STUDY QUESTIONS: What is the extent of variation in patterns of ambulatory care practice across one state's Medicaid program once case mix is controlled for? How much of this variation in resource consumption is explained by factors linked to the provider, patient, and geographic subarea? DATA SOURCES/STUDY SETTING: Practices of all providers delivering care to persons who were continuously enrolled in the Maryland Medicaid program during FY 1988 were studied. A computerized summary of all services received during this year for 134,725 persons was developed using claims data. We also obtained data from the state's beneficiary and provider files and the American Medical Association's masterfile. Each patient was assigned a "usual source of care" (primary provider) based on the actual patterns of service. The Ambulatory Care Group (ACG) measure was used to help control for case mix. STUDY DESIGN: This was a cross-sectional study based on the universe of continuously enrolled Medicaid enrollees in one state. PRINCIPAL FINDINGS: After controlling for case mix, the variation in patient resource use by type of primary provider was 19 percent for ambulatory visits, 46 percent for ancillary testing, 61 percent for prescriptions, and 81 percent for hospitalizations. Across Maryland counties, comparing the low- to high-use jurisdiction, there was 41 percent variation in case mix-adjusted visit rates, 72 percent variation in pharmacy use, and 325 percent variation in hospital days. At the individual practice level, physician characteristics explain up to 17 percent of ambulatory resource use and geographic area explains only a few percent, while patient characteristics explain up to 60 percent of variation. CONCLUSIONS: Since a large proportion of variation was explained by patient case mix, it is evident that risk adjustment is essential for these types of analyses. However, even after adjustment, resource use varies considerably across types of ambulatory care provider and region, with consequent implications for efficiency of health services delivery.
STUDY QUESTIONS: What is the extent of variation in patterns of ambulatory care practice across one state's Medicaid program once case mix is controlled for? How much of this variation in resource consumption is explained by factors linked to the provider, patient, and geographic subarea? DATA SOURCES/STUDY SETTING: Practices of all providers delivering care to persons who were continuously enrolled in the Maryland Medicaid program during FY 1988 were studied. A computerized summary of all services received during this year for 134,725 persons was developed using claims data. We also obtained data from the state's beneficiary and provider files and the American Medical Association's masterfile. Each patient was assigned a "usual source of care" (primary provider) based on the actual patterns of service. The Ambulatory Care Group (ACG) measure was used to help control for case mix. STUDY DESIGN: This was a cross-sectional study based on the universe of continuously enrolled Medicaid enrollees in one state. PRINCIPAL FINDINGS: After controlling for case mix, the variation in patient resource use by type of primary provider was 19 percent for ambulatory visits, 46 percent for ancillary testing, 61 percent for prescriptions, and 81 percent for hospitalizations. Across Maryland counties, comparing the low- to high-use jurisdiction, there was 41 percent variation in case mix-adjusted visit rates, 72 percent variation in pharmacy use, and 325 percent variation in hospital days. At the individual practice level, physician characteristics explain up to 17 percent of ambulatory resource use and geographic area explains only a few percent, while patient characteristics explain up to 60 percent of variation. CONCLUSIONS: Since a large proportion of variation was explained by patient case mix, it is evident that risk adjustment is essential for these types of analyses. However, even after adjustment, resource use varies considerably across types of ambulatory care provider and region, with consequent implications for efficiency of health services delivery.
Authors: Sara E Evans-Lacko; Susan Dosreis; Elizabeth A Kastelic; Cristiane S Paula; Donald M Steinwachs Journal: Prim Care Companion J Clin Psychiatry Date: 2010
Authors: Amy K Rosen; Susan A Loveland; Jennifer J Anderson; Cheryl S Hankin; James N Breckenridge; Dan R Berlowitz Journal: Health Serv Res Date: 2002-08 Impact factor: 3.402