This past week, Google had its annual developers conference, Google I/O. One of the more provocative talks, called “The End of Search as We Know It,” was by Amit Singhal, who is in charge of search for Google.
The vision, as described by Amit, is that instead of typing words into a box on a website or mobile app, we will have conversations with Google, enabling a much more personalized, refined experience. The holy grail, of course, is that Google analytics become both predictive and prescriptive, serving you content that is just right for you and anticipates your needs.
It seems there is a race on now to achieve this vision. One could argue that Amazon, Apple, Facebook, Pandora and others are all in the same mode. Best I can tell, the promise these companies are floating to advertisers is that their ads will be served up to that focused slice of the population that will find their product relevant in the moment.
If you apply this thinking to healthcare, several controversies/topics come to the fore.
Is Google competing with IBM’s Watson? Undoubtedly yes. On the other hand, I’m guessing Google is disenchanted with the consumer health space after the demise of its personal health record (PHR). And IBM seems to be focused on clinician decision support. So early in the game, with respect to healthcare anyway, maybe there is not much competition. The path for clinician decision support is clear and the market obvious, whereas the path and market for consumer health decision support are blurry.
To achieve their vision of having a conversation with you at the level of search, Google has to know lots and lots about you. Some people are spooked by what they track now. Will there be a way for them to track all of the other necessary data points without running afoul of the privacy lobby (see the exchange between Wally and Catbert below)?
Can we achieve the same vision for consumer health information (i.e., make it highly personalized, motivational, caring, and eerily anticipatory)?
At the Center for Connected Health, we are banking on it. A big part of our research agenda for the next several years will be in this area. Current work in type 2 diabetes and cancer pain control represent a start in this area.
Perhaps the most penetrating question is whether there is an economic imperative for personalized prevention that is as compelling as the one for personalized search. We know the economics of advertising and the ultimate goal of targeting every ad to someone who is a qualified, interested customer – someone who really views the product as a solution to their problem.
I have been thinking about the counterpart in health….it is not as clear to me. Stated another way, why have we not managed to devote the kind of resources Google is devoting to personalized search to personalized prevention? For instance, if I can use your genetic analysis to predict your risk for type 2 diabetes and create a uniquely motivating self management program for you, who will find that economically appealing? You? The government? Your employer? Your health plan? Your doctor? It’s not so clear.
Of course we should all want to be healthier, but we all know how that goes. Short-term discipline for long-term reward? It’s a tough one. We are not as eager consumers of health information as we are information on clothes, shoes, concerts, etc.
Who else should be interested? Payers? Hard to imagine Aetna, United or Wellpoint competing with Google in this area. Even though they pay for care, they are agents of employers. I’m guessing HR executives will be more sensitive to the privacy issues than Google is.
It is a quandary. The pay off of personalized prevention will be breathtaking – you’ll manage your own prevention and your own chronic illnesses largely without the help of a healthcare provider. This will result in improved health as well as decreased costs. Who should step up to the plate to build the war chest that will attract the talent from the ‘recommendation engine’ and/or financial industry?
Your thoughts?
Is Design Important In Healthcare?
I’ve been in the healthcare delivery business now for 30 years and I can confidently say we’ve paid almost no attention to design. Systems are built around physicians’ desires and workflows, and physicians tend to be quantitative, content-focused and able to handle multiple variables at once. Perhaps that is why we have not put a premium on design. As we get closer to a world where we are engaging patients in their own health outcomes, design will be much more critical.
I am not talking about designer drugs — using highly specific molecules based on genetic signatures to attack cancers and the like — which are truly amazing. I also saw one of the first real-world examples of designer organs that I’m aware of recently where a toddler born without a trachea had a transplant of a plastic tube lined with her own cells grown in the lab.
Rather, I’m talking about design in systems used for day-to-day delivery of care. First, a couple of examples where design is wanting: A little over a year ago, Medicare put in place a payment rule that penalized hospitals if a patient is readmitted within 30 days of discharge for the same condition. CMS published a listing of what these penalties might be, based on the readmissions rates of each hospital the year before. It is a 59-page PDF file. The information is all there, but the design leaves a lot to be desired. By contrast, the folks from Health Recovery Solutions took the same information and presented it in a much more digestible way at checkmypenalty.com. Here you enter your hospital’s name and get a clear, one-page output that visually shows your exposure on readmissions, which can be understood at a glance.
The second example is the Blue Button initiative. A bold and exciting move by the Department of Veteran’s Affairs, the idea is that clicking on a blue button on a web page allows a Veteran to download his or her health information. The downside, at least in the beginning, was that the resulting file was an ASCII text file. This file format combined with the ‘medical-ese’ made for a rendering that was less than ideal for the Vets.
At the Center for Connected Health, we strive to bring healthcare into the day-to-day lives of patients as a continuous function. In doing so we’ve discovered that small design principles can make a big difference. To illustrate this, I first want to point out the relationship between engagement and outcomes. This is best illustrated in a study we published (J Diabetes Sci Technol 2012;6(6):1328-1336) on our Diabetes Connect program. In this particular paper we observed that patients who upload their glucometer readings more frequently have better outcomes. Likewise, if a primary care practice enrolls patients in the program, those that upload earliest after enrollment do better. Our interpretation is that patients who care enough to test their glucose early and often probably pay close attention to other aspects of their care (diet, exercise, medication adherence) and thus frequency of glucometer uploads is a proxy for adherence.
Here’s an analogy for you. If you shop at Amazon or Apple, you see different designs that decrease the friction of commerce (one-click purchasing or mobile in-store purchasing) in order to motivate sales. It is just easier to shop with these two vendors than many others. Likewise, if we know that frequency of uploads predicts better outcomes, we should design our programs to make it easier for patients to upload.
We had a chance recently to compare the same to variables noted above (time to first upload and frequency of uploads) using two different in-home data-gathering hubs. In one case, patients had a phone-modem-based solution where they had to plug their sensor into the modem and push a button to upload. In the second case, using a wireless device, they had only to plug the glucometer in with a cable (skip pushing the button). The results of this comparison were impressive. The design which did away with pushing the button was correlated with earlier time to first upload and more frequent uploads over time – both were statistically significant. (J Diabetes Sci Technol 2013;7(3):623–629)
Sometimes what seem like simple design tweaks make a big difference.
We have a long way to go. That also means there is great opportunity for those in the information design community to have a profound effect on the quality of healthcare delivery. In the U.S., where the price-to-quality ratio of healthcare delivery is embarrassing, that should result in exciting, improved outcomes.
Serendipity, Personalization and Connected Health
In a world where most folks change jobs every couple of years, people are often amazed to hear that I started work in connected health 19 years ago. Implicit in their amazement is a question, “Why did you stick with it so long?” The quickest response is that the work we started back then is not yet finished. A more accurate answer is a bit more complex. First, we have never wavered from a clear vision of how healthcare delivery needs to change. We’ve been able to recruit a truly stellar team of people to work with. And, we’ve had some good luck along the way. In this post, I want to tell the story of a couple of those serendipitous moments, what they meant to us and how we moved beyond them.
The first goes back to circa 2000 when the Medicare payment structure for home health changed to prospective payment. The Center for Connected Health was formed in the mid 90’s with a vision that connected health could be an antidote for capitation. Capitation never really took hold in the mid ‘90s, and thus connected health stayed largely in the realm of pilots and experiments. But, when prospective payment hit home care, the leadership of what was then called Partners Homecare saw telehealth as an opportunity to succeed under this new payment model. We then had our first real customer.
The serendipity comes in when we sat down with our home care colleagues to talk about where to focus our efforts. We chose congestive heart failure (CHF), because it was costly, prevalent and the source of a lot of home nursing visits. What we didn’t appreciate was just how suited CHF is to connected health as an intervention. CHF management is mostly about fluid management. Weight is a pretty good proxy for fluid status, and dietary interventions (largely around salt and fluid intake) are usually effective at controlling fluid. Thus the feedback loop of daily weight combined with just-in-time nursing interventions, usually by phone, resulted in a reproducible improvement of patient care, a decrease in hospital admissions and a net lowering of the cost of care. With this early success, we formed the connected health initiative and began to work in other areas of chronic illness management including hypertension, diabetes, mental health and asthma/COPD. Some of these programs were more successful than others, but the impressive results that we saw in CHF spurred us on to work in these other areas. Had we started with COPD, where the relationship between any one variable and clinical deterioration is much less clear, we probably would have not gone in this direction at all.
One of the insights from this early work was the power of feedback loops. Measuring weight in those CHF patients and following it up with just-in-time teaching about salt and fluid led to a remarkable trend in improved patient self-care. We began to design programs around the idea that objective, numeric feedback is an important part of connected health program success. The development of wireless sensors, mobile apps and shared APIs created a technical infrastructure that allows us to easily gather numeric data from our patients and feed it back to them as well as to a clinician. We showed great success with this method in hypertension and diabetes. The Quantified Self movement was born. A number of companies sprung up using various trackers to monitor health and fitness information (Runkeeper, MapMyFitness, Fitbit and our own Healthrageous are just a few examples). In each case, and in our own research, we followed outcomes for a group of participants defined from a disease perspective. And we were happy that we could increase engagement to 60-70% over the course of a program. We felt good about the combination of objective feedback loops and just-in-time clinician involvement. But something was not quite right. If our approach was so good, why weren’t we hitting closer to 100%? And why did we never do better than 60-70% at a population level? The dilemma is illustrated in the figure below.
I’ve been traveling to meetings and conferences for the last couple of weeks and heard a lot about the power of social interactions to change behavior. Lots of companies are coming up with social-based interventions – ask your friends to encourage you to get more active; compete as a team; set goals as a group, etc. I heard others say with confidence that incentives are the key to behavior change. Still others say health coaching is key and of course there are the proponents of making healthcare a game to encourage engagement. If you look at the chart above and the population-level outcomes for any of these interventions, I think you can objectively conclude they are all RIGHT and they are all WRONG. Right because for some significant part of the population, each respective strategy resonates, but wrong because for others it does not. Let’s face it: we’re far too complex and different as individuals to have a one-size-fits-all solution.
Having started with CHF, diabetes and hypertension – conditions where a physiologic measurement can be used as a tool to drive engagement — we in turn learned about how no single motivational psychology resonates with every individual.
We continue to observe, to learn and to adapt our thinking, as most recently as this past week. I was chatting with a friend the other day and she confided in me that she is underwhelmed by the use of tracking as a method for healthcare improvement. At this moment another insight crystallized for me. Some people just don’t respond to quantitative information. This friend told me how she could really see the opportunity for social media to increase engagement, but just was not convinced that self-tracking was that relevant.
This opens up an interesting new vista for us. Not only do we have to work on personalizing our motivational interventions, but we have to figure out how to engage people in other ways than with feedback loops.
Our understanding of the value of feedback loops is that they provide measurable, objective data (there is lots of good evidence that we are generally terrible at estimating quantitative information); they allow for goal setting and measurement of achievement; they generally increase awareness of the variable being followed (e.g., weight, steps, blood pressure, etc.) and they allow for some reality-check or accountability (it turns out self-deceit is quite popular when it comes to health).
So let’s hear from all of you non-quantified-selfers – those of you who do not really see quantified information about your health as interesting or motivating in the least. How do we achieve the goals of objective measures, goal-setting and accountability without using self-tracking?
Note: Situated in the heart of Boston, and within the Partners HealthCare network, we at the Center could not help but be personally impacted by the tragic events at last week’s Boston Marathon, as were citizens around the world. Much has already been said about the bombings, the incredible and inspirational reaction of our police, first responders, bystanders and healthcare professionals, and the outpouring of compassion and support for the victims and their families. There is not much more I can add, except to welcome the thoughts and prayers of our friends and colleagues, as we join together, to move forward, heal and stand strong for Boston.
It has been 3 years since the cHealth blog entered the blogosphere, and it’s been a lot of fun. I recently looked back to see if there are highlights or trends in my posts and/or if my thinking or communication had changed dramatically in 3 years. Although not dramatic, if you look at these writings over time, you can see a meaningful shift in thinking. With that background in mind, here is an attempt to stitch together a narrative of the last 3 years worth of posts.
Provider-centric use of connected health
The early themes hit upon just-in-time care, and patient accountability and ownership. Here are a few quotes from that first year:
- Accountability can be viewed as reining in freedom and no politician wants to remind their constituents that individuals are in control of unhealthy behaviors. It’s much easier to be in favor of programs that guarantee care to victims…..
- Imagine a world where a provider can tell you with near certainty how you will respond to a certain medication (typically we hear things like ‘you have a 60% chance of X outcome’) or with great certainty what your risk of contracting a certain condition is…. For so much of our health forecasting, knowing our specific genetic make up will change that….Now consider the implications of connected health. By collecting objective measurable information about you and sharing it with you to inform your self-knowledge and improve your health behaviors, we are improving patient adherence, better engaging patients in their care, and helping clinicians improve clinical outcomes. Here again, highly individualized information, in this instance, is changing how we deliver care….
- We recently did some phone outreach to those individuals who were asked by their doctor to join the program, but did not participate…. I was fascinated by some of the reasons these folks gave for not adhering to their doctor’s advice. Here are some notes from the staff member who called these individuals:
- “Patient said it is just too much work for him to do. The doctor checks on him anyway.”
- “Patient said that she does not think it is necessary for her to upload (she has done it before) because Dr. XXX does not look at the readings, and she sees him every 8 weeks anyway.”
- “Spoke with patient, who said she just “hasn’t gotten to set it up yet.” She just forgets….”
A new and durable theme appeared in a post in October of 2010, that of Automated Care. This idea — that we must extend our provider base across more patient demand, emphasizing that patients must feel cared for and outcomes must improve — has persisted as a theme since:
- Let’s call this phenomenon Emotional Automation. Let’s start a dialogue about it. Is it far fetched to think that we could parse provider work flow into those actions that truly require a real-time interaction with a provider and delegate others to technology? Can we set up systems that are extensions of our providers that will allow patients to feel cared for by their doctor but be interacting with a piece of software or a robot?….
- Our work in connected health has shown that engagement with measured information about oneself is variable and unpredictable. On the one hand, we have the ‘quantified selfers’ who are forever measuring data about their health, reporting it out and adjusting lifestyle to be more healthy – very thermostat like. On the other, we have patients that we’ve cared for at our Center who are completely disinterested in measuring and learning from biometric information, regardless of how much their doctor recommends it or how much their health will benefit.
- Can we take advantage of some of those “Darwinian buttons” that deceive us into believing we’re interacting with a person rather than a technology and combine them with some of Cialdini’s persuasive techniques with the hope of delivering compelling, motivational health-related messaging to individuals?
- In a study conducted at the Center for Connected Health where we measured adherence to activity goals, individuals who had three times/week meetings with a computerized relational agent (courtesy of Tim Bickmore, Northeastern University) had almost three times the adherence to their step count goals as did those in a control group…..
- I think the right healthcare slogan should be the same as Esurance, “technology when you want it, people when you don’t.” But our charge as connected health entrepreneurs is to create solutions that offer our patients such a compelling experience that they choose technology over people, whenever it makes sense….
Introducing analytics and genetic data into the conversation
The shift to our current thinking on connected health design is seen in writing that starts mid 2012 and continues today:
- Connected health does this too. It is the ‘phenotypic map’ that corresponds to the detailed ‘genotypic map’ the geneticists come up with. Consider if we have a population of workers and we want to incent them to be more active. Connected health can provide, at a minimum, a very precise measurement of the outcome. It enables folks who are investing in the program to see — both at a population and individual level — whether the program is resulting in increased activity.
- To sum it up then, wear and forget sensors generating automatically uploaded data provide feedback loops and insights whereas customized motivational psychology tools create the environment for sustained behavior change.
This thinking is still in force at CCH as we design the next generation of connected health programs:
- I recently became intrigued with free text analysis as a tool to learn more about an individual’s health. Tools like Wordle and Mirror.me are among the many that enable you to create interesting graphics of text. The words that are mentioned more frequently in any block of text are featured more prominently in the graphic…. I’m intrigued because of my belief that analyses like this are probably a more accurate barometer of someone’s health than what we’d learn if we asked anyone of us a series of questions. We want to look good for others and we tend to report ourselves as being healthier than we are. This phenomenon is well studied and referred to as Social Desirability Bias….
- Feedback loops plus varied, contextual messaging are much more powerful than feedback loops alone.
- Our current understanding of the power of connected health as an agent of behavior change includes the power of feedback loops as tools to drive awareness, accountability and goal setting. Layered over these feedback loops are targeted, personalized motivational messages that will drive behavior change, adherence and engagement. Various inputs such as an understanding of one’s state of motivation, one’s phenotypic data such as real-time biometric inputs and one’s genetic data will be combined and analyzed to create these personalized motivational messages.
- If we do it right, we will be able to create a world where patients and consumers can and will take charge of their own health, improving population-level health, consumer health satisfaction and lowering the societal expenditure on health.
It’s been fun blogging. I look forward to many more years and I thank each of my loyal readers for your comments and insights. I’ve learned a great deal from communicating with you through this medium.
Is Facebook a Predictor of Your Health?
I’ve written before about the power of analyzing what you write in order to predict things about your health. As more and more ‘big data’ companies and projects get publicity, it is fascinating to see how rapidly the field is growing. We are at the point where each one of us experiences the web differently when we open our browser, as our clicks and other data available on our habits and preferences are constantly analyzed. It’s all designed to get us to click or tap the ‘buy’ button. I’d like to see some efforts made to use the same approach to motivate us to click/tap the ‘get healthier’ button.
With that in mind, I was fascinated to read a recent article in the Wall Street Journal (the news appeared in other outlets as well) citing a paper published in the Proceedings of the National Academy of the Sciences that studied how using the ‘like’ function on Facebook can reveal details about your personality. Researchers first collected a lot of information from participants using standard personality inventories and psychological tests. They then looked at patterns of clicking ‘like’ on Facebook to see if they could predict any personality traits or other identifiers.
The results were startling. The researchers found, for example, that ‘likes’ for Austin, Texas, “Big Momma” movies, and the statement “Relationships Should Be Between Two People Not the Whole Universe” predicted drug use. But “likes” for swimming, chocolate-chip cookie-dough ice cream and “Sliding On Floors with Your Socks On” were part of a pattern predicting that a person didn’t use drugs. It gets better. Patterns of using ‘like’ accurately distinguished between democrats and republicans 85% of the time, between black and white people in 95% of cases and between homosexual and heterosexual individuals 88% of the time.
You might be thinking that this is both technology and science run amuck. “Why can’t you just ask me?” you might say. It is worth pointing out the preponderance of evidence that, if asked about your health, you will predictably exaggerate those facts that make you look healthier and minimize those facts that make you look less healthy (this is predictable for some other types of questions too). This is called “social desirability bias” and the phenomenon explains so much of what goes on both in the doctor’s office and when we talk casually to our friends and family about our health.
At the Center for Connected Health, we’re conducting a study now where we ask participants to fill out a short questionnaire on their readiness to move their behavior to a healthier state, followed by a period where we track their actual success (as measured by activity level). Almost to a one, we’re finding participants overestimate where they are at the beginning of the study and need to be downgraded after one month of tracking.
There are many other examples of this phenomenon in the literature. I find it liberating to think that there is a future (not too distant) where we’ll be able to calculate a health profile indirectly from records of your online and mobile behaviors (not just likes, but texts, emails, GPS data, etc.). The ‘likes’ study is a great example of how that might work.
I can just hear the voices of the privacy advocates, the volume of their objections growing as they read this. Rest assured, privacy will be preserved. This health profile might be shared with you only, before it leaks to anyone else, even your doctor or your loved ones. And if you object because it feels threatening to have someone tell you a version of the truth that might be too hard to hear, rest assured again. A message that falls on deaf ears is of no real value in improving your health. The message has to be customized to your state of readiness. That is something we’re working on too.
At some point, there may be implications for your healthcare premium costs in this sort of model. Exercising your libertarian right to privacy may lead you to have to choose higher insurance premium costs. That is if we can show conclusively that this objectively-derived healthcare profile is accurate and predictive of your costs to the healthcare system. There is a growing tension between ‘big data’ analytics and privacy. When it’s designed to induce you to hit the ‘buy’ button, I think the argument to guard privacy is an easy one. When we get it so that the design will be to induce you to improve your health, both for your own sake and for the sake of society, I’m not so sure.
What do you think?
Will ACOs Fail?
The cHealth Blog is coming up on its 3rd anniversary and during that time I’ve taken my share of pot shots at organized medicine. Most implementations of connected health are in some way disruptive to the status quo, so I can’t help but point out those opportunities and barriers.
So I was surprised at my reaction to a recent editorial in the Wall Street Journal by Clayton Christensen and colleagues. The main premise of the piece is that the concept of Accountable Care Organizations (ACOs) is misguided and that these organizations will be more likely to fail than not. I’ve been an admirer of Christensen’s work over the years and as you might expect, there is a lot to like in his WSJ piece. However, as I watch healthcare providers in our integrated delivery system deal with the challenges of payment reform and accountable care, I am more optimistic than dismayed. Let me go through the WSJ piece point by point and offer some thoughts.
For me, the first red flag is the assertion that ACOs are ‘latter day health maintenance organizations.’ The biggest difference in HMOs, as we experienced them, and ACOs are the locus of execution. The HMO was a health plan-based organization. With the exception of some of the early staff-model HMOs, these organizations were constructed to give providers financial guard rails without any tools for delivery reform. The ACO is provider organization-centric. I guess we can argue that this amounts to a ‘fox in the hen house’ situation, but as I see our own strategy unfold, I am optimistic. I see doctors saying things like, “This is how I always wanted to practice medicine.” Giving provider organizations the financial risk is step one. They can use internal payment structures to motivate doctors to care for patients in more effective, but more efficient ways. The IT tools at hand to facilitate this population-based payment management are breathtaking, from registries to data analytics to connected health. We’re using all of them at Partners and our goal is to transform care delivery.
The next logical misstep is saying that ACOs will not succeed without changing doctors’ behavior. No argument with this premise on the surface. But I was puzzled by the lack of discussion of payment reform strategies (shared savings, bundled payments, full capitation) and how these will motivate changes in provider behavior. It may surprise a number of people to learn that many of the doctors at Partners HealthCare (and elsewhere) are tired of the mouse’s wheel of fee-for-service reimbursement and welcoming of the opportunities to re-think care delivery. I see MD thinking and behavior changing. Whether it will change fast enough is still an open question.
Most ACOs are large organizations, and healthcare organizations are by nature conservative because we are overshadowed by our “do no harm” Hippocratic oath. I’d love to speed up the process here at Partners, but I am comfortable that the process is well designed and will lead to the right changes in physician behavior.
Further along is the statement that we’ll not succeed without changing patient behavior. This is where I find myself in resounding agreement with the WSJ authors. If there is a flaw in the thinking of our government ACO architects, I’d say it’s here. It’s not so much that we as providers can’t hold patients accountable for their health, though we can’t. But rather that the prevailing sentiment nationwide is that chronic illness is an accident that leaves patients as victims to be cared for and covered by insurance. The most effective way to move patient/consumer accountability forward would be to acknowledge that lifestyle plays an enormous role in our most prevalent and costly illnesses.
Further, giving citizens healthcare premium rewards for evidence of improved health, and penalties for evidence of disregard for health would be a great step forward. Connected health tools enable that vision. However, I don’t see any stomach for this in our politicians, so we will need to work within the current policy framework. Some health plans are making progress in this arena. And just because we don’t have this lever to pull, it does not mean we should abandon the whole concept.
As to the argument that ACOs will not save money, I think we have a chance. At Partners there is a deliberate and sustained effort underway to make us more lean and efficient. I’m sure other providers are going through similar processes. Is it enough? Probably not, but it’s a start and movement in the right direction.
I have to admit that were it not for my near complete agreement with the authors’ prescription for success, I probably would have ignored the piece and moved on. But the thinking articulated in the last half of the editorial is so compelling I had to write about it.
The first answer the authors propose is more use of tools such as retail clinics. Couldn’t agree more. The good news is that it appears this will happen with or without ACOs, because of consumer demand for simpler healthcare that is easy to access. However, some provider organizations, notably the Geisinger Health System, are putting a big investment into the expansion of retail clinics. I look forward to watching their results and in turn watching others adopt.
At the risk of having my telehealth colleagues throw rotten eggs and tomatoes at me, I’m going to suggest that the focus on interstate licensure as a barrier is overblown. If you look at the market share penetration of telehealth and remote monitoring programs within any given state (patients managed in this way divided by total patients managed), I think you’ll see what I mean. There is more than enough opportunity for expansion within a given state. Of course, the ideal system would not be state based, but I don’t see this as a true barrier to telehealth adoption or to success of ACOs.
Last but by no means least is the nod that the authors give to the importance of connected health. Of course, I agree with that! Seriously, if we are going to use our current provider work force to care for the current patient base, plus the millions of new folks coming into the system due to access reform, we will have to expand our tool set beyond the tired old office visit. Work at the Center for Connected Health has been building toward this vision for the last 10 years or so and we’ve managed to demonstrate better outcomes for chronic illness management, better patient engagement, and more efficient use of provider labor all with connected health. In fact, we were quite proud to see our work featured in a recent report from The Commonwealth Fund.
On balance, I congratulate the authors of the WSJ editorial for daring to buck the establishment. And I suspect if we sat down over a cocktail, we’d agree far more than we’d disagree. But in the end, I’m an optimist. I’m also pleased to see how our leadership and our providers are responding to the ACO call here at Partners, so I see greater chances of success than Christensen et al.
Where do you come down on this one?
Connected Health Predictive Analytics: A Long Road Ahead
We’re spending a lot of time at the Center for Connected Health (CCH) these days thinking about and experimenting with algorithms. It’s part of our general interest in micro-segmenting the population and creating unique, engaging health messaging for each individual that will keep them on the path to better health. Healthrageous is working fast and furious on this as well. Of course, we’re not the only ones. A number of other labs and firms are on the same journey. The vision is compelling.
However, today when you get health related messages from your insurer or another source, they are typically public health focused. Stop smoking! Get your mammogram! Get your flu shot! These three messages illustrate the challenge. I’ve been the recipient of all of them recently. I’ve never smoked, clearly do not need a mammogram and was vaccinated for influenza in early October.
I always thought our friends on the consumer web side were doing better. The first time you experience Amazon’s or Netflix’s recommendation engines, they tend to raise eyebrows. Over time, the experience is less salient. And let’s face it, it’s got to be easier to guess which type of movie I might want to watch or a book that might interest me than to predict what a really engaging health-related message might be.
At CCH we’re in the middle of an interesting trial funded by the McKesson Foundation, where we collect three types of data (a measure of readiness to change, ongoing activity data and location data) and use an algorithm to generate motivational messages based on these variables. It’s ongoing now, so I don’t know how it will turn out, but we’re excited about the possibilities. Still, it’s only three variables and only one (activity level) is continuous. My instinct is that we have a long journey ahead of us.
I got some confirmatory evidence when I read a front-page article in the Boston Globe last week. A Harvard professor picked up a sad fact about Google’s algorithms. If one searches for an African American sounding name such as Trevon, Rasheed or Tamika, you are more likely to get an ad offering you an opportunity to run a background check on that individual than if you put in a more Caucasian sounding name such as Brad, Cody or Amy.
Imagine if we goofed like this in our ‘highly customized’ health care messaging. What a disaster. We’d lose patients and consumers by the legion.
Health messaging is complex. It needs to be highly individualized, motivational, caring and of course accurate.
Indeed we have a long way to go. Although our friends in consumer, retail and financial services industries have not quite conquered this complex challenge (witness the Google gaff noted above), they are way ahead of where we are in healthcare. So, I hope some of those bright algorithm scientists can be convinced to turn their attention to connected health. We need the talent. The problem is intellectually challenging enough and the rewards will be great.
