Prediction of higher cost of antiretroviral therapy (ART) according to clinical complexity. A validated clinical index.

Background: The financing of antiretroviral therapy (ART) is generally determined by the cost incurred in the previous year, the number of patients on treatment, and the evidence-based recommendations, but not the clinical characteristics of the population.

Objective: To establish a score relating the cost of ART and patient clinical complexity in order to understand the costing differences between hospitals in the region that could be explained by the clinical complexity of their population.

Methods: Retrospective analysis of patients receiving ART in a tertiary hospital between 2009 and 2011. Factors potentially associated with a higher cost of ART were assessed by bivariate and multivariate analysis. Two predictive models of "high-cost" were developed. The normalized estimated (adjusted for the complexity scores) costs were calculated and compared with the normalized real costs.

Results: In the Hospital Index, 631 (16.8%) of the 3758 patients receiving ART were responsible for a "high-cost" subgroup, defined as the highest 25% of spending on ART. Baseline variables that were significant predictors of high cost in the Clinic-B model in the multivariate analysis were: route of transmission of HIV, AIDS criteria, Spanish nationality, year of initiation of ART, CD4+ lymphocyte count nadir, and number of hospital admissions. The Clinic-B score ranged from 0 to 13, and the mean value (5.97) was lower than the overall mean value of the four hospitals (6.16).

Conclusions: The clinical complexity of the HIV patient influences the cost of ART. The Clinic-B and Clinic-BF scores predicted patients with high cost of ART and could be used to compare and allocate costs corrected for the patient clinical complexity.