Nonmelanoma skin cancer is the most common malignancy. Multiple therapies prevent recurrence but vary widely in cost. The most common therapies are local destruction, excision, and Mohs surgery (histologically guided tumor removal). Clinical variables that may affect treatment choices can be identified, but little is known about how clinicians choose among therapies.
The objective of this study was to learn if variations exist in the treatment of nonmelanoma skin cancer in different practice settings.
Prospective cohort study.
Subjects consisted of consecutive patients with nonmelanoma skin cancer at a university-affiliated private dermatology practice and the dermatology clinic at the nearby affiliated Veterans Affairs (VA) medical center.
We studied data from medical records and patient surveys.
Overall, 1777 nonrecurrent nonmelanoma skin cancers were diagnosed in 1375 patients. Compared with the VA site, patients at the private site were younger, more likely to be female, and less likely to be poor, and their tumors were smaller and less likely to be on visible areas of the body. Treatments varied between the 2 sites (P <0.001). The proportions of tumors treated at the private and VA sites, respectively, were 23% and 19% for destruction, 25% and 48% for excision, and 37% and 25% for Mohs surgery. In multiple clinical subgroups, Mohs surgery was more likely to be performed at the private site than at the VA. Moreover, in multivariable models controlling for clinical features that may have affected treatment choice, tumors at the private site were more likely than tumors at the VA to be treated with Mohs surgery (odds ratio, 2.39; 95% confidence interval, 1.54–3.70).
Care for nonmelanoma skin cancer varied at 2 academic practice sites that are near each other and that share some clinician staff. These findings raise questions not only about overuse or underuse of procedures at the 2 sites, but also about systematic differences in patient preferences and/or physician incentives in prepaid and fee-for-service settings.