Doctors flunk quiz on screening-test math | ScienceNews →

This headline depresses me.

It comes from a research letter recently published in JAMA Internal Medicine. In this study, the researchers asked a convenience sample of 61 attendings, residents, and med students the following question:

If a test to detect a disease whose prevalence is 1 out of 1,000 has a false positive rate of 5 percent, what is the chance that a person found to have a positive result actually has the disease?

Depending on how stringently you define a ‘correct answer’ [1], at best, only 40% of the respondents got it right. The most common answer was 95%. [2]

Why do I find these results depressing? Because understanding this concept is fundamental to the practice of medicine. We are talking about test characteristics, more specifically positive predictive value (PPV). PPV is dependent on the prevalence of a disease—the higher the prevalence, the greater the PPV. This is why we gather patients’ histories and physicals. By asking a few questions and examining a patient, we can identify risk factors—that place the patient in a group with higher prevalence of a condition—that allow us to choose appropriate tests with a good chance that a positive result will actually mean the patient has the disease.

I truly believe if we had a better understanding of this concept, we would order fewer diagnostic tests and save both some anguish and money.

  1. That is, if you define having a correct answer as being exactly correct, only 14 (23%) provided such a response. The researchers, in their Methods, have a little bit looser (but, in my opinion, acceptable) definition of the correct response.  ↩

  2. To be fair, this was not the most rigorously conducted study. It wasn’t a random or complete sample, nor was the sample very large. The question they posed was not validated, nor did they pose more than one question about this concept. But, how rigorous of a study do we need? The best data might be on biostats and epi questions from board exams, except reviewing such concepts is a routine part of prep for those exams.  ↩

Nurses Are Not Doctors | NY Times →

Sandeep Jauhar:

And to attract more of the best and brightest in our medical schools [to become primary-care physicians], let’s find a way to increase their pay.

While pay disparities between specialties contribute to the dwindling numbers of med students choosing primary care, they are not the core problem.

The core problem is who we select to become doctors.

The premed curriculum focuses exclusively on basic sciences. Admissions committees emphasize basic science education by favoring applicants majoring in a basic science or those with research experience. [1] These fields focus on using advanced technology to find a single answer to a narrowly-defined question. Is it any wonder that when we select students from these backgrounds—students who spend 4 years doing such work and are deeply interested in it—that they choose to go into subspecialties the focus on a narrow set of problems and employ high-tech tools to find answers?

Certainly the money helps, but I think students are far more interested in happiness and doing work that suits their interests. Our admissions criteria biases our selection of candidates towards those predisposed to enter the ever-growing number of highly technical subspecialties. Until we adjust these criteria, we will continue to have primary care physician shortages, no matter how compensation is adjusted.

An excellent, complementary discussion of how medical education erodes ‘primary care skills’ written by med student Michael Bradfield was recently posted over at KevinMD.

  1. I know, I know…it doesn’t matter what you major in and you don’t have to have research experience. The truth is, you improve your chances of admission if you do.  ↩

Doximity raises $54 million but what value will they add →

On one hand, I’m glad to see these guys continue to raise money and continue their development.

On the other hand, I’m disappointed that we don’t have a better physician-centric social network. While they have been successful at signing up doctors, it seems (at least anecdotally) few are engaging with the network. [1] I have connected with many of my classmates and some physicians I know on the network. I have never interacted with any of them through Doximity.

The article quotes LinkedIn co-founder and Doximity board member [2], Konstantin Guericke:

I think a lot of doctors will have a LinkedIn profile and Doximity profile. But the key is which part is really going to get ingrained in their lives.

The key question is—what value does Doximity provide over other, non-physician centric social networks? More plainly, what is going to make me open up Doximity on my iPhone instead of my favorite Twitter client?

The current answer to that question is: nothing.

In their smartphone app, the news feed features medical journal articles from the likes of NEJM, JAMA, Lancet, etc. It is unclear exactly how these are selected, but quite clear they are not tailored to my interests. [3] Twitter, on the other hand, provides a constant stream of thoughts and articles related to my interests because of the people I’ve chosen to follow. Doximity’s ‘Groups’ feature helps this a little by providing focused streams around topics. Unfortunately, I am a member of only one group and it looks like you can’t add additional groups from the app. [4] A quick perusal of a few groups shows that many of these are dominated by a single person, not robust discussions amongst large groups.

Where Doximity has truly missed a golden opportunity is messaging. Instead of eschewing the most tired of physician tools—the fax machine—they have embraced it as a core component of their messaging platform. The call their messages ‘Fax+Mail’. You can send and receive both email-like messages and faxes. While this may appeal to the current generation of physicians, nobody graduating med school today wants to deal with fax machines (virtual or otherwise).

More specifically, Doximity so far has missed the opportunity to become the de facto text messaging service for physicians, residents, and med students. If they offered a robust messaging service similar to something like Tiger Text, they would not only have more than 40% of physicians sign-up but daily engagement. This would create a ‘halo-effect’ and drive usage of their other services.

Hopefully, this new injection of capital will help them refine their network. I’m cautiously optimistic based on what they done with $27 million so far.

  1. I would like to quote some engagement figures here, but I’ve never seen any from Doximity.  ↩

  2. I don’t know if I consider Doximity and LinkedIn to be direct competitors, but close enough that having their co-founder on my board would seem uncomfortable.  ↩

  3. Even if you find an article in this stream that you want to read, it would take at least 3 taps to get to the full version and you are likely going to hit a paywall (though you can setup a proxy, but most people don’t have access to a proxy and don’t know how to set it up). Also, within an article the sharing button contains only two options—’View in Safari’ and ‘Share email link’. No clear option to share it within Doximity and no sharing options with other apps or saving to read later services.  ↩

  4. In fact, there seems to be a large disconnect between their app and their website. The app seems to be missing some key functionality. ↩

A unique perspective on EHR dissatisfaction | MedCity →

An elegant description not only of what better EMR design should entail, but also what it will accomplish for physicians' practices.

The goal should not be to ‘master’ a tool, but to use tools to help providers function at high efficiency. Efficiency does not mean clicking buttons and mouse, it means amplifying strengths.

The Medical Chart: Ground Zero for the Deterioration of Patient Care | The Health Care Blog →

As I have said many, many, many times on this blog—the problem with electronic medical records is not that they are electronic, but that they are poorly designed.

This post from Dr Val seems to tacitly admit this [1], but still laments electronic records and yearns for the return of paper charts. I respectfully disagree with her conclusions.

To her specific complaints:

Medical notes are no longer used for effective communication, but for billing purposes.

Medical records have always been used for billing purposes. [2] The real issue here is that both EMR design and EMR purchasing are conducted by non-physicians.

Few doctors know much about design and even fewer know how to program, so very few have been involved in the actual creation of EMRs. When they are, it generally seems to be quite successful. Unfortunately, we don’t have a stable of brilliant physician-designers. We need to be actively recruiting med students with backgrounds in design and programming to build up these resources.

EMRs are enterprise software. This means, among many other things, that the end-users (i.e.—physicians, nurses, therapists, etc) don’t directly control purchasing. Executives and IT departments with their own interests and motives control those decisions. I have never been privy to such deliberations, but I would surmise that improving billing workflows is a highly prized feature. Usability and communication features are much lower priorities.

No one talks to each other anymore.

I agree; a lot could be accomplished through better face-to-face or verbal communication. However, two things conspire against such communication:

  1. Due to both billing and legal requirements, everything must be charted. While a verbal order may be the clearest method for communicating something, it must also be entered in the chart. Why double your work and do both? [3]
  2. There is a cultural shift away from verbal communication. Just look at the proliferation of online ordering for pizza and take-out food. Although I have no research to back this up [4], I would guess that the current generation of med school graduates feels much more comfortable and amenable to ordering something through a computer interface than picking up a phone.

It’s easy to be mindless with electronic orders.

Again, I agree; it is easy to mindlessly check boxes. Again, I also think this is a design problem.

Somewhere along the line, EMR developers decided that clicking was waaaaay better than typing. In some EMRs, you can write entire notes and complete orders without ever touching the keyboard. This design decision is most likely rooted in programmers desire to produce as much structured data as possible because structured data is a lot easier to deal with than free text. [5] This helps with billing as well.

But, we don’t need to go back to paper records, orders, and prescriptions.

What we need is better free-text recognition and autocompletion. Ever filled out an online form with your address that subsequently analyzed your input and suggested a corrected form that was more complete (e.g.—changing “St.” to “Street” or adding the last four digits of your zip code)? We need systems that can take free-text orders and prescriptions and suggest completed, corrected forms. This would be faster and more accurate.

For example, let’s say you want to write a prescription for amoxicillin. [6] You type:

amox 500 PO BID x 10d

This would translated into a full prescription:

amoxicillin 500 milligrams
Sig: Take one tablet by mouth two times per day for 10 days.
Dispense: #20
Refills: 0

This would be faster than clicking individual boxes and provide more complete, understandable instructions. The full prescription could then be electronically sent and simultaneously dropped into the note.

The current state of electronic medical records is quite frustrating. You don’t have to look very far to find posts similar to Dr Val’s expressing such frustration. But, instead of raging against the EMRs, I think we need to work together to build the next generation of great EMRs that fulfill all of our expectations for technology aiding better patient care.

  1. In a post-script, she praises one EMR that “was created by physicians for supporting actual thinking and relevant information capture.”  ↩

  2. I just recently heard a story of a rural Kansas doctor working many years ago whose clinic notes consisted only of a ledger with a diagnosis and the charge. It would seem his records were only used for billing purposes!  ↩

  3. A bit of devil’s advocate here. As I said, I think verbal communication is good and face-to-face communication is great. When I start my internship is a couple months, I am going to strive to do as much face-to-face communication as possible.  ↩

  4. Maybe Papa John’s or Domino’s has some figures…  ↩

  5. As a researcher, I have been guilty of this in designing databases and survey forms. Very guilty.  ↩

  6. Many current EMRs have individual entry boxes for each prescription component (conservatively, up to 10 boxes). Many doctors seem to do whatever they can to eschew the carefully designed boxes and cram everything into the only free-text box available (usually the “Sig” box for instructions). ↩

It’s Insanely Easy to Hack Hospital Equipment | Wired →

In hospitals increasingly composed of the Internet of Things, how secure are the ‘things’?

In a study spanning two years, Erven and his team found drug infusion pumps–for delivering morphine drips, chemotherapy and antibiotics–that can be remotely manipulated to change the dosage doled out to patients; Bluetooth-enabled defibrillators that can be manipulated to deliver random shocks to a patient’s heart or prevent a medically needed shock from occurring; X-rays that can be accessed by outsiders lurking on a hospital’s network; temperature settings on refrigerators storing blood and drugs that can be reset, causing spoilage; and digital medical records that can be altered to cause physicians to misdiagnose, prescribe the wrong drugs or administer unwarranted care.

[…]

“There are very few [devices] that are truly firewalled off from the rest of the organization,” he says. “Once you get a foothold into the network … you can scan and find almost all of these devices, and it’s fairly easy to get on these networks.”

Probably even easier now, given the recent revelation of a big Internet Explorer exploit.

They found a number of infusion pumps that have a web administration interface for nurses to change drug dosage levels from their workstations. Some of the systems are not password-protected, while others have hardcoded passwords that are weak and universal to all customers.

So, while these infusion pumps are under physical lock-and-key in the patient’s room, they remain unlocked through their web portals?

And the most damning revelation:

“The vendors don’t have any types of security programs in place, nor is it required as part of pre-market submission to the [Federal Drug Administration],” Erven notes. “There’s no security assessment before it goes to market.”

(h/t Kevin Wang)

Anti-Net-Neutrality “Fast Lanes” Are Bullshit | Marco.org →

Marco Arment:

Everyone in this discussion has been led, most likely by talking-points marketing by the FCC and ISPs, [1] to describe the destruction of net neutrality as allowing ISPs to “create fast lanes”.

This language was carefully constructed to sound like a positive, additive move: It’s building, not destroying or restricting. They want to offer faster service, not reduce the speed or priority of all existing traffic. Who could possibly be against that? They’re building fast lanes, like a highway! Everyone loves fast lanes! U-S-A! U-S-A!

Naturally, this doesn’t reflect reality at all. Only a fool would believe that the ISPs would actually create any new capacity, higher speeds, or consumer value in this process, leaving their existing service untouched…

This is not making anything faster—it’s allowing ISPs to selectively slow down traffic that they don’t strategically or financially benefit from, and only permit traffic from their partners to run at the speeds that everything runs at today.

It is a money grab by the network owners, pure and simple. Like all such money grabs, consumers lose in the end. The Internet will be less free and the barrier to entry for small, disruptive companies will be greater.

Read more about net neutrality in this great Atlantic piece and sign the White House petition to maintain true net neutrality.

  1. FCC = Federal Communications Commission; ISP = Internet Service Providers (Comcast, Time Warner, Verizon, etc)  ↩

Jay Parkinson Reddit AMA | Kevin Wang →

Kevin Wang has put up an excellent selection of excerpts from Jay Parkinson's Reddit AMA today.

I have followed Parkinson's work for many years now and he always has great insights and this list is a greatest hits. Love his thoughts on wearables and consumer health technologies such as a Scanadu; I share his skepticism.

Genentech writes a check and PatientsLikeMe agrees to share “de-identified” data for 5 years | MedCity →

Apparently, PatientsLikeMe’s ‘Openness is a good thing’ philosophy doesn’t apply to their own business—the amount of the payment from Genentech hasn’t been disclosed.

From their ‘Openness Philosophy’ page:

Furthermore, we believe data belongs to you the patient to share with other patients, caregivers, physicians, researchers, pharmaceutical and medical device companies, and anyone else that can help make patients’ lives better.

Shouldn’t it be up to the patients participating in the site to decide who they share the data with? Shouldn’t the patients themselves profit from that sharing?

The Quest To Predict Flu Outbreaks Moves From Google To Wikipedia | Fast Co Exist →

Follow-up for my previous post on Google Flu Trends and big data.

You know what may be better than Google or Wikipedia searches—UpToDate data for both influenza-related searches and oseltamivir indications/dosing queries.

In my limited experience, doctors go to UpToDate when they have a patient in front of them they highly suspect has a specific condition and they want to verify either (1) criteria for diagnosis or (2) treatment options and/or dosing. Seems like you would get a high signal-to-noise ratio looking at UpToDate for tracking influenza.

The only hurdle is that pesky problem of getting a private company to turn over their data…

Google Glass is now ‘try before you buy’ | VentureBeat →

Kia Kokalitcheva:

Since Glass is obviously not quite “socially acceptable” yet, this could be yet another small attempt at making it seem … normal.

Back in February, Google posted a friendly list of “dos and don’ts” for Glass Explorers, a first step in combating the anti-social aura of Glass.

[…]

Google is trying so hard to make “Glass” happen.

So are many physicians…

Quantified Self Public Health Symposium | Susannah Fox →

Interesting discussion in the comments section of this post about the need for physicians who have a deep understanding of data.

Dr Bryan Vartabedian (of 33 charts fame) writes:

Just like we need physician scientists and physician creatives, we need physician data scientists. I suspect these are unicorns…

As usual, Dr V’s analysis is spot on. We do need physicians who understand data and they do seem to be quite rare.

However, I object to adding more tech/media jargon to the discussion. [1] What we need are physician-biostatisticians.

These are people with a MD and a PhD in biostatistics. The field and programs to go into the field already exist at schools of public health around the country. Unfortunately, they don’t seem to be very popular. Most MD-PhDs seem to pursue basic science fields like molecular biology, biochemistry, physiology, physics, etc and we need them in those fields to advance the science of medicine. But, with more and more large data sets being created through the spread of electronic medical records, we are going to need some really good physician-biostatisticians.

I briefly considered this path. Ultimately, I decided it was too long of a road when I was already starting my medical career a little late. It is a very difficult path. Of the 5 MD-PhD candidates in my original med school class, one dropped out of the PhD program and another one is unhappy with their PhD work. We need some strong, smart people (and there are many in medical schools around the country) to take up this yoke.

  1. I don’t know what a ‘data scientist’ is. I know what a statistician and biostatistician are.  ↩

If you want a CT scan that costs $802 less, go to Canada | Vox →

It’s no secret that many (most?) health care costs are lower in other developed countries. The real question is—why?

Is the $802 difference solely due to higher reimbursement paid to radiologists in the US to read the scan? [1] Or is it due to a relentless pursuit of advanced technology for a competitive edge in America’s capitalist health care system that means we are using unnecessarily advanced CT scanner compared to Canada? Or is it pure profit motives? Or is some of the increased cost due to health insurance overhead and malpractice costs?

While it’s interesting to look at these 15 charts comparing costs in the US to other developed countries, it provides very little insight into the actual problems.

  1. It’s unclear from this data what the figures actually include (e.g.—are physician fees included in these costs) and if the figures are even comparable across countries due to differential bundling of facility and physician fees.  ↩

The Mayo Clinic's New Doctor-in-an-iPhone | Fast Company →

For approximately $50 a month, the Mayo Clinic is offering unlimited access to the famed hospital's nurses through a smartphone app.

If that's the opening line of your article, how can you possible justify the use of 'doctor' in your title?!?

Terrible, hyperbolic headline.

How Information Overload Sabotages Our Observation Skills | 99u →

Bob Hambly:

It's our job to be better observers. We've become lazy about it.

In this talk, Hambly is addressing designers but I think his lessons are not only pertinent but prescient for physicians.

The basic job of a doctor is to take in information and make observations in order to generate creative ideas. Then, order those ideas in terms of their likelihood and again generate creative ideas for solutions. The medical colloquialisms are 'differential diagnosis' and 'treatment'. But, don't be fooled—these are creative enterprises.

In the world of electronic medical records, molecular testing, advanced imaging, and genomic data, we face the threat of information overload. Reams of information are at our fingertips. No longer do we have to be exhaustive historians, staunch observers of signs and symptoms, or careful practitioners of the physical exam. We've become lazy observers.

Yet, valuable information not contained in the records lies within the history and physical. The subtle symptom overshadowed by more salient problems; the travel history or environmental exposure not previously asked about; the close family member with a similar problem. How do we improve our observational skills?

For designers, Hambly suggests documenting observations in various media. For physicians, I think the key is long-form narratives of the History of Present Illness and careful documentation of the physical exam (our versions of documenting our observations). In our quest for efficiency (and more patients and more pay), we've reduced the HPI to a grocery list of signs and symptoms and copied templates of physical exams. While these may fulfill billing requirements, they are not useful later on when things don't turn out how we anticipated and we need to revisit the original patient presentation. Holding ourselves to a high standard also pushes us to gather enough information from the patient to form a thorough narrative. You'll see how much detail you've missed when you try to form a comprehensive, readable narrative.

The next best tool for avoiding information overload to improve our observational skills...don't look at the chart before seeing a patient. Gather all the information yourself. Don't rely on the admitting HPI or previous hospitalization notes for the basis of your narrative. Take the patient's own words; then use the other records and clinical data to enrich the narrative.

Such process takes time, a valuable commodity in medicine. But, practice breeds proficiency and speed. We (especially those of us still in training) need to ensure we are practicing the correct process.