#LeadingVBHC – We talked to Jonathan Pearce of Singletrack Analytics after his presentation, How Physician and Analysts Can Support Each Other.
Bre Cunningham: Hi there, Bre Cunningham, and I am here with Jon Pearce, who is the conference resident data expert. Jon you just gave an amazing presentation on Medicare claims data and how you utilize that data.
First question I have for you is, if I wanted to get Medicare data, I’m a hospital and I want to see my claims data, I want to look at my readmissions, how do I get that data? And secondly, what kind of format is it in and how do I-
Jonathan Pearce, CPA: Oh jeez-
Cunningham: Exactly. How user-friendly is it?
Pearce: The first thing is that the data we’re talking about today is available to participants to in these government programs, the Bundled Payment for Care Improvement program, the Comprehensive Care for Joint Replacement program, and whatever programs might appear after that. It’s not available if you’re not a participant in those programs. So, if you didn’t voluntarily participate in BPCI or you’re not in one of the mandatory metropolitan statistical areas for CJR, you’re out of luck. You’re not going to be able to get your hands on that stuff.
If you are, then there are certain hoops you got to jump through with CMS to get access to it, you have to apply for it, you have to sign certain documents and make certain commitments about maintaining privacy, and things like that. If you do all of that right, you can go to a CMS portal, log in to the portal, and we deal with it on a monthly basis for our BPCI clients, we do it on a quarterly basis for our CJR clients, and we’ll just go download the stuff.
What it is, it’s very detailed. It’s claim level data. So, for each provider whose proved services to a patient, in this episode, we have all the claims. Anything Medicare paid for, we have all the claims for that patient, during that initial admission and the 90 days afterward, in most cases.
Cunningham: Even if they were readmitted, let’s say, to a different facility.
Pearce: Into a different facility? We’re going to have that. It’s really the only way that some hospitals would find out that a patient was readmitted somewhere else. One of the examples that was given earlier in the conference, is a specialty hospital, it’s a specialty orthopedic hospital, and we deal with orthopedic patients.
Cunningham: They wouldn’t know.
Pearce: If a patient’s admitted for congestive heart failure, they’re admitted someplace else. They wouldn’t know otherwise, unless they got that information anecdotally from the patient, or from the physician, or something else.
Cunningham: One thing you had mentioned in your talk was that, this data you’re getting retrospectively, so you can’t really use it for real-time management of your program. When you say retrospectively, how far out are is that? Let’s just say, I’m looking at a patient and I want to know if they were readmitted, I’m at one of these specialty orthopedic hospitals, when, realistically, would I find out if they were readmitted?
Pearce: The BPCI data is released from CMS monthly, the CJR data is only released quarterly. CJR is really bad, you can only really use CJR data way, way, way after the fact. You can’t even use it for the most recent quarter, because some of the episodes in the most recent quarter aren’t finished. Right now we’re getting CJR for the quarter ended December 31st. If a 90-day episode started on December 15th, that episode’s not even finished yet, let alone, having all the claims come in after that. You really can’t do anything with it. CJR is really bad, unfortunately CMS should be sending this data monthly. They haven’t yet acquiesced to that. In BPCI it’s a little bit better, we get data for the episodes in the previous month from which it was reported. So, we can usually tell that there was a readmission one or two months previous to when we got the data. In no case is this stuff usable contemporaneously. It’s way, way after the fact. It’s probably two months after the event happened.
In my session, what I said it’s useful for, is assessing the landscape of what services are provided to these patients and then designing a strategy around that. So, if you’re using inpatient rehab facilities more extensively than you think you should have, after you’ve looked at the data, you can design a strategy around that. Then, later, after you’ve implemented the strategy, you can see whether that works or not. But you’re not going to be able to tell if a patient went to an inpatient rehab facility yesterday, because you’re not going to have those claims for another two months.
It’s not useful for any type of contemporaneous care management. It’s useful for strategy design and then strategy review, feedback, and recalibration.
Cunningham: What do you tell your clients, let’s just say I’m not a CJR hospital right now, but chances are, in the future, we’re going to be. Bundles are pretty much here to stay. We’re seeing these programs gain traction and expand. I want to be proactive. I want to see where I’m at right now. I want to get a baseline of where I’m at, so I can prepare to be successful in the future. How does one go about that?
Pearce: There are options, CMS does sell a large dataset of claims data that they, basically, can buy, the biggest dataset is 100% of hospital post-acute provider claims and 10% of the physician claims, they never sell 100% of the physician claims, because they think it’s too much data. Which 10 years ago it was, but with the size of the discs we can buy nowadays, it’s not. Organizations, like the Datagen group [but the postle 05:44] association of New York, who is presenting also on our session, actually buy that data, it’s quite expensive, it’s in the tens of thousands of dollars, they will go build episodes for a hospital that’s not in a CJR region, if that hospital wants to engage them to do that. If you’re not in a CJR region, you could say, “Tell me what happens to my patients after they leave. Are they going to skilled nursing? Are they going home? Are they going to home health agencies,” and things like that. You could do a lot of the analysis that we’re doing with that type of data.
That data’s also not up to date, I think it’s either 2014 or 15 has just been released. That’s also somewhat delayed, and that affects its usefulness, because you might not have the same doctors working with you that were working with you several years ago. It’s unfortunate, it would be nice if hospitals could request data for this type of thing, for their own episodes, and get it and be able to plan. Because, as you say, especially for something like joint surgery, I think it’s a no brainer that that’s going to be somewhat ubiquitous in Medicare, probably within three or four years. It’s all going to be paid on an episode basis.
Cunningham: Absolutely. I would 100% agree with that. With these bundles, it’s forcing all these parties to work together. You used to be operating in silos, you have your physician group, the hospital, then all the post-acute care that goes on. There’s an old joke that says, “Where did the patient get a bed sore, at the hospital or at the SNF?” And the answer is, “In the ambulance.” Now, this is actually showing, and you can point fingers to where these complications or these events happened. How are you seeing this data actually utilized? You present this data, you show the trends in it, what kind of information are you finding surgeons and hospitals seem to be able to glean from the data?
Pearce: It’s really interesting to get with the surgeons and how to look at their own data, because a lot of cases they only have anecdotal information about what happens to their patients after they discharge. Especially with joint surgery, you’re not following this patient closely for a long period of time. The episode is a 90-day episode. This patient might have something happen two months after surgery. It counts as part of your episode, but you really wouldn’t know about it. The guy’s leg was fine, he’s able to walk, the joint replacement has worked well, and you’ve moved on. Then, 60 days later, he’s got a heart attack. A heart attack is related readmission and all of a sudden this patient’s back and you didn’t even know what happened. He might not even come back to your hospital. It’s interesting, number one, to see how physicians react to knowing these things about their patients that they didn’t know.
That’s on an individual basis, you can also look at somewhat of an institutional basis. For example, we can look at a skilled nursing facility that receives a certain number of referrals from the hospital throughout the year, and see that in that skilled nursing facility, and we’ve actually found this a couple of times, you’ve got a lot of patients who are staying for a long period of time. Then, that curve will flatten out in terms of how long they’re staying. We’ll find a whole bunch of patients, a large percentage of them, you can see it usually on a curve, that are staying for 20 days. The reason they’re staying for 20 days is the patient co-pay, for a skilled nursing facility, kicks in on the 21st day. The 21st day, the patient goes from paying zero to $160 a day. What’s happening in that case, and we found a couple of these, is that the SNF, Skilled Nursing Facility, is holding on to the patient until they would start to complain. Has nothing to do with the patient’s medical necessity or anything like that, and the docs have no idea. The docs know the patient went to a SNF but they don’t know how long they stayed there, they’re due in surgery. They’ve never had the concise, compiled, organized data for this. We can look at that.
We can also look at things like how many patients were readmitted from that facility. Is this a case where if a patient starts sneezing, they get sent to the ER. That’s one extreme. Or the other extreme is, some of the SNFs have snippets, they have in-house physicians who take care of that and who will deal with a non-emergent condition, and the patient stays in the skilled nursing facility, and doesn’t get sent to the ER, and readmitted back to the hospital. We can tell that type of thing too.
Pearce: This is all new, because the hospitals have never really had this data before. They don’t have data for what happens outside your walls. They mostly get it from the payer.
Cunningham: I think it’s exciting. I think it gives me a lot of hope in things that we can do better and improve the overall delivery and quality of care.
Pearce: It’s nice for me as a data guy. I mentioned in the conference that my dad was a physician and my brother’s a physician, and they really have helped people. A lot of people can hear because of the ear surgery my dad on them, and I’m sure that Steve has kept people alive through surgery that perhaps wouldn’t have worked with a less skilled anesthesiologist. So, what do I do? What I can do, is try to help providers and healthcare organizations improve care. If I can point out a situation where a patient’s sitting in a skilled nursing facility, where nobody wants to be in, past the time it’s medically necessary for it, and somebody else can take action and improve that patient’s care, that makes me feel good even though I’m not a doctor like my dad and my brother were.
Cunningham: I love it. I have one final question for you. Ten minutes till kickoff. Who’s going to win, the Packers or the Falcons?
Pearce: I’ve always told people, anyone who doesn’t believe in God has never seen Aaron Rodgers play quarterback.
Cunningham: He’s 007 these days.
Pearce: That’s it.
Cunningham: He’s crushing it. But, I don’t know, it’s going to be a good match up. Let’s get to that game.
Pearce: Matt Ryan is terrific.
Cunningham: He is. Matty Eyes.
Pearce: I think Ryan’s going to prevail.
Cunningham: All right, let’s go watch that game.
Pearce: Sounds terrific.
Cunningham: We’re out. We’re back with our next interview.