How video can reduce medical errors and improve patient care

About eight years ago I was desperate to improve my golf game. I just couldn’t straighten out my drives or hit my irons crisply. (Yes, I’m fully aware that this is a first world problem). I decided to try golf camp in Palm Springs for a few days.

My sensei, a crusty ex-touring pro named Artie McNickle, watched me hit several dozen balls on the driving range, video recorder running. “So, did you figure it out?” I asked with hint of sarcasm after my last shot. I thought I was a hard case.


“How long did it take you?” I asked.

“One or two swings. But you looked like you were having a good time, so I didn’t have the heart to stop you.”

Artie patiently told me what I was doing wrong. Though it made sense in theory, when I tried to follow his directions, I didn’t get very far.

“Let’s look at the video,” he offered. “Whose swing do you really admire?”

I named Ernie Els and Tiger Woods, two pros with silky smooth yet powerful swings.

“Fine,” he said. On a video screen, I appeared on the left side. As if by magic, Ernie Els was on the right.

“OK, let’s see where your club is about 18 inches into your backswing.” My hands were low, and the club formed a straight line with my arms; my wrists hadn’t begun to cock. He then showed Els at the exact same point in his swing. His club was facing skyward, forming an acute angle with his wrists. This wasn’t a subtle difference in technique; it was an enormous one.We then reviewed my follow-through. Once again, about two feet after I’d struck the ball, my club extended straight out from my arms. Conversely, Els had his wrists swiveled about 45 degrees counterclockwise, his right hand rotated powerfully over his left.

Tiger’s swing was slightly different, but similar in all the ways that mattered.

My trek to Palm Springs shaved 2-3 shots off my score (and they haven’t come back). I’m convinced it was the video that did it.

This kind of video coaching has become standard procedure in major league sports. My son Doug, who works in baseball operations, tells me that every major and Triple A minor league game is video recorded. Not only do players watch themselves to see what they’re doing right and wrong, they watch how other teams play pitchers or hitters who resemble them in style and build.

Yet we hardly ever use this extraordinarily powerful tool in healthcare. Thankfully, that’s beginning to change.

Earlier this year, Johns Hopkins surgeon Marty Makary published a JAMA article entitled “The Power of Video Recording.” It’s a thoughtful and eye-opening piece, well worth a read.

Makary reviews several ways video can be used for peer review, quality improvement, and coaching. I’ve previously described the use of video monitoring for hand hygiene: a study performed at Long Island’s North Shore Hospital found a staggering uptick — from 7 to 82 percent — in hand hygiene performance in ICUs that were monitored with cameras pointed to the sinks and gel dispensers. The key was that there was someone in Bangalore, India reviewing the video feed every hour and sending back compliance data, which was posted on an electronic tote board — real-time feedback, either positive or negative. Makary also cites a study in which gastroenterologists whose colonoscopies were video recorded improved their performance by one-third. As Makary observes, the obstacle to doing this is not a technical one: there’s plenty of video equipment in the operating room.

“Procedures ranging from cardiac stent placement to arthroscopic surgery are performed using sophisticated video equipment; however, the record button is turned off.”

Adding to this literature, one of the most impressive health services research studies in recent memory was published last week in the New England Journal of Medicine. In it, John Birkmeyer, a surgeon and researcher at the University of Michigan, described the results of a study in which 20 bariatric surgeons submitted videos that demonstrated their surgical technique. The recordings were rated by several peers on a 1-5 scale, where 1 was the skill expected of a general surgery chief resident who hadn’t yet performed this complex operation, 5 was that of a master bariatric surgeon, and 3 was that of the average bariatric surgeon. I’m not sure why, but I would have naively guessed that — though the recordings might reveal a quirk or two — everybody’s technique would be pretty good, and not terribly dissimilar.

I would have been dead wrong. The reviewers, who were blinded to both surgeon and institution, used a lot of the terrain on the grading scale. Surgeons in the top quartile averaged a 4.4 (on 5 domains, including exposure, flow, and gentleness), while the lowest quartile surgeons had a mean rating of 2.9. The ratings did not correlate with years in practice, fellowship training, or teaching vs. nonteaching hospital.

Instead, they correlated strongly with surgical volume: lowest quartile surgeons averaged 106 bariatric procedures in the prior year, while highest quartile performers averaged 241. (Of course, this doesn’t answer the age-old chicken vs. egg question of whether better performers get more cases, or more cases make better performers. But it does support the use of volume as a proxy for quality, at least until video is more readily available.)

Here’s the amazing thing: after adjustment for any patient differences that might have influenced outcomes, surgeons who were rated in the top quartile technically had far better outcomes than those in lowest quartile. The better technicians got through their cases more quickly (98 vs. 137 minutes) and had lower infection rates and lower overall complication rates. Their patients required readmission, return to ER, or reoperation less than half as often as the patients of their less skilled colleagues. Finally, their mortality rate was one-fifth as high (0.05% vs. 0.26%), all significant differences.

Importantly, all of the participating surgeons were volunteers, and the videos were selected by the surgeons themselves. This makes it likely that the variations observed in the study might understate the real world differences. Scary stuff.

In 2011, Atul Gawande, in one of his wonderful New Yorker pieces, wrote about coaching. Gawande described how he invited a senior colleague, a retired Brigham surgeon named Robert Osteen, to observe him performing a thyroidectomy, a procedure that he had done roughly a thousand times. One piece of Osteen’s advice bore a remarkable resemblance to what I heard from Artie McNickle on the driving range:

Osteen also asked me to pay more attention to my elbows. At various points during the operation, he observed, my right elbow rose to the level of my shoulder, on occasion higher. “You cannot achieve precision with your elbow in the air,” he said.

Yet as helpful as Gawande found the coaching, it was awkward to have the observer in the room — seen by peers, other staff, and especially patients. One patient, seeing Osteen in the corner, asked, “Who’s that?” Gawande called Osteen “a colleague,” adding, “I asked him along to observe and see if he saw things I could improve.” After seeing a look on the patient’s face “somewhere between puzzlement and alarm,” Gawande added, “He’s like a coach.” The patient did not seem reassured.

Just think how much easier it would have been if Osteen had been watching a video feed. (In fact, Gawande’s Brigham group has been experimenting with just that.) Similarly, in the famous Northern New England Cardiovascular Study, cardiovascular surgeons traveled to each other’s hospitals to watch their peers perform surgery, providing honest feedback about matters ranging from surgical technique to teamwork. The result: a 24% decrease in surgical mortality. As fabulous as these results were, the study — performed more than 20 years ago — has not been replicated, undoubtedly because of the hassle and expense of the intervention. Here too, video could make such observation and feedback far more routine.

I had the chance to try Google Glass a few weeks ago, through a company that Google is working with to identify “use cases” in healthcare. Well, here’s one: how about if novice surgeons — or all surgeons — periodically did operations that were observed, in real time, by certified experts, who then provided them rapid, perhaps even real time, feedback.

While the primary use of such information should be for coaching and improvement, after watching the videos (examples that illustrate good and poor technique accompany the Birkmeyer article), I would not want a lowest quartile surgeon rummaging around my abdomen. (Even I, a complete novice when it comes to surgical technique, could easily distinguish between the assured, polished motions of the experts and the hesitant, clumsy moves of the lower performers.) Putting on my ABIM hat, this study suggests that we need to move briskly into measuring the technical proficiency of proceduralists … and perhaps everyone else. One could easily imagine differences like these in the quality of history taking, physical examination, and end-of-life discussions.

As with all quality measures, the primary use should be for improvement. But the surgeon whose technical performance remains poor even after feedback and practice should not be certified, at least for that procedure. In a recent interview, Birkmeyer endorsed this stance, while pointing out the many knotty issues it raises, such as where to set the threshold.

The instinct to go to the video is an area in which the young ‘uns have a big advantage over geezers like me. My wife did a story in the New York Times last year about the digital revolution in doctoring. The Times brought along a videographer to follow one of our VA-based teams on rounds. When the team was visiting one of its first patients, the 78-year-old man had a grand mal seizure. Several of the team members gathered around the patient to attend to his airway and circulation. An intern stood at the foot of the bed and promptly pulled out his cell phone.

Did he go on Epocrates to check the dose of Lorazepam? On UptoDate to find the management algorithm for new onset seizures? No, he began video recording the seizure. A Times reporter asked him why. “I wanted to record his activity…. So rather than describe to [the neurologist] what took place, I can just show them a video of what took place, and they’ll be able to assess better and treat the patient.”

With tools like smart phones and Google Glass, the technical obstacles to the widespread use of video are beginning to melt away. Of course, other barriers — patient privacy, clinician pride, archiving, cost, the “eewww” factor – will remain. Let’s work through these quickly, so that we can take full advantage of this remarkable tool to improve our patient care.

Bob Wachter is professor of medicine, University of California, San Francisco. He coined the term “hospitalist” and is one of the nation’s leading experts in health care quality and patient safety. He is author of Understanding Patient Safety, Second Edition, and blogs at Wachter’s World, where this article originally appeared.

Comments are moderated before they are published. Please read the comment policy.

  • Mika

    You let random interns just go around recording unconscious, seizing and helpless patients at their most vulnerable, on their personal cell phones?

    How does the hospital then control what happens to that data: where it is stored, where it ends up getting transmitted to, how secure the device is, etc?

    What do we do when the intern accidentally or otherwise hits a button and sends said video to everyone in his contact list, or to his Twitter/Facebook friends, or to YouTube?

    I don’t think most patients would be impressed that there’s video of them floating around on the subways and in the streets, in the personal cell phones of people who aren’t even their doctor, people they might not even have known were going to be filming their medical adventures.

    • EE Smith

      “You let random interns just go around recording unconscious, seizing and helpless patients at their most vulnerable, on their personal cell phones?”

      It’s a bit of a worry, isn’t it.

      Also a worry is the attitude that patient privacy is a “barrier” to be “overcome”. I’m pretty sure most patients wouldn’t see it that way.

  • Margalit Gur-Arie

    Yes, let’s work through other “barriers” like patient privacy and cost “quickly”, so we can have “someone in Bangalore, India reviewing the video feed every hour and sending back compliance data”, particularly videos of “history taking, physical examination, and end-of-life discussions”.

  • Tore Nicolai Fjelldal

    I would really like that one of the staff recorded a seizure to show it to the spesialist. I also hope the hospitals would start using video for improvement and education. The more knowledge the staff has got the more safe would any treatment become.

  • guest

    A grand mal seizure is a grand mal seizure. Unless you are doing simultaneous EEG telemetry, video of the seizure is of no clinical utility to a neurologist or any other doctor. Kind of surprising that Dr. Wachter doesn’t know this.

    • FEDUP MD

      He would have been better off with a stopwatch, actually, to help them. No one can ever seem to say how long a seizure lasts…except for maybe experienced nurses and ED docs.

      • guest

        What a good point. You would think there would be some sort of best practices standard where anyone witnessing the onset of a seizure is expected to check their watch and record the time. You would think someone in academic medicine would be working on developing that, rather than goo-ing and gah-ing over how neat it is that the “young-uns” have a compulsion to whip out their cell phones and record every little thing that happens around them.


    I’m sure in today’s litiginous, corp-med, insurance controlled world, that this data would only be used for good. Not for one more way fof bean counters of all kinds to further erode physician control over their own practice of medicine.

    • Guest

      Also, if there were a bad outcome, and the patient’s lawyer found out that various doctors and onlookers involved or even not involved in the patient’s treatment had been recording the patient on their personal electronic devices, wouldn’t they be able to subpoena all these personal electronic devices, and any iCloud/Google accounts the videos might have been uploaded to?

  • southerndoc1

    “I am 100% confident this doc meant well with this post”

    Actually, if you become familiar with his writings and his work, you see he really doesn’t mean well. He is completely dedicated to subsuming the needs of individuals (both patients and physicians) into those of the all-powerful corporate/administrative/governmental system.

    It’s really pretty frightening when you think about it.

    • Claire

      I stand corrected! I took your advice and read some of his other writings. I agree with a few things he says but, taken as a whole, his interests seem to me to be aligned with something other than patients and the physicians who care for them. Patient privacy and autonomy as well as physician influence and independent decision making seem to get tossed out the window. I am not surprised that he is affiliated with a teaching hospital after reading his other posts as they often are the biggest offenders in this arena. Thanks for the clarification!

  • karen3

    Wonderful article. It is very difficult to acknowledge weaknesses and improve without meaningful and fair feedback. And that seems to be lacking in medicine.

    But the point below is fair. I do not want a doctor to have a film of me on his or her personal Iphone. Period. It needs to be appropriate equipment and not uploadable to youtube.

  • Brad White

    I suspect that many of the objections to this article are rooted in clinician pride (or insecurity?) masquerading as patient privacy concerns. The point of the article – that the use of video can provide powerful feedback – is quickly dismissed because video could be misused if poorly managed. Rather than concentrate on finding solutions to these concerns these commentators would rather shut the whole idea down and throw the good out along with the hint of the bad. And we wonder why healthcare is stuck in the 80′s.

    • guest

      Actually most of the objections have to do with the author’s uncritically sanguine attitude regarding the unconsented videotaping, with a non-secure device, of a patient experiencing a seizure. Speaking as a patient myself, I would be appalled to find that I had been recorded on someone’s personal device while unconcious and without my prior permission.
      I would add that videotaped sessions with patients were a standard part of my residency training so I don’t have a problem with that application. Of course, those patients were consenting participants in the process. Many training programs these days use video, it’s not a novel approach at this point. The practice of casually recording on personal devices, on the other hand, is increasingly common, and does concern many of us.

      • Brad White

        Good job making all the comments about a single paragraph in the article! You are actually making my point rather nicely. I of course agree with your assessment – my point is that the healthcare professionals that are commenting seem to want to pay attention only to that one piece and ignore the fact that they, like any skilled operator in any other industry, could benefit greatly from recorded feedback.

        Glad to hear they use the technique in medical school. The only problem is, once you are out of school the real learning starts!

Most Popular