Outcome Measurement in Value-Based Payments.

Valueinhealthcare,thebalancebetweenoutcomesthat mattertopatientsandthecostsrequiredtoachievethem, isbeing increasinglyrecognizedasapathtohealthcarereform.TheDepartmentofHealth andHumanServices recentlyannouncedits intentiontotie50%oftraditionalfeefor-service payments,made by the Centers forMedicare andMedicaid Services (CMS), to value or quality through alternativepaymentmodels(APM), includingaccountable careorganizations(ACOs)andbundledpayments,by2018.1 The Centers for Medicare and Medicaid Services is also payingspecialattentiontocancercare,sinceitaccountsfor nearly$125billioninmedicalspendinganddatarevealswide variations in the cost of care deliveredwith no relation to survival.2 AsAPMsbecomemore commonplace andvaluebecomes the global metric certain challenges arise. To address the numerator of the value equation we must develop, test, endorse, and use meaningful outcomes measuresfortheserviceshealthcaresystemsprovide.Currently, there are an overabundance of validated process measures, and apaucity of outcomemeasures. As policy experts have noted before, a quality measurement approach that iswholly reliant onprocessmeasuresmisses 2 important facts:patients caremoreabout the resultsof theircarethanhowtheseoutcomesareachieved,andprocess measures might contribute to but are not surrogates foroutcomes,andomit factors suchasstaffingpatterns, interdisciplinary communication, supportive infrastructure, and transitions of care.However,measuring and reporting outcomes that matter to patients has proventobedifficult,andareespeciallychallengingincancer, as evidenced by the slow pace of measure development and adoption over the past 2 decades. In1999,theInstituteofMedicine(IOM)publishedEnsuringQualityCancerCare,3an influential report thatgeneratedmuchfervor. Itdescribedanaspirationalcancercare system and issued 10 recommendations to address pervasivegapsinthedeliveryofqualitycare.Asof2014,these recommendations remained largely unfulfilled.3 Further evidence comes from theDepartment ofHealth andHuman Services, which commissioned the National Quality Forum (NQF) in 2010 to identify areas where outcome measureswereneededbutnotyetdeveloped.Theresulting gapanalysis found that very fewoutcomesmeasures for cancer hadbeen endorsed, and those that havewere focused primarily on end-of-life care.4 Toupdate that analysis,we comparedqualitymeasures for colorectal, breast, prostate, and lung cancer, endorsed by the NQF, ASCO’s Quality Oncology Practice Initiative (QOPI), and the International Consortium forHealthOutcomesMeasurement (ICHOM), aswell as those that are being tracked by several major APMs including Medicare Shared Savings ACOs, the PPSExempt Cancer Hospitals Quality Reporting Program, and the Oncology Care Model (OCM). As shown in the Table, all of QOPI’s measures, and themajorityofNQF’smeasures representprocessmeasures. Moreover, none of the APMs plan to track actual outcome measures. The quality measures that will be used inMedicare’supcomingMerit-Based IncentiveProgram are yet to be finalized, so theywere not included. Insomerespects,itisnotsurprisingthatidentifyingand measuringmeaningful outcomes in oncology is difficult. Cancer represents a wide spectrum of heterogeneous diseases, and a detailed outcomemeasuremay be applicable to a limited set of patients in anygiven timeperiod. In addition delivering cancer care is complex owing to its multi-disciplinarynature(thoughsomecancercentersare startingtocollocatemultiple treatmentandpalliativedisciplines).Furthermore,canceroftendoesnotfollowthelinearprogressionofdiseasemorbidityofother chronicdiseases,suchasheart failureordiabetes,butratheradheres toawindingcoursethatcanabruptlyshift inacuity.Finally, qualitymeasures thatwouldapply topatients in theadjuvant settingmay no longer applywhen patients develop treatment-refractorydiseaseandarepursuingcomfortcare. Butalthoughdevelopingsoundoutcomemeasures is hard, it isnotimpossible.TheICHOMhasusedglobalteams ofphysicianleaders,outcomesresearchers,andpatientadvocategroupstoproposecancerdisease-specificstandard outcomemeasures(Table).ComparedwithNQF,QOPI,and APMmeasures, ICHOMplacesa largeremphasisonPROs, previouslyhighlighted intheNQF’sgapanalysis.Whereas theASCOQualityofCareCommitteehasfocusedonusing PROs to assess pain and chemotherapy-induced nausea andvomiting, ICHOMhasgoneastepfurthertoassess importantPROsspecific foreachcommoncancertype,such assexualdysfunctionandincontinenceinprostatecancer, to address theheterogeneity of this disease.5 Notably, the OCM requires a documented care plan that contains the 13 components of the IOM’s caremanagementplan,which include itemssuchasgoals of treatment, estimated out-of-pocket costs, defining physician responsibilities (oncologistvsprimarycarephysician),expected effects on quality of life, and posttreatment surveillance plans.6 Although care planning is not outcome measurement per se, this care plan represents anopportunityto improvepatientengagementandmovecloserto thevalueagenda.TheinclusionofthecareplanintheOCM doesrepresentCMMI’sability toadoptmeasuresnotpreviouslyvettedbytheNQFandCMSinapilotprogramthat may set the way for more rapidly testing true outcome measures in upcoming CMMI programs. It remains to be seenwhethersuchacceleratedadoptionofmeasuresthat proffers theopportunity toquickly iterate andpivot is ultimatelyabetterstrategythanrigorously testingandvalidatingmeasures in controlled experimental settings. What is needed now is an accelerated path for canceroutcomemeasurestobetestedandendorsedthrough theNQF-convenedcollaborations.Totheircredit, theNQF VIEWPOINT