Studies to measure drug use will vary from setting to setting. The nature and design of such studies will depend on many factors, which include: the specific information needs of health managers; the types of record-keeping systems available in health facilities; the types of providers whose behaviour is to be characterized; and the resources available to carry out the work. In general, however, drug use studies by means of indicators will fall into four broad categories:

• Describing current treatment practices: Such a cross-sectional survey is done by taking specific measures of treatment practices from carefully selected groups of facilities and patients.

• Comparing the performance of individual facilities or prescribers: Rather than being primarily concerned with summarizing the treatment practices of the group as a whole, such a study seeks to compare practices between individual facilities or prescribers, or between groups.

• Periodic monitoring and supervision of specific drug use behaviours: After the broad outlines of drug use behaviour are known, the indicators can be used to identify facilities or providers whose performance falls below a specific standard of quality, so that they can be targeted for intensive supervision.

• Assessing the impact of an intervention: Specific indicators can be used to evaluate the effectiveness of an intervention designed to change prescribing practices, by providing the capability for reliably measuring practices both before and after the intervention, and in both an intervention and a control group.

The first step in designing a drug use study is to specify the objectives clearly. The size of the samples required, the design of the sampling process and the complexity of data analysis will vary greatly depending on the specific objectives. This manual will focus primarily on the design and implementation of a basic indicators study to describe treatment practices, and on how this basic design should be changed to allow for a comparison of regions or facilities. Ideas for using the indicators for supervision will also be presented briefly.