Restructuring Payment to Improve Quality and Efficiency Improve Quality and Efficiency Harold S. Luft, PhD Harold S. Luft, PhD Palo Alto Medical Foundation Research Institute U i University of California San Francisco it f C lif i S F i Payment affects quality, as well as efficiency • Amounts paid may be inadequate for quality • Payment structure, however, is more important This leads us to ask: • How does one assess the quality of what is purchased? • How does one value that which does not h happen? ? Classic Fee‐for‐Service Classic Fee for Service • Unit Unit of transaction is the individual service of transaction is the individual service • There are certain expectations regarding the quality of what is provided (i.e., fraud) • Occasional concerns about whether what was Occasional concerns about whether what was done was egregiously in error (i.e., malpractice) • Little focus on whether what was provided was necessary (i.e., caveat emptor) • But, with third party payment it is unclear who the assessor is the assessor is Classic Capitation Classic Capitation • The unit of transaction is “a year’s worth of necessary medical care” y • The focus is on an undefined package of services • Implicitly, this requires an assessment of what is I li i l hi i f h i necessary and should be provided • Risk adjustment is needed for payment to be fair • But who should assess what the “package” yields B t h h ld h t th “ k ” i ld in terms of outcomes and patient satisfaction? Problems with the Classic Models Problems with the Classic Models FFS • Focus is on the inputs • Providers are only paid for what they do for what they do • There is no payment for coordination of care coordination of care • Fee schedules are skewed • There is no natural focus for assessing quality Capitation • Incentives to avoid risk • High quality attracts risk • Risk adjustment is weak Risk adjustment is weak • Impossible to capitate i di id l individual providers id • Incentives to do less or to refer to others • Micro‐management What are the problems we’re trying to solve? • Patients do not have the expertise to assess exactly what services they need y y • Capitation needs a way to transfer to providers the incentives plans have for coordination the incentives plans have for coordination • FFS (with third party intermediaries) can overly incentivize provider production • People need insurance, but this is because: People need insurance but this is because: – Of occasional very expensive medical care needs – Chronic illness coverage isn’t available ex post Four major types of care j yp • Major acute and interventional (MA/I) ~47% – Hospitalization and its equivalents – Infrequent, expensive, short; managed by teams • Chronic illness management (CIM) ~24% – Ongoing outpatient management g g p g – No clear endpoint, expensive; complex referrals • Minor acute Minor acute ~25% 25% – Frequent, relatively low cost – Primary and at home care; often Primary and at home care; often ‘first first contact contact’ • Preventive ~4% – Sometimes very beneficial, sometimes not A mixed model for payment A mixed model for payment The “occurrence risk” is borne by the plan The “production The production risk risk” is shifted to providers is shifted to providers Patients may pay premiums & minor acute care This separates the major risk‐bearing aspects of insurance from low cost, routinely needed care , y • Focus is on “units” the patient can understand • The patient bears manageable, marginal costs • Use income Use income‐based based subsidies for equity subsidies for equity • • • • Episode payment for Major Acute and Interventional (MA/I) Care • • • • • • • Include pre‐admission and some post‐discharge Payment to a Care Delivery Team (CDT) ( ) A voluntary organization of a facility and clinicians Excludes physicians who would order admissions Episodes are analogous to expanded DRGs (EDRG) Episodes are analogous to expanded DRGs (EDRG) Payment includes all services the providers order Members of the CDT decide their governance and b f h d d h d how to share the revenue amongst themselves Setting the payment for an EDRG Setting the payment for an EDRG • Avoid externally imposed quality measures • Instead, incentives for continuous improvement Instead, incentives for continuous improvement • Set payment for each EDRG at the average for those CDTs with above‐average h C ih b outcomes – Medicare model would use adjusted costs – Private sector would use negotiated charges • Until Until outcome measures are set, assume equality outcome measures are set assume equality across CDTs Beyond Lake Wobegone y g • • • • • • Focus on above average outcomes, not processes g , p Outcomes should be those patients care about This may mean collecting post‐discharge data Data from the top half determines the $ amount Data from the top half determines the $ amount All CDTs get paid this average figure Some CDTs will claim they have sicker patients – They can put forward better risk adjustments They can put forward better risk adjustments • Some will want to tout their high quality scores – They can self‐identify and provide details to support Incentives to Improve • EDRGs cover the average cost of those with above average outcomes (no outlier add‐ons) • CDTs make their money by reducing costs CDTs make their money by reducing costs – Collective cost reduction lowers future payments • CDTs gain business by showing they are “stars” b b h h “ ” • Business opportunities for consultants to learn pp best practices and teach to others • New technologies that are cost reducing and N t h l i th t t d i d quality constant (or enhancing) will flourish • CDTs will demand evidence before purchasing Chronic Illness Management Chronic Illness Management Offer yearly CIM payment to primary care groups Offer yearly CIM payment to primary care groups Initially it covers just outpatient care and drugs Includes specialty care associated with the illness A small degree of risk sharing may be added for A small degree of risk‐sharing may be added for MA/I events due to the chronic illness • Use FFS claims and an agreed‐upon grouper • • • • – Deduct from the CIM payments those charges that Deduct from the CIM payments those charges that group into the chronic illness (after the fact) • The practice gains if it lowers its overall costs Th ti i if it l it ll t Adding quality in chronic illness care Adding quality in chronic illness care • • • • Payers can solicit measures of good outcomes Probably more patient, than clinically, focused Probably more patient, than clinically, focused Initially use these data for “star” ratings Eventually, test to see if better care is more costly – If so, this may require higher payments and If so, this may require higher payments and premiums – If not, then begin to set CIM payments at the average of the payments to the above average PCPs (groups) of the payments to the above average PCPs (groups) • Risk adjustment will again be needed – But this can follow the lead of the volunteers The Bottom Line(s) The Bottom Line(s) • Bundling Bundling payment can be a powerful way to payment can be a powerful way to improve production efficiency • Measures of quality, however, are critical • Simply expanding the DRG model won Simply expanding the DRG model won’tt work work – Specialty societies and interest groups will focus on ( b bl (probably self‐serving) process measures lf i ) – Offering higher payments just to those meeting criteria puts pressure on the “cut points” – That “fix” fails to encourage dynamic improvement Instead Bundling with Floating Payment Instead, Bundling with Floating Payment • • • • • • • • Focus on patient‐centric outcome measures Allow flexibility in the use of inputs Allow flexibility in the use of inputs Set payments at the average of the “best” Cut‐points are of minimal importance Foster self‐identification Foster self identification of of “positive positive deviants deviants” Seek volunteers for the new payment approach Make data public for multiple assessments Providers will demand knowledge transfer Providers will demand knowledge transfer