1. Implementing CG-CAHPS: Issues and Strategies Dale Shaller, MPA Shaller Consulting Group September 18, 2011

2. 2 Forces Driving Use of CG-CAHPS Public Reporting –AF4Q and CVE initiatives –State mandates –Possible use in PhysicianCompare ACOs and Value-Based Purchasing Patient-Centered Medical Home HRSA Bureau of Primary Health Care American Board of Medical Specialties Rising consumer and patient expectations

3. Profile of CG-CAHPS Users 12-Month Version Public reporting initiatives in CA, MA, and other markets Some health plans and systems (CA, MI, WI, MA) Medical home evaluations Department of Defense Visit Version Public reporting initiatives in MN, WI, MI, ME, and other markets Growing numbers of medical practices (including UHC and 6 safety net clinics in CA) Vendors such as Press Ganey, NRC, Avatar ABMS for MOC (Doctor Communication items) 3

4. 4 CG-CAHPS Database Composition (as of December 2010)

5. 5 Key Implementation Issues Survey version Patient populations and languages Unit of sampling and reporting Source of sample frame Sample size Data collection mode Data aggregation, analysis, and reporting

6. 6 Survey Version Selection of survey version driven by user objectives, e.g.: –Internal improvement –External reporting 12-month version –Works well for assessing experiences that transcend individual visits –Commonly used for external reporting Visit version –Preferred by many clinicians for internal improvement

7. 7 Patient Populations and Languages Primary/specialty care Adults/children Commercial/Medicaid/Medicare/Other Patients with chronic conditions English-speaking patients or other

8. 8 Sampling and Reporting Unit Units of sampling and reporting include: –Individual clinician –Clinic or practice site –Medical group or health system –Community/state/region/other Sampling and reporting units are often not the same –Users may sample at clinician level for internal use but report results externally at higher levels

9. Sample Size 9 CAHPS guidelines: 45 completes per provider 300 completes per medical group ~ 220 completes per practice site (based on MN pilot) New estimates for site- level samples are under development NCQA recommendations for PCMH survey at site level:

10. 10 Data Collection Modes: Outbound Mail Telephone –Landlines –Cell phones Interactive Voice Response (IVR) –Touchtone IVR –Speech-enabled IVR In-office distribution –Paper survey –Kiosk or other electronic modes Email distribution

11. 11 Field Period May depend on sampling method –Continuous –Point in time Same field period needed for comparability of results –Ex: 3 rd quarter of the year

12. 12 Regional Implementation Models Centralized Model –Single vendor –Sample frame drawn from combined files of health plans or medical groups –Examples: MHQP, PBGH, CHECKBOOK Decentralized Model –Medical practices use their own vendors –Integrate CG-CAHPS into current surveys –Aggregation of multiple data sets through a neutral vehicle (CAHPS Database) –Examples: MN, Detroit, Maine, and WI

13. 13 Minnesota: Leveraged Model 18 medical groups, 110 clinic sites 3 different vendors (PG, NRC, PRC) Common administration protocol –Sampling –Administration (mail mode) –Field period CAHPS Database merged files and produced clinic- level results for reporting

14. 14 Massachusetts: Centralized Model Over 500 practice sites Single vendor financed by health plans Results reported privately to systems, then publicly (every two years) Systems collect own data internally more frequently, using same or different survey instruments

15. 15 Implementation Models: Pros and Cons

16. 16 Challenges Ahead Reconciling multiple survey requirements –Internal improvement –External reporting Reducing cost of implementation to achieve sustainable business models –Using one survey and administration for multiple requirements –Lowering administration costs through new data collection technologies