Bundled Payment Models for Actinic Keratosis

Open Access
Sciacca Kirby, Joslyn R
Graduate Program:
Public Health Sciences
Master of Science
Document Type:
Master Thesis
Date of Defense:
October 30, 2014
Committee Members:
  • Douglas L Leslie, Thesis Advisor
  • Actinic keratosis
  • utilization
  • management
  • treatment
  • cost
  • efficiency
  • value
  • alternative payment model
  • bundled payment
Importance: Actinic keratosis (AK) is a skin condition induced by ultraviolet light and has the potential to progress into skin cancer. AK affects approximately 7% of the United States (US) population. In 2004 the US health care spending for AK management was $1.2 billion. Some of the procedural treatments for AK were targeted by CMS for reduced reimbursement, potentially due to over-utilization. Alternative reimbursement models are being investigated in other therapeutic areas to reduce costs, but there is little investigation for dermatologic conditions. Objectives: To assess AK-specific healthcare utilization and costs and develop eight bundled payment models. Methods: A retrospective cohort was obtained from a large private insurer for the mid-Atlantic region (January 2010 to December 2012). This sample was used to describe utilization and cost, and served as a test and a validation sample for bundled payment development. A random sample from a nationally representative dataset, the Truven Health MarketScan® database, was used as a second validation sample. Results: The total cost of AK-related care was $40,719,495 for 95,294 patients with AK. Of this cost, prescriptions accounted for 8.6% and outpatient care for 91.4%, of which 44.9% was due to destructive procedures. The use of extensive destruction (CPT 17004), prescription medication therapy, male gender and age were associated with a higher mean three-year cost, although the cost data were right-skewed and had a large variance. Winsorization was applied at the 1st/99th and 5th/95th percentiles, which reduced the mean 3-year cost by $34 and $64, respectively. Several percentile- and mean-based bundled payments were developed, with and without adjustments, and most predicted savings over historical reimbursements. All of the models predicted that some patients and providers would have costs above the bundled payment. Conclusions and Relevance: Bundled payments appear theoretically feasible for AK management, but further work is needed to determine the method that is most equitable for providers and patients. Gender and history of non-melanoma skin cancer (NMSC) were used to adjust payments as these contributed importantly to variation in cost. More research is needed to establish the value of various treatment methods in order to guide practitioners’ and patients’ decisions. In addition, the utility of bundled payments for AK management needs to be implemented and compared to other reimbursement models.