At a time when issues around taking on risk-based contracts are becoming more complex and challenging for physician groups, some physician organizations are moving ahead to lay the foundations for success in risk-based contracting, all of which require strong health risk assessment processes around broad covered populations, and the creation of care management and population health management strategies for the long haul.
Among the physician organizations that has taken on these challenges and begun to lay strong foundations for risk-based contracting has been Community Medicine, Inc., which is the largest group of employed primary care physicians—330 in all—connected to the vast UPMC health system in Pittsburgh. The president of that group is Francis X. (Fran) Solano, M.D., an internal medicine physician, who is also vice president of the Physician Services Division at the umbrella UPMC system.
While taking on potential downside risk has been years in the making—Dr. Solano is currently working to develop a risk-based contract between Community Medicine, Inc. and the UPMC Health Plan—he has spent 15 years helping to lead a quiet, data-facilitated revolution around improvements in care quality and utilization management.
At the core of that effort has been the program Template for Change, which has laid a firm foundation for the taking on of financial risk in contracting, even as Template for Change has focused on initial shared-savings goals between and among the medical group constellation, the hospital system, and the health plan. The program, which has been applied to medical care delivery across all insurance lines (Medicare, commercial, Medicaid), has been implemented across the 330 primary care physicians in the Community Medicine constellation of groups, as well as 30 PCPs at Renaissance Family Practice based at St. Margaret’s Hospital, and 60 internal medicine physicians in a third group.
The focus, Solano says, has been on providing monthly performance metrics to every physician in the program, at the individual, practice, and specialty level, for the following core statistics: inpatient admissions per 1,000, inpatient days per 1,000, ED visits per 1,000, primary care visits per 1,000, specialty visits per 1,000, diagnostic imaging rates per 1,000, and medical expense ratio. The physicians in the program have also been receiving data on their “HEDIS gaps”—gaps in care conveyed by HEDIS (Health Effectiveness Data and Information Set) data. Solano himself and his team look at a monthly 400-page printout of data, but, he notes, he learned early on that sending out massive amounts of raw data to the physicians in practices only led to mass confusion and frustration. The key has been that providing the doctors with these data points has helped them find “the patients who are high utilizers but are not yet in care management.
Francis X. Solano, M.D.
What’s more, Solano says, the monthly sharing of clinical performance data, along with coaching sessions for physicians who are struggling, has made a huge difference. Still, “The biggest contribution” to physician behavioral change has been the average $20,000 in physician compensation that has been tied to individual physician performance on the various metrics being measured. Also very big has been the fact that the shared-savings program has involved sharing savings with the hospital system, an important factor, given that intended reductions in inpatient admissions and ED visits inevitably will affect the hospitals’ bottom lines.
Dr. Solano spoke this summer with Healthcare Informatics Editor-in-Chief Mark Hagland, as Hagland prepared the September cover story on physicians and risk. Below are some additional excerpts from their interview. This is part one of a two-part interview.
Tell me a bit about the background and context for what you’ve achieved so far at Community Medicine, Inc. This is quite pioneering work. What is the landscape of managed care like right now in western Pennsylvania?
Actually, there is very little true risk-based contracting going on in western Pennsylvania at this time. Most programs really only have upside potential, and not much in the way of downside potential. The Pioneer ACOs [accountable care organizations] can do either upside risk or two-tailed risk, involving some potential downside. We have a shared-savings program with our health plan, which is mainly based on hitting a medical expense ratio. We started our first year in 2013 with that program. Meanwhile, we over 30,000 Medicare members, 28,000 commercial members, and 30,000 Medicaid patients, or the largest [aggregation of such groups of patients by a medical group] in western Pennsylvania.
So Medicare has several models; it has the ACO model. And a lot of Medicare Advantage programs have developed their own ACOs, and that’s what this is. We’ve been extremely successful, getting this many physicians engaged to care about cost; and we’ve always cared about quality at UPMC. Our first year was interesting in that we were successful. Those programs take three years to become successful usually; but we’ve learned a lot about variability in practice patterns, and also how people approach access to primary care. One of the biggest challenges is to reduce the number of admissions, towards what I call appropriate admissions.
And in western Pennsylvania, where there are a lot of Medicare patients and a lot of Medicaid patients; and our emergency department utilization, urgent care utilization, and inpatient hospitalization, are all off the charts in western Pennsylvania.
Why have those utilization rates all been so high in your region?
I wish I could answer that question. I think some of it is cultural. I know there are patients who feel like they’re bothering me when they call me at night, and don’t want to bother me, so they just go to the ER. And some of it is the default that physicians have on their answering machines or services saying, go to the emergency room. So we’ve changed that to say, always talk to the physician. One example is chest pain. From a primary care office, the admission rate for chest pain is extremely low; from the ER, it’s extremely high, because the ED doctors don’t have an established relationship with the patient and they don’t want to risk legal liability.
So those are some of the big things that we focused on in our program. And our admission rates declined substantially. Most Medicare admission rates are 310-320 admits per 1,000, and we got ours down to 240 within two years. It’s based on changing the way doctors practice. We have a team of nurses who go into the practices and spend a week or two sitting at every staff position in the office, training people to do things differently. It’s called the Template for Change. Program on how to triage patients, how workflow should work, improve gaps in care access, etc. And we’ve expanded access in our offices through walk-in clinics, extended hours at night, even weekend hours. And these walk-in ….
And, in that, you and your colleagues clearly made a calculation that investing strategically in certain positions, such as nurse case and care managers, would pay off in terms of admissions and readmissions, correct?
Yes, there is a real ROI in having people not get readmitted. We didn’t even focus on specialty over-utilization; we focused on more of a concierge format for Medicare patents.
Tell me about the centrality of the pursuit of data analytics in all this?
We have a lot of data coming from our health plan. And one of the challenges in dealing with the federal government versus commercial plans is that it’s really hard to get data from Medicare. In fact, one of the reasons that ACOs are dropping out of the Pioneer ACO program has been the challenges of dealing with the data issues from Medicare. And we have the advantage within This Template for Change applies to Medicare, Medicaid, and commercial, all three. But most of the money is from Medicare. If you’re spending $1,000 PMPM [per member per month] for care delivery under Medicare, and commercial and Medicaid are half of that, you can see where the money is. So we’ve focused most of our efforts on Medicare patients, but there’s a Hawthorne effect on the other patients as well. But the whole program is designed to work across all insurance lines.
And, specifically, what has the role of data analytics been in success to date of the Template for Change program?
We do analytics looking at, say, how successful our practices have been in getting their medical expense ratio down, versus those that haven’t been as successful. Every insurance company has a different medical loss ratio it calculates to break even. Many insurance companies set a MLR of 12 percent; our insurance company is a lot more efficient than that. But when we looked at the practices that seemed to be successful, we looked at their admit and ED rates. Those that are successful in lowering their admissions also tend to have lower ED utilization rates. So we have performance metrics that we share monthly by practice. And we have our doctors grouped into pods by region, so that the data is more powerful.
And you have to have at least 200 patients in a product so that the data is a reliable. So we’re broken down into pods for adequate numbers in each pod. For example, in our Medicare product, we have 1,6000 people in the Bedford pod, close to 6,000 in our central pod around Oakland, Shadyside, and around 7,000 patients in our southern pod, etc. And we’re looking at admissions per 1,000, days, per 1,000, ED visits, per 1,000, primary are visits, specialty visits, per 1,000, diagnostic imaging rates, and medical expense ratio. So when we look at all these metrics, we actually educate our doctors on how to do look at them.
Do you engage in the use of dashboards?
Yes, we do it by practice. Every practice at a primary care practice sees monthly their dashboard, including their medical loss ratio for their line of products. Every month, we get a 400-pate printout of data for our population. So we can find the unplanned care people who are high utilizers. And the true key has been helping our physicians and their teams to identify the patients who are high utilizers but are not yet in care management. Those patients are typically the ones that I jokingly refer to as our Seagram’s 7&7, because these are the people with seven diagnoses, who are on seven medications, and are seeing seven specialists—and they really need to be identified and care-managed.
And you apply incentives to all this work, too, correct?
Yes, absolutely. It’s a combination of things: we share monthly clinical performance data with them, we engage in coaching sessions with physicians who are struggling, and the biggest contribution to physician behavioral change has been the average $20,000 in physician compensation that we’ve tied to individual physician performance on the various metrics being measured. Also very big has been the fact that the shared-savings program has involved sharing savings with the hospital system, an important factor, given that intended reductions in inpatient admissions and ED visits inevitably will affect the hospitals’ bottom lines.
Sharing savings between the physicians and hospitals helps align the hospitals with us, because the hospitals take the biggest hit when you reduce admissions and readmissions. So when you break down the silos by sharing money across the continuum, that helps. And now the passage of the MACRA [Medicare Access and CHIP Reauthorization Act of 2015] law has everyone aware of where this is going. They had never had cost data until recently, and charge-based data is very hard for physicians to understand. But with actual cost data from our health plan, doctors can use that, and they will start looking at the cost differential between a hospital-based colonoscopy and an ambulatory surgery center-based one, for example. And they are starting to pay attention to which specialists are over-utilizers, in terms of whom they refer their patients to.
In part two of this interview, to be published next week, Dr. Solano will share his broad perspectives on where physician practice and the healthcare delivery system are headed, and what CIOs and CMIOs need to do to move their organizations forward in the emerging risk-based world of contracting.