AEGiS-13IAC: Underreporting of AIDS cases in Brazil: applications of capture-recapture methods.

13th International AIDS Conference

Durban, South Africa - July 9-July 14, 2000

Underreporting of AIDS cases in Brazil: applications of capture-recapture methods.

Int Conf AIDS 2000 Jul 9-14; 13:(abstract no. ThOrC667)

Oliveira MT, Caiaffa W, Mingoti S, Fernandez H, Macedo LM, Mafra A, Paiva L, Picinin I, Santos L, Silva R
M. T. Oliveira, Rua Aimors, 2480 sala 604, 30140-072 Belo Horizonte, MG, Brazil, Tel.: +55-31-275-2413, Fax: +55-31-275-2413, E-mail: jelihovs@net.em.com.br

BACKGROUND: It is crucial to estimate the total number of AIDS cases accurately, aiming to prioritize health actions. In order to know the underreporting of AIDS in adults residents in a City located in Southeast Brazil during 1995-1996, a capture-recapture methodology was used.

METHODS: Five data sources were used: (1) The AIDS Surveillance National Data (SINAN); (2) The National Hospitalization Registry (SIH);(3) The Death Certificate Data (SIM); (4) The Tuberculosis Surveillance System (TB-AIDS) and, (5) AIDS patients diagnosed by private physicians. The matching criteria were based upon name initials, date of birth, gender and mother initials name. Probabilistic models were used assuming that capture probabilities vary with time and behavior (Mtb) and time and heterogeneity among individuals (Mth).

RESULTS: According to the data source, 618 AIDS cases eligible to the study were reported to SINAN, 726 to SIH, 398 to SIM, 102 to TB-AIDS and 81 diagnosed by private physicians. A total of 1211 distinct individuals were identified using the five sources; 561 (46.3%) were unknown by the official AIDS Surveillance National Data (SINAN). Preliminary estimates using Mtb resulted in 1752 (95%CI: 1551-2074) AIDS cases with probabilities to be captured ranging from 0.04 to 0.35. Using Mth resulted in 2071 (95%CI: 1918-2256), with probabilities to be captured ranging from 0.03 to 0.37. One case of AIDS had the probability of 0.30 to be reported to the official AIDS Surveillance National Data (SINAN), during this period.

CONCLUSIONS: Instead using active surveillance involving high costs, capture-recapture methods using available data from official sources are likely to provide a means to accurately assess the total number of AIDS cases. Also, this methodology is simple, fast and less expensive to be used in public health. * Sponsored by: National STD/Aids Program, Minister of Health, Brazil

Keywords: AEGIS, Acquired Immunodeficiency Syndrome, Death Certificates, Registries, Data Collection, Probability, Models, Statistical, Hospitalization, Tuberculosis, Hospitals, Case-Control Studies, Brazil, Adult, Human, methods, prevention & control
000709
ThOrC667