Long-term impact of highly active antiretroviral therapy on HIV-related health care costs

Context: Highly active antiretroviral therapy (HAART) is associated with de creased opportunistic infections, hospitalization, and HIV-related health c are costs over relatively short periods of time. We have previously demonst rated that decreases in total HIV cost are proportional to penetration of p rotease inhibitor therapy in our clinic. Objective: To determine the effects of HAART on HIV health care use and cos ts over 44 months. Setting: A comprehensive HIV service within a Veterans Affairs Medical Cent er. Design: A cost-effectiveness analysis of HAART. Main Outcome Measurements: The mean monthly number of hospital days, infect ious diseases clinic visits, emergency room visits, non-HIV-related outpati ent visits, inpatient costs, and antiretroviral treatment costs per patient were determined by dividing these during the period from January 1995 thro ugh June 1998 into four intervals. Viral load tests were available from Oct ober 1996. Cost-effectiveness of HAART was evaluated by determining the cos ts of achieving an undetectable viral load over time. Results: Mean monthly hospitalization and associated inpatient costs decrea sed and remained low 2 years after the introduction of protease inhibitors (37 hospital days per 100 patients). Total cost decreased from $1905 per pa tient per month during the first quarter to $1090 per patient per month in the third quarter but increased to $1391 per patient per month in the fourt h quarter. Antiretroviral treatment costs increased throughout the entire o bservation period from $79 per patient per month to $518 per patient per mo nth. Hospitalization costs decreased from $1275 per patient per month in th e first quarter to less than $500 per patient per month in each of the thir d and fourth quarters. The percentage of patients with a viral load <500 co pies/mL increased from 21% in October 1996 to 47% in June of 1997 (p = .014 ). The cost of achieving an undetectable viral load decreased from $4438 pe r patient per month to $2669 per patient per month, but this trend did not reach statistical significance (p = .18). Conclusions: After an initial decrease, there was an increase in the total monthly cost of caring for HIV patients. Cost increases were primarily due to antiretroviral treatment costs, but these costs were offset by a marked decrease in inpatient-related costs. Increases in costs were not related to antiretroviral treatment failures as measured by the proportion of patient s with low or undetectable viral loads, The cost of achieving an undetectab le viral load remained stable despite increases in the cost of procuring an tiretroviral agents.