Data quality of the Drug Abuse Warning Networks.
The purpose of this article was to assess the quality of data collected by the Drug Abuse Warning Network (DAWN), which reports drug abuse emergency department visits.
The results of quality assurance studies at 36 sites were reviewed and interpreted.
Data collection procedures are not consistent among hospitals and, along with personnel, regularly change within a hospital.
Trained investigators reabstracted DAWN report forms at 24 sites and determined that only 57.4% of the cases that met DAWN case definition criteria had been reported ; one of five cases had been reported at one site.
The technique used in 11 (47.8%) of 23 hospitals to screen for potential DAWN cases detected only 36% of the cases found when all medical charts are examined.
The investigators found discrepancies between reported and actual cases in 81.3% of the report forms reabstracted, with an average of 2.3 errors per form.
Information as to the drug (s) involved was incorrect in 36.3% of the forms.
Due to underreporting of drug abuse emergency department visits and poor quality data in DAWN report forms, DAWN estimates of drug activity must be viewed with caution.
Furthermore, estimation of trends is risky, due to differences between emergency departments as to reporting systems and changes over time.
Mots-clés Pascal : Alcoolisme, Toxicomanie, Service santé, Mortalité, Hospitalisation, Epidémiologie, Etude critique, Collecte donnée, Méthode statistique, Homme
Mots-clés Pascal anglais : Alcoholism, Drug addiction, Health service, Mortality, Hospitalization, Epidemiology, Critical study, Data gathering, Statistical method, Human
Notice produite par :
Inist-CNRS - Institut de l'Information Scientifique et Technique
Cote : 96-0372899
Code Inist : 002B18B04. Création : 10/04/1997.