Working Paper: NBER ID: w19956
Authors: Jason Abaluck; Leila Agha; Christopher Kabrhel; Ali Raja; Arjun Venkatesh
Abstract: We develop a model of the efficiency of medical testing based on rates of negative CT scans for pulmonary embolism. The model is estimated using a 20% sample of Medicare claims from 2000- 2009. We document enormous across-doctor heterogeneity in testing decisions conditional on patient risk and show it explains the negative relationship between physicians' testing frequencies and test yields. Physicians in high spending regions test more low-risk patients. Under calibration assumptions, 84% of doctors test even when costs exceed expected benefits. Furthermore, doctors do not apply observables to target testing to the highest risk patients, substantially reducing simulated test yields.
Keywords: medical testing; pulmonary embolism; physician behavior; diagnostic imaging; healthcare efficiency
JEL Codes: I01; I12
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| physicians' testing frequencies (I11) | test yields (C52) |
| physicians in high-spending regions (I11) | test frequencies of low-risk patients (I11) |
| physicians' testing frequencies (I11) | costs exceeding expected benefits (H43) |
| observable risk factors (J28) | targeting testing towards highest-risk patients (I11) |
| misweighting of risk factors (D91) | expected benefits of testing (C52) |
| previous diagnoses of similar conditions (I12) | overtesting of patients (I11) |