How to disrupt the pager


The pager is an ancient piece of technology by most standards, but much to the bafflement of many, we in medicine continue to use this device today. Why? Reliability. The pager shines brightest in large inpatient environments, where cell reception is often questionable. I’m not sure if cell signal just doesn’t penetrate through that much concrete or if there’s another reason, but I don’t know anyone who would consider their personal phone a reliable communication device when in a large hospital. Even WiFi is not always guaranteed (that’s an entirely different can of worms). On the other hand, if I page someone, I know that unless their pager is off or out of juice, they’re going to get my message. That’s reason #1 why making HIPAA-compliant messaging apps isn’t actually very useful: it negates the primary hardware advantage of a pager.

Here’s reason #2: pagers get handed off constantly. Many of the current batch of HIPAA-compliant messaging apps assume that I really care about messaging a specific person. Wrong. I don’t care about messaging a specific Joe Shmoe. I DO care about messaging whoever-is-currently-covering-that-service; whether that’s Joe, Moe, Doe, or another physician on their team is completely irrelevant to me. I should never need to know who’s on call that day; I should just be able to call the [insert service name]’s pager (which gets passed around on their end) so I can talk to whoever is responsible at that current time. For you software engineers: it’s kind of like the doctors’ equivalent of blackboxing.

To achieve this blackboxing with smartphones, a hospital would have to invest in one or both of the following:

  1. A ton of smartphones + plans, ideally one per team/service
  2. Employee(s) to keep track of everyone’s schedules (which change constantly) and maintain a database of the right person to message at all times.

…do you see now why people would rather just stick to our current paging system?

Let’s get into the nuts and bolts. When I page someone, I want to open a line of communication that, ideally, encompasses three components:

  1. My contact number.
  2. The message itself.
  3. An idea of how urgent the message is.

Classic paging systems only have 1. I send out my contact number, the person receives it, and then they call me back to get 2 and 3. Text paging systems, which are less common, have 1 and (sometimes) 2, but 3 is still lacking.

AFAIK, there is no paging system that conveys urgency; phone apps don’t address this either. This is more a hardware than software issue. Amazingly enough, no one seems to have figured out that it might be useful for a pager to occasionally do something other than continuously screech an ear-splitting pitch until you acknowledge it. This is a huge problem because by default, this means you have to assume the page is of high urgency, which dilutes true signal-to-noise ratio considerably.

Could we develop a new device with a tiered system of notifications?

  1. Low urgency: an LED light that blinks unobtrusively.
  2. Medium urgency: a short beep that repeats every 10 minutes.
  3. High urgency: what we currently have, i.e. unrelenting cacophony

What if we could eventually develop hooks into our EMRs? A normal lab gives a low urgency notice to let the hospitalist know things are ok, a slightly abnormal value gives a medium urgency notice, and a flagrantly abnormal value gives a high urgency notice. The possible applications with EMR integration are endless.

But first, we need a complete refresh of the pager. And it will not come in the form of an iOS or Android app. It will come in the form of a new piece of hardware that retains the reliability we need but comes with optimizations to its communicative abilities.

Photo credit

Increasing medical student autonomy

One of my greatest frustrations as a medical student: increases in competency were rarely rewarded with meaningful increases in autonomy.

For example, I learned that an EKG is a great thing to ask for in patients with chest pain, but I was never granted the ability to order this basic diagnostic tool. Unless I grabbed the machine and did the EKG myself[1], my knowledge base and desire to help the patient had, for all intents and purposes, no actual effect.

It was the same story with the BMP, CBC, troponin, urinalysis, and numerous other basic tests. I could learn when these tests were necessary/extraneous, and even how to interpret them[2], but in the end, I still had to chase down a resident or attending and ask them to put in the orders for me. If the service was busy that day, who knew when that would be; the patient would essentially just sit there until I could get ahold of someone on my team with the right letters after their name[3].

I get it: when a med student first starts clerkships, it’s necessary to do some hand-holding. Beginners need significant oversight, particularly when patient health is involved. But if a med student doesn’t know when a BMP is appropriate after a few weeks/months of clinical training, then either that student needs (serious) remediation or the school needs to reevaluate its curriculum. It really doesn’t make sense to wait until a med student becomes an intern to give them the power to write for basic orders; this implies that walking across the stage magically grants some monumental difference in knowledge. Let’s be honest and admit it doesn’t, or else we wouldn’t dread July so much.

Nota bene: I’m not saying med students should have the ability to order everything; that would be irresponsible. I’m saying that if we’re training them appropriately, we should have faith in their ability to learn when the most common diagnostic tools are appropriate, and validate this learned competency with the ability to order these tests. This controlled escalation of learner autonomy/responsibility strikes me as a much more balanced approach to training a future physician than suddenly flipping the switch when the learner has “M.D.” after their name. Every good teacher knows that at some point, you have to stop holding the student’s hand. We’re waiting too long to let go.


[1] I did so, frequently, much to the annoyance of some and the delight of others. I got away with it because I could always point out to a frowning resident that the success of a GI cocktail does not, in fact, rule out cardiac pathology.

[2] Admittedly, I never learned to read EKGs well while in med school. Thankfully, residency is changing my ignorance.

[3] To be fair, the patients I was allowed to see on my own before someone more senior stepped in were low enough acuity that it didn’t make much difference in their clinical course; that said, knowing this did little to alleviate my frustration.

Residency hack: energy in a bottle

Ever since I started clinical rotations, I’ve experimented with a fair number of different powder combos in my quest to come up with something that’s clean, lean, and delicious. I’ve been using this formula for several months, and it’s transitioned well to residency, so I figured I’d share (obviously, your mileage may vary).

energy in a bottle


From left to right:

  • Gatorade powder (pictured: lemon-lime flavor)
  • Optimum Nutrition’s Amino Energy (pictured: fruit punch flavor)
  • whey protein powder of choice (pictured: Dymatize Elite, gourmet vanilla flavor)
  • a blender bottle (smaller/bigger variations exist; I use the 28oz Sport Mixer)

Into the blender bottle, I put in:

  • 1 scoop of Gatorade powder
  • 2-4 scoops of Amino Energy (2 scoops = 100mg caffeine)
  • 1-2 scoops of whey protein
  • the springy metal ball thingy

I make this every morning/evening and throw it in my bag on the way to work. When I’m feeling low on energy or it looks like I’m going to have to skip lunch, I just add 20-24 oz of water, shake the bottle, and bam, in 30 seconds I can chug down an instant clean-and-lean boost of electrolytes, carbs, protein, caffeine, and hydration. It’s not a true meal replacement, but it’s always managed to carry me through the next 4-5 hours, and the flavors I use combine to taste like a fruity vanilla ice cream, which is fantastic. With my usual 4 scoops of Amino Energy and 2 scoops of whey, the macros for the brands shown end up being approximately:

80 cal/scoop of Gatorade + 114 cal/scoop of whey + 5 cal/scoop of Amino Energy = 328 calories, 3 g fat,  29g carb, 50 g protein

An M.D. does not a CMO make

One of the more concerning trends I’ve noticed in recent years is the poor selection of Chief Medical Officers in health startups. It seems like more and more often, the nearest person with the right letters after their name is set up as CMO without any vetting or seeking out of actual expertise.

While it’s nice to have the veneer of clinical legitimacy and give the sense of “Oh hey look, this one doc has clearly approved of our product, so other doctors should too”, when a product is clinically faulty or the claimed use cases don’t make sense, it’s painfully obvious that the CMO’s medical credentials are questionable at best.

To be fair, a great deal of the fault also lies with the M.D. who accepts such a position without (a) investing a significant deal of energy and relevant clinical acumen into the product and/or (b) actually working in a field (at least somewhat) related to goal of a device/app. I would argue that this behavior, particularly if done for profit, inches into the realm of being unethical.

So…what to do? I’d like to suggest some guidelines for health startups:

  • As a bare minimum: your doctor should have done residency. Medical school is otherwise known as “undergraduate” medical education. Just getting an M.D. is, in the medical world, the equivalent of getting a B.S. in any other knowledge realm. Residency is known as “graduate” medical education, and completion is the mark of someone who has undergone the required training to practice independently. If you get someone who never did residency, you’re essentially trusting the clinical judgment of a noob.
  • Better yet: get someone clinically active. They call it the “practice” of medicine for a reason. If you can get someone who’s still working with patients, your products will benefit. I would argue that much like pro sports, the validity of a person’s medical expertise drops with the number of years they’ve been out of practice–I don’t think someone who last touched a patient 10 years ago is the person whose advice you want to be relying on for a practical clinical product (exceptions exist, but are exceedingly few and far in between).
  • Your CMO doesn’t have to be a doctor. Ten times out of ten, I would happily trust the opinion of a well-trained-and-currently-practicing nurse over an inexperienced/out-of-practice doc. Many of the so-called “auxiliary” healthcare staff can, in fact, be your greatest resources, and yet they are constantly and inordinately belittled. I don’t get it. What’s wrong with calling on the expertise of an RN, NP, PA-C, CRNA, PT, etc? If they have the requisite knowledge, experience, energy, and desire to advocate for their patients, the letters after their name shouldn’t be the most important part of their role.

I’ve said this before: I love the talent health entrepreneurs (MDs and non-MDs alike) bring to the table, and I have enormous respect for their disrespect of the status quo. But it hurts to see the less scrupulous among them try to game the healthcare space for a quick buck. It’s also not fun to see non-healthcare professionals who are actually trying hard to make a difference jump on the fast track to failure because they hire a CMO primarily for degree and not value. I really hope the CMO-in-a-box trend starts reversing.

Photo credit

An Open Letter to Health Entrepreneurs

Originally written for the RockHealth blog, cross-posted here for future reference.

Dear health entrepreneurs,

Throughout my journey in medicine, I’ve been asked countless times why healthcare moves so slowly. Many of the same technological challenges have existed for years, if not decades, and despite numerous attempts to invade/disrupt/revolutionize the healthcare space, these efforts have altogether been more amusing than fruitful. Despite having so much bait around, most clinicians just don’t bite.

Regulation and funding issues aside, this apparent apathy when it comes to novel health products is at least in part due to two fundamental differences between how you and I think. If we’re going to work together, it’s about time we acknowledge that these differences exist, establish why, and figure out how to deal with them.

First, my threshold for allowing error is insanely high. My job doesn’t give me the luxury of beta testing or iterating. When your work is faulty, your site might go down, customers might get angry, you might lose money. When my work is faulty, people die. Call me OCD, call me conservative, call me whatever you want, but if you want me to use something new, it needs to have an error rate at least equal to, if not significantly better than, the status quo, preferably with data to back up that assertion.

Solution: Lengthen your product timescale, and make sure everything has a level of polish that puts doubts to rest, particularly with clinician-targeted apps. The “move fast and break things” motto that worked for Facebook does not apply to healthcare. “Be deliberate and test everything” is probably more appropriate. When you give demos, take the time to explain and prove that you went the extra mile. For example, if it’s a secure messaging app to eradicate pagers, explain basic cryptography and show both the completely incomprehensible garble that results from your encryption as well as the flawless retrieval at the other end. Reassure us that your servers are HIPAA-compatible with backups ready at an instant’s notice. Make sure the safeguards to prevent data leak in the event of a lost phone/laptop are front and center in your presentation.

Second, if you’re going about product development by thinking “this could be cool”, instead of “This fulfills a pressing clinical need better than anything else out there”, you’re doing it wrong. Just because a piece of hardware/software can be made does not make it clinically useful or interesting. It’s admittedly quite nifty that for under two benjamins we can now create a device that measures my heart rate, temperature, and oxygen saturation in 10 seconds by putting it to my forehead. Clinically, I couldn’t care less, and I don’t think the vast majority of patients should either. Anyone can get their heart rate in 15 seconds for free by putting two fingers on their wrist/neck while looking at a clock and doing some basic math. They can get their temperature at the same time with a $10 thermometer from Walgreen’s. If they don’t have a lung condition and they feel ok, their oxygen saturation is almost guaranteed to be 98-100%. Even being in the low 90s generally feels incredibly uncomfortable, and healthy adults don’t go that low without serious problems, so if someone feels this way, it would be wiser to just call 911 instead of wasting time seeing what a device says, particularly since an O2 monitor can be fooled by emergent conditions such as carbon monoxide poisoning.

Solution: Make your product so clinically compelling that physicians want to use it and maybe even “prescribe” it. While pricey, the Withings Scale, Runkeeper app, and Sanofi’s iBGStar device are fantastic examples of this. For a less expensive example, watch this TED talk on a simple diagnostic test for anemia. If your product clearly provides meaningful health benefits in realistic scenarios, I will want to use it without much convincing, and chances are I’ll be more than happy to become an evangelist.

I hope this has helped clarify why I view just about every new health startup with a cautious eye. I’m not against change at all; in fact, I have tremendous respect for your abilities. It’s just that, if you come barging into my domain without adequately considering patient safety and product quality, I will lose all respect for you. I get it; we both belong to peculiar, and somewhat arrogant, fields, and we both want what’s best according to our worldviews. But let’s try to meet eye-to-eye, or else we’ll get nowhere.


You can’t cheat sleep

Sorry, but I feel compelled to burst this absurd bubble that people keep blowing.


This popular desire to push off sleep to “allow more more productivity” has at its core a pivotal flaw in logic: it assumes that greater time spent awake necessarily equals greater productive output.

…if you’ve ever pulled a desperate all-nighter, you know this is a blatantly false (and stupid) assumption.

And yet, it survives. People abuse stimulants ranging from caffeine to modafinil to ritalin to stay awake longer and purportedly do more work. I don’t get it. As fatigue sets in, the quality of your work drops significantly, you work slower, and you make more mistakes. The time you spend compensating for this crappy work later on almost always outweighs any minuscule benefit it may have given you.

Here’s the deal: assuming you have no underlying medical conditions, sleep is a need-based “drive” state much like hunger or thirst, meaning you only feel sleepy if you need sleep. Try this simple test: put yourself in a really boring situation.  If you start to doze, you’ve probably been shortchanging yourself. Why? Because contrary to popular belief, boredom does not cause sleep; it merely allows for a nice lull in which sleep can happen. If you’re bored and well-rested, you’ll be antsy, not dozy.

Granted, a few select people are genetically programmed such that they legitimately need less than the average of 7-8 hours of rest a night. However, the overwhelming majority of people who get by with less than 7-8 hours of sleep have simply trained themselves to do so with some combination of a regular schedule, an alarm clock, mental fortitude, desperation, exercise, caffeine/energy drinks/other stimulants, etc. This isn’t natural. Nor is it particularly healthy in the long run. The studies on sleep deprivation make a few things clear:

  • You cannot beat biology. You are not going to beat the evolutionary process that took tens of thousands of years and resulted in your need for sleep.
  • The “special”, “ground-breaking”, “magical” supplement/diet/exercise/power-napping-regimen/whatever found on late-nite TV specials or websites-that-try-a-bit-too-hard-to-sell-you-something have no basis in sound science.
  • There is no healthy way to decrease the amount of sleep your body demands. No such exercise/food/supplement exists.
  • Every available peer-reviewed study shows that sleep deprivation has extraordinarily bad long-term consequences on health; depriving someone of sleep is straight-up torture, increases stress levels, and will lead any healthy human to insanity and inevitable death.
  • Every available peer-reviewed study shows that getting ample sleep helps build long-term memories, lowers the risk of various diseases, keeps stress levels in check, and delivers multiple other benefits.

Don’t get me wrong; you can abuse your body pretty heavily, particularly while young, and in the short-term, you might not see any major damaging effects. But in the long run, your body’s needs will always win. It’s far wiser to respect those limits to begin with and just get the sleep you need than to pay the piper later.

Site Makeover, June 2013

Whew. It’s been a long ride through med school. Haven’t touched this blog since 2011, and most of my writing time has been spent on Quora answering questions about medicine. I’ve started backing off from Quora lately, so I need another outlet for my writing inclinations, and I figured I might as well use this opportunity to more or less start from scratch. I’m not sure what direction I’ll go in terms of new content for this blog, but my general interests lie in education, research, ethics, informatics, and UI design, so I’ll probably stumble through some stuff related to those topics.

I’ve decided to remove comments entirely from the site as well; aside from the aesthetic detractions seen by having the comment form being prominently displayed on the screen, it’s generally not worth the effort to keep up with them. Even with services like Akismet, the form just invites spammers, so signal-to-noise ratio is pretty low as well. If you’ve got any feedback you’d really like me to know about, just use the contact form in the menu and it’ll go straight to my inbox (no guarantees on replies, obviously).

If you have a suggestion for a post, by all means, please shoot me a message.
If you’re looking for older posts of mine and can’t find them, I apologize, some of them got lost when I rebuilt the site–on the bright side, if you shoot me a message, I can write a new post on the topic!