• HealthITAnalytics.com > Articles > Tailored analytics bring power to population health management

    Tailored analytics bring power to population health management

    Author | Date February 3, 2014
    Population health management requires providers to pinpoint the needs of individuals by reading the patterns created by tens of thousands of similar cases using algorithms and intelligent prediction engines churning away somewhere behind the scenes.  While the first tools designed to aid this science were geared towards data scientists and informaticists, the seismic shift towards accountable care and patient-centered medicine has necessitated a change in the way physicians, nurses, and care coordinators interact with this valuable cache of patient data.
    Analytics tools aren’t always cheap, and they aren’t always easy for clinicians to wrap their heads around. Dr. Peter Edelstein, Vice President and Chief Medical Officer of Health Care with MEDai, a LexisNexis Company, spoke to HealthITAnalytics about how population health analytics tools are being redesigned to fit the needs of the modern clinical world.
    What are the changes in healthcare that make population health management a necessity?
    It’s kind of a fun time to be in population health management, since it’s at the forefront of the changes associated with healthcare reform.  We have a whole new group of players trying to figure out, first of all, what the heck is population management, and secondly, how they can do it.  If you ask ten people to define population health management, you’ll get twelve definitions, as we say.
    I’m a surgeon by training, and now I’ve spent more than a decade in the business community, but, honestly, the bedside providers like myself really weren’t thinking in terms of population management until a few years ago.  We had plenty to do just taking care of the population right in front of us.  But what’s happened has been a big shift in how we think of healthcare.
    Now, we think it should be like the rest of the economy, where you actually pay for a high quality outcome.  You don’t pay more for lousy outcomes.  Healthcare has been backwards for some time on that.  If I don’t do a good job in the operating room and you need another operation, and then you’re readmitted and you go to the ICU, people get paid and hospitals get paid, and the payers pay for all that.  And at some point someone said, “Gee, that just doesn’t make a lot of sense.”  I think foundationally everyone sees the importance of that paradigm shift, but now it’s legislated to happen very quickly.
    How do analytics intersect with population health management?
    There’s a whole new philosophy in healthcare, which is to help people all along the healthcare spectrum help identify groups of patients where you have an opportunity to intervene over the next 12 to 24 months to prevent something from happening by giving them better quality care.  That makes them happy and the physicians and nurses happy and ultimately saves the system money.
    And that’s a challenging field, because that’s not just analytics.  Clinical analytics says, “Here’s what happened with your patients over the last 6 or 12 months.”  That’s important to understand, but it’s sort of like having a sign on the road after you hit a speed bump that says, “You just hit a speed bump.”  It doesn’t help you much.
    Predictive analytics says, “Hey, given all the things that we know about your patient population from the past couple years, there are a number of ways to slice and dice the information based on what’s important to you.”  Is it preventing emergency room admissions?  Is it preventing a lot of money spent here and there?  Is it looking for patients who need extra contact?  Depending on what your goals are, you can say, “Here’s a manageable size group of patients who haven’t had the bad stuff happen yet, but we think it will happen in the next 12 months.  So intervene now, before the dialysis or the diabetic wound.”
    And that’s what the predictive part does, and that’s been our goal for years.  We’ve had products out there for years, but we have a new product that’s just being released this quarter that’s really tailored for a lot of new users.  It used to all be data analysts.  There was some guy in a building for Blue Cross, and he said, “Here are 5,000 people who we could save money on.  Best of luck.”  Now it’s care management teams.  Nurses and social workers and physicians are working with patients, and they have very different skill sets and needs, and so that’s what our new Population Health Monitor product can do.
    How do new analytics tools need to rethink the way patient populations are presented to providers?
    Most of the population health management tools on the market were made for data analysts.  So, you’d say you have 1.6 million patients in your population.  Let’s start with your 55,000 diabetics.  You would ask who spent the most money last year, and you’d end up coming up with a number of like 15,000 diabetics.  

And there are two problems with that.  One is there’s no single care management team or group of physicians who can handle 15,000 patients.  It’s an insane number.
    Second of all, there’s a problem with just saying, “Who spent a lot of money last year?” which is the most common way groups do it.  What if those patients spent money because they had renal failure and now they’re on dialysis?  There’s not a lot of intervention opportunity there.  What if they spent a lot of money last year because they had diabetic foot ulcers and now they have amputations?  

You can’t intervene on people who are all the way at that end of the spectrum in the same way you can for people who are less severely ill.
    So, we’ve created these predictive tools, and the user can decide how to weigh those predictions in picking a population.  Let’s say that my hospital or my group of doctors or my payer doesn’t want to just look at 55,000 diabetics.  They want to look at who had the highest predictions of emergency room use next year or unanticipated hospital admissions.
    We can tell you who, in that group, are the least motivated based on their history.  We’re not interested in seeing who spent the most last year.  The users themselves can customize our predictions weights to meet their needs.  And let’s say you do that and the tools says that there are 1,200 patients out of the 55,000 that we should go after.  

Twelve hundred is still a lot if you only have seven nurses.  What’s next?  In a little box, you say, “Okay, let’s look at 300,” and it takes the top 300 that most meet your requirements.  So it allows an off-the-shelf type product to be extremely individualized for your hospital, your doctors, your network.
    And the truth is, not all of our clients have the same goals right now in their population health management programs.  They just don’t.  So it’s very powerful, and it’s very personal for them, and instead of just saying, “Here’s a large group of patients,” we can say, “Here is the right size in terms of number that you can deal with based on the predictions that matter to you.”
    It’s a big, big feature compared to how the tools have been built.  When I first came into the business, they said, “Well, would you use the current tool as a physician?”  I was honest.  I said, “No, it wasn’t built for me.”  This one’s built for me and for my care management nurses, and it’s just really quite powerful.
    Do providers need to be convinced about the value of these tools, or is it more about the expense of procuring the right technology?
    You know, the biggest challenge isn’t that they don’t get it.  They do get it, and they do want it.  But right now they tend to be so far behind on EHRs, CPOE, an Meaningful Use.  It’s a matter of asking how they’re going to prioritize it.
    It’s very uncommon for us to leave a room of providers or their business team and they go, “Yeah, I don’t see any role for this.”  They’re all excited, but then someone will say, “Yeah, but we still are implementing Epic, and that’s another 11 months.”  So, the real battle, honestly, is trying to show them that there is so much potential ROI here.  We ask if there is any way they can do this sequentially with some of the other stuff, or even ahead of it.  And it just depends on the provider, but that’s going be the biggest challenge over the next 24 months as providers are trying to catch up with everything.  There’s a lot to do, and they really have to watch their nickels and dimes.
    Related White Papers:
    Related Articles:





    Leave a Reply

    Sign up for our free weekly newsletter and stay up to date with tips and advice on:

    • Population Health
    • Clinical Analytics
    • Business Intelligence

    Our privacy policy