Examining Cost Differences in Malignant Brain and Central Nervous System Tumor Treatment Across US Military Health System Care Settings.
AI interpretation is pending for this paper.
Open original publication →What the AI sees
Not AI summarized yet.
Research significance
Pending deeper interpretation.
Source abstract
GOAL: The objectives of this study were to determine whether the cost of care for malignant brain and other central nervous system tumors (MBT) in the US Military Health System differs between the Department of War-operated direct care (DC) network and the affiliated civilian-operated private sector care (PSC) network, as well as to identify the treatment types (i.e., surgery, radiation therapy, chemotherapy) driving any observed cost differences. METHODS: Linked cancer registry and administrative-claims data from the Military Cancer Epidemiology database-including 574 patients ages 18-64 diagnosed with MBT between 2003 and 2014-were used to describe and compare patient characteristics and costs in 2022 US dollars associated with primary treatment in each setting. Because of the skewed distribution of cost data, the possibility that the effect of the care setting may have varied for patients with different levels of total cost, and to allow greater flexibility in modeling associations between independent variables and cost, nonparametric and semiparametric methods were used for the primary aims of this study. First, we used Wilcoxon-Mann-Whitney tests and Hodges-Lehmann difference estimator 95% confidence intervals (CIs) to compare DC and PSC median total cost and individual treatment type cost per patient. Next, we used quantile regression to compare DC and PSC costs at five levels between the 10th and 90th percentiles of total cost and individual treatment type cost per patient, while controlling for administrative, demographic, health, and tumor characteristics. PRINCIPAL FINDINGS: Most patients were male (69%), younger than 29 years (28%), or older than 50 years (30%). Glioblastomas were the most commonly diagnosed tumor type (38%). Overall, 270 patients (47%) were treated in DC and 304 (53%) were treated in PSC. The median unadjusted total cost of treatment per patient was significantly higher for PSC at $87,652 compared to DC at $55,361 (Hodges-Lehmann difference estimator 95% CI [-49,544,-10,086], p < .001). While median unadjusted surgery costs were similar, radiation therapy and chemotherapy costs were significantly higher in PSC than in DC. After controlling for confounders, total cost per patient was significantly higher in PSC compared to DC for patients near the median, 75th, and 90th percentiles of the cost distribution, with the additional total cost per patient associated with PSC increasing from $31,443 (95% CI [4,216,58,669], p = .024) near the median to $305,200 (95% CI [211,412,398,897], p < .001) at the 90th percentile. After controlling for confounders, radiation therapy was significantly more costly for patients at each quantile in PSC compared to DC, while surgery costs were similar for patients in PSC and DC at each quantile of the cost distribution. PRACTICAL APPLICATIONS: The decision to internally deliver care or outsource services for complex and costly diagnoses such as MBT involves multiple strategic considerations, including treatment costs. Using data from the universal-access Military Health System, this study demonstrated that providing care within system-owned and operated facilities may be more cost-efficient than outsourcing care to affiliates, especially for patients whose treatment costs are above the median.