Premium
Estimating the size of an illicit‐drug‐using population
Author(s) -
Simeone Ronald,
Holland Lynn,
ViverosAguilero Roman
Publication year - 2003
Publication title -
statistics in medicine
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.996
H-Index - 183
eISSN - 1097-0258
pISSN - 0277-6715
DOI - 10.1002/sim.1528
Subject(s) - population , confidence interval , statistics , computer science , population size , sampling (signal processing) , sample size determination , selection bias , interval (graph theory) , econometrics , medicine , mathematics , environmental health , filter (signal processing) , combinatorics , computer vision
Abstract This paper describes a new method for estimating the size of an illicit‐drug‐using population. It is designed to overcome certain limitations of registry‐based techniques that require both comprehensive site coverage and unique case identifiers, and which do not typically provide estimates of the number of drug users who are currently active. The approach involves collecting retrospective self‐report data on the careers of individuals who appear at drug treatment programmes. A model is developed that corrects for the selection bias introduced by the sampling plan, and which allows us to estimate the rate at which drug users generate treatment admission events during spells of use. The size of the drug‐using population is estimated by dividing the estimated total number of treatment admissions that are generated during some fixed interval of time by the estimated rate at which individuals generate such events. The technique is tested in a series of simulation studies which demonstrate that accurate estimates of the size of the drug using population can be obtained in this manner. Analytical expressions for confidence intervals about the population estimates are derived as part of the exercise. Limitations of the approach and other potential applications are discussed. Copyright © 2003 John Wiley & Sons, Ltd.