Application of the extended fuzzy discrete event systems theory to HIV/AIDS treatment regimen selection

Selecting an appropriate treatment regimen for an HIV-infected patient is a significant problem. We previously developed a regimen selection system for the first-round HIV/AIDS treatment based on the fuzzy discrete event systems (FDES) theory where type-1 fuzzy sets are used to represent experts' consensus. In clinical practice, different experts may have their own distinct knowledge and expertise and it is often difficult for them to reach a consensus. In a recent paper, we presented our extended fuzzy discrete event systems (EFDES) theory in which type-2 fuzzy sets were utilized to capture and represent diverse expertise of different experts. In the present paper, an EFDES-based HIV/AIDS regimen selection system is constructed and its performance is evaluated using 35 patients. We focus on three different scenarios to equip the new system with the capabilities of handling higher levels of uncertainties, which is impossible under the FDES framework. The regimens selected by the EFDES-based system match the actual prescribed regimens for these patients over 80% of the time. While the matching rates are about the same as achieved by the FDES-based system, they are obtained under more realistic conditions, and hence more practical.