Fourth Year & The Match

Here’s a new book. It’s called Fourth Year & The Match, and you can get your copy by using this form to (at least temporarily) sign up for my new planned very infrequent/sporadic email newsletter:

Get your free book download (ebook and PDF) of Fourth Year & The Match.


If you’d like the book but aren’t interested in hearing from me, just click the instant unsubscribe link at the bottom of the download email. I don’t want to pester you.

And if you want to learn more about this project, keep reading:


Continue reading

Explanations for the 2018 Official Step 1 Practice Questions

Here are my explanations for the new NBME 2018 USMLE Step 1 Sample Test Questions. This year there are 51 new ones (marked with asterisks).

Like in years past, the question order here is for the PDF version (not the FRED-simulated browser version). This facilitates using these explanations in future years when they change the available question set (because the old ones are always available via The multimedia explanations are at the end.

Prior sets/explanations can be found here.
Continue reading

Review: The Unremarkable reMarkable

Man, I really really wanted to like the reMarkable.

The reMarkable, if you haven’t heard of it yet is a large format e-ink device the bills itself as a Paper replacement. It’s billed as a large touchscreen enabled Kindle with a fast refresh rate and a bundled non-battery-powered stylus that supposedly mimics the texture of paper.

And if the software running this thing was anywhere close to the level of polish in a Kindle or your typical smartphone device, then this thing would be pretty awesome.

But it’s not, so it’s pretty much not.


The packaging was slick and got my hopes up a lot.


Screen and Hardware

The screen is pretty nice. A big almost notebook size e-ink display, though without the backlight that folks may have gotten used to on recent Kindle generations. The 300 PPI of the high-end Kindles is substantially better than the 226 PPI of the ReMarkable, so it doesn’t even look as crisp as other modern e-readers. As a radiologist, I was hoping the device would be good for reading image-rich books and journal articles, but the images aren’t crisp enough and the contrast is washed out. The refresh rate when drawing is impressive but the delay when redrawing the screen to change pages in a book is sloooow.

The surrounding is less nice. The plastic frame does jive with the lightness of the device but it belies a build quality far below, say, an iPad or even a Kindle. The front buttons, in particular, are flimsy and loose without firmness or a satisfying click. The little-nubbed pen feels like a cheap but works exactly as billed. It has a satisfying paper scratchiness on the screen.


The software is incredibly weak. The design is hobbled to function as a serviceable paper sketching/drawing app with a buggy, slow, underpowered e-reader bolted on than a real e-reading device. Page turns are slow. Documents are barely searchable. Annotations only sorta work. Unlike Kindle, you can’t, say, export your highlights—they’re basically saved an image overlay, so useless for most people’s purposes unless you are editing/marking up a PDF (as if it were actually on paper). They also have a tendency to shift on the page, making them nonsensical. Underlines should not be strikethroughs. Only non-DRM ebooks in epub or pdf formats work, and the reMarkable doesn’t yet support the new epub3 format. It also doesn’t parse epub files properly, mutilating the majority of the formatting. No links, no footnotes.

You can only sync ebooks via the app, and the app is buggy. I found I was routinely unable to add files to the reMarkable via the app normally, but that I was usually able to add via the iOS “open in reMarkable” extension (which means the files are compatible, just the platform sucks). The organization schema are just a few folders, and book covers aren’t even displayed (the last page visualized is), which makes seeing your collection at a glance a homogenous mess.

There is no in-book search. No parsing of the table of contents. No internal linking.

It shows PDFs like a stack of pictures, not complex files combining images and—of course—text.


Someone who wants to sketch digitally but wants a more paper-like feel than an iPad or a Wacom might enjoy this device.

Anyone hoping for large-format super Kindle is going to be extremely disappointed.

So now I’m using Bear

There were two developments that led to me implementing new workflows to get my writing done. I’ve been interested in different writing environments and different tools for longer than I’ve actually done any significant writing but always fell into the typical trap of spending more time researching what to use than I ever spent actually using anything other than the built-in post editor for WordPress and Microsoft Word.

The first was the birth of my son 3 years ago. I don’t know what I used to do with all of that time, but it became clear that with priorities of being a father and then being a physician that I would need to be more deliberate in carving out a niche for writing.

The partially-related second development is that I started to do a lot more writing on my iPhone. To write effectively on the phone, you need a tool that allows you to get words down quickly and keep your snippets and thoughts organized. Another big plus is solid syncing so that you can switch back and forth on the same project on other devices seamlessly.

Last year, I had a two-app system. WorkFlowy for my snippets, lists, and braindumps. I previously wrote about how I used this awesome free service (with completely optional paid upgrades) to write my second book. Then when it was time to sit down and really flesh out that manuscript I used Ulysses. I even had my entire to-do list in Workflowy for a while.

This worked great. But I found that my post snippets and drafts were numerous enough that they started to get cluttered and lost in my tangle of multi-layered WorkFlowy outlines. I use the app for so many different things that it took time to find the right place to put the right snippets, and I didn’t have the time to sit down and use the Web app to re-organize frequently enough. Workflowy is awesome for unstructured data entry, but it was less good for building up multiple drafts simultaneously, which was especially tricky because I was using it for so many different things that I had to navigate through frequently. These extra seconds sometimes meant either losing organization or losing a thought I wanted to transcribe.

The simple solution was staring me in the face the whole time: Bear.

Now, it didn’t actually have to be Bear. It could have been any decent note app. Byword, iA Writer, and even the built-in notes app would all work the same. Bear just happens to be a particularly well-crafted one that is super responsive, a pleasure to use, and has a robust tagging system (and was perhaps coincidentally the one I was testing out when I figured out my new plan). Bear is completely free to use and $15/year for premium features like cross-device syncing.

What I needed to optimize for my phone writing workflow was an app for long projects which is well organized and synced to my computer (Ulysses). An app for lists and brain-dumping and especially for building up lots and lots of snippets for longer works (WorkFlowy). And—and this was the surprise/revelation—an app just for blog posts.

I think most people intuitively want to use fewer apps. The view of so many app reviews and so much productivity writing is that if you just find the right system or best workflow that everything in your life magically falls into place. The truth is that putting the hours in and developing good habits are what gets results. The tools are just lubrication.

So originally, I wanted to find the right thing that works for everything. Ulysses is basically perfect for that. But I wear different kinds of writing hats and write different kinds of things on my phone so having Bear dedicated to fleshing out a limited number of brewing blog posts is extremely efficient for me, even if it also results in a little additional clutter on the home screen.

ABR simplifies Core Exam scoring

After years of pretending that people could actually fail (“condition”) individual exam sections other than physics in its convoluted two-stage exam scoring process, the ABR has decided to simplify things going starting this year in 2018.

From now on, there are three scoring outcomes:

  • PASS if you get a score of 350 or higher when averaging all sections together (and specifically pass physics)
  • CONDITION if you pass the overall exam but score less than 350 on physics
  • FAIL if your overall score is less than 350 when averaging all sections together

Conditioning physics means re-taking just physics. Failing means re-taking the whole thing.

This means that your performance on any individual section (except physics) is irrelevant so long as the average score across all sections meets the passing threshold of 350. No surprise there. For followers of last year’s mammography kerfluffle, you’ll remember that the ABR acknowledged that the results of the mammography section in isolation literally had no bearing on a single examinee’s passage result. Whether or not it was really technically possible to condition a non-physics section, no has ever conditioned a section other than physics since the Core Exam’s inception.

Scoring is still cloudy, however, because the passing threshold of 350 is a meaningless number without any measure of the preparation required or the percentage of questions you must answer correctly in order to achieve that score. It’s purportedly derived from the sum of the Angoff method scores for each section based on what the expert panel believes a “minimally competent” radiologist should know. So, whatever. This does mean, however, that strong sections can make up for weak sections. Consider this is your license to ignore GI and GU fluoroscopy.

While this sounds like a big positive development, I believe this is basically just a paper change. The ABR is just acknowledging outright the reality on the ground for the past several years: The large gap between overall passing performance and the true failure threshold for all non-physics sections is so large that in practice no one could actually fail an individual section.

Frankly, I wouldn’t be surprised if the one person per year who should have conditioned a non-physics section was just given a score of 200 on the offending set in order to pass via an informal secretive score floor. Who knows.

But at least it’s simpler and more straightforward now.