Study design

The study was "longitudinal", designed to follow and study drugs in their natural course through the distribution system to health facilities. It was designed in such a way that the worst case situation (with regard to conditions known to adversely affect drug quality) was likely to be represented. Drug batches were sampled from stocks in central medical stores and retained. At hospitals and health centres, samples were collected according to two different criteria, for the reasons given:

1. Oldest batch of each drug - most likely to show degradation if the drug was unstable; this was usually also the batch with longest storage time at facility, and nearest to expiry.

2. If present, samples from the same batch of each drug as those retained at central medical stores - testing the same batch should then reflect any changes after the time of sampling in central medical stores.

Selection of sampling points: central medical stores and health facilities

Both central medical stores in Zimbabwe were included in the study since each had a catchment area or distribution system with unreproduceable features.

Five out of 55 districts were chosen in the "lowveld" (low lying regions with hot climate in the major river valleys in Zimbabwe), namely Hwange, Binga, Kariba, Chipinge and Beitbridge (Figure 1). This study area included the most remote parts, farthest in the distribution chain in Zimbabwe, The majority of facilities were in this "worst case" region. However, a proportion were in cooler areas within the same district, the medical store's catchment area or the distribution route as the worst case facilities. The majority of facilities were health centres (first level of care) where drug supplies were managed by nurse-prescribers. District level hospitals were also included. There, drug supplies were managed by qualified pharmacy personnel and a wider range of drugs was prescribed by doctors.

Figure 1: Map of Zimbabwe

Selection of drugs

Drugs were selected on the basis of two criteria: (1) known or suspected quality problems in adverse conditions; and (2) relevance in public health care (drugs on the national essential drugs list, with high turnover in the public sector).

13 drugs were selected for the study (Table 2). All of these drugs are used at hospitals staffed with doctors. Nine of them are used at primary health care (PHC) facilities as well. The drugs included all commonly used antibiotics on the essential drugs list in Zimbabwe, with the exception of co-trimoxazole, which is considered stable.

Table 2 - Selection of drugs and shelf life by manufacturer

Note:

^a Shelf life extended from two to three years during the period of study.
^b From "International Price Indicator Guide 1992", Management Sciences for Health, Boston.

Pharmaceutical manufacturers

During the period of study each drug was procured from manufacturers in Zimbabwe. In addition to these, two drugs were imported from foreign manufacturers.

Table 2 shows that seven of the 13 selected drugs were supplied by a single manufacturer. The remaining six drugs were supplied by two or more manufacturers. In reporting results, manufacturers were distinguished by a code (Manufacturer "A", "B") without explicit mention of name. Where possible, the manufacturers' results were compared.

Shelf-life of drugs

On the basis of manufacturer's data on accelerated stability studies, the drug regulatory authority in Zimbabwe assigns a provisional shelf-life of two years (at the time of drug registration). Manufacturers are required to submit further stability studies to justify an extension of shelf-life. This is not done in most cases, as apparent in Table 2, with the exception of amoxicillin capsules and doxycycline tablets, where the shelf-life was extended from 2 years to 3 years during the period of study.

Indicators for quality assessment

The level of active ingredient (expressed as percentage of stated content) was used as the unit of measurement In addition, dissolution characteristics were investigated in the case of ferrous sulfate tablets.