Pay-for-Performance: A Decision Guide for Purchasers

Guide Prepared for:

Agency for HealthCare Research and QualityU.S. Department of Health and Human Services

Prepared by:R. Adams Dudley, M.D., M.B.A.University of California San FranciscoMeredith B. Rosenthal, Ph.D.Harvard School of Public Health

2

Pay for Performance:A Decision Guide for Purchasers

Electronic Copy of Guide and other AHRQ P4P Resources: Electronic Copy of Guide and other AHRQ P4P Resources: http://www.ahrq.gov/qual/pay4per.htm

3

Overview

Not a users manual: too little data Addresses:

Developing an overall strategy Incentive design and measures selectionImplementationEvaluation and revision

4

Is Our Community Ready?

Do we know what we are trying to accomplish?

Do we have enough influence? Are the providers ready?

5

Strategic Issues: Getting Started

Voluntary vs. mandatory: Voluntary: easier, may only attract high-

performing providersMandatory (i.e., written into all contracts):

creates uniform incentives, but may need high market share

Bonus program is in between: “mandatory”, but providers are free to ignore it

Phasing in: start with volunteers, or “pay for participation”/“pay for reporting”

6

Strategic Issues: Getting Started

Which providers to target?: Hospitals and/or physicians Large vs. individual/small group

Contribution of hospitals vs. physicians to quality and cost varies from region to region

Measurement issues favor larger groups but incentives may be stronger for individuals

System view of quality improvement suggests higher level

Choose the provider target for which you can get the biggest bang for your buck

7

Increasing Inclusion of Specialists and Hospitals in Pay-for-Performance

96%

46%

32%

100%

75%

50%

0.0%

20.0%

40.0%

60.0%

80.0%

100.0%

120.0%

PCP Specialists Hospitals

2003 2006

8

Choosing Measures

National measure sets should be considered firstTestedAcceptedAlready being collected

Some sources: AHRQ (Inpatient Quality Indicators), National Quality Forum, Hospital Quality Alliance, Ambulatory Care Quality Alliance, NCQA, Leapfrog Group

9

Incentive Design Challenges

All P4P programs are not the same Design choices matter First critical question is orientation:

Quality improvement across all providers, patients?

Rewards for the best only? E.g., Premier Inc./CMS demonstration

10

Explicitly or Implicitly Rewarding Quality Improvement

P4P programs that reward top group (e.g., 20%) or set a benchmark for reward that all must meet do not uniformly encourage improvement

These features should result in more QI:Rewarding improvement explicitly (i.e.

change rather than/in addition to level)Multiple levels of rewards (partial credit)Payments tied to each patient treated well

11

Case Example: Hudson Health Plan: Rewarding Quality Diabetes Management

Measure Reward

Blood pressure $15 for screening and $35 for BP<130/80 or $20 for <140/90 or $15 for ≥10 mmHg decrease in one and goal in the other

Smoking cessation counseling $15

A1C testing and control $15 for screening and $35 for A1C<7 or $20 for A1C<9 or $15 for a 1% or more reduction

LDL-C testing and control $15 for screening and $35 for LDL<100 or $20 for LDL <130 or $15 for evidence of drug tx

Documentation of albuminuria; ACE/ARB treatment if positive

$15 for screening and $35 for negative test, evidence of drug tx, evidence of contraindication, or nephrology consult

Retinal exam $15 for exam with documentation of result

Pneumococcal vaccine $10

Flu shot $10

12

Key Design Issues: How Much Money?

To be effective, bonus should be commensurate with cost of effort

Little good information about what it takes to reach quality targets

Most P4P programs for physicians 5-10% of associated fees; hospitals 1-2%

13

Planning Ahead for Evaluation

You spent all that time and money…shouldn’t you assess what you accomplished?

Aspects of implementation can facilitate evaluation Collecting data during a measurement (i.e. non-

payment) year will allow before/after comparison Implementing P4P for some providers before

others may create a natural comparison group

14

What Types of Effects to Look For

Data collection should not only track intended consequences but also monitor potential side effects:Patient selection/dumping (changes in

case-mix, excessive switching)Diversion of attention away from other

important aspects of careWidening gaps in performance between

best and worst

15

Summary

Pay-for-performance can facilitate improved patient care, cost-efficiency

Best practices still unknown Careful matching of goals and

mechanisms will most likely lead to best results

In light of uncertainties about design, evaluation is key