Big data’s big puz­zle: Now what?

Modern Healthcare - - NEWS - By Beth Kutscher

HHS’ part­ner­ship this month with Op­tumLabs is likely to add mo­men­tum to the push to use big data to lower health­care costs and im­prove the qual­ity of care de­liv­ery.

But the an­nounce­ment also high­lights the di­vide be­tween health sys­tems that are ahead of the curve in work­ing with so­phis­ti­cated data an­a­lyt­ics and those that are still learn­ing how to col­lect and use that in­for­ma­tion.

For decades, many in­dus­tries, such as fi­nan­cial ser­vices, have used data an­a­lyt­ics to de­ter­mine their strate­gies for ev­ery­thing from mar­ket­ing to main­te­nance sched­ul­ing. But the health­care sec­tor as a whole has lagged in us­ing an­a­lyt­ics to learn about its cus­tomers’ be­hav­ior and how to in­flu­ence it.

“Clearly there are some places that are really in­no­va­tive, but most sys­tems are just start­ing this jour­ney,” said Mur­tuza Mukadam, global head of health­care strat­egy and so­lu­tions at con­sult­ing firm Vir­tusa.

Through the new part­ner­ship, HHS agen­cies will gain ac­cess to Op­tumLabs’ big-data re­sources, in­clud­ing de-iden­ti­fied med­i­cal claims and clin­i­cal data. Op­tumLabs is a re­search and in­no­va­tion cen­ter es­tab­lished by Unit­edHealth Group’s Op­tum health­ser­vices di­vi­sion and the Mayo Clinic.

The first re­search project, to be led by the Agency for Health­care Re­search and Qual­ity, will com­pare the re­sults of the gov­ern­ment’s Med­i­cal Ex­pen­di­ture Panel Sur­vey to Op­tumLabs’ real-world claims data.

Large sys­tems like Mayo and Geisinger Health Sys­tem, which in 2013 launched its own tech­nol­ogy and an­a­lyt­ics group, xG Health So­lu­tions, are start­ing to ap­ply their re­search find­ings to chang­ing their stan­dards of care. Geisinger has spent $200 mil­lion on in­for­ma­tion tech­nol­ogy over the past two decades. By us­ing big data to iden­tify its sickest pa­tients, it has re­duced hos­pi­tal ad­mis­sions by 27%, the sys­tem re­ported at a 2014 health IT con­fer­ence.

Other sys­tems are just start­ing to dab­ble in data an­a­lyt­ics. Un­til re­cently, much of health­care’s IT ef­forts have fo­cused on adopt­ing elec­tronic health records, or their dig­i­tal in­fra­struc­ture, as the gov­ern­ment has pushed that tran­si­tion through mean­ing­ful-use pay­ments, Mukadam said.

It’s only re­cently that providers have turned their at­ten­tion to data an­a­lyt­ics. As more health­care providers take on fi­nan­cial risk for care out­comes, they are fi­nally see­ing the fi­nan­cial in­cen­tives for making the an­a­lyt­ics in­vest­ment.

“Peo­ple are wak­ing up,” said Mun­zoor Shaikh, a di­rec­tor in the health­care prac­tice of con­sult­ing firm West Mon­roe Part­ners. “Peo­ple are just start­ing to understand the prob­lem a bit bet­ter.”

Yet one of the big­gest bar­ri­ers to ef­fec­tively us­ing data an­a­lyt­ics is con­fu­sion over how much and what type of data to col­lect, ac­cord­ing to a sur­vey from Stoltenberg Con­sult­ing, which col­lected re­sponses at this year’s Health­care In­for­ma­tion and Man­age­ment Sys­tems So­ci­ety meet­ing.

About half of re­spon­dents (51%) cited “not know­ing how much or what data to col­lect” as the big­gest hin­drance to us­ing data an­a­lyt­ics, while 33% said their or­ga­ni­za­tions did not know what to do with the data they col­lected. Even the term “big data” elic­its con­fu­sion, be­cause it can cover any­thing from ad­min­is­tra­tive data, to clin­i­cal data, to pa­tient-gen­er­ated data.

Health sys­tems are gath­er­ing ter­abytes upon ter­abytes of data each day through mul­ti­ple sources, Mukadam said, from claims in­for­ma­tion, to physi­cians’ EHR notes, to data from con­sumers’ mo­bile and wear­able de­vices.

“If there’s no good, low-cost way to cap­ture that data, it’s go­ing to be point­less,” he said. “I don’t think hos­pi­tals are even there yet. A lot of them don’t even have the right skill set.”

A num­ber of ven­dors have stepped in to help providers man­age the load. Mukadam pointed to Hadoop, an open­source plat­form, which is making it cheaper and sim­pler to build data lakes.

Quest Di­ag­nos­tics is tar­get­ing the physi­cian mar­ket through a part­ner­ship with Ino­valon, an an­a­lyt­ics soft­ware com­pany. Quest reaches about half of the coun­try’s physi­cians through its lab­o­ra­tory ser­vices.

The part­ner­ship’s on-de­mand Data Di­ag­nos­tics ser­vice goes live at the end of the year, said Dan Rizzo, Ino­valon’s chief in­no­va­tion of­fi­cer.

“You’re still in this world of big-data ap­pli­ca­tion,” he said. “It’s amaz­ing how much data are still in a dis­con­nected EHR. Where you will see it pick­ing up is when you have en­ti­ties bring­ing to­gether dif­fer­ent sources.”

While big­ger health sys­tems and in­sur­ers tend to have more data, they’re not nec­es­sar­ily us­ing it more ef­fec­tively, Shaikh said. For or­ga­ni­za­tions both large and small, the best place to start is to zero in on a sin­gle ques­tion they’re try­ing to an­swer, such as iden­ti­fy­ing the cost of care for a given con­di­tion in the East­ern re­gion of the U.S., for ex­am­ple.

“We call it small data,” he said. “You have to prove the value in a small and di­rected way. Don’t in­vest in ev­ery­thing all at once; in­vest in some­thing where you can really make a dif­fer­ence.”

“It’s amaz­ing how much data are still in a dis­con­nected EHR. Where you will see it pick­ing up is when you have en­ti­ties bring­ing to­gether dif­fer­ent sources.”

DAN RIZZO, Ino­valon’s chief in­no­va­tion of­fi­cer

GETTY IM­AGES

Newspapers in English

Newspapers from USA

© PressReader. All rights reserved.