As the leading patient care organizations in the U.S. get deeper into population health management, priorities are becoming far more patient-centered than ever before, according to a panel of speakers at the iHT2 (a sister organization to Healthcare Informatics) New York Health IT Summit on Sept. 29.
More than half of healthcare managers surveyed expect to recoup their investments in population health management programs within three to four years, according to a recent survey by KPMG LLP in New York. The growth in the number of accountable care models, value-based contracts, and other reimbursement models is driving much of the emphasis toward population health. As such, the panel discussion at iHT2 New York centered around how organizations are measuring the impact of population health management on the patient, population, and health of the organization delivering care.
On the panel were: Michael Udwin, M.D., executive director, physician engagement at McKesson Corporation; Luis Tavares, senior vice president and CIO at the West Orange, N.J.-based Barnabas Health; Paul Wilder, CIO at the New York eHealth Collaborative (NYeC); and Mark Hagland, Editor-in-Chief at Healthcare Informatics. Moderating the session was Judy Hanover, research director, IDC Health Insights. Below are key experts from that panel discussion.
Hanover: What were best practices for getting started with population health and how did you make decisions on technology to support those strategies?
Tavares: We have ACOs in north and central New Jersey. For our central ACO, we are one of the few in the country where we had an organization who wasn't part of our [network] as part of the ACO. That was an interesting twist. We started with the Medicare Shared Savings Program (MSSP) three years ago. The marketplace is different in the north and central part of the state. There a lot of different physicians. When we started looking three years ago, there was no solution that met our needs, so HealthEC co-developed with us a solution for our needs, and they have since sold to other organizations across the U.S. Our [population] has gone from 30,000 to well over 200,000, so we have to decide if we need something bigger for us.
How has the technology landscape as well as priorities changed as we get deeper into population health management?
Udwin: Most folks are focusing on the quality element, in terms of readmissions, and identifying the highest risk patients. The next step is to look back one level and see if we are identifying the patients who aren't the sickest, but could be at risk in the future. Utilization is the next step after that, and then pharmacy, radiology, and imaging issues.
Wilder: Well one thing that has changed in the last few years is meaningful use. We all were doing it for years, and now the dialogue is waning a bit, so we are accepting the pain that brings with it. We do think it's worth it in the end. Standards and the ability to move data around has improved, though it's not perfect. The idea is to build a larger tent with broader constituents, and you can't do that for a population that is not directly at risk for us. New York has one of highest Medicaid percentages in the country, so it's a different kind of risk and a different model. There are different health systems and different data. There is a real challenge in dealing with non-standard clinical data. Behavioral and mental health data from a security standpoint has been tough. That's a big part in improving the population, and there is a lot we could learn from the Veterans Administration (VA) if we integrated with them and better understood some things that they do. Less people are resistant to sharing data now though; they know if they don't share it, they don't get it back.
Hagland: I just interviewed someone from a health system in Delaware, and I learned that putting all of the pieces together is really fascinating. You have risk assessment analytics, you have to stratify and figure out who you have. Organizations are realizing that the rising risk enrollees are the key people you have to figure out. Highest risk patients are obvious to providers as they see them often. The care management tied to analytics and all processes being linked is key, though. That requires process integration and operational integration, and doing HIE, but doing all those things at once is challenging.
Tavares: We wanted to make sure that we put together an operating plan for the year. Patient stratification was the first thing—4 percent of our patient population in our ACOs represent 26 percent of our costs. From a cost perspective, how do you manage that? That is risk category 1. Our focus was on risk category 2 patients, and trying to move them down to a lower category, while also keeping them from getting to category 1. So for us, management was often in that middle layer.
Hagland: Organizations are finding that you have to catch those people earlier. Claims and clinical data need to be married— you have to connect Mrs. Smith who is pre-diabetic with a care coordinator and manager in a way that makes sense. This requires patient engagement too, since those people are not realizing the risk they are at.
Udwin: Now, what's the best model to deal with those patients? Is it a hybrid or central approach? Do you want to do it a practice level, overarching level, or some sort of hybrid of both? Regardless, you have to make sure it works with physicians' workflows and doesn't impede anything, and you have to identify all of the gaps in care.