Last week Wired Magazine hosted a Wired Health Conference (subtitled "Living by the Numbers") in NYC. As there was lots of buzz about Dr. Eric Topol's talk, "Information into Action," I watched it online after the fact. It's a very interesting interview with Wired's Thomas Goetz, in which Topol describes a handful of new technological tools that he feels doctors should be using. In fact, he states that "we're trying to get a lot of things out into the real world of everyday care."
Now, I'm a practicing doc interested in finding and adopting technology that will improve geriatric care. So whenever I hear about a new tech tool recommended to doctors, two questions come to my mind:
- Does this sound like it will improve the clinical care of my patients (frail and vulnerable elders), at a reasonable cost?
- What would it take for a doc such as myself to integrate using this tool into my workflow?
For those who haven't heard of Eric Topol: he's a well-respected cardiologist and expert in using genomics and digital technologies to personalize prevention and management. Among other things, he's famous for using a nifty iPhone add-on called AliveCor (turns your iPhone into a mobile ECG) on a plane to diagnose a case of acute coronary syndrome. He also no longer uses a stethoscope, preferring to use a Vscan (a pocket-sized ultrasound) to visualize valves as part of the physical exam. His book, "The Creative Destruction of Medicine: How the Digital Revolution Will Create Better Health Care," is often referred to as a "must-read" for those interested in the bright healthcare future that technology will bring to us. (Confession: haven't read it yet, but probably will some time soon.)
The tech ideas described by Topol and Goetz during this talk (my thoughts on each idea are in purple):
- Pharmacogenomics-based prescribing. How exactly might you personalize medical care based on someone's genome? Topol proposes pharmacogenomics as a good place to start. The FDA has a list of drugs with known genomic interactions (I found it here; 117 drugs on it). For instance, according to the talk, Plavix (clopidogrel) doesn't work in 1 of 3 people due to a genomic factor that can be identified through www.23andme.com.
- Wow, between stents and strokes, many older adults have an indication for Plavix. I'd heard that some are resistant, but if a commonly used drug really doesn't work for a third of people, then considering pharmacogenetics would improve clinical care for my patients. Cost-effectiveness as of yet unclear to me.
- But how am I supposed to know which drugs require special dosing? Hard to keep track unless a specific genomic testing recommendation is going to start popping up in the e-prescribing systems. (Vague warnings that there is a genomic interaction without further specifics on what to do are not very helpful to providers.)
- Added twist: Uptodate.com, the clinical decision-support tool I use on the fly, has a topic page on clopidogrel resistance, but the summary recs "suggest against routine testing of patients for "clopidogrel resistance", whether by in vitro testing of platelet function or by genetic testing for CYP2C19 poor metabolizers (Grade 2C)."
- Uh-oh. Discrepancy between what I hear from Wired Health and Uptodate. I could dig through the literature to arbitrate, but in general, I am just going to go with the clinical decision support tool I trust.
- Point-of-care genotyping. Topol explains that a barrier to using genomic information is the usual week-long delay in getting test results. He proposes that one is more likely to act on genomic information if one can get it at the point of care (POC), rather than having to wait a week. He's been working with a company called DNA Electronics, which is developing a rapid POC genotyping test. They are envisioning pharmacists doing the test (customer gets to shop in the store for 15 min while awaiting results), and then being able to help match up "the right person with the right drug and right dose."
- Hm. Unclear whether the test would have to be ordered by a doc, would get ordered by insurance, or would be up to the patient to request. There are pluses and minuses to all three options; suspect the issue will be heavily wrangled in next few years.
- The value of increased access to genomic testing would really depend on what was being genomically tested for, and whether acting on that information has been shown to improve outcomes. (As you can see from the example above, there is clearly debate on when genomic testing adds value.)
- Workflow question: when test results are available after 10-20 min, does the pharmacist call a doc, or dispense based on a protocol? How are the docs notified?
- Nanosensors in the bloodstream. This one's a little further off in the future, but this source of Big Data's coming within a few years to a bloodstream near you. Scripps Digital Medicine has a grant from Qualcomm Foundation to work on these nanosensors (90 microns), that would be carried by the bloodstream and embed in a capillary somewhere. The sensors could track a bloodborne signature of pathology, like signs of an impending heart attack or autoimmune attack, and send the information to a nearby smartphone.
- This is intellectually intriguing, but in practical clinical terms, I've already written about my concerns with apps sending data to doctors (we're not currently equipped to deal).
- There is also the question of what kind of continuous data monitoring is appropriate for the geriatric population; what I think would help me provide better care is the ability to query a data stream related to symptoms and behavior.
- AliveCor. This is the iPhone-into-ECG technology mentioned above. Topol uses it all the time, despite it not yet being FDA-approved. (Doctors, you can get one here if you pretend to be a veterinarian.)
- I do have a clinical need for something like this. I see a lot of homebound older folks who otherwise do not have easy access to an ECG. And I often hear irregular heart sounds, and wonder if they are having ectopy versus afib.
- But assuming the FDA eventually approves something like this, who will help me interpret the rhythm strip? There was a time during my UCSF residency when I got really good at reading ECGs. (I once pissed off a bigwig attending by questioning the ECG machine's automated interpretation, which the attending had accepted. The cardiology fellow sided with me, but I was still left in deep diplomatic doo-doo.) Eight years later, I'm hardly ever called upon to interpret a rhythm strip on the fly, and I don't feel comfortable doing much more than identifying afib. I would certainly feel better using this if it came with a feature allowing me to send to someone who could confirm the interpretation, or some other form of clinical decision support. I don't need more work or uncertainty to deal with.
- Wearable continuous glucose monitor. This one uses a 27 gauge needle and you can continuously follow your blood glucose on your smartphone. Unclear to me how it's different from what some diabetics currently wear (other than talking to the smartphone). Topol sounded like he's envisioning regular people wearing these and continuously getting feedback on how their diets are affecting blood glucose.
- This probably has some value for the exploding population of diabetics and prediabetics. For my own older patients however, I can only envision recommending this to the truly brittle diabetics, or others in which continuous monitoring is shown to improve outcomes.
- This is yet another source of Big Data, so we again need to consider what patients will expect of physicians, and how we can equip physicians to deal with this data stream.
- Vscan. "Why would I listen to lub-dub?" asks Topol, when a handy pocket ultrasound can just show you just what the heart valves are up to. Topol believes that this should be "part of the physical exam." He points out that this could save a lot of money by replacing some of the thousands of $800 echocardiograms that are done every year. Note to medical educators: Topol says that Mt. Sinai's med school is now giving Vscans to the entering students, rather than stethoscopes (um... and the lung sounds?).
- This is basically offering patients a POC abbreviated echo. For frail elders who are already struggling with too many appointments and getting to clinics, this could be beneficial, and probably cost-effective (unless all the POC echos generate follow-up complete echos).
- But how often will I find something using Vscan that changes management without requiring a full echo?
- And how do I get to the point of using a Vscan? How much training would I need to feel comfortable making a preliminary interpretation of what I see? Who is available for a backup read? How do I document the findings in my EHR?
- Who's going to pay for me to have this Vscan anyway? I'm in private practice, as are many primary care docs. Am I supposed to pay to get this because this will bring more patients into my office? (Um, a geriatrician always has more than enough demand for her services, and there's a shortage of primary care docs in general.)
My conclusions on Topol's talk:He's got a lot of interesting ideas on how we doctors might be using technology soon. He is certainly promoting these ideas to consumers before we know if these technologies will improve health outcomes. I can't blame him for this; the usual methods of conducting peer-reviewed research can't possibly keep up with the pace of technological innovation. A little consumer pressure might be what we docs need to up our game and move into the 21st century.
Still, before society invests a lot of money into healthcare driven by these technologies, the outcomes issue is important. Technological advances are one of the primary drivers of increasing health care costs, and historically have not always generated cost savings, although that's not to say the right technologies couldn't do a better job of this in the future. Needless to say, it's not the job of tech companies or most tech evangelists, including Wired, to ensure that adopting new technologies bends the cost curve. They provide the tools and spread the word. Policymakers and payors have to figure out what to do with them.
Plus there are special considerations related to geriatrics and the care of medically complex adults. Geriatrics in particular is riddled with examples of excess technology and diagnostic efforts leading to harmful excess care, and unimproved outcomes. So I find it hard to not feel cautious when thinking about implementing these technologies in the Medicare population, whose fate is about to make or break the nation's financial outlook.
Last but not least, there are the nitty-gritty issues related "getting this into the real world of everyday care," which means getting everyday docs to use these new technologies:
- How do we keep up with an exploding array of new diagnostic and therapeutic techniques? It's not just about rebooting medical school; how will we reboot the million doctors already at work? We need viable strategies that allow us to keep up with the changes (assuming the experts agree on what new technologies we should be using). The week of paid CME time that many get is not enough.
- How do we get time and infrastructure to help patients with the data that they will apparently be bringing to us?
- How does society plan to make it viable from a financial and workflow perspective, for us to adopt this neat new diagnostic equipment in our day-to-day practice?
Actually, I thought about ordering the Alivecor last night. I have a homebound and confused elderly patient who could use an ECG, but I haven't yet found a mobile ECG service to give me some info on what his heart is up to. He is anxious and desperately wants to avoid going to the ER, so pursuing an ECG under current conditions would impose a significant burden on him and his family.
I'll keep thinking about it for now. I do have a cardiologist friend who would read the strip, if I'm able to securely email it to her.