Objective: To determine the extent to which geographic variation in invasive cardiac procedures can be explained by the variable use of diagnostic testing.
Design: A population-based cohort study using Medicare Part B data (physician services).
Setting and subjects: Procedure data for all Medicare beneficiaries in northern New England.
Main outcome measures: Twelve coronary angiography service areas were constructed for Medicare beneficiaries in northern New England. Age- and sex-adjusted utilization rates were developed for three procedure categories: total stress test, coronary angiography, and revascularization. Total stress tests were further stratified into nonimaging and imaging procedures (eg, thallium). Tests performed in follow-up to invasive procedures were excluded (eg, stress test following revascularizations). Linear regression was used to assess the relationship between procedure categories.
Results: A tight positive relationship was found between total stress test rates and the rates of subsequent coronary angiography (R2=0.61, P<.005). Most of the variance was explained by imaging stress tests (R2=0.50, P<.02). A strong relationship was found between coronary angiography and revascularization (R2=0.82, P<.001). Finally, a clear relationship between total stress tests and subsequent revascularizations was also found (R2=0.55, P<.006).
Conclusion: The population-based rates of diagnostic testing largely explained the variance associated with subsequent therapeutic interventions. Our results suggest that local testing intensity is an important determinant of the variable use of invasive cardiac procedures.