Nepal - Survey on Hard Drug Users in Nepal 2006-2007

Usual survey methodology (general household/population survey) is not recommended to estimate the prevalence of hard-to-reach hidden population such as hard drug users, commercial sex workers, homeless people, etc. for some reasons. Sampling frame is not usually available and also very difficult or impossible to construct. Due to the low prevalence rate as compared to the total population, it may require larger sample size and relatively more resources. More importantly, respondents may not be willing to respond because of the sensitivity of the topic itself.

An alternative approach of estimation technique has to be used to get the most reliable picture of the real situation. For this, we have selected the Multiplier Method of indirect estimation technique.

The multiplier method has two elements in common: the benchmark and the multiplier.

The benchmark (B) is the data source that captures the number of hard drug users who are under treatment in the rehabilitation centers in the reference year. The multiplier (M) is an estimate of the proportion of current hard drug users who have experienced the event recorded by the benchmark, i.e. the proportion of such drug users who have been in treatment in the rehabilitation centers in the reference period. This information is obtained independently of the benchmark data. The inverse of that proportion is the multiplier (M), which is an indirect estimate of the proportion of the total population of the hard drug users represented in the benchmark data.

The prevalence is calculated by multiplying the benchmark by the multiplier (B x M). Hence,

N = B x M = B x (1/p)

where,
N is the total number of hard drug users,
p is the proportion of the hard drug users who have visited the rehabilitation centers in the reference period.

Sample size:

In each of the selected area, sample size for the interview with current hard drug users were fixed at 100 respondents. This number is derived with following assumptions:

Estimated proportion of drug users visiting rehabilitation centers = 7%

Margin of error in estimation = 5% and

Level of confidence = 95%

The total achieved sample size was 1319.


100 percent


Sample weights were calculated for each of study area. The weight was calculated as the ratio of drug users captured in the rehabilitation centre in the past year to rehabilitation visit rate i.e. number of drug users taken as numerator and rehabilitation centre visit rate taken as denominator.

In the HDUS 2006/07 dataset, the weighting coefficient is called WEIGHT.