USING CLINICAL AND FUNCTIONAL DATA FOR QUALITY IMPROVEMENT IN OUTCOMES MEASUREMENT CONSORTIA

Background: Using standardized measures, American Group Practice Assoc iation care providers compiled a national database from which patients could be tracked, allowing for epidemiologic comparisons among treatm ents, sources of care, and results. Within five years, the consortia e xpanded form 6 to 55 clinics and from a focus on total hip replacement surgery to eight different health conditions. Data collection process : Reflecting areas of significant public concern, high prevalence, hig h cost, or research needs, both patient- and provider-source data are collected at group practices at standardized intervals: 6 months, 12 m onths, and annually thereafter. Outcomes data management and reporting : A readily adaptable database infrastructure allows data-collection t ools to adapt individual questions to changing conditions in the healt h care environment while maintaining the integrity of the whole struct ure. Aggregate-level reports complements individual clinics' own inter nal analysis efforts by providing a context for the interpretation of results. Case studies: In a total hip replacement consortium (includin g more than 2,300 patients), early findings have shown that patients d o not fully recover from surgery as quickly as they themselves anticip ated. In the cataract consortium, data scores on near, distant, day, n ight, glare, and overall vision scales improve considerably after cata ract surgery giving clinics a new tool to monitor and improve performa nce. In the asthma consortium, one clinical reevaluated the distributi on of peak flow meters and the reasons for their underuse after noting the low number of patients in the clinic that had peak flow meters. L essons learned: Provider participation in study design, instrumentatio n, data analysis, and feedback is important, and physician buy-in and support are critical to the success of any outcomes initiative. Missin g data are the greatest limitations of a longitudinal data set and are difficult to collect through follow-up. There is still much to be lea rned about what functional status and well-being measures can slow abo ut the relationship between health care services and patient health.