The MemorialCare Medical Foundation, a physician organization in Southern California with more than 2,000 physicians and mid-level practitioners, is a core element in the four-hospital MemorialCare Health System, headquartered in Fountain Valley (and whose hospitals are Long Beach Medical Center, Miller Women’s and Children’s Hospital, Orange Coast Medical Center, and Saddleback Medical Center). The integrated health system encompasses more than 200 sites of care and more than 14,000 employees and contractors, and, according to its 2018 annual report, the system encompasses 1,353 licensed beds, 267,000 patients in value-based health plans, 570,000 medical group visits, and 246,000 emergency department visits. The report notes that “MemorialCare offers more value-based products than any other health system in Southern California. Our Accountable Organizations (ACOs),” the report notes, “include a customized, direct contract between MemorialCare and Boeing called the MemorialCare Health Alliance, as well as innovative plans like Aetna Whole Health, and Vivity offered by Anthem Blue Cross. And it notes that, between 2017 and 2018, the MemorialCare Health Alliance “grew enrollment by 28 percent; reduced total cost of care by 4 percent for plan members who opted for the ACO and participated for the entire year; exceeded patient satisfaction targets; [and] scored number-one in quality for all Boeing ACOs.”
Earlier this month, Adam Solomon, M.D., the chief medical officer of the MemorialCare Medical Foundation—which encompasses both the MemorialCare Medical Group and the Greater Newport Physicians IPA—co-presented with two other integrated health system leaders, at the APG Annual Conference, held at the Manchester Grand Hyatt Hotel in San Diego. On Thursday, April 11, Dr. Solomon co-presented in a pre-conference session entitled “How to Move from Upside-Only to Downside Risk.”
During that session, Dr. Solomon said that “Each [value-based] contract will be different, and there will be different things to focus on to be successful. Outside of the contract, the most important element to consider is your network. Who will you include? The basis is your primary care physicians,” he emphasized. “If you don’t select primary care physicians who naturally manage patients well—who utilize properly and don’t refer excessively—you won’t be successful. Physicians won’t alter their care, so thinking about that is essential.” Meanwhile, he added, the specialty physicians who participate in your network need to “utilize appropriately, utilize the right locations for cost-effective care, and communicate well with primary care physicians and patients. And the last element, of course, is facilities. You may have your own facilities that are part of the structure. But also, what is the ability of those facilities to collaborate and coordinate with us?”
Recently, Dr. Solomon spoke with Healthcare Innovation Editor-in-Chief Mark Hagland regarding the range of issues and opportunities facing leaders like himself who are helping to lead their physician-led organizations into risk-based contracting. In part one of this two-part interview, Dr. Solomon spoke about his organization’s journey into risk, what he and his colleagues have been learning, and some of the particular challenges of ACO-based payment. Here now is part two of that interview.
Are things more challenging in terms of working with Medicare? We hear consistently that the Medicare Program provides data that tends to be very late.
Exactly. We were in the Next Generation ACO Program in 2016 and 2017, and got out in 2018. There are problems with how it’s run. But also, they would send us the full file. And there’s information in each of the separate files, but you have to cross-reference between and among files to get the full picture. But they did give us the full data sets; so I can get the information out of there if I want, via the EDW. But it was so delayed that it was hard to do anything with it. And it was worse if you weren’t able to manipulate the data itself. And usually, it was three quarters of a year gone before you got the data, and then the population changes. The Medicare data is too delayed. The commercial plans are better.
For example, Anthem has a portal; it’s quite usable. Not only do they give us the information in a nice format that makes it useful; they allow us to download the data to manipulate it. And they’ll give us a live ADT feed, so we’ll know the next day if a patient has been admitted. And for Boeing, where BCBS Illinois is the third-party administrator, we don’t necessarily know if those patients are admitted to a facility outside of our network, so we use Experian Member Match, and it covers most hospitals. And often, they’ll use Member Match, and if it’s on our roster, we immediately get a ping to know our patients are in an ER.
Tell me about your contract with Boeing?
Yes, we’re the preferred partner for them in Southern California; that contract involves more than 7,500 covered lines.
What has your organization’s journey been like around data analytics?
It has been a learning experience. One of the things we realized in Next Gen was that we really needed to manage this data. In 2014, our original ACO was with Anthem; then we started conversations with Boeing at the same time we began participating in Next Gen. On our HMO side, we’re delegated, we’re the ones paying the claims, so we have that information and could manage that data. But as we moved into this ACO world, with Boeing and then moving into Next Gen, we knew that a lot of the data was going to come from the outside. So we bought a solution from [the Salt Lake City-based] Health Catalyst for its tools that help you dig into that data in a variety of ways. Like so many other similar tools, there are the pretty-useful out-of-the-box tools they provide, but they don’t go all the way there, so you have to modify them.
And Health Catalyst didn’t have a tool for ambulatory pharmacy management in an ACO setting, or quality of care in ambulatory. So our team has developed tools for identifying care gaps, and patients who need preventive measures for services. And to help us identify which patients are on high-cost drugs—which drugs are the problem and which patients need to be monitored—we developed those parts on our own, through HealthCatalyst; they didn’t really have something that did that. So our pharmacists worked with HealthCatalyst to develop an app around pharmacy analytics. And they took an existing community care app, and turned it into a quality fulfillment app.
And they have a tool that helps with attribution; another looks at PMPM [per member per month] costs by specialty and provider. So we can look at our cost for this ACO and that ACO. So a variety of tools were available for the management of populations.
And when we have other ACO contracts where they send us reports or expect us to use their portal, none of that data is in there. So I’ll meet with one of our ACOs, and they’ll show us potential opportunities where procedures being done in a hospital that could be done in a surgery center. And I may have 40,000 people in an ACO, and 13 of those procedures may have been done, but they’re all one-offs, this doctor did one, that doctor did one.
So if I can combine all my ACOs together and look for patterns, that would be good, but we need to get all the data in one place and look at an overall population. We’re struggling on that, but moving forward.
And we have about 35,000 patients with Aetna and another 30,000 with Anthem, and another 7,500 with Boeing, and we just signed another contract with United, so that will be another 10,000 with them, and as those numbers grow, I will have more data to use. And for every one of those, I have to provide a roster of PCPs and specialists.
Over time, you’ll determine which PCPs and specialists perform best in terms of resource utilization and outcomes?
Correct. And if I identify outliers, I can ask an outlier to explain their anomalies. Or maybe how they’re practicing is such that I need to move the patients. And to me, that’s the same approach I’d follow in a managed care environment; there, I would simply take them off my referral list. In the PPO world, I would need to monitor. And the ACO arrangement is a form of PPO—that’s important to stress. There’s a lot that’s similar; they overlap, but they’re not the same. My level of control is different, how I manage the network is different, how I get information is different; there are subtle differences. Probably the greatest difference goes to how I talk about the attribution process. In the HMO world, I know those patients know they’re in a managed care product that’s delegated to one of our two medical groups; so when my team reaches out on behalf of those doctors, the patients aren’t surprised. But on the PPO/ACO side, patients don’t necessarily think that way, so trying to designate their PCP, oftentimes, they don’t want to designate a PCP. And then when they’re attributed to me, they don’t even know that they are. That’s a huge problem in the Next Gen ACO. In their minds, those patients don’t have any relationship to us. They don’t think of themselves as part of a system, so when we have outreach, it’s a little bit more challenging. If they’re in the product model, and especially if that product model incentivizes them to stay in the network, that changes things.
The attributed model is the least controllable one; the middle of the road is the designated model. Overall, the options for engagement include a standard PPO network with no additional provider oversight/collaboration; a full, open network with an ACO option so that the plan shares claims data with MemorialCare and MemorialCare has some incentives to help improve costs and quality; and a narrower, high-performance network (designated ACO) with incentives for employees to get all care within that network and, again, this data and risk sharing with MemorialCare. When we’re able to participate in a plan involving a narrower network, that provides us an advantage, because the employees know what’s going on, and the employees are incentivized to stay in the network. Boeing has both models, an attributed and a designated model. And the designation supports us in managing the costs for the organization, because we know the employees are being encouraged to stay in the network. And then one step further is an EPO model, where there’s no benefit at all for going outside the network; they’re locked into the network. So it’s far easier to manage those costs.
How do you see everything evolving forward the next few years?
Even though there’s been a long history of managed care in California, there’s been a steady decline in enrollment here; and the movement towards high-deductible PPO plans has not slowed healthcare inflation. In fact, costs are going up. So that’s why we’re seeing this growth in ACOs. So we’re learning how to be successful with all of this. And to have that direct-to-employer contracting, where the employer can work with a group like us to manage their costs, can be excellent. And the thing is, they go to the health plans and say, our costs are going up, what can we do about it? And the health plans really don’t know. These employers are paying all this money, and don’t feel they have any control. But when they have a relationship with a structure like ours, where we have skin in the game, it’s great. And we’re the first organization in California to really have that capability to fully manage a direct-to-employer contract.
And that arrangement is better for the employer; it helps them manage their costs and know the quality is better, and it helps them to manage costs and improve employee presentism, and to know they’re sharing in savings, that’s a win for them, and a win for us. And in a direct-to-employer contract, it makes total sense because it results in lower premiums and costs to the employer, and we’re now helping the employer and ourselves, and it gives us the ability to grow and manage our patients.
