BACKGROUND: Data on drug consumption obtained from emergency room clinical records have been used for various epidemiological purposes. However the validity and reliability of these data remain unknown. This paper assesses the reliability and validity of an Emergency Room Toxicological Register (HMR) which has collected information on drug misuse from emergency room clinical records since 1979, and examines the implications for epidemiological applications. SUBJECTS AND METHODS: An Emergency Room Survey (ERS) was carried out in a Barcelona Hospital including opiate or cocaine users identified by the physician and a systematic sample of other patients age 15 to 49 years old. Data on clinical records of interviewed patients were also reviewed. Episodes from identified drug users (686) and HMR (676) for the same study period were linked and validity and reliability were analyzed. RESULTS: Sensitivity ranged between 63 and 86%, and specificity was 98%, Kappa index higher than 0.72 and intraclass correlation coefficient was 0.99. CONCLUSIONS: Information about drug users included in emergency room clinical records proved to be valid as an information system for drug use surveillance. However data about patterns of less heavy users, as cocaine use, are underreported.
BACKGROUND: Data on drug consumption obtained from emergency room clinical records have been used for various epidemiological purposes. However the validity and reliability of these data remain unknown. This paper assesses the reliability and validity of an Emergency Room Toxicological Register (HMR) which has collected information on drug misuse from emergency room clinical records since 1979, and examines the implications for epidemiological applications. SUBJECTS AND METHODS: An Emergency Room Survey (ERS) was carried out in a Barcelona Hospital including opiate or cocaine users identified by the physician and a systematic sample of other patients age 15 to 49 years old. Data on clinical records of interviewed patients were also reviewed. Episodes from identified drug users (686) and HMR (676) for the same study period were linked and validity and reliability were analyzed. RESULTS: Sensitivity ranged between 63 and 86%, and specificity was 98%, Kappa index higher than 0.72 and intraclass correlation coefficient was 0.99. CONCLUSIONS: Information about drug users included in emergency room clinical records proved to be valid as an information system for drug use surveillance. However data about patterns of less heavy users, as cocaine use, are underreported.