Imaging Room and Beyond: The Underlying Economics Behind Physicians’ Test-Ordering Behavior in Outpatient Services

• Published in: Manufacturing & service operations management Vol. 19; no. 1; pp. 99 - 113
• Main Authors: Dai, Tinglong, Akan, Mustafa, Tayur, Sridhar
• Format: Journal Article
• Language: English
• Published: Linthicum INFORMS 01.01.2017; Institute for Operations Research and the Management Sciences
• Subjects: Analysis; Consumer behavior; Decision making; Diagnostic tests; Economic aspects; Economic incentives; Health care industry; Health insurance; healthcare management; incentives in healthcare operations; Insurance coverage; Medical imaging; Physicians; Practice; queueing games; Queuing theory; Reimbursement; service operations; Services; Waiting period
• ISSN: 1523-4614; 1526-5498
• DOI: 10.1287/msom.2016.0594

Motivated by a collaborative study with one of the most comprehensive ocular imaging programs in the United States, we investigate the underlying three-way trade-off among operational, clinical, and financial considerations in physicians’ decisions about ordering imaging tests. Laboratory tests may be processed in parallel and thus have a limited effect on patients’ waiting times; imaging tests, by contrast, require patient presence and thus directly influence patients’ waiting times. We use a strategic queueing framework to model a physician’s decision of ordering imaging tests and show that insurance coverage is the key driver of overtesting. Our further analysis reveals the following: (i) Whereas existing studies hold that lower out-of-pocket expenses lead to higher consumption levels, we refine this statement by showing the copayment and the coinsurance rate drive the consumption in different directions. Thus, simply expanding patient cost sharing is not the solution to overtesting. (ii) Setting a low reimbursement ceiling alone cannot eliminate overtesting. (iii) The joint effect of misdiagnosis concerns and insurance coverage can lead to both overtesting and undertesting even when no reimbursement ceiling exists. These and other results continue to hold under more general conditions and are therefore robust. We enrich our model along two extensions: one with patient heterogeneity in diagnostic precision, and the other with disparities in health insurance coverage. Our findings have implications for other healthcare settings with similar trade-offs.