Background: Retail pharmacies are important sources of advice on pharmaceuticals in developing countries, where many purchasing decisions are unmediated by medical professionals. For childhood diarrhea, choice of drug sales in pharmacies has been found to be consistently poor, whether with or without prescription, as evidenced by a low use of effective oral rehydration salts (ORS) and high use of marginally effective or ineffective products such as antimotility agents, adsorbents, and antimicrobials. Little information is available about factors influencing prescribing by pharmacy personnel in these settings. This paper reports results of an analysis which examined the influence of rural versus urban location, neighborhood socio-economic status (SES), and clinical knowledge of pharmacy assistants on quality of prescribing in retail pharmacies in Kenya.
Methods: Using WHO treatment guidelines, we measured quality of prescribing in childhood diarrhea as: (1) the proportion of client encounters in which ORS is prescribed (indicating higher quality practice); and (2) the proportion of visits in which an antidiarrheal product is prescribed (lower quality practice). We obtained data on prescribing in a simple case of childhood diarrhea using surrogate patient visits to 91 pharmacies located in the major urban area of Kenya (Nairobi) and four rural towns. Multivariate logistic regression was used to examine the association between pharmacy predictor variables and the quality of prescribing.
Results: No consistent relationship was found between the clinical knowledge of pharmacy assistants and quality of prescribing. Controlling for number of products prescribed, gender of the assessor, and level of knowledge, multivariate analyses indicated that in comparison to rural pharmacies, the odds of prescribing ORS were significantly higher in pharmacies located in high SES (OR = 4.7; 95% CI = 1.4,16.1) or middle SES (OR = 4.5; CI = 1.9,10.6) urban neighborhoods; the odds in low SES urban pharmacies were similar to those in rural areas (OR = 2.0; CI = 0.6,6.2). The odds of prescribing antidiarrheal products did not differ significantly in any geographic area.
Conclusions: Location of a retail pharmacy in a rural area or in a low-income urban neighborhood was associated with suboptimal quality of prescribing as evidenced by lower use of ORS. No relationship between clinical knowledge and quality of practice was detected. Future research is needed to examine such factors as the influence of intra-pharmacy authority structure, perceived efficacy of pharmaceuticals, local patterns of physician practice, and economic incentives on the quality of prescribing.