Google Glass Is Finally Dead. Ish. | Slate

Will Oremus:

"Moving even more from concept to reality” is a pretty fantastic PR euphemism for discontinuing a product, moving the team to a different division, and going back to the drawing board. While the company promises that there will be future versions of the device, it has offered no timeline.

Some consolation for Google Glass enthusiasts–Tony Fadell, previously at Apple working on the iPod and then started Nest, is reported to be taking over the project.

The most damning part of this announcement is the lack of details or timeline for next steps. If Google didn’t think Glass needed serious work, the next announcement would have been plans for retail sales.

I expect Google Glass to undergo serious changes that will be on a time scale of years not months. I remain bearish on Google Glass, even in medical settings.


The dumpster fire (part 2) | Controversies in Hospital Infection Prevention

Dr Mike Edmond boldly writing about the crisis in recruiting new ID doctors:

But ethical mentoring now requires that we have frank discussions about the relatively low pay of ID physicians with young physicians who are in the process of career discernment. I tell would-be ID physicians that they need to come to terms with the fact that they will work harder and make less money than their peers who are hospitalists. And the issue isn’t just about money, it’s about how valued you feel.

Many (maybe most) subspecialists maintain their generalist board certifications. Why not trade your overworked, underpaid job for shift work as a hospitalist?


Should Surgeons Keep Score? | Backchannel

James Somers:

In order for surgeons to improve, they have to know how well they’re doing. In order to know how well they’re doing, they have to know how well their patients are doing. And this turns out to be trickier than you’d think. You need an apparatus that not only keeps meticulous records, but keeps them consistently, and throughout the entire life cycle of the patient.

Great article on the need for performance measurement and feedback. This article focuses on surgeons, but we need this everywhere in medicine. Unfortunately, as stated above, any efforts require meticulous and expansive data. This is one of the promises of electronic medical records, but we are far from that goal right now.

I would love to have a system like this for diagnostic accuracy for myself, especially as I train.


Don’t Homogenize Health Care | NY Times

Sandeep Jauhar:

We have to get smarter about how we try to improve medical care. I believe the next phase of quality improvement will be a move away from homogenizing care and toward personalizing it, perhaps with the help of genomic research. Neither the old approach, in which seemingly every patient was treated differently, nor the new one, where we try to treat them all the same, has worked well. Medicine needs another way.

A great idea except genomic medicine has almost no evidence supporting it. We have excelled at uncovering the genetic components of diseases and physiology. Unfortunately, we have not excelled (so far) at applying that knowledge to changing the practice of medicine.

A quick example, we know that opioids are metabolized with the help of an enzyme encoded by the CYP2D6 gene. We also know that this gene is highly variable in humans. We know how to detect this variability using genetic testing. We do not know what to do with this information [1]. I could have a patient’s CYP2D6 profile in my hands but I would not have a precise estimate of how I should adjust their morphine dosing…because no such studies have been done.

Genomic medicine is likely part of the answer, but we are going to need the translational research first and then guidelines on what to do with that research.

  1. …other than to tell people at the extremes (very poor metabolizers and very rapid metabolizers) they should likely avoid opioids all together.  ↩


Atul Gawande responding to CIA Torture Report | Twitter

Yesterday morning, Atul Gawande posted a series of 12 tweets detailing and condemning the medical professionals complicit in the CIA’s torture of detainees. The details are appalling and the doctors’ actions reprehensible.

I find it interesting that Gawande used Twitter and a tweetstorm format to publicize this important criticism. While he has over 108,000 followers, he also has access to The New Yorker and The New York Times. With 12 tweets worth of content, clearly he had enough for a longer form [1]. Perhaps this was only the first public debut of his thoughts and we will see more in the next few days in one of those prominent publications.

  1. Maybe we need to hook him up with a tumblr account…  ↩


How residency programs are training doctors to waste money | Vox

(It’s difficult to link to an article with such a ridiculous title, but here we are.)

An interesting piece written about a new JAMA study that shows residency impacts and continues to impact costs of care for physicians once they are practicing.

I initially read through the Vox article, then read the abstract from the JAMA [1] and came to the conclusion that training in a high-spending region seems to mean that you will have higher costs of care when you go out in practice (regardless of spending patterns in your region).

From the Vox article:

Residents who train in regions with high health care costs (that is, the places that err on the side of more scans and specialists) continue to practice expensive medicine decades beyond graduation — even if they move to low-cost parts of the country.

While proofing my original post, I realized that this is not what the paper says at all [2]. There is never a direct comparison made among residents trained in high-spending areas versus residents trained in low-spending areas. The comparisons (summarized in Table 2) are all made within the training spending levels.

The data (as presented) actually shows the opposite of what Vox concludes. For residents trained in regions with high health costs, their average cost patient expenditure is $9,482 when practicing in a high-spending region versus $7556 when practicing in a low-spending region [3].

The comparison we want to look at is between the high-spending trainees and low-spending trainees and their costs within their specific practice spending strata. So, for a resident trained in a high-spending region who goes to practice in a low-spending region, his average cost is $7,556 (as mentioned above). For a resident trained in a low-spending region who goes to practice in a low-spending region, his average cost is $6,882. This represents a difference of $674—not chump change—but no direct statistical comparison is made, so it is difficult to say if this is a real difference.

This data is correlational and speculative at best. It’s hard to tell the direction of the effect.

Vox concludes:

The JAMA paper suggests a tantalizingly easy way to save money in American health care: train more residents in low-cost areas of the country.

While I don’t think that is a bad idea, I wouldn’t base huge programs to shift residency training to lower-cost regions based on this paper.

  1. I have to jump through about 27 hoops to get a full-text article from JAMA through my institution while at home.  ↩

  2. Which subsequently had to be pulled because it was all sorts of wrong.  ↩

  3. P < 0.001  ↩


Blood, Simpler | The New Yorker

Theranos and Elizabeth Holmes are utterly fascinating. To develop a diagnostic testing company in complete secrecy over 10 years that is now valued at $9 billion is unreal.

Speaking of Theranos’ valuation:

Investors have valued the company at more than nine billion dollars, comparable to the two major diagnostic labs.

Comparable valuations, but not comparable scale of business:

Quest runs about thirty full-service laboratories around the country; it performs six hundred million tests of all kinds annually—Theranos’s aim is to hit one million blood tests in 2015—and owns four thousand vehicles for picking up samples.

Valuations never make sense to me.

When I first read about Theranos earlier this year, the proposition—hundreds of laboratory tests from a few drops of blood obtained by finger-stick—seemed too good to be true. Conversely, why not? Why do we have to draw so much blood?

I am not a pathologist and definitely not qualified in any way to answer that question. I think the juxtaposition between Theranos’ promise and the eventual reality is what makes it such a fascinating company. As mentioned several times in the article, transparent peer review needs to be conducted before we can fully buy into Theranos’ promise. Comparison to 23 and Me is quite apt; Theranos will undoubtedly go through growing pains which will include government oversight.

I can’t wait to see what happens. I would prefer, both for my patients and myself, easier diagnostic testing, but it has to live up to current standards for accuracy.


Surgeons and GPs more likely to cause car accidents than other workers | The Guardian

Difficult to judge the validity of this study, but the results are chilling nonetheless:

For every 1,000 surgeons who drive a car, 361 made an at-fault claim in the past five years, compared with just 3.5 building society clerks, the research found.

I think about this whenever I drive home from the hospital late at night after a long day. Be safe on the road.


E is for...Enterovirus 68 | HAI Controversies

Eli Perencevich:

Never a dull moment. Just as facilities have begun wrapping up their Ebola preparation plans, there’s gathering evidence that several US states may be facing large clusters of acute respiratory illness associated with human enterovirus 68 (EV68).

Great post on the recent enterovirus outbreak (including a discussion of a Morbidity and Mortality Weekly Report from 2011 discussing an outbreak of EV68).

I have been working in the ED during this period and we have certainly been very busy.

What bothers me is the lack of information about the extent of this outbreak. Our current understanding is limited to what individual hospitals and health departments are investigating on their own. We have invested billions of dollars in health information infrastructure, yet there is no easy way for an entity like the CDC to quickly pull microbiology data from hospitals [1]. At a bare minimum, all decent sized academic medical centers run their own micro labs and organize patient data in databases—if not full EMRs—and we should be able to link those together and query them.

Unfortunately, we probably won’t know the full extent for quite some time as we use proven, but antiquated epidemiological tools.

Enterovirus–68-like Respiratory Virus Sickening Children Across the US | NEJM Journal Watch

  1. At least not that I’m aware of and nothing has been reported from such a system.  ↩


Doctors Raise Concerns Over Apple HealthKit | Forbes

An interesting and nuanced analysis of the possible impact of HealthKit for doctors.

I will propose a more mundane and broad analysis: HealthKit won’t matter.

Doctors won’t trust the data and won’t know what to do with most of it. Additionally, the data will remain siloed within patient’s phones despite high-profile partnerships with EMR companies because most physicians don’t use those EMRs and even those that do often run outdated versions.

The current pursuit is novel ways to collect data. Apple has taken the next step and is trying to collate and organize it. However, the crucial step—one that will require research [1]—is making sense of the data and providing actual guidance. So far, nobody seems to be taking this step (at least not at scale).

  1. Yes, research (!), that long and nasty process of trying to understand something.  ↩


Should Twitter be Regulated at Medical Meetings? | 33 charts

Bryan Vartabedian MD:

We don’t need our meeting dialog subjected to legislation.

Using terms like “regulation” and “legislation” creates a straw man. Of course we don’t want meeting organizers to formally control the use of social media at meetings. However, we should be careful in our use of social media at meetings.

The biggest danger I see is the public broadcasting of non-peer reviewed research that may be deeply flawed.

Scientific meetings are the beginning of the vetting process for research findings. Meeting organizers do a cursory review of research prior to accepting it for presentation at a conference. This review is based on submitted abstracts and mostly involves ensuring the research is relevant to the conference; it does not involve scientific review. Thus, when a researcher stands at the podium promulgating their conclusions, few people have had a chance to critically review their methods and findings.

In contrast, research published in scientific journals has been through a formal peer review process and found fit for public distribution.

We are also in danger of over-simplifying complex research findings in our dash to encapsulate a speaker’s words in pithy 140-character snipits. We tend to broadcast select conclusions that reflect our own biases and may misrepresent the researcher’s own thoughts. 140 characters does not allow for qualification or explanation of a result.

While I agree with Doctor V that “regulation” or “legislation” are wholly unnecessary, I would implore conference attendees to think twice about what they tweet at meetings.


Virtual surgery gets real | Rémi Rousseau

Researchers in Paris used two synchronized GoPro cameras to capture a total hip replacement from the surgeon’s perspective and then played it back in 3D on an Oculus Rift as a training tool.

When you’re a surgeon in training, you always have a task to do during the surgery; it becomes difficult to see what the main surgeon is doing. Being able to live a surgery in the surgeon’s shoes thanks to the Rift is very useful to replay the surgery in detail, pause, fast forward or backwards. It can also be an inspiring way to share and learn new techniques among surgeons.

Maybe a bit of an overstatement, but viewing complex surgeries from the lead surgeon’s point-of-view has undoubted value for both surgical residents and experienced surgeons learning new techniques.


✚ A tale of two nights—the problem with resident work hours

A few weeks ago, I worked Friday/Saturday overnight shifts on the general inpatient pediatrics team. The two nights were complete opposites.

On Friday night, when I arrived at the hospital my team was just finishing up 3 admissions. I immediately went to work helping out with those admissions. We admitted 4 more patients overnight. We were steadily busy throughout the night without much of a break.

On Saturday night, we had zero admissions [1]. I received a couple of pages regarding minor issues. I relaxed most of the night and got several hours of good sleep.

Due to the resident work hour restrictions, I had to be out of the hospital by noon following each of those nights, regardless of how hard I worked or how much sleep I got.

We restrict resident work hours because work hours are easy to count [2]. They are not indicative of what we are really trying to avoid—fatigue. Unfortunately, fatigue is very difficult to measure and even more difficult to predict. Until we make progress in this area, we will continue to see minimal impact of work hour restrictions on patient safety.

  1. According to the senior residents, busy Friday nights followed by slow Saturday nights are quite common. Their speculation—people take their children to the doctor’s office on Friday before the weekend, the doctor doesn’t feel comfortable with the patient and can’t see them in follow-up the next day, so they send them for admission. On Saturdays, parents tend to try to wait it out until Monday to see their regular doctor. I have no empiric or even much anecdotal evidence to support this theory.  ↩

  2. While easy to count, work hours are by no means an accurate measure. All residency programs I am aware of rely on self-report of work hours.  ↩


Progress notes are a poor tool for doctor-patient collaboration | KevinMD

Steven Reidbord MD:

[Physicians’ progress notes] assist in performing and archiving medical work, much as a scientist’s laboratory notebook records the design, data, and results of experiments. Progress notes were not designed to be a legal defense against malpractice suits, justification for third-party payment, quality-assurance tools for health institutions, or educational handouts for patients. Yet these notes now serve many masters…

But even in the best-case scenario, progress notes are a poor tool for doctor-patient collaboration. By nature they are shorthand, telegraphing complex medical reasoning in a few words…Written material designed specifically for patients is better suited for reminders about what was discussed and how to take medications as prescribed.

We need to fundamentally re-think clinical documentation.

The current structure serves no one well. Physicians don’t like their EMR generated progress notes because they are junked up with extras for billing purposes and become click-laden nightmares to create. Insurance companies don’t like them because of the amount of unstructured data. Patients don’t like them because they are polluted with medical jargon and difficult to navigate.

In the transition from paper to electronic records, nobody stopped to fundamentally examine the role of doctor’s notes and the myriad of purposes they now serve in modern medicine. EMR designers simply took the SOAP note format and made it electronic instead of trying to re-think how an electronic format could facilitate a new approach to documentation.

All parties are seeking more from our documentation. Without a re-design, all parties will continue to be frustrated.