Interactive neural network-assisted screening: An economic assessment

To apply clinical effectiveness estimates of interactive, neural network-assisted (INNA) screening to economic cervical cancer screening models to assess the economic impact of using this technology. Estimates of the sensitivity of INNA screening were drawn from a recently completed comprehensive synthesis of the INNA literature and applied to the Computer Model for Designing CANcer ConTROL Programs-based Cervical Cancer Screen economic model. The economic analysis was conducted from a modified payer perspective using costs borne by payers combined with patient deductibles and copayments. Costs of treating cervical cancer were updated to 1997 values using the medical care component of the Consumer Price Index. The model was run for a cohort of women starting at age 20 and screened on a triennial schedule through age 75. In the primary analysis (sensitivity of unassisted manual examination assumed to be 85%), the ratios found in this investigation varied from approximately $35,000 to $80,000 per life year saved, with the preponderance of ratios < $50,000 per life year saved. These results were sensitive to estimates of sensitivity of unassisted manual screening but not to estimates of treatment costs. This investigation applied accuracy data on INNA rescreening to a model of the cost-effectiveness of cervical cancer screening. The results support the use of INNA rescreening as an appropriate expenditure of resources to identify missed cases of cervical epithelial abnormalities and potential cervical cancer.