Hos­pi­tals ques­tion use of HCAHPS in scor­ing for value-based pur­chas­ing

Modern Healthcare - - Front Page -

Hos­pi­tals ques­tion use of HCAHPS scores for value-based pur­chas­ing

Even as hos­pi­tals brace for the un­cer­tain im­pact of link­ing pa­tient-sat­is­fac­tion scores to Medi­care pay­ments start­ing in Oc­to­ber 2012, some hope for tweaks in the sys­tem be­fore that time. The CMS plans to base 30% of hos­pi­tals’ scores un­der the value-based pur­chas­ing ini­tia­tive on pa­tient re­sponses to the Hos­pi­tal Con­sumer As­sess­ment of Health­care Providers and Sys­tems sur­vey, or HCAHPS, which mea­sures pa­tient sat­is­fac­tion. That over­all score could af­fect at least 1% of a hos­pi­tal’s to­tal an­nual Medi­care fund­ing.

But some hos­pi­tal ad­vo­cates and re­searchers have warned that the sur­vey has un­der­gone lit­tle peer-re­viewed val­i­da­tion and does not ac­count for ap­par­ent pa­tient bi­ases.

“We have to be care­ful about how we hold hos­pi­tals re­spon­si­ble,” says Dr. James Mer­lino, chief ex­pe­ri­ence of­fi­cer at the Cleve­land Clinic, which has con­ducted re­search on pa­tient-sat­is­fac­tion sur­veys. “We should be hold­ing hos­pi­tals re­spon­si­ble for things that they can ac­tu­ally im­prove.”

The few pub­lished analy­ses of HCAHPS sur­veys have found their re­spon­dents gen­er­ally dis­fa­vor or­ga­ni­za­tions such as large, aca­demic hos­pi­tals in north­ern re­gions of the coun­try that treat large num­bers of pa­tients with ei­ther de­pres­sion or com­plex and se­ri­ous ill­nesses. Those poor pa­tient scores come re­gard­less of the high qual­ity of the clin­i­cal care that other mea­sures have found those in­sti­tu­tions pro­vide.

Among the prac­ti­cal im­pacts of such per­ceived bi­ases is that no hos­pi­tal in the nation with 500 or more beds has scored in the 90th per­centile for such ba­sic mea­sures as physi­cian com­mu­ni­ca­tion or nurse com­mu­ni­ca­tion, one Cleve­land Clinic study found.

The cu­mu­la­tive im­pact of the HCAHPS bi­ases, ac­cord­ing to hos­pi­tal ad­vo­cates, in­di­cates that small com­mu­nity-based hos­pi­tals in the South­east per­form best on the sur­veys, based on their cur­rent de­sign. The re­gional dis­par­i­ties con­sis­tently re­sult in hos­pi­tals in the South and Mid­west rank­ing higher and those in North­east­ern states lower, ac­cord­ing to Medi­care’s Hos­pi­tal Com­pare web­site.

Dif­fer­ing ex­pec­ta­tions of hos­pi­tal care in var­i­ous re­gions of the coun­try pro­duce pa­tient-sat­is­fac­tion re­sults that some­times run counter to re­ports on the qual­ity of the hos­pi­tals’ clin­i­cal care, ac­cord­ing to pa­tientsat­is­fac­tion re­searchers.

How­ever, some hos­pi­tal of­fi­cials say on­go­ing dis­cus­sions with the CMS and the Agency for Health­care Re­search and Qual­ity, which de­signs the sur­vey, could pro­duce some changes in the method­ol­ogy of the sur­veys in an at­tempt to ac­count for the per­ceived bi­ases.

Among the changes re­port­edly un­der dis­cus­sion are mod­i­fi­ca­tions to the method­ol­ogy of the sur­vey to add risk ad­just­ments and pop­u­la­tion ad­just­ments.

Of­fi­cials at the CMS did not con­firm or

Hos­pi­tal ad­vo­cates cite short­com­ings in use of pa­tient-sat­is­fac­tion sur­vey data.

deny whether dis­cus­sions of changes to HCAHPS are un­der way.

An­other com­pli­cat­ing fac­tor on the im­pact of the pa­tient-sat­is­fac­tion sur­veys are some in­di­ca­tions that their bi­ases may be off­set by other com­po­nents of the value-based pur­chas­ing pro­gram.

Nancy Fos­ter, vice pres­i­dent of qual­ity and pa­tient safety with the Amer­i­can Hos­pi­tal As­so­ci­a­tion, says that ear­lier re­views of such data found teach­ing hos­pi­tals that per­form poorly on pa­tient sat­is­fac­tion tend to re­ceive higher scores on process-of-care mea­sures than non­teach­ing hos­pi­tals. She says that is one of the many is­sues the AHA plans to track as the value-based pur­chas­ing pro­gram rolls out, which also will in­clude iden­ti­fy­ing the hos­pi­tals that are strug­gling or ex­ceed­ing ex­pec­ta­tions un­der the pro­gram and why that is hap­pen­ing.

“For the ones that are suc­ceed­ing, our hope is to find that se­cret sauce and make sure ev­ery­body knows how to bet­ter serve their pa­tient pop­u­la­tion,” Fos­ter says.

Wider im­pli­ca­tions

The un­in­tended bi­ases of hos­pi­tal pa­tientsat­is­fac­tion sur­veys also could arise in other health­care set­tings where the CMS is re­port­edly plan­ning to add sur­veys, in­clud­ing physi­cian of­fices and out­pa­tient hos­pi­tal care.

Out­pa­tient treat­ment cen­ters could present some of the great­est ac­cu­racy chal­lenges, ac­cord­ing to hos­pi­tal of­fi­cials, be­cause am­bu­la­tory health­care ser­vices dif­fer so widely. A sin­gle sur­vey could strug­gle to cap­ture per­cep­tions of care qual­ity among pa­tients vis­it­ing a pri­mary-care clinic and those re­ceiv­ing com­plex treat­ments, such as chemo­ther­apy.

“How you cre­ate a sur­vey that works for all of those dif­fer­ent kinds of clin­ics and is use­ful and help­ful in un­der­stand­ing what’s go­ing right and what’s not go­ing as well as you in­tend and then en­able com­par­isons across set­tings is a sig­nif­i­cant chal­lenge,” Fos­ter says.

An­other po­ten­tial com­pli­ca­tion also emerg­ing from new pa­tient-sat­is­fac­tion sur­vey analy­ses is that high marks for per­cep­tions of care may have lit­tle con­nec­tion to high qual­ity clin­i­cal out­comes.

One re­cent anal­y­sis of Medi­care pa­tientsat­is­fac­tion sur­vey data and mor­tal­ity sta­tis­tics, con­ducted and pub­lished this month by USA To­day, found hos­pi­tals given the high­est pa­tient rat­ings also had high death rates.

The find­ings didn’t sur­prise Mer­lino of the Cleve­land Clinic be­cause pa­tients are not trained to un­der­stand the fac­tors that con­sti­tute high-qual­ity care.

“They can mis­take that for how they are treated in the en­vi­ron­ment, which is not a good sur­ro­gate marker for the med­i­cal care that is be­ing de­liv­ered to them,” he says.


Newspapers in English

Newspapers from USA

© PressReader. All rights reserved.