The cen­tral­ity of health IT to ACO suc­cess

Modern Healthcare - - Q & A -

Achiev­ing the goals of ac­count­able care—bet­ter co­or­di­na­tion, im­proved out­comes, lower costs—re­quires a ro­bust health in­for­ma­tion tech­nol­ogy in­fra­struc­ture and the abil­ity to an­a­lyze large vol­umes of pa­tient data.

Mod­ern Healthcare re­cently hosted a we­bi­nar that fea­tured a lead­er­ship panel will­ing to share their views on the im­por­tance of health IT sys­tems, the chal­lenges as­so­ci­ated with im­ple­men­ta­tion and their ad­vice for other ac­count­able care or­ga­ni­za­tions.

Mod­ern Healthcare New York Bureau Chief Me­lanie Evans mod­er­ated a dis­cus­sion with Jason Dinger, CEO of Mis­sionPoint Health Part­ners, a Nashville­based sub­sidiary of As­cen­sion that fo­cuses on ac­count­able care; Kim Kauffman, vice pres­i­dent of value-based care at Sum­mit Med­i­cal Group, a 220-physi­cian group based in Knoxville, Tenn.; and Katie White, as­sis­tant pro­fes­sor of health pol­icy and man­age­ment at the Univer­sity of Min­nesota School of Public Health. This is an edited tran­script of that dis­cus­sion.

Katie White: We are liv­ing in ex­cit­ing times for health IT. It’s clear that in­cen­tives for ex­pand­ing IT sys­tems are in place, but there’s a lot of learn­ing to be had be­fore we see sys­tems used to their full po­ten­tial.

In our study of the ACOs in the Medi­care Shared Sav­ings Pro­gram and the CMS In­no­va­tion Cen­ter’s Pi­o­neer ACO pro­gram, we found that more ad­vanced in­for­ma­tion tech­nol­ogy and data an­a­lyt­ics did not nec­es­sar­ily en­sure suc­cess. For ex­am­ple, we had an in­te­grated de­liv­ery sys­tem with a sin­gle elec­tronic health-record sys­tem, so­phis­ti­cated value-an­a­lyt­ics ca­pa­bil­i­ties, with years of per­for­mance in risk-based man­aged-care con­tract­ing that could not achieve shared sav­ings.

On the other hand, we had a siz­able physi­cian group part­nered with a hos­pi­tal sys­tem with lit­tle risk-based or per­for­mance­based con­tract­ing ex­pe­ri­ence that had mul­ti­ple EHR sys­tems but achieved shared sav­ings in the first year with lit­tle IT sup­port. They tracked pa­tients with Ex­cel spread­sheet-based reg­istries and did ex­ten­sive health coach­ing and care co­or­di­na­tion through­out their prac­tices.

The bot­tom line seems to be that truly adding value in health IT will come from de­vel­op­ing the abil­ity to an­a­lyze big data, to un­der­stand pat­terns in those data, and to act on trends where it mat­ters most. We are just be­gin­ning to un­der­stand what ex­actly fully func­tion­ing health IT for value-based pay­ment sys­tems means.

Mod­ern Healthcare: What are the key lessons you iden­ti­fied as you looked at the role of IT among Medi­care Shared Sav­ings ACOs?

White: When you en­ter pro­grams like this, you come in with a set of sit­u­a­tions and a con­text and you try to adapt based on what­ever you have to­day and what­ever your ex­pe­ri­ence is. Some of these ACOs, frankly, were just pretty lucky. They’ve been on this jour­ney for some pe­riod of time. Other ACOs found that they weren’t pre­pared and they couldn’t achieve those sav­ings. Much of that had to do with not hav­ing the data at their fin­ger­tips to make re­ally good de­ci­sions to put them into the right tra­jec­tory so that they could suc­ceed.

Kim Kauffman: Sum­mit Med­i­cal Group formed 20 years ago and is cur­rently home to 220 physi­cians and about 140 ad­vanced prac­ti­tion­ers. It’s physi­cianowned and pri­mary-care driven, and 99.9% of our el­i­gi­ble sites have achieved Pa­tient-Cen­tered Med­i­cal Home Level 3 recog­ni­tion. A full 35% of our pa­tients are un­der a value-based con­tract, in­clud­ing one Medi­care Ad­van­tage con­tract with up­side and down­side risk and con­tracts with our two largest com­mer­cial health plans.

Our strat­egy in­cludes aligned in­cen­tives, full trans­parency and var­i­ous track­ing tools. We also use dash­boards and side-by­side re­ports to demon­strate the provider’s progress on qual­ity mea­sures, ex­pense man­age­ment mea­sures and other key per­for­mance in­di­ca­tors such as ad­mis­sions per thou­sand, emer­gency depart­ment vis­its per thou­sand, generic pre­scrib­ing rates and read­mis­sion rates.

MH: For or­ga­ni­za­tions that are new to data an­a­lyt­ics, what do you see as the first steps?

Kauffman: A low-cost way to be­gin—and most or­ga­ni­za­tions have the ca­pa­bil­ity to do this in­house—is to start with a

sim­ple cre­ation of dis­ease reg­istries and look for those his­tor­i­cally high-utiliz­ing, high-cost pa­tients, or those pa­tients who have prob­lems but per­haps haven’t seen their pri­ma­rycare physi­cian in the year to date, and en­gage those pa­tients.

Then there’s the op­por­tu­nity to move on to hind­sight. Again, it’s the eas­i­est in­for­ma­tion to wrap your hands around. That is the his­tor­i­cally high-cost uti­liza­tion. This pre­sumes that past be­hav­ior is the best in­di­ca­tion of fu­ture be­hav­ior. And then as your or­ga­ni­za­tion ma­tures, en­deavor to iden­tify your ris­ing-risk pa­tients.

MH: It sounds as if trans­parency re­ally played a role in achiev­ing the out­comes you were seek­ing. How did you ap­proach com­mu­ni­ca­tion with physi­cians?

Kauffman: Years ago, when we first started down this path and mul­ti­ple pay­ers were ap­proach­ing us with a list of 24 and 36 and 47 dif­fer­ent qual­ity mea­sures they wanted us to track, we de­cided that we needed to pick a fi­nite num­ber of mea­sures, de­velop work­flows and pro­cesses and point-of-care re­minder tools within the EHR for a sub­set of those mea­sures and fo­cus on those mea­sures. And now, in­stead of a health plan ap­proach­ing us with a myr­iad of mea­sures, we ap­proach them and say, “These are the mea­sures we are pre­pared to knock out of the ball­park. Let’s fo­cus our work around these, and then we’ll add a cou­ple more next year and a cou­ple more the year af­ter that.”

The dash­board is fully trans­par­ent. Any of our providers can go on to our In­ter­net site and look at their ag­gre­gate stars on these Healthcare Ef­fec­tive­ness Data and In­for­ma­tion Set mea­sures. Sim­i­larly, on a quar­terly ba­sis we send out a sideby-side re­port that shows each in­di­vid­ual pri­ma­rycare physi­cian the num­ber of pa­tients they have in a par­tic­u­lar con­tract, their med­i­cal-ex­pense ra­tio, their risk score and their key per­for­mance in­di­ca­tors.

The new­est ad­di­tion to that sta­ble of trans­parency tools re­lates to the dis­tri­bu­tion of bonus or pay-as-you-go per­for­mance dol­lars. Any time any of those dol­lars are dis­trib­uted, ev­ery provider sees the same re­port with the provider name right there and the amount and the rea­son for which they were re­ceiv­ing in­cen­tive dol­lars.

Jason Dinger: Mis­sion Point Health Part­ners started with 10,000 mem­bers in 2012. We’re now man­ag­ing the needs of over 250,000 mem­bers, and we’ve clin­i­cally in­te­grated with more than 7,200 providers. We’ve learned a lot by be­ing in dif­fer­ent geogra­phies and see­ing quite a bit of vari­a­tion.

The first step is get­ting a his­tor­i­cal view of your pop­u­la­tion. As you know, a small per­cent­age of peo­ple gen­er­ate most of the cost, and that’s one of the big chal­lenges for ACOs. We have found, as we go on our IT jour­ney, is that be­ing able to strat­ify pa­tients to make sure we’re al­lo­cat­ing the right amount of time to each per­son and en­gag­ing them in the right set­ting to re­ally help them and their fam­i­lies is so im­por­tant.

We are start­ing to do a lot of work around pre­dic­tive mod­el­ing and ma­chine learn­ing, putting more and more data into kind of our data repos­i­tory and let­ting that data get smarter and smarter about which in­ter­ven­tions are work­ing and which ones aren’t. For ex­am­ple, re­cently we were look­ing at some data and found that our sec­ond call with the mem­ber is by far the most pre­dic­tive of im­prov­ing out­comes and low­er­ing costs.

MH: Could you give us an ex­am­ple of how data strat­i­fi­ca­tion al­lows you to al­lo­cate re­sources ef­fi­ciently and in an ap­pro­pri­ate set­ting?

Dinger: The one that comes to mind is de­pres­sion. His­tor­i­cally, we would have looked at folks with de­pres­sion as asthma pa­tients or ac­tive can­cer pa­tients and re­late to them as such. But we now know that un­less we can re­ally help them through their de­pres­sion, all the other con­ver­sa­tions are just not go­ing to have the same im­pact. By do­ing some scor­ing di­rectly with mem­bers and work­ing through our providers, we can get a lit­tle bit closer to the root cause.

MH: What would you rec­om­mend as first steps for or­ga­ni­za­tions that are new to data an­a­lyt­ics?

Dinger: I would find a part­ner to just help clean and stan­dard­ize your data. There are a num­ber of lower-cost so­lu­tions now on the mar­ket, and find­ing a part­ner can take a whole bunch of things off your plate as well as kind of re­duce the num­ber of po­ten­tial er­rors. Then I’d lis­ten and watch that data and re­ally kind of soak your­self in what it can tell you about the peo­ple you’re serv­ing. And then I’d cus­tom­ize and slowly add to that data set and just let it get richer and richer for you and your part­ner­ing providers.

Katie White As­sis­tant pro­fes­sor of health pol­icy and man­age­ment, Univer­sity of Min­nesota School of Public Health

Jason Dinger CEO Mis­sionPoint Health Part­ners

Kim Kauffman Vice pres­i­dent of value-based care, Sum­mit Med­i­cal Group

Newspapers in English

Newspapers from USA

© PressReader. All rights reserved.