Evaluating the Quality of Drug Users' Information from Health Treatment Centers and

Abstract

Evaluating the Quality of Drug Users' Information from Health Treatment Centers and

Police Stations Sources in Bangkok under Capture-recapture Method

Busaba    Supawattanabodee    BNS,MSc (Biostatistics)

Clinical Epidemiology Unit, BMA Medical College and Vajira Hospital

Objective: To evaluate the quality of data from health treatment centers and police stations sources.

Study design: Survey research.

Subjects: Drug dependencies from 61 health treatment centers and narcotics cases from 88 police stations in Bangkok.

Methods: The data were taken from 61 health treatment centers and 88 police stations    in Bangkok during the period of 3 months from October 1 to December 31,              2001.

Main outcome measures: Quality of data were evaluated by the following three criteria;   the first criterion was the accuracy of drug users' information which remained on the list after duplicate records were removed, the second was completeness of key variables such as name, surname, sex, age, type of drugs/cases, place of residence/place of occurrence, and the last criterion was investigation from percentage of under-reporting number of drug users form two sources comparison with the number of drug users estimated by capture-method.

Results: In health treatment centers source, 26,703 cases were recorded in the list, 11,222 cases remained on the list after duplicate record were removed or the accurate record rate was 42% For the second criterion, all key variables were recorded for all persons (100%). In police station sources, 12,389 cases were recorded in the list, 12,012 cases remained on the list after duplicate record were removed or the accurate record rate was 97%. For the second criterion, some key variables such as place of recorded only 5 % As for the last criterion, the values of percentage          under-reporting from two sources were over 97%.

Conclusion: In health centers source, the accuracy of drug users' information was only   42%. However, the completeness of key variables such as name, surname, sex, age, type of drug, and place of residence were recorded in all cases. As for the police stations source, the accuracy of narcotics cases' information was 97%. However, only some key variables such as name, surname and sex were recorded in all cases. As for the percentage of under-reporting from two sources were over 97% because the estimate of drug users in match type A+B+C+D was 427,933 which deviated far from the number of drug users reported from two sources.

Key words: Drug users, capture-recapture, dual records, data quality.

Vajira Med J 2004 ; 48 : 135 - 143