Lauren Patrick is president and CEO of Healthmonix.
Tell me about yourself and the company.
I am an IT engineer kind of person by training. I spent about 15 years working for consulting companies such as Capgemini and E&Y. I founded Healthmonix when I moved to Philadelphia about 15 years ago and wanted to do something more meaningful than selling tickets on the internet or compiling part lists for engineering firms.
Healthmonix is a healthcare analytics company that primarily focuses on quality metrics, with the MIPS and MSSP programs in particular. We got our start developing quality measures before everybody came to know what they were. I started working with the University of Pennsylvania to put together some CME programs. To figure out where the gaps in need were, we started using the Epic system to figure out what physicians knew and maybe what they didn’t know, then worked to provide performance improvement.
We work on quality metrics and now cost metrics as well, then help providers improve in those areas.
How do practices work and think differently in a value-based care model?
Rather than looking at fee-for-service — where every time you provide a service to a beneficiary or a patient, you are figuring out how to bill for that – it is focused on how to best take care of that patient, how to best serve that patient’s needs, and how to understand what that patient wants from their care.
How do you support practices that may have a mix of value-based and fee-for-service, or that work with multiple value-based care programs?
We take in all of the data for all patients. For the Medicare quality reporting program in particular, you have to report on all of your patients. No matter who the payer is – self-pay, private, Medicaid, or Medicare — all of those patients have to be included in your quality metrics panel in order to report to Medicare and get the incentives or avoid penalties. We track what their insurance is and can partition that data. We can show you quality metrics of your Humana patients versus straight Medicare fee-for-service patients.
We’re bringing all that data in, putting it in a repository, and then saying, here are the quality metrics that Medicare really cares about. We can show you those metrics from that data. Because of the way that we’ve built the software, we can also take that data and show you quality metrics that perhaps apply to Humana. We can take that same repository of information and pull it out in all the different ways for different reporting that is required. That’s quite a challenge, but that’s part of the beauty of coming at it from a data-based perspective. We pull all that data in and then build those quality metrics to help these folks report out with as little burden as possible.
How do physician behavior and education fit in?
While payment drives a lot of the participation in our programs, I started out working with UPenn trying to figure out how we could help providers improve. At the heart of what I want to do personally is to make a difference and help these people improve. But sometimes doctors will see these quality metrics put up on a screen somewhere and they will say, “No, these are wrong.”
We say, OK, let’s drill into them. Let’s take a look. This is the quality metric. You agree that this should be the standard of care, right? Let’s look at how your patients are or are not adhering to that. From there, you have to softly get these physicians into it. We’ve taught our physicians that they are knowledgeable and are out there making day-to-day decisions. To say to them, “Maybe you need to do something different“ is really a little bit of a shock to their system. It’s a standard cycle where they don’t believe, then they accept it, and then they have to figure out how they can change.
Is it enough to provide convincing data, or does change require having people on the ground to nudge them?
It depends on the personality of the physician. It also depends on the metric. You have to make sure that the measures that you’re putting in front of them are meaningful and something that the doctors can buy into.
When we started, we were looking at A1Cs. Let’s get the A1Cs down under fill in the blank – some doctors feel that nine is appropriate, some feel that seven is appropriate. Figuring out what that metric is, getting everybody to agree to that metric, and then having them work towards that. If they feel like it’s the right quality measure, then they are much more willing to work towards it.
When we do this, we look at process measures, which would be like filling out a prior auth or making sure that you put the meds in the EHR or whatever. But it’s the outcome measures that we are all striving for. Let’s make sure that our patients’ diabetes is in control. Let’s make sure their blood pressure is in control. Let’s make sure that they can walk out of the emergency department healthy. Let’s make sure that the patient’s objectives, in terms of what they want, are being adhered to. If you put the right metrics in front of these docs, they are much more willing to buy in, but you read journal articles all the time about how doctors don’t like a lot of the metrics that are being imposed by some of these programs.
What challenges are involved with collecting data from multiple systems and then packaging it together so that it is reliable?
That’s probably one of the biggest challenges industry-wide. We work very hard to pull data out of a variety of systems. Part of the challenge that we have now is that we might be reporting for not just an individual practice, but for an accountable care organization, which is a group of doctors that have banded together to say, “We are going to take responsibility for making sure that Mrs. Jones is healthy.” We have to pull all the data from all of those various practices and put it into one dashboard. We have to say, these are the outcome measures for Mrs. Jones, and who is working on that?
It’s hard to pull that data together because some of it is in an Athena system, some is in an Epic system, and some is in a billing system. Bringing it all together is one of the biggest challenges. We don’t just bring in a file, dump it, and say we’re done. We work with providers to understand where they are putting the data. A lot of times one doctor will put it in one field and another doctor will put it in a different field, so we have to understand that we have to get it from both fields. We spend a lot of time on data integration.
Has 21st Century Cures and broader interoperability improved that, or will it in the future?
That’s the dream. Everybody keeps saying FHIR, bulk FHIR, and all the regs that have come out. But some of the EHRs are kicking and screaming. They don’t want to share their data. Some of them just don’t have it together. Some doctors don’t put the data in the right fields for a standardized mechanism for data integration to be effective.
What progress has been made with accountable care organizations?
Everybody says that’s the brass ring. That’s what we’re striving for. But I heard somebody from Intermountain say that it’s a 30-year journey. We are all working towards figuring out how to do accountable care.
CMS was a little stifled by the pandemic for a few years and the growth of ACOs didn’t occur 2020 through 2023. We are hoping that we are back on track. We see more and more patients being involved in some sort of accountable care relationship. That’s good. That’s what we want. We want somebody to be in charge of that patient’s health and to be looking at the whole patient. What we at Healthmonix are trying to do is to bring all that data together so they can see the picture of the whole patient.
Does having information available from multiple systems create new opportunities?
Yes, absolutely, and not just with EHRs. Social determinants look at where the patient lives. What sort of life does that patient have in terms of a support system? Are they in a food desert? Are they getting the sort of social support that they need? Then, combining that in. As we move forward, we’re integrating more and more sources of data so that when that patient walks into that care facility, a provider can get a much better picture of what is going on with that patient.
How do providers use the social determinants screening? Do any of the quality measures have it built in?
To get physicians started with using them, the Medicare programs, the ACOs, are giving providers bonuses for tracking those metrics. We call that pay for reporting. Then as we go forward, they are starting to factor those into measures. A lot of the measures are what we call risk based, where we take in social determinants or other patient history and give the provider credit for the fact that it’s a harder patient to take care of.
Is MIPS the only program that looks at these measures?
MIPS is a CMS program. It adjusts the payment that providers get from Medicare for fee-for-service. If you don’t participate in MIPS, you’ll get a 9% penalty on every single claim that you turn in to Medicare the following year. That’s a big hit. If you turn the data in that we compile — if we turn the data in for you, essentially — and you do well, then you can get up to a 2% to 5% incentive on every bill that you put into CMS next year.
CMS drives a lot of it, but there’s a whole rulemaking process that we participate in. We will draw up some of the quality measures that are in the MIPS and CMS programs, and then CMS will decide if they think it’s a good thing or not. Once CMS adopts it, it funnels out to private payers because there’s a certain standard of care that you want to adhere to for patients. A lot of what we do is based on science.
How does the cost analytics part of MIPS work?
MIPS decided that part of the score that you get from Medicare is based on this cost component. CMS is looking at how much is it costing to take care of your patients for certain episodes of care. When you have a knee surgery, what is it really costing CMS in terms of all the claims that come in for that knee surgery? That includes X-rays, anesthesiologist, the surgery if you’re in an ambulatory surgery center, the post-acute care that happens for 30 days afterwards, and complications. We look at that as a whole to say that the total cost of that knee surgery was X. We look at everybody across the US and figure out what the average was. If you did better than the average, then we say, yay, you’re doing great, and we give you an incentive. If you’re doing worse, then CMS will ding you for that.
What factors will be most important for the company over the next few years?
Data integration is a huge one. Can they really work to build better data interoperability? Because that will help a great deal. But I still think that we are going to need to spend a lot of time on data integration.
The other thing is that CMS is by far the leader in terms of where we’re going in terms of value-based care. Looking at the programs that they put forth, it will be interesting now with Mr. Trump to see how much he supports or does not support the movement into value-based care. I didn’t see a lot of changes in the four years of his prior administration, but there will probably be some changes in the next four years since he feels like he’s got a little bit more of a mandate. Those sorts of things will impact where we go with this.
Everybody’s favorite term right now is artificial intelligence. We report data for 50,000 providers across the US. To do these cost metrics, we gather a lot of claims data from CMS as well. We have a pretty big repository of healthcare data. Now we are digging into that data to understand the correlation between patients with great outcomes, both in terms of cost of care and in terms of quality, and all the other factors that are in there. We are trying to use AI to see if using this medication for this patient is associated with better outcomes. If you go to this kind of post-acute facility versus that kind of post-acute facility, does it impact the cost of care? I am hugely interested in exploring this as we go forward so that we can form this feedback loop with our providers to say, you’re doing really well here, and here are some areas where you can improve based on our analysis of data.
Thanks, appreciate these insights. I've been contemplating VA's Oracle / Cerner implementation and wondered if implementing the same systems across…