Why so much variation?

Stud­ies dif­fer on causes of Medi­care cost dis­par­i­ties

Modern Healthcare - - THE WEEK IN HEALTHCARE - Me­lanie Evans

Medi­care spends rad­i­cally more in some places of the coun­try than oth­ers. In­tense de­bate over what’s be­hind that variation— and how to save money by nar­row­ing it— erupted again last week with new re­search con­clud­ing that dif­fer­ences in pa­tients’ health are pri­mar­ily to blame.

That runs counter to long­stand­ing work by re­searchers at the Dart­mouth In­sti­tute for Health Pol­icy and Clin­i­cal Prac­tice. In­stead of at­tribut­ing re­gional cost dif­fer­ences to dif­fer­ences in pa­tient health, Dart­mouth re­searchers con­tend that over­sup­ply of providers and tech­nol­ogy, per­verse fi­nan­cial in­cen­tives and dif­fer­ences in doc­tors’ ap­proaches to medicine lead to un­nec­es­sary care at greater cost.

The Dart­mouth work has been widely cited. Peter Orszag, who was di­rec­tor of the Of­fice of Man­age­ment and Bud­get as health re­form took shape, de­scribed the re­search as “one of the best ex­am­ples that we have of aca­demic work in­form­ing the pol­icy process” in an in­ter­view with the school’s Dart­mouth Now. But it’s also been sharply crit­i­cized.

Pow­er­ful stake­holder in­ter­ests will be af­fected by who turns out to be right—or who is be­lieved. If in­ap­pro­pri­ate prac­tice variation, in­ef­fi­ciency and waste—not health dif­fer­ences—are to blame, Congress could squeeze Medi­care pay­ments where spend­ing is high and re­ward low-cost com­mu­ni­ties for bet­ter qual­ity at a lower cost, said Peter Hussey, a se­nior health pol­icy re­searcher with RAND Corp. Tax­pay­ers in cheaper com­mu­ni­ties would no longer subsi- dize the waste of pricier spots, he said. But re­sis­tance to such a change could be pow­er­ful from law­mak­ers rep­re­sent­ing high-spend­ing states. “It’s a very po­lit­i­cal ques­tion,” he said.

Still, as Congress grap­ples with ris­ing health­care costs and fed­eral deficits, Medi­care spend­ing looms as an unavoid­able tar­get.

The spread be­tween the least and most ex­pen­sive lo­cales is jar­ring. The new study, pub­lished in the jour­nal Med­i­cal Care Re­search and Re­view, found Medi­care spends $6,612 per per­son, on aver­age, in the cheap­est re­gions and $11,643 in the most costly, af­ter ac­count­ing for the cost of liv­ing. At least three-quar­ters of that dif­fer­ence can be ac­counted for by health, ac­cord­ing to the study.

One of the au­thors, James Reschovsky, a se­nior fel­low at the Cen­ter for Study­ing Health Sys­tem Change, said Dart­mouth’s re­search fails to ad­e­quately ac­count for how ge­o­graphic dif­fer­ences in health in­flu­ence spend­ing.

Reschovsky, work­ing with Jack Hadley, a health ad­min­is­tra­tion pro­fes­sor at Ge­orge Ma­son Univer­sity, and Dr. Pa­trick Ro­mano, a pro­fes­sor of medicine and pe­di­atrics at the Univer­sity of Cal­i­for­nia, Davis, an­a­lyzed Medi­care spend­ing and di­ag­noses dur­ing the fi­nal months of life, a pe­riod dur­ing which Dart­mouth re­searchers con­tend all pa­tients are equally sick.

But Reschovsky and col­leagues found the gap be­tween aver­age Medi­care cost in the least and most ex­pen­sive ar­eas nar­rowed to $3,333 from $20,514 once they fac­tored in di­ag­noses, a roughly 84% dif­fer­ence.

Of course, di­ag­noses, taken from bills sub­mit­ted to Medi­care might be sub­jec­tive or ma­nip­u­lated by physi­cians. To ad­dress that risk, re­searchers an­a­lyzed spend­ing for di­ag­noses with no room for dis­cre­tion, such as heart at­tacks or hip frac­tures. A fi­nal anal­y­sis also con­trolled for mar­ket forces such as com­pe­ti­tion.

Hussey noted, how­ever, that di­ag­noses con­tained in Medi­care bills are a rel­a­tively crude way to iden­tify pa­tients’ health sta­tus.

In a writ­ten re­sponse to Reschovsky’s new study, Dart­mouth’s Jonathan Skinner, an econ­o­mist and re­searcher on spend­ing variation, ar­gued that di­ag­noses from claims forms can be bi­ased by doc­tors who overtreat pa­tients. Skinner ar­gued that the anal­y­sis by di­ag­noses did not side­step those bi­ases for di­a­betes, which can be sub­ject to doc­tors’ dis­cre­tion.

In an in­ter­view, Skinner also ar­gued Dart- mouth has been un­able to repli­cate find­ings by Reschovsky and his col­leagues us­ing sim­i­lar data.

Re­sults reached by Reschovsky could be at­trib­ut­able to method­ol­ogy, Skinner said. If the mix of physi­cians sam­pled by the re­searchers in­cluded a greater per­cent­age of spe­cial­ists, that could have in­creased the like­li­hood of more acutely ill pa­tients be­ing rep­re­sented in the study. In con­trast, Dart­mouth sam­ples pa­tients, not physi­cians, he said.

Reschovsky de­fended his re­search. “Our sam­ple was bench­marked against ad­min­is­tra­tive data from CMS, so I am con­fi­dent that our re­sults are not the re­sult of how our ben­e­fi­cia­ries were sam­pled,” he said in an e-mail.

Oth­ers also have chal­lenged the ex­tent to which Dart­mouth at­tributes spend­ing variation to waste­ful care. “It is not dif­fi­cult to iden­tify waste­ful prac­tices,” wrote Dr. Richard Cooper, a for­mer pro­fes­sor at the Univer­sity of Penn­syl­va­nia’s Leonard Davis In­sti­tute of Health Economics, in a blog post late last year. “The ques­tion is whether they ac­count for as much as 30% of health­care spend­ing.” Cop­per in­stead ar­gued that poverty—and the greater preva­lence of dis­ease as­so­ci­ated with low in­comes—is the cause of ge­o­graphic dif­fer­ences in health spend­ing.

De­spite this con­tin­u­ing de­bate, prac­tice vari­a­tions, in­ef­fi­ciency and waste­ful care re­main widely ac­knowl­edged as con­trib­u­tors to Medi­care’s high costs and sig­nif­i­cant re­gional spend­ing dif­fer­ences.

Reschovsky praised Dart­mouth’s work for high­light­ing this phe­nom­e­non, but said dis­crep­an­cies may not be re­gional as much as spe­cific to each provider’s prac­tice.

Karen Davis, di­rec­tor of the Roger C. Lipitz Cen­ter for In­te­grated Health Care at the Johns Hop­kins Bloomberg School of Pub­lic Health, agreed. Davis said Reschovsky’s work sug­gests that anal­y­sis of variation should fo­cus more finely on per­for­mance of hos­pi­tals and providers in­stead of broader ge­o­graphic blocks.

The In­sti­tute of Medicine’s Com­mit­tee on Ge­o­graphic Variation in Health Care Spend­ing and Pro­mo­tion of High Value Care said in an in­terim re­port re­leased this year that “sub­stan­tial” variation could not be ex­plained even af­ter an ad­just­ment for health sta­tus “con­sid­er­ably” nar­rowed variation.

Stephen Zuck­er­man, a se­nior fel­low with the Ur­ban In­sti­tute and co-di­rec­tor of its health pol­icy cen­ter who pre­vi­ously pub­lished re­search on health and spend­ing variation with Hadley, said the sup­ply of providers and dif­fer­ent prac­tice pat­terns do play a role. “There is variation af­ter you con­trol for health sta­tus ad­e­quately,” he said. “But it’s also the case, once you con­trol for health sta­tus ad­e­quately, this might be a less im­por­tant source” of variation than Dart­mouth sug­gests.

Newspapers in English

Newspapers from USA

© PressReader. All rights reserved.