Syndicate content

Biting back at malaria: On treatment guidelines and measurement of health service quality

Arndt Reichert's picture

Growing up in a tropical country, one of us (Alfredo) was acutely aware of mosquito-borne diseases such as dengue and malaria. For many years now, vector-control strategies were—and still are—promoted by government- and school-led campaigns to limit the spread of these diseases. Consequently, it is somewhat alarming to know that diseases spread by mosquitoes remain an enormous challenge facing large parts of the developing and even developed world, particularly sub-Saharan Africa. It is perhaps less surprising that our shared interest in the health sector has resulted in a joint paper on assessing the overall quality of the health care system via compliance with established treatment guidelines.


In Paloyo and Reichert (2016), we propose and demonstrate the use of compliance with prescribed treatment guidelines as a measure of the overall quality of medical care. Our application uses the adherence of health service providers to international and national treatment guidelines for uncomplicated childhood malaria. The novelty is that we leverage the use of standardized household survey data, particularly the Demographic and Health Surveys (DHS), which is available in over 90 countries. Although the DHS was not designed for this purpose, we argue that it is nevertheless effective and is cheaper than conducting health-facility surveys, administering clinical vignettes, or relying on simulated patients (for an overview of quality measures of health facilities, see Lindelow and Wagstaff, 2008). The possibility to control for characteristics that are likely related to both health care provider choice and effective malaria treatment is another advantage of this rich household survey database.

Using DHS data for eight West African countries (Benin, Burkina Faso, Ghana, Guinea, Liberia, Mali, Niger, and Nigeria), the paper specifically examines the differences in treatment practices between self-treatment, traditional healers, private health care providers, and the public sector by comparing the use patterns of antimalarial drugs and fever-abatement rates of children below five years of age who had fever within the last two weeks. At a time when presumptive treatment of childhood malaria was the prescribed protocol, the likelihood of taking effective antimalarial drugs for febrile children can be regarded as an indicator of overall quality because such adherence to treatment guidelines is indicative of an understanding, appreciation and commitment to evidence-based medicine.

Under a presumptive-treatment regime, antimalarial drugs should be administered to children if they show a fever even without a clinical confirmation of malaria. The rationale behind such a regime in regions like Sub-Saharan Africa is that malaria is endemic to the area, with transmission rates that are quite high. This implies that the most likely underlying cause of a fever episode in children is malaria, which should be treated immediately to improve the likelihood of being cured.

Current WHO guidelines recommend “test and treat” (i.e., confirm the presence of the malaria parasite in a febrile child), but this is only possible where rapid diagnostic test kits are available or where access to trained personnel can lead to a confirmation. This change in treatment practice is in response to new evidence, declining prevalence, and the increasing availability of rapid diagnostic tests (RDT). Under both regimes, however, treating malaria should be through the use of artemisinin-based combination therapies (ACT) because of the resistance of the malaria parasite in that part of the world to the earlier drug, chloroquine.

It is arguably straightforward to tailor our approach to the updated malaria treatment guidelines and protocols for other diseases as well. For instance, we might think to include the likelihood that a child had blood taken from a finger or heel for testing as another quality indicator. This information is readily available for many countries surveyed in the DHS. Use patterns of antimalarial drugs can then be separately compared for children who got tested and children who were not, with the caveat that a low likelihood of receiving any antimalarial drug would not necessarily be a negative result because these drugs should not be given to those who have fever but do not test positive for malaria.   

Relative to self-treatment, we find that febrile children who seek treatment at public and private hospitals, other public and private facilities, and pharmaceutical shops have a higher likelihood to take any antimalarial drug at all. This means that most service providers adhered to the prescribed treatment protocol that was in place during the period of analysis. Notably, going to a traditional healer is no different from self-medication. With respect to receiving good antimalarial treatment (conditional on having received any antimalarial drug), we only observe significant associations for public hospitals and other public facilities. This may be explained by national and international efforts to make ACTs available on a large scale through public channels.

As to the probability of taking only chloroquine conditional on antimalarial treatment, we find significant correlations for public and private hospitals, other public facilities, and pharmaceutical shops. While other private facilities such as private doctors and private maternity homes do not perform any better than self-treatment in terms of good antimalarial drug and chloroquine-only use, they are associated with significant fever relief.

Navigating health care systems in developing countries can be a difficult task for the uninitiated. Weak institutions, for instance, can stifle efforts at getting an accurate picture of the quality of the system itself. This can generate inefficiencies in the program-deployment strategies of multilateral aid. Promptly and easily measuring the overall quality of health care provision, therefore, becomes an important task in this context where getting more nuanced and more accurate measures would be ideal. We hope that the proposed tool may be helpful in this regard.

Add new comment