In Part 2 of the interview, Bradley speaks to what makes analytics in healthcare unique, the breadth and integrity of provider data, and how better financial interactions with patients may improve their perception of healthcare providers
For the first time in our ongoing C-Suite Q&A series, Healthcare Analytics News speaks with a Chief Data Scientist. ZirMed is an analytics company trying to refine the process of provider reimbursement through the use of predictive modeling. Their chief data scientist is Paul Bradley, who spoke with us in July.
In Part 2 of the interview, Bradley speaks to what makes analytics in healthcare unique, the breadth and integrity of provider data, and how better financial interactions with patients may improve their perception of healthcare providers overall. Part 1 can be read here.
You’ve done this sort of work in other spheres before, so what are the particular differences and difficulties of doing it in the healthcare space?
A lot of my previous life in modelling was in the ecommerce space, where people are coming to an online store and browsing and buying…while web browsing data and ecommerce data can be challenging in terms of finding the right pieces of information to extract for modelling, there really isn’t an issue of how it’s collected. It’s collected by machines, there’s not really any integrity issues at the low level of data collection.
As I moved into healthcare and I worked a lot with the administrative and claims data, there’s a little bit more variance in how the data is collected and how it’s used. Primarily claims are put together to get reimbursement from the payer, but secondarily they’re also a great source to look at historic care patterns and trends. When you get to that level, and you’re looking at claims across a wide geographic area with different payers, sometimes the data integrity that was filled in on the claims can differ across regions.
One of the challenges that my team works on is trying to bring that data to a consistent level across a hospital system, so we’re able to accurately see care the might be given in two different regions and speak about it, for lack of a better word, in the same language.
I wanted to go back and talk a bit further about data integrity. How do you go about ensuring that you can get the highest level of integrity in your data, and how do you validate the result from it?
One thing I do want to point at, at least when it comes to the charging data that is put onto a claim, typically across the industry hospitals are about 98% correct. 98% of the time, they’re not missing billable things being on the claim. As a data scientist, what that means to me is that there’s actually a fair amount of signal around the charging practices of a hospital or a provider’s location. What we do is we leverage the signal in that 98% of the data to identify these historic patterns and trends around charging behavior. To give an example: talking about charges for a pacemaker, I might have a cardiac event and I go into the hospital for a few days and I end up with a pacemaker. So 98% of the time, that pacemaker is correctly billed, and we can find patterns and trends around that, like what are the other drugs and procedures that have been given to similar patients in this scenario. We essentially start to build a profile of what common things occur when a pacemaker is provided to a patient.
There’s a lot of information in the data for us to derive that, and then we’re able to exploit it for the other 2% of the time. It typically that there’s usually a person in the process who is deriving codes to go in the claim, they’re either taking dictated notes from a doctor and looking to derive codes from them or they’re looking to text of a medical record. Because there is a person in this process, there’s just variance in how the coding happens. Analyzing this 98% signal in the data, we’re able to really zero in on cases where a number of the drugs and procedures that are often done in, say, that pacemaker procedure are on a claim, but actually that $15,000 charge for the pacemaker isn’t…the statistical models were raise a pretty big flag and say ‘this looks odd.’
Thinking about the problems that you’re trying to solve, let’s imagine that your company and companies like it get their way entirely, a total market dominance, all goals achieved. What does the payment landscape in healthcare then look like, at that endpoint?
If we were able to provide these analytics and make sure that every claim submitted was properly coded, I think what you’d be seeing is a lot less of the “churn” around claims and denials in healthcare. If we can more automatically take the information that is captured in the providing of care, and derive the language to go on to a claim and put it in, well, the likelihood that a claim is going to be denied is much lower, so the likelihood of being paid is much higher. That’s where I would see things going in the near term, if these systems were all in place and all functioning really well. I don’t often go online to Amazon and have to think about the financial exchange between me and Amazon, it just kind of happens. In healthcare, it’s a more complex thing, but we can march along that process to make it a little more standard and, with some of the efforts like ICD-10 and more common code sets, or more common ways to represent clinical care, that’s just one step along the way.
Do you think it would improve the way patients perceived health systems?
I think, when it comes to the financial aspect of healthcare, yeah. Because patients are bearing more of the burden of their healthcare, they’re now being talked to and addressed in certain ways and, going back to that self-pay predictive model I explained, that gives providers some predictive intelligence to connect with patients who are in that group. For me as a patient, either I paid my bill or I was getting a collection notice. Now, we’re starting to have a conversation, which I think is just going to continue down the road to having a closer relationship between the provider and patient. And I think that also extends not only to the financial relationship between the provider and patient but also to the clinical relationship between doctor and patient, when you start to think about coordinated care and some of the places where healthcare may be going. This is just a part of that conversation, bringing the patient and the providers together in a tighter relationship to improve things in the long run for both groups.
There’s certainly volatility as to what healthcare payment policy is going to look like in a year, or two years, or however long down the line. How much does that level of uncertainty impact you?
It definitely has an impact, but quantifying the uncertainty of what payment plans might look like 12 months or 24 months away might impact what’s going on with ZirMed today…I’m not really able to give you a quantifiable answer. We are continuing to march forward with leveraging our predictive modeling technology and pulling patterns and trends from historic data to better capture charging practices. A lot of the result of that modeling does bring in additional net money to our clients, but in a sense too it’s also an automated safety net regarding correct coding of the care that they’re given.
Today there’s a monetary return on that, but it’s also setting up a provider for a great baseline when it comes to historic coding practices. If payment models start to change or do change, a provider knows where they’re at and how that change may impact their reimbursement going forward. We also provide some analytics around that scenario I just talked about, called Contract Management Modeling, there we are looking at changes in a contract between a provider and payer. If there’s a change to that contract, what it would look like on the provider side. That’s another piece of analytics that’s giving the provider insight when they’re going in to discuss possible changes with the payer.
What do you get out of this, what do you value most about this field?
That’s a great question, and also one I wasn’t expecting. First, I’ll put my researcher cap on: the predictive modeling problems we are looking at are some of the most challenging I’ve seen across industries. Healthcare is a complex thing: people are not products, and what works for me might not work for a similar condition in a different patient. There’s a little bit of the uniqueness of healthcare that makes it challenging. There’s also something in that we all, as people, interact with that system, so I have some knowledge of what happens there.
At the end of the day, helping providers to be properly reimbursed…there is a feel-good there. We’ve had stories of some hospitals who might be partially funded by local government, and the additional revenue that we’ve brought them has allowed the local government to lower tax rates. Every now and then I get to hear from a provider about using the extra revenue that we bring them, and we’ll hear it’s allowed a provider to put more nursing resources into a certain area, or something like that. At the end of the day, I think we’re trying to make the system more efficient and even a little bit more fair, in the sense of making sure that a provider is being paid properly for the care they have given their patient.