The 2nd P in P4P: Towards a Hybrid Data Strategy Irene Fraser, Ph.D.. Director Center for Delivery, Organization and Markets Presentation to Mini-summit on HIE and P4P National Pay-for-Performance Summit March 10, 2009

Accurate Measures and Data Are Fundamental for P4P System Reform and Redesign for Quality & Value Incented by payment and other incentives Documented Results Quality Patient Safety Efficiency Access ROI For ALL Americans Informed by evidence and models of successful design strategies Facilitated by improved HIT

Achieving Good Measures and Data Requires Multiple Bridges Between… HIT World EMR World Clinical data Health Info. Exchanges Measure developers Medicare and Medicaid data And… Quality world Claims data world Claims data State data orgs. Data developers Private payer data

AHRQ Work to Strengthen EMR Measures and Data Work w/ NQF to add measures to EMR Supporting HIEs Grants to improve ambulatory care measures Facilitating use of e-prescribing Jonathan.White@ahrq.hhs.gov

Working from the Other Side: AHRQ Work to Strengthen Claims Data Healthcare Cost and Utilization Project Quality Indicators Efficiency Measures Public Reporting Template Portals with software for local use

A Voluntary Federal-State-Private Sector Collaboration The HCUP Partnership: A Voluntary Federal-State-Private Sector Collaboration WA MT VT ND ME MN OR NH ID WI SD MA NY WY MI RI IA PA NV NE CT OH UT IL IN NJ CO DC CA WV MO VA KS DE KY NC MD TN AZ OK One of the most fundamental things to understand about HCUP – one of the things that distinguishes it from most other data collection efforts – is that it is a partnership, a voluntary partnership. This partnership model is unique and is an incredible strength. We (AHRQ) leverage the very substantial investments that are made by each of the statewide data organizations that contribute their data to HCUP. So for a relatively small investment, we are able to work in unison to create a national data resource. But this also adds complexity. First, and foremost, it means that we work within the constraints set by our partners’ needs, rules and requirements. AHRQ and the statewide data organizations sign a MOA, which documents both of our responsibilities, as well as the intended and appropriate uses of the data. While AHRQ provides the resources (get back to this), ($ and staff) to support the project, HCUP must also provide value to our state partners. Statewide data organizations participate for a variety of reasons, but each must see a benefit to belong in the partnership. The benefit they see often resonates with their own capacities, mission, visions and budgets. We interact with our state partners regularly, and we bring them together yearly to update and discuss the project in person with them. (skin in the game) Describe slide. We often are asked why not all the states – more than one answer to the question. NM AR SC AL GA AK MS Legend TX LA HCUP Partner FL Potential Partner Does Not Collect Statewide Inpatient Data 40 states in partnership 90% of all discharges HI Not a Partner

HCUP Databases State Ambulatory Surgery Databases (SASD) State Inpatient Databases (SID) State Ambulatory Surgery Databases (SASD) State Emergency Department Databases (SEDD) As a quick review – you can see from this slide that the NIS is produced from HCUPs state-level databases. The NIS contains a stratified, random sample of community hospitals from all the SID in a single data year. All of the discharges (not a sample of discharges) from the selected hospitals are included. This database approximates a 20% sample of community hospitals in the U.S. Nationwide Inpatient Sample (NIS) Kids Inpatient Database (KID) Nationwide Emergency Department Sample (NEDS)

What is HCUP? And what is it not? HCUP is... HCUP is not... Discharge database for health care encounters A survey All payer, including the uninsured Specific to a single payer, e.g. Medicare Hospital, ambulatory surgery, emergency department data Outpatient visits, pharmacy, laboratory All hospital discharges A sample Accessible multiple ways: raw data, regular reports, on-line (HCUPnet) Just another database It is important to understand what HCUP is...and, what it’s not. (Read slide) The raw data are the ascii flat files of the NIS and SID. The regular reports are the SBs, Fand F, methods series technical reports Online is the HCUP tools and software, HCUPnet, HCUP-US database, the TA

AHRQ Quality Indicators (QIs) Developed at behest of state partners Use existing hospital discharge data Incorporate severity adjustment Current modules: Inpatient, Patient Safety, Pediatric and Neonatal, Prevention Include composites Growing use for reporting and P4P NQF endorsement for 40+ so far CMS using 9 under new Inpatient Payment rule Evidence-based public reporting template available 14 states use AHRQ QIs for public reporting

14 States Use AHRQ QIs for Public Hospital Reporting Wisconsin (parts of state) Vermont Iowa NewYork Oregon Ohio Massachusetts New Jersey Kentucky California Utah Colorado Florida Texas

Next Frontier: Measuring Cost and Efficiency Evidence Review on Efficiency Measurement Workshops on hospital efficiency measurement and physician-level public reporting HSR issue on “Improving Efficiency and Value in Health Care” New efficiency chapter in Nat’l Healthcare Quality and Disparities Reports National admissions and costs for aggregate PQIs Relative hospital cost efficiency index

Study Shows Cost-effective Enhancements to Admin. Data* Assessed impact of incrementally adding more complex clinical information Administrative data can be improved at relatively low cost by: Adding present on admission (POA) modifiers Adding numerical lab data on admission Improved coding AHRQ Awarded pilots in VA, FLA, MN, planning contract in WA to show proof of concept We conducted a study under contract with Abt and Michael Pine and Associates, to identify the most cost-effective clinical data elements that could be added to administrative data for the purpose of publicly reporting hospital quality of care. What we found was that: Adding POA information (CMS rule – October 2007) Adding lab data from the day of admission and Improved ICD coding Added nearly as much information to mortality models as full clinical information at a fraction of the cost. * Pine M, Jordan HS, Elixhauser A, et al. Enhancement of claims data to improve risk adjustment of hospital mortality. JAMA 2007, 267 (1): 71-76. Jordan HS, Pine M, Elixhauser A, et al., Cost-effective enhancement of claims data to improve comparisons of patient safety. Journal of Patient Safety 2007, 3(2): 82-90. Fry DE, Pine M, Jordan HS, et al. Combining administrative and clinical data to stratify surgical risk. Annals of Surgery 2007 (forthcoming).

Administrative/Hybrid Data: The Future Improve timeliness Provide on-line all-payer market-level data on cost, quality, efficiency, price. Add clinical detail, data links for accuracy, credibility Expand outpatient reach (e.g. physician, episode) Pilot cross-site data, new data links New tools for expanded data Additional states, as feasible Develop, validate, maintain, deploy measures in priority areas Expand data elements to align with levers of change Tools for change Some ideas that AHRQ will be exploring for this “joining forces” effort are: Use HIT to improve the timeliness of the reporting of data to state Partner oraganizations Add clinical detail to improve quality measurement, including an indicator of whether each diagnosis was present before admission to distinguish complications of hospital care from patient comorbidities. And a recent AHRQ-funded study shows that there are 20 lab test values that would greatly enhance the accuracy of the claims data for hospital quality reporting. Expand statewide outpatient data, develop new data analysis tools for outpatient and clinically enriched data, and do this with continued vigalence of patient privacy and data security,

Frontier in All-Payer Claims Data: Data ACROSS Sites: Hospital Data ED Data A. Surg. Data Cross-site Data

QI Reporting Template(s) Challenge: Presentation Matters!! Approach: Two Model Templates Composite scorecard Health topic/disease Report Sponsors Choose: Overall approach Topics, composites, & measures to report How scores will be calculated The medium to be used Model reports are based on: Literature review and analysis Interviews with experts Focus groups with different populations Cognitive interviews The idea for the reporting template originated as a request from the QI user community. These reports are designed specifically to report comparative information on hospital performance based on the AHRQ QIs. There are two model reports, one developed in mid-2006, that groups individual indicators together and organized them into Health Topics. This first model is known as the Health Topics Model Report. The second model, known as the Composite Model Report was developed later in 2006, and is based on a set of four composite measures using a substantial portion of the IQIs, the PSIs, and the PDIs. These composites were created on the basis of substantial statistical analyiss as well as an expert review. 15

Preventable Hospitalization Costs: A County-level Mapping Tool Potentially Preventable Hospitalizations cost over $30B a year. Maps show the admission rates for health problems by county. Calculates cost savings if admissions are reduced. Can add information about local populations show number of persons at greatest risk for health problems in each county.

NEW! Portal for States, Communities, Others to Display, Analyze Data Query paths currently in HCUPnet Results based on AHRQ QIs Preventable Hospitalization Costs mapping tool New ways of presenting information Beyond tables QI Reporting Template Other AHRQ tools? Other information? Strategies for Improvement The first version of EQUIP, which should go into beta-test early next year will contain some of the most popular AHRQ tools: The query paths currently in HCUPnet Results from the AHRQ QIs And the Preventable Hospitalization Costs mapping tool The intent is to present the information in an accessible, user-friendly way that relies on helpful use of graphics and reporting characteristics based on the best evidence currently available. We hope to add more AHRQ tools in the future – such as CAHPS – and perhaps other information such as process measures from Hospital Compare.

Sites for Implementing Measures and Data: Chartered Value Exchanges

Coming Up Next: Starting some Bridge-Building Janet Marchibroda – National View on Improving HIE J. Marc Overhage – HIE and Pay for Performance Linda Davis – Direct Submission of Clinical Data Discussion