Variation in Private Insurance Payments

Health insurers pay vastly different sums for the care they purchase based largely on geography.

Understanding payment variation in health care services is essential to improving quality and reducing costs. Yet, few studies have examined price variation outside of Medicare and Medicaid. One recent article in Health Affairs evaluated national data from the Truven Health Market Scan Commercial Claims and Encounters Database. This dataset includes claims for approximately 33 million covered lives. The authors analyzed “allowed amount” payments for office payments (established and new patients), office consultations, and preventative visits (established patients).

To evaluate payment variation, the authors examined the mean, median, 5^th, 10^th, 90^th, and 95^th percentiles of allowed physician payments. The investigators used the ratio of payment amounts for the 95^th-to-5^th percentile to provide a single metric encompassing the range of variation while remaining robust to outliers. Additionally, the authors used analysis of variance to determine the proportion of payment variation explained by patient (age, gender), physician (speciality, location, and “network” membership), and plan characteristics (PPO vs. POS plan).

This study yielded two primary findings. First, for the vast majority of services, patient, physician and plan characteristics explained less than 1% of the variation in payment amount. In contrast, one-third of variation was explained by geographic location as measured by metropolitan statistical area (MSA). Second, there was wide variation within-MSAs. In some MSAs the highest paid providers (highest 95^th percentile) received over 200% more than the lowest paid providers (lowest 5th percentile). However in other MSAs, the highest paid providers received only 60% more than the lowest paid. The article did not identify MSAs by name owing to confidentiality provisions in their data use agreements. Because the study included only patients in PPO and non-capitated POS plans, the results may not be widely generalizable or nationally representative.

Commentary

With the expansion of health insurance under the Affordable Care Act, this study is particularly timely. As with many preliminary studies, this article mainly adds to the existing literature by characterizing a source of significant variation. Perhaps the study’s greatest value derives from its dataset and choice of endpoints. Unlike many other studies, this paper evaluates actual transaction amounts, not mere charges. Further, the endpoints by which they assessed price variation in private insurance markets was quite insightful. By using the ratio of the 95^th-to-5^th percentile, they were able to capture the variation in a single number while remaining robust to outliers.

As with much thoughtful research, this study may ask more questions than it answers. The most critical of which would be to determine if price variation in private insurance is associated with quality of care. The authors also provided other avenues for further research, namely to investigate the impact of illness severity, market power, and search costs on price variation.

Baker, L, et al. Health Affairs. 2013; 32 (9):1583-1590.

by

Twitter: @DrAGonzMD