AIDS Forecasting

Pursuant to a congressional request, GAO reviewed 13 existing national forecasts projecting future cumulative numbers of Acquired Immune Deficiency Syndrome (AIDS) cases to identify: (1) prediction variations and uncertainties; (2) types of forecasting models used; (3) the quality of the main types of data used in the models; and (4) a realistic range of estimates for the cumulative number of AIDS cases through 1991.

GAO found that forecast model approaches using: (1) extrapolation estimated 200,000 to 325,000 cumulative cases through 1991; (2) back-calculation predictions of individuals already infected with the human immunodeficiency virus (HIV) estimated 120,000 to 295,000 cases; (3) macro-level assumptions about the epidemic's future course predicted about 160,000 to 400,000 cases; and (4) micro-level examination of individual behaviors that contribute to the epidemic estimated about 25,000 cases. GAO also found that all forecasts relied on Centers for Disease Control (CDC) national AIDS surveillance data, which had quality problems deriving from: (1) limited definitions of AIDS; (2) diagnostic errors and restrictions; and (3) inaccurate or late state and local reporting to CDC. In addition, GAO found that: (1) micro-level simulations were the most comprehensive but least empirical type of model; (2) extrapolation models were the least comprehensive; (3) some models were based on questionable assumptions; and (4) few models adequately adjusted for the identified data problems. GAO believes that, if the existing forecasts were adjusted to compensate for data biases, the realistic range of forecasts would be 300,000 to 480,000 cases.

Recommendations

Our recommendations from this work are listed below with a Contact for more information. Status will change from "In process" to "Open," "Closed - implemented," or "Closed - not implemented" based on our follow up work.

Director: Team: Phone: