End of year review: Malaria is declining, and IE should help address the remaining challenges


This page in:

A conference on access to malaria medicine recently held at the World Bank offered many substantive studies – and I will discuss some in detail in the new year. However with my last post of 2011 I’d like to end the year on some good news (even if the news is only partially related to impact evaluation).

As reported in many fora, the world is making very significant gains against malaria. One way to track gains is to note fairly substantial declines in child mortality in certain malaria endemic countries. Another less publicized metric is the decline in the prevalence of the malaria parasite in population surveys.

Carlos Guerra and others review thousands of community-based malaria prevalence surveys (specifically, these are surveys of the parasite Plasmodium falciparum (Pf), one of the most common and most deadly variants of malaria). It is difficult to draw careful distinctions across time periods since the community data is not necessarily spatially consistent over time. However the authors find a remarkable decline in the mean prevalence of pF among children – from 37% over the 1985 to 1999 period to 17% over 2000 to 2007.

Echoing these gains, Valerie D’Acremont and others review 39 studies of causes of fever cases taken over the past 20 years in Sub-Saharan Africa. They find that the median proportion of fevers associated with Pf has declined from 44%, when measured over the period 1986 – 1999, to 22% from 2000 to 2007. Declines of similar magnitude were observed for both the under 5 and over 5 populations, in rural or urban areas, and among patients presenting to either a primary clinic or a hospital.

Clearly the renewed attention that malaria has received globally, coupled with efforts to introduce new prevention technologies such as long-lasting insecticide treated bed nets, has achieved important gains. Another factor has been the introduction of a new and largely efficacious treatment for malaria – Artemisinin Combination Therapy, known by its acronym ACT.

But this good news of declining burdens also presents several challenges for cost-effective policy. For one, malaria has had a history of resurgence so even if it is nearly eliminated in the coming decades, vigilance must be maintained. Another challenge concerns the curative practices for suspected malaria infections. Traditionally in much of sub-Saharan Africa, all fever has been presumptively treated as malaria. But of course many fevers are due to other causes, and the proportion of fever due to malaria is likely to be falling dramatically. Hence presumptive treatment in an environment of falling malaria prevalence not only “wastes” resources (ACT is comparatively expensive) but increases the relative likelihood of missing other potentially fatal diseases.

Fortunately this challenge may also have a solution. There are now new reliable malaria rapid diagnostic tests (mRDTs) that enable proper diagnosis of malaria at all levels of the health system. Results are obtained within 20 minutes from a minimally invasive capillary prick. (In terms of mRDT quality, WHO helpfully evaluates all marketed mRDTs).

Sylla Thiam and others report that in late 2007 Senegal introduced universal parasitemia based diagnosis with mRDTs in all public health facilities. The rate of parasite-based diagnosis for febrile cases increased from 4% to 86% by 2010. Over the same period the prescription of ACT dropped from 73% of all fever cases to 32% and 520,000 courses of inappropriate ACT prescription were estimated to be averted.

Valerie D’Acremont and others assess the effect of introducing RDTs on the prescription of anti-malarials in urban Tanzania through a variety of methods including before-after analysis with administrative data and with specially collected survey data. The authors also implement a small-scale pair-matched randomized trial at the clinic level. The results across all 3 methods were highly consistent – a significant reduction in the proportion of fever cases receiving anti-malarials (for example a three-fold decline from 60% to 22% in the randomized study). Their study additionally suggests that mRDT use should be integrated with training on the management of other fever causes to prevent overuse of anti-biotics.

So there is great promise from the widespread use of mRDTs. But there are wrinkles as well. One challenge is to assure prescription adherence to a test result in a widespread culture of presumptive treatment. Another key factor – the variations in care-seeking across endemic countries. In Senegal, approximately 75% of patients with fever go to public facilities, so introducing mRDTs and ACT through the public sector will reach a great deal of the population. However 74% of anti-malarials bought in Tanzania are bought from the private sector (mostly from drug stores and general stores) and well over 90% of anti-malarials in Nigeria were purchased through the private sector. We need smart policies that assure affordable medicines and access to diagnostics, even when the patient presents to a private shop owner with little or no formal health training.



Jed Friedman

Senior Economist, Development Research Group, World Bank

Join the Conversation

December 22, 2011

Hi Jed,

The decline in pf- related illnesses described in the Guerra and D'Acremont studies is certainly interesting, but I don't think there's enough information here to make assertions that ACT, bed nets and increased attention on Malaria is the primary driver here. We don't see past 2007, and many of these changes in global public health priorities and technologies occur quite late in this time frame (when I arrived in Malawi in 2006, artesunate was still considered quite new).

Sure, we have the Global fund's anti-malaria work kicking in in 2002-03, but given the current (fairly arbitrary) parsing of the results, we still can't say much about causation here - not all of these countries would have received the same attention from donors, so we could be stumbling into "everyone is treated" Jeff Sachs territory here. Meta-studies like these are always constrained, but still I'd like to see some attempt at a time series.

Good news, yes, but let's not start handing out trophies yet.

Jed Friedman
December 23, 2011

Hi Matt, thanks so much for the thoughtful comment. You are absolutely right that the observed declines can be due to many factors (and likely are due to many factors). Only some of these factors relate to introductions of new malaria control technologies such as LLINs and ACTs. Other factors, perhaps, include general health system strengthening, better management of malaria control activities including IRS, obviously local weather can play a role, etc.

I certainly didn’t mean to imply that these gains were solely due to new technologies introduced around 2005 or so. And I agree that more careful time-series studies would be illuminating. In fact one example the analysis you request is this study from Zanzibar:


that relates the decline in parasite prevalence to the sequential introduction of ACTs and then LLINs. I am also aware of work in progress that relates relative declines in child mortality at the local level to the relative coverage of LLINs – hopefully this work will be circulated soon. Of course these studies are associative in nature, but it would be odd to think the introduction of new effective technologies (LLINs, ACTs) have no impact. And the introduction of ACTs occurred as early as 2002 in some areas. Determining the relative contribution of these interventions to prevalence reduction, of course, is much trickier as you indicate.