Unwarranted Variation in Healthcare: Impacts of Test-Order Practice Variation on Care-Delivery Cost

Abstract

Motivated by payment reform model policies that aim to reduce practice variation, this study explores whether and how lower variations in clinical practice relate to hospital resource usage. Specifically, we focus on practice variation in test-ordering (e.g., laboratory and radiology tests) stage and use the statistical process control as a theoretical lens to hypothesize its negative impacts on subsequent care-delivery costs. We define practice variation as all variation not resulting from patient mix, and construct a measure using inpatient discharge data for each patient cohort having an identical medical condition. The high underuse practice variation in test-ordering implies that a hospital is more likely to fail to order proper tests for some patients. Thus, such a hospital may spend extra efforts in care-delivery stages to alleviate potential adverse effects of the test underuse. We also consider the intervening effects of two quality initiatives on the relationship: process quality (i.e., how well a hospital adheres to evidence-based medical guidelines to diagnose and treat patients) and experiential quality (i.e., external perceptions of care quality from a patient’s perspective). Based on a comprehensive six-year inpatient data from New York and Florida states, we find that higher underuse variation in the test-ordering lead to higher care-delivery cost. Interestingly, this phenomenon is even stronger when a hospital provides a higher-quality patient experience because such a hospital tends to provide more responsive care, which is often resource-intensive. Therefore, hospitals may be improperly rewarded for quality improvements if practice variation is ignored, implying that incentives and penalties for hospital operations should be designed to account for such effects. We discuss our findings and policy implications.