EKGs and Artificial Unintelligence | 33 charts

As a follow-up to my previous post, Dr Bryan Vartabedian talking about applying artificial intelligence to EKG interpretation and medicine in general:

Machines will evolve to do ‘mindless’ things like identify heart rhythm disturbances. As that happens our work as doctors will be redefined around the things that only we can do as humans. Those things involving, as [Deloitte’s John] Hagel suggests, “imagination, creativity, curiosity and emotional and social intelligence.”

For the record, I never look at automated EKG reads. I’ve never been able to trust them because of all the reasons Dr John Mandrola cites.

Permalink

A.I. versus M.D. | New Yorker

Excellent piece from Siddhartha Mukherjee on the state of advanced computer learning in medicine.

While this piece is very long, it is well worth the read. Mukherjee takes care to highlight the promise of computer-aided diagnosis as well as the potential pitfalls.

Sebastian Thrun, formerly of Stanford’s Artificial Intelligence Lab and Google X who has worked on machine-learning for medical diagnosis, discussing the impact of artificial intelligence in medicine:

“I’m interested in magnifying human ability,” Thrun said, when I asked him about the impact of such systems on human diagnosticians…"The industrial revolution amplified the power of human muscle. When you use a phone, you amplify the power of human speech. You cannot shout from New York to California”—Thrun and I were, indeed, speaking across that distance—“and yet this rectangular device in your hand allows the human voice to be transmitted across three thousand miles. Did the phone replace the human voice? No, the phone is an augmentation device. The cognitive revolution will allow computers to amplify the capacity of the human mind in the same manner. Just as machines made human muscles a thousand times stronger, machines will make the human brain a thousand times more powerful.” Thrun insists that these deep-learning devices will not replace dermatologists and radiologists. They will augment the professionals, offering them expertise and assistance.

We need such augmentation in medicine. The current practice of medicine is incredibly labor intensive, not only from the well known burden of paperwork and administrative tasks, but also the fundamental process of diagnosis and treatment. For complex diseases, physicians must integrate a long patient history and disease course with hundreds of clinical data points. This process is cumbersome and error-prone. The complexity of modern medicine is only going to grow and with it our need for augmented medicine.

Permalink

Peer Into the Post-Apocalyptic Future of Antimicrobial Resistance | WIRED

Well-written piece but not substantially different than other post-antibiotic doomsday narratives that seem to come out about once a month now.

While antimicrobial resistance is, as the World Health Organization has said, a global public health threat and deserves attention (it is what I will be studying and working on in my own fellowship), I’m not sure what function pieces like this serve. Ostensibly they alert the public to the persistent threat. But at around 2,700 words and chocked full of alarmist language, I’m not sure this piece reaches the appropriate audience to accomplish that goal. Like climate change, messaging for antimicrobial resistance [1] is a delicate ballet.


  1. The term “antimicrobial resistance” represents the fine line between accurate and effective messaging. The general public has no idea what an “antimicrobial” is. While it is the utmost correct term, at best we can hope the general public misreads it as “antibiotic” because that’s what they (in their minds) get from their doctor for an infection.  ↩

Permalink

A Google-backed health insurer wants to disrupt insurance by ... limiting patient choice? | Vox

Direct scheduling might not sound that innovative; it isn’t hard to build software that lets users book anything from yoga classes to plane flights.
But it is far, far from the industry standard in health care right now. Usually patients have to go to their insurer to look up who is in the network and then start calling doctors to see who has a slot available.

Probably one of the most insane aspects of modern medicine. Why are we chasing after telemedicine and fancy mobile health solutions when patients can't even book their own appointments online?!? 

Permalink

Duration of Antibiotic Treatment in Community-Acquired Pneumonia | JAMA Internal Medicine

This is a noninferiority randomized clinical trial out of Spain that showed 5 days of antibiotics for community-acquired pneumonia (CAP) were enough and validated the IDSA/ATS guidelines for CAP in terms of duration of therapy.

The findings themselves are enough for a link, but really I want to draw attention to this study because of this quote in their “Discussion”:

Determining the duration of antibiotic treatment based on clinical response appears to be a better strategy than using arbitrary treatment lengths.

True and ballsy.

For those who don’t know, most of the recommended durations of antibiotic treatment have never been studied. They are based on our best guesses. I suspect that if we did studies like this one we would find that we could safely treat for shorter durations. This will be an intense area of research for antimicrobial stewardship if anybody will fund it.

Permalink

The Royal Children’s Hospital asks that Trainers don’t drop Pokéstop Lures | Stevivor

Steve Wright:

“We have lots of entertainment and distractions for our patients, who are confined to wards and unable to move about the hospital. Placing lures around the hospital, when children cannot leave their rooms, may create unrealistic expectations, and subsequently, much disappointment.” [Royal Children's Hospital spokesperson]

Simply put, a Pokéstop just out of a child’s reach will have the opposite effect the well-intended Trainer was hoping for.

“While we understand everyone’s good intentions, we would prefer if people did not place ‘lures’ at the RCH,” the spokesperson concluded. “We know everyone means well, and appreciate that the kids are in their thoughts.”

This is in contrast to the story last week out of C.S. Mott Children’s Hospital in Michigan that is using Pokemon Go to help get patients out of their rooms.

Children’s hospitals are created and designed specifically for kids, not only in the medicine they provide but also the environment in which care is delivered. The child life specialists [1] will figure out creative and appropriate ways to use new technologies like Pokemon Go for patients.


  1. Child life specialists are people who work in children’s hospitals that do many things to make a child’s hospital stay better. One of their responsibilities is providing developmentally appropriate and safe play.  ↩

Permalink

The SmartPhrase as Medicine’s Old Technology | 33 charts

Bryan Vartabedian:

But packaged as part of a dangerous copy-and-paste trend in medicine, the SmartPhrase has earned the reputation as the processed meat of health information. More evidence, they say, of medicine’s dysphoric slide.

But I used SmartPhrases long before EPIC.

The truth is that my ink scratch on paper during the analog phase of my career was nothing more than an endless daisy chain of recurring bits of language scribbled again and again and again. My mind was a SmartPhrase generator, creating one after another with a Uni-Ball Signio 207 Bold (still my weapon of choice when paper calls).

Great point by Dr Vartabedian. Just because there is a new technology that facilitates something, that doesn’t mean that we hadn’t always been doing something similar all along.

Permalink

How This Technology Is Making Doctors Hate Their Jobs | Fast Company

Christina Farr:

[Electronic medical records were] supposed to reduce inefficiencies, make doctors’ lives easier, and improve patient outcomes. The only problem? Many hospitals spent millions (and sometimes, billions) on systems that weren’t designed to help their providers treat patients. “Frankly, the main incentive is to document exhaustively so you cover your ass and get paid,” says Jay Parkinson, a New York-based pediatrician and the founder of health-tech startup Sherpaa.

I think Jay Parkinson is primarily referring to physician incentives in the use of EMRs. The larger problem not addressed in this piece is that hospitals are the biggest buyers of EMRs. Hospitals, above all, want an integrated and efficient billing system. Ideally, they’d like an automated system so they could get rid of billers [1] and generate bills automatically from doctor’s notes in realtime. Physician workflow is only a tertiary consideration.

Follow the money. Until doctors are the primary buyers of the electronic systems, other priorities for those holding the money will predominate.


  1. Hospitals employ people for the sole purpose of “translating” a doctor’s note into a bill they can send to an insurer or patient. These people require specific training in medical coding. Thus, they are expensive employees and it takes time for them to “translate” the note, time that the hospital is not getting payment.  ↩

Permalink

Detecting shunt failure in hydrocephalus without imaging or surgery: ShuntCheck | Vector

This device chills CSF fluid within the shunt by placing an ice cube over the tubing externally and then checking for temperature changes downstream to detect blockages [1] . What a great application of simple principles for a complex and frequent problem in pediatric EDs!


  1. It’s slightly more complex than that, so check out the full articles and the links within it. Promising technology nonetheless.  ↩

Permalink

Toward an AIDS-Free Generation: Can Antibodies Help? | NIH Director's Blog

Dr Francis Collins:

…researchers discovered that a single infusion of the antibody reduced levels of HIV in the bloodstreams of several HIV-infected individuals by more than 10-fold. Furthermore, the study found that this antibody—known as a broadly neutralizing antibody (bNAb) for its ability to defend against a wide range of HIV strains—is well tolerated and remained in the participants’ bloodstreams for weeks.

This is (hopefully) really great news, especially in light of many patients who develop resistance to current antivirals.

Permalink

Using Phones to Connect Children to Health Care | NY Times

This article in the Times goes on for 15 paragraphs about ideas and examples for using mobile technology to improve adolescent and young adult health communication [1]. However, it not once mentions sexual health.

Perhaps the largest and most significant gap for most teens and young adults in their health maintenance is sexual health. It’s a taboo subject in general and one that this age group (like many of their older counterparts) don’t readily talk about. The smartphone represents a huge opportunity to initiate the conversation asynchronously in a private setting [2] that can then translate into an honest conversation with a doctor (whether that happens on the phone or in person).


  1. Truthfully, I don’t know why this article has “children” in the title. It mostly talks about teens and this is the age group where there is the most potential.  ↩

  2. For example, a simple app could take a sexual history through a series of on-screen questions and then transmit this to the patient’s doctor. Their doctor could then quickly ask any clarifying questions during a visit and provide the appropriate care. This would “grease” the initiation of the conversation on both sides.  ↩

Permalink

Apple Watch heart monitor saves teen’s life | Cult of Mac

Remarkable story about how tachycardia on a teen's Apple Watch led to the diagnosis of rhabdomyolysis.

I am generally fairly skeptical of continuous monitoring for healthy people, but there are cases all the time where it has provided clear benefit.

Permalink

Not “Just a Nurse” | Slate

Nurses don't just grease the wheels of medical care, they are the wheels. Virtually nothing happens in a hospital or clinic without the help of a nurse. They are invariably hard-working, dedicated, passionate people who deserve our respect and gratitude. I am grateful every day for the great nurses I'm privileged to work with.

In terms of what was said on that talk show—why are any of us surprised that somebody said something stupid and insensitive on a talk show? If anybody had a nuanced, thoughtful comment that would be more newsworthy.

Permalink

Computerized order entry: preventing spoiled doctors? | Edwin Leap

Edwin Leap:

I don’t understand why it’s so much better for me to enter orders than for the secretary, or even nurse in some instances, to enter them. Other people still vastly outnumber doctors in most departments. Further, I still work in two places where I check the box on the form, or tell the nurse/secretary, and they enter it into the computer.

Why was that so bad? Were there that many errors? Were the secretaries that overworked? Is the use of language, as in ‘Mr. Schwartz in room 5 needs CT angio to rule out PE,’ so utterly fraught with confusion and uncertainty? […]

And in an age of nurse empowerment, are nurses so unworthy of our confidence that they can’t do anything until it’s in the computer? I swear, I expect to someday say ‘start CPR’ and be asked, ‘did you put it in the computer yet?’

The answer to the question is not to allow more verbal orders, but to make the process of CPOE quicker and more convenient for physicians, namely by making it mobile and context-sensitive. It would be great working the ER, walk into a patient’s room, assess them, and then pull out my smartphone which would automatically pull up the particular patient (based on the fact that I’m standing in their room) and a menu of relevant orders based on their chief complaint with easy search for any other orders.

There is no reason why physicians can’t enter their own orders; we just need to make it quicker and easier.

Permalink

How To Speak Startup, Part Deux | TechCrunch

In a list defining terms frequently used by start-ups:

“We have lots of unstructured big data.” — We have a big pile of shit in a bucket.

Physician notes = unstructured big data

Permalink

Dropout Docs: Bay Area Doctors Quit Medicine to Work for Digital Health Startups | KQED

File this under #notSurprisingAtAll.

A particularly damning indictment of clinical medicine:

“I loved working with patients but I looked around me and realized that I didn’t want the jobs of anybody who had ‘succeeded’ as a clinician,” said Rebecca Coelius, who graduated with an MD from UCSF.

I fear more and more medical students and residents are feeling this way. Though there is no shortage of excellent clinical mentors, their career and lifestyle may not be appealing to the current generation of trainees.

Coelius goes on to note:

“Tech culture is very appealing when juxtaposed against the hierarchy and myriad hoops to be jumped through in clinical medicine,” she explained.

What hoops exist?

The most familiar to the public include performing well in an undergraduate program, doing well on the MCAT, applying for and getting into medical school, then completing a residency.

There are many, many less familiar hoops that include the four USMLE licensing exams [1], the process of applying for residency (which includes traveling all around the country), applying for licensure in a given state as a resident then later as a practicing physician [2], becoming credentialed at a hospital (forms for which may be up to 100 pages long and a process that takes months) [3], applying for a myriad of ‘numbers’ such as a National Provider Identifier and DEA number, obtaining malpractice insurance, creating/maintaining an academic portfolio (which can take up entire 3" three-ringed binders, passing medical specialty boards [4], and maintaining board certification through the Maintenance of Certification programs. This is not an exhaustive list but represents only the bare minimum [5].

The medical students detailed in this article who have jumped to startups had to do none of these things and will likely make more money working fewer hours in their positions than they ever would have as doctors.

However, those people as do not get to experience the deep satisfaction of practicing medicine. Helping patients is an indescribable feeling. It’s something I really didn’t even experience as a medical student; you only start to get a real flavor for it as a resident.

I fear two things for this trend:

  1. We are losing very smart people who would make tremendous clinicians and further the practice of medicine itself.
  2. Startups are attracting people with MDs who have never actually practiced medicine and do not truly understand medicine [6].

Both of these issues need to be addressed by (1) trying to tackle all of those ‘hoops’ listed above while ensuring doctors are well-trained and qualified and (2) developing career paths for practicing doctors to work with startups on creating ways to make technology and medicine work in harmony.


  1. There are three parts to the USMLE Step exams (i.e.—Step 1, Step 2, Step 3). However, Step 2 is two parts—a written test and a practical test where you examine and diagnose actor-patients—so there are really four licensing exams that must be complete.  ↩

  2. Some people may go through this process more than just two times. In my case, my residency operates in two states so I will undergo it twice for residency, another time for fellowship (assuming I go to another state for further training), and then again for when I get a ‘permanent’ position somewhere. That means a minimum of 4 times. If should ever switch jobs to something in another state, I will have to go through it all over again.  ↩

  3. A process that is so arduous and time consuming that a startup called CredSimple has just raised $1.5 million to take care of this for physicians.  ↩

  4. The USMLE licensing exams which are completed in medical school and the first year of residency are only to obtain a medical license. To become “board certified” in a speciality (often a requirement to be hired by a hospital or private practice) you have to complete a residency/fellowship in the speciality (ranging from 3 - 9 years) and pass their exam at the end of your residency/fellowship.  ↩

  5. I’m sure I’ve even missed some ‘hoops’ in this list.  ↩

  6. I would argue that you don’t begin to really practice medicine until your residency and even residency is not truly practicing medicine. A true practitioner of medicine has full responsibility for a patient’s care and deal with the weight of that responsibility.  ↩

Permalink

ResearchKit Asthma app developer praises framework in first official blog post | 9 to 5 Mac

I’m very excited about ResearchKit and the potential it has to revolutionize how research is conducted. It is certainly not without its problems, most notably the bias it inherently introduces by only including iOS users. The graph at the bottom of this post highlights another problem—initial excitement but precipitous drop-off in engagement. This developer had a 50% dropoff in less than a month. [1]


  1. This graph doesn’t look too different than such graphs I’ve seen for most apps. People try things out and only a small fraction stay engaged. Unfortunately, that introduces a huge bias when talking about research.  ↩

Permalink