The aim of this study was to create new measures of quality that combine individual service measures. Using an all-or-none approach, we identify 5 levels of care reflecting the extent to which optimal patterns of service were obtained by patients with asthma, diabetes, and heart failure. We also assess the feasibility of these levels-of-care measures and their potential value in quality improvement efforts. The study was designed to analyze claims data to reflect patterns of services used in a single metropolitan market of about 1 million residents in the northeastern United States. More than 80,000 patients insured over 4 years (1994–1997) had claims with 1 or more of 3 chronic conditions. The analysis showed that the measures discriminated effectively among groups of patients with the 3 chronic conditions and highlighted areas to target quality improvement efforts. Although the numbers vary by year, for two of the diagnoses, most patients were in the lowest categories (59%–75%), and for the third, 40% were in these categories. Few were in the highest category. Most patients were in the same category from one year to the next. The levels-of-care approach to quality measurement can help caregivers and policy makers find methods for avoiding unnecessary utilization and expenditures while raising—not lowering—the probability that utilization patterns will conform to condition-specific recommended care.