Method: Surveillance Epidemiology and End Results (SEER)-Medicare cases diagnosed in 2004 who were enrolled in Medicare’s fee-for-service program from 365 days prior to the time of diagnosis to three years (1095 days) after the date of diagnosis were compared with matched controls. Frequency matching was based on SEER-region, sex, race, age, and Charlson comorbidity index category. Quality indicators were drawn from previously reported research; indicators represent follow-up, monitoring, and continuing care for comorbid chronic conditions (including diabetes, coronary artery disease, chronic heart failure, stroke/transient ischemia, anemia and GI bleed, COPD/asthma, and depression), avoidance of serious clinical manifestations, therapeutic intervention, and workup at initial diagnosis. We used simple linear regression and multivariable linear regression controlling for SEER region, sex, race, Charlson comorbidity index, and the census tract median income and the proportion of individuals in a census tract with high school (or less) education only to assess the relationship between quality indicators and total costs from days 366 through 1095. Only those eligible for each quality indicator were included in the relevant comparison.
Result: Frequency matching yielded two controls (N=17322) for each case (N=8661). The matched controls had a statistically significant relationship between 19 of the 36 quality indicators using simple linear regressions while the cancer cases had 13 statistically significant relationships. In multivariable regression analyses, there were 17 and 12 statistically significant relationships respectively. Measures of appropriate continuing care (i.e., appropriate frequency of visits) for chronic conditions and appropriate workup at initial diagnosis were associated with higher overall costs. Avoidable serious clinical manifestations accounted for larger positive incremental costs. Appropriate post-inpatient care was associated with lower costs for the small number of statistically significant relationships. More frequently than not, incremental changes in costs associated with measures of quality were larger (more positive or more negative) for the matched controls than for cancer survivors.
Conclusion: We found some associations between better care quality and lower costs, particularly for those who avoid serious clinical manifestations and who obtain appropriate post-inpatient care. Relationships tend to be stronger for non-cancer controls compared to cancer survivors. In other cases, providing higher quality care, particularly ongoing care of chronic conditions, is associated with higher costs.
See more of: The 33rd Annual Meeting of the Society for Medical Decision Making