捕获再捕获法估计四川乐山市中区吸毒人群基数

ESTIMATION OF THE POPULATION SIZE OF DRUG USERS IN SHIZHONG DISTRICT OF LESHAN CITY,SICHUAN:A CAPTURE-RECAPTURE METHOD

目的:使用捕获再捕获法估计乐山市中区现有吸毒人群基数。方法:利用强戒所建所以来资料确定平均捕获期和平均捕获间隔期及两次捕获起始点并收集相应时间段内的收戒人数。利用Lincoln-Petersen模型进行估计,定性和定量研究结合以解释说明估计结果等。结果:平均捕获期3.02月,平均捕获间期12.59月,两次捕获期起止时间点为2002年9月30日-2003年1月31日和2004年1月31日-2004年5月31日,对应收戒人数为234人和251人,重复捕获34人。不分层点估计数1742人,标准误240.45,95%CI1220-2163(人)。分层后合计点估计数为1645人,标准误225.46,95%CI1203-2087(人)。结论:75%位数捕获期和平均捕获间期为4个月和12个月,可满足分析需要。Lincoln-Petersen模型简单,费用低,可操作性强。估计结果存有一定程度的低估,可作为估计底限值。选择适当的捕获再捕获模型可提高估计精度。估计数为公安部门同期提供现有吸毒登记数的1.5倍。

Objective： To estimate the population size of active drug users in Shizhong District of Leshan City, Sichuan Province, with a two -sample capture -recapture method. Methods： Data were collected from one local compulsory detoxification center. The average capture period, the average capture interval, initiation points and end points of periods of the two captures and the numbers of active drug users who were in the corresponding period were picked up from the document of those drug users, The result was calculated by Lincoln -Petersen model. Moreover, qualitative study and quantitative study were combined together to explain the study results. Results ： The average capture period was 3.02 months and the average capture interval was 12. 59 months. The first capture comprised 234 persons from September 30,2002 （the initiation point） to January 31, 2003 （the end point） and the recapture 251 persons from January 31,2004 （the initiation point） to May 31,2004 （the end point）, There were 34 persons that included on both lists, yielding an estimate of 1742 （ SE = 240. 45, 95% CI = 1220 - 2163 ） active drug users in Shizhong District of Leshan City in May,2004. A stratified analysis by gender was conducted, yielding a point estimate of 1645 total active drug users （ SE = 225.46, 95% CI = 1203 - 2087 ）. Conclusion： In this study, the 75 -percentile capture period and the average capture interval are considered to be 4 months and 12 months, respectively. Lincoln - Petersen model is simple and practicable because of its lower costs and avoidance of difficult access to drug users in practice. The conclusive results combined to qualitative study trend to be underestimated, but we can regard them as the bottom line of estimation value. Suitable capture - recapture models should be selected based on the data available to increase the precision of estimation. The study shows the number of active drug users is about 1.5 times as much as the registered number from public security departments during the same period.