisions.

All that said, I believe strongly that more young docs and docs-in-training need the material in this book. Many if not most graduating students don’t understand their loans or even have the basic financial literacy foundation to make sound decisions about them (or any other financial decisions for that matter). That is why I wrote the book, and I’m glad I did!

Overall, it was a big fun project that took way longer than I’d planned. On to book #3!