The Economic Impact of Artificial Intelligence Usage in Screening Mammography
A group of researchers evaluated the potential economic impact of artificial intelligence (AI) implementation on screening mammography from the perspective of treatment facilities and payers. The researchers built an excel-based model to assess the impact of two types of AI solutions: augmentative and autonomous. Augmentative AI detects suspicious lesions and generates a case recommendation, which is reviewed by a radiologist. This tool has been shown to reduce read times. Autonomous AI has similar capabilities but can also triage “low suspicion” or “normal” cases, enabling radiologists to focus on reviewing high-priority exams.
The researchers used base-case assumptions for a facility that screens 10 000 women per year. The findings showed that the use of augmentative AI could result in time savings of 52 to 72 hours of mammography exam readings per year. According to the authors, these time savings could lead to a financial savings of $71 100 per year. However, the anticipated impact to payers is neutral due to the facility’s existing reimbursement structure and base care model assumptions.
For the autonomous AI, the triage feature greatly impacted the amount of savings in time and revenue. For instance, if “low suspicion” cases would not be eligible for professional reading fees, the facility could save 217-229 hours per year. These saved hours could then be used to increase throughput for clinics with backlogs or provide staff more time to complete additional radiologist tasks, according to the authors. However, there could be a decrease in revenue of up to $46 300 per year. If there are no additional reimbursements for mammography AI services, the savings could be $412 500 per year for payers, although the lack of reimbursement poses a potential disincentive to the facilities.
Per the model’s findings, the authors suggest that the implementation of AI in screening mammography could “help mitigate radiologist shortages, relieve current workflow pressures, and increase radiologists’ confidence and accuracy.” Overall, the use of augmented AI led to an increase in savings and flexibility for facilities and had a neutral impact on payers. Although autonomous rule-out AI showed savings for payers, the authors note that reimbursement for AI mammography will greatly impact the adoption of such tools in treatment facilities.
Reference
Olsen J, Alivisatos E, Asai E, Knox E, Pohlman S. A framework for evaluating the economic impact of artificial intelligence for screening mammography: implications for facilities and payers from the US perspective. Presented at: International Society for Pharmacoeconomics and Outcomes Research Conference; May 5-8, 2024; Atlanta, GA, and virtual; Abstract EE54.
