news analysis advocacy
AfricaFocus Bookshop
New Gift CDs
China & Africa
tips on searching

Search AfricaFocus and 9 Partner Sites



Visit the AfricaFocus
Country Pages

Burkina Faso
Cape Verde
Central Afr. Rep.
Congo (Brazzaville)
Congo (Kinshasa)
Côte d'Ivoire
Equatorial Guinea
São Tomé
Sierra Leone
South Africa
South Sudan
Western Sahara

Get AfricaFocus Bulletin by e-mail!         More on politics & human rights | economy & development | peace & security | health

Print this page

Visit AfricaFocus Bookshop US | UK

Africa: Dramatic Anti-Malaria Results

AfricaFocus Bulletin
Feb 5, 2008 (080205)
(Reposted from sources cited below)

Editor's Note

New anti-malaria interventions, when applied together, can have dramatic results, according to a new World Health Organization study. The study reported declines in cases in children under five of 60% in Ethiopia, 64% in Rwanda, 29% in Zambia, and 13% in Ghana, between the period 2000-2005 and the year 2007. The greater impact in Ethiopia and Rwanda was clearly associated with massive campaigns of free distribution of long-lasting insecticidal-treated bednets.

This AfricaFocus Bulletin contains excerpts from the study, which was released on January 31. For additional news and background on malaria, see and Previous AfricaFocus Bulletins on health issues are available at

Another AfricaFocus Bulletin sent out today contains an update on President Bush's budget proposals and rival Congressional proposals to increase the budget allocation for global health. While President Bush highlighted renewal of spending on HIV/AIDS in his State of the Union Message, activists and congressional advocates say that his budget proposal "flat-lines" funding at current levels and does not respond adequately to the needs.

New on

5 full-length transcripts of interviews for book:

Mary Jane Patterson -
Ben Magubane -
Robert Van Lierop -
Geri Augusto -
Dumisani Kumalo -

++++++++++++++++++++++end editor's note+++++++++++++++++++++++

Impact of long-lasting insecticidal-treated nets (LLINs) and artemisinin-based combination therapies (ACTs) measured using surveillance data, in four African countries

Preliminary report based on four country visits

31 January 2008

Submitted by: World Health Organization, Global Malaria Program Surveillance, Monitoring, and Evaluation Unit

[Excerpts. For full report, including tables and figures, and other related information on malaria visit]


Background and methods:

In collaboration with The Global Fund, the World Health Organization evaluated the impact of recent investments in malaria control by conducting field evaluations in four countries (Zambia, Ethiopia, Ghana, and Rwanda) in November- December 2007. The main interventions were nationwide distribution of long-lasting insecticidal nets (LLINs) and
artemisinin-combination therapy (ACTs) medicines.

The principal method was review of clinical data at rural hospitals and health centers geographically distributed in each of four countries. In Zambia, we reviewed data at the national level for all hospitals and clinics. The main impact indicator was percentage change in the number of in-patient malaria cases and deaths in children <5 years old prior to and after nationwide implementation of LLINs and ACTs. The weighted average percentage decline of in- patients in children <5 years old in in-patient facilities visited in Ethiopia was 60% for cases and 51% for deaths, and, in Rwanda, was 64% for cases and 66% for deaths. Zambia national data showed decline of 29% in cases and 33% in deaths in children <5 years old. In general, non-malaria cases and deaths remained stable or increased, except in Ghana. The median decline of in-patient malaria in Ghana was 13% for cases and 34% for deaths, but non-malaria cases and deaths declined more than those from malaria (40% and 42%).


We found strong initial evidence that the combination of LLINs delivered during mass distributions to all children <5 years or all households and nationwide distribution of ACTs in the public sector was associated with widespread decline of >50% in in-patient malaria and deaths throughout Rwanda and Ethiopia. The main difference between Ethiopia and Rwanda with dramatic impact, compared with Zambia and Ghana with more limited impact, was sufficient quantities of LLINs delivered in mass distributions in 2005 or 2006.

About This Report

This preliminary report is due on 31 January 2008 to the Global Fund from WHO about visits to four African countries.[Ethiopia, Rwanda, Zambia, and Ghana] ... For this preliminary report, we concentrated on in-patient ("hospitalized") malaria cases and deaths in children <5 years, the age group with the highest mortality rate due to malaria. As we extend our analysis, we will add more information on older age groups, out-patient cases, laboratory data, measures of dispersion, and other more in-depth analyses.


Conversations, field trips, and research reports indicated impact from long-lasting insecticidal nets [LLINs] and artemisinin-based combination therapies provided by national governments and international partners, but systemic documentation was lacking. Surveys alone were not providing sufficient and timely impact data for advocacy or to optimally inform management decisions at district, national, and international levels. Therefore, the Global Fund and WHO used routine surveillance data to measure impact in several African countries. Because most countries did not have strong surveillance and logistic information systems in place, it was necessary to make field visits to districts, hospitals, and health centers to assess surveillance and logistic data. Five countries were chosen to be visited by WHO malaria personnel--Zambia, Ghana, Ethiopia, Rwanda, and Tanzania. This preliminary report covers four countries that had visits during November and December 2007.


Countries were chosen based on their early (2003-2006) introduction of LLINs and ACTs and qualitative assessment by Global Fund and WHO staff about reasonable nationwide distribution. The visits took place during November-December 2007 and lasted two weeks. A written protocol was followed by all teams for selection of districts and health facilities, and data collection. Both Ministry of Health and WHO personnel were involved in data abstraction. In Ethiopia and Rwanda, we attempted to mostly select districts with stable malaria as well as widespread geographical representation. In Zambia and Ghana, all districts have stable malaria. In each selected district, we planned for interviewers to visit one hospital and one out-patient health facility. Interviewers abstracted data either from health-facility copies of national surveillance forms, other health information forms, or from patient registers. Data was collected from the district health team about the starting date of distribution of insecticide-treated nets (ITNs), LLINs, and ACTs in the district; and the quantity that was received by month. At least two persons visited each district for at least two days. We attempted to abstract monthly data starting in 2000. In in-patient facilities, we collected data on in-patient malaria and all-cause cases and deaths for two age groups- -<5 years and H5 years. In out-patient facilities, we collected data on out-patient malaria and all-cause cases for two age groups (<5 years and H5 years) and malaria laboratory testing data where available (number of suspected malaria cases, number tested, number laboratory positive). Additional health information data was collected at the national and district level about surveillance and malaria interventions.

Selection of districts.

We planned to visit 4 districts in each country and to examine national surveillance data if that was available. In Ghana, 4 districts in different parts of the country were selected based on the knowledge of reasonable malaria program operations by the national malaria program. Previous knowledge of impact measures was not used to make selections. In Zambia, we examined quarterly national health management information system (HMIS) data from 2000 to the second quarter of 2007. More than 900 health facilities report in-patient data and approximately 1300 health facilities report out-patient data. In addition, 4 districts were selected for visits that had high percentage decline in inpatient malaria cases in children <5 years old based on HMIS data. HMIS data by health facility was available, including third quarter 2007, for two of the districts visited. In Ethiopia, Ministry of Health officials wanted to expand the number of districts (weredas) from four to eight and the health facilities to 13 to cover four major Regions--Ormoya, SNNP, Amhara and Tigray (these regions have areas with moderate and unstable malaria). Two districts (one health centre and one hospital) were selected from each Region. Selection of the districts was mainly based on knowledge of malaria burden and epidemiological risk factors (such as altitude, water bodies, etc). In Rwanda, despite the original plan to cover 4 districts, the scope of the evaluation was extended (at the request of the national malaria programme) to include all five provinces. We selected two districts randomly per province; hence covering 10 of 33 districts. We selected one hospital and one health center per district, covering 9 out of the 39 hospitals and 10 out of the 439 health centers. All heath centers had in-patient data and all outpatient departments had data on malaria laboratory testing. One health facility was excluded from analysis because of incomplete data.

Number of health facilities included in the analysis.

In-patient data came from approximately 900 facilities in Zambia, 4 facilities in Ghana, 7 facilities in Ethiopia (6 hospitals and 1 health centre), and 19 facilities in Rwanda (9 hospitals, 10 health centres). Out- patient data came from approximately 1300 facilities in Zambia, 13 facilities in Ethiopia (6 hospitals, 7 health centres), and 19 facilities in Rwanda (9 hospitals, 10 health centres).

Pre-intervention and post-intervention time periods.

We estimated the percentage change in malaria cases and deaths by comparing the average annual number of cases and deaths before large-scale distribution of LLINs and ACTs (usually 2000-2005) with the number of cases and deaths in the latest post-introduction period (2007). We used the same period of analysis (for example, January to October or January to November) for both baseline and post-intervention (2007) periods. The baseline period (2001-2005) was constant in Rwanda. In Ethiopia, we used different baseline periods depending on the availability of data. For in-patient data in Ethiopia, data was missing for 2001 for 3 of 7 in-patient facilities, for 2002-2003 for 2 of 7 in-patient facilities, and for 2003 for 1 or 2 of 7 in-patient facilities. Instead of imputing data, we reduced the baseline period to include years in which data was available for 6 or 7 of 7 in- patient facilities.



Impact--Ethiopia and Rwanda. Percentage decline of in-patient malaria cases and deaths in children <5 years old in 2007 compared to 2005 was 64% for cases and 66% for deaths in Rwanda, and 60% for cases and 51% for deaths in Ethiopia. Figure 1 and 2 show trends of in-patient malaria and non-malaria cases in children <5 years by year for Ethiopia and Rwanda. In-patient malaria cases in children decline markedly while nonmalaria cases remain stable (Ethiopia) or decline only slightly (Rwanda). ...

Impact--Zambia. From national data, percentage decline was 31% for in-patient malaria cases and 37% for in-patient deaths of all ages, and was 29% for cases and 33% for deaths in children <5 years. Non-malaria in-patient cases and deaths remained stable, but out-patient cases increased 48%. The median percentage decline from HMIS data for the 4 districts that were visited was 73% for in-patient malaria cases in children <5 years old and was 76% for in-patient malaria deaths in children <5 years old. In two districts that we visited that had mass distribution of LLINs in 2005 or 2006, percentage decline was 71% in cases and 33% in deaths of in-patients <5 years old in Kalomo district, and was 53% in cases and 85% in deaths in Kaoma district.

Impact--Ghana. The median percentage decline was 13% for in-patient malaria cases and 34% for in-patient malaria deaths in children <5 years old. However, non-malaria cases declined even more--40% for non-malaria in-patient cases and 42% for non-malaria in-patient deaths in children <5 years old.


Table 3 shows available national-level information about LLIN and ACT distributions by country. Ethiopia Ministry of Health (MOH) conducted two mass distributions of LLINs--one in 2006 and one in 2005--both targeting one LLIN per 2 persons. ACTs were first distributed in the public sector in 2005. Rwanda MOH introduced LLINs and ACTs nationwide within a 2-month period (September-October 2006). The MOH conducted mass LLIN distribution to children <5 years in September 2006 during the measles campaign. ACTs were introduced quickly in October 2006 to public-sector health facilities. No nationwide mass distribution was conducted in Zambia in 2005-2006 (nationwide mass distribution was mostly completed in 2007). A nationwide mass distribution of LLINs to children <24 months was conducted in November 2006 in Ghana. There was stock-out of LLINs for routine distribution at antenatal care clinics in Ghana in the districts that we visited for nearly all of 2007. LLIN use. Survey data indicated use of insecticide-treated nets in children <5 years old of 23% in Zambia in 2006, 55% in Ghana in 2007, and 60% in Rwanda in 2007.



This report documents for the first time marked, geographically widespread impact in medium- and large-sized countries using large- scale distribution of LLINs and ACTs. Our investigation showed that declines of malaria cases and deaths were dramatic in Rwanda and Ethiopia (>50%) and occurred within 12-24 months of nationwide 11 distribution of LLINs and ACTs. In fact, declines in in-patient cases and out-patient laboratory-confirmed cases occurred within 60 days of nationwide distribution in Rwanda (Figure 4). In both Rwanda and Ethiopia, similar declines (>50%) occurred for impact measures that required malaria laboratory testing--out-patient laboratory-confirmed cases and malaria slide positivity rate. In Rwanda, all 19 health facilities performed malaria smears on all suspected malaria cases. The decline in in-patient and out-patient laboratory-confirmed malaria cases occurred in the face of increases in out-patient and inpatient non-malaria cases in most countries during 2001 to 2004- 2005 due to introduction of health insurance schemes, resolving civil conflict, and improvement of health services.

In Rwanda, there was a difference in percentage declines of inpatient cases and deaths, and out-patient laboratory-confirmed malaria cases in the 10 health centres compared to 9 hospitals. We are investigating this difference with further analyses.

In Ethiopia, indoor residual spraying (IRS) has been a wellestablished vector control intervention for a long period. It is applied in a focalized manner by targeting villages at risk for malaria epidemics. All districts that we visited had been applying IRS in a limited way while deploying LLINs to all populations. We were not able to evaluate contribution of IRS to the decline.

The nationwide decline in in-patient malaria cases and deaths in children in Zambia by approximately one-third is a significant achievement. However, the nationwide decline in Zambia appears to have been lower than in Rwanda and Ethiopia. The key difference between Zambia compared to Rwanda and Ethiopia appears to have been insufficient LLINs to distribute nationwide in 2005 or 2006 in Zambia. In addition, our visits to districts and health facilities showed frequent stock-outs of ACTs occurred at health facility level in Zambia during the 2006-2007 malaria seasons. However, in contrast to the moderate impact nationally, decline of in-patient malaria cases in children in two districts with mass distribution of LLINs in 2005 or 2006 that we visited in Zambia was similar to the decline in Ethiopia and Rwanda.

The lack of definite impact associated with LLINs and ACTs in Ghana is unexplained. The decline was 13% for in-patient malaria cases and 34% for deaths in children <5 years but the declines in non-malaria cases (40% and 42%, respectively) was greater. This is consistent with general improvement in general health services, but it is difficult to confidently ascribe the moderate declines in malaria cases and deaths to the malaria interventions. The short period of data available (2005-2007) at the hospital level limited our analysis. Both malaria and non-malaria out-patient malaria cases were rising in 2005-2007, probably due to effects of health insurance in 2006 and 2007. However, it was clear that the decline in Ghana did not approach that of Rwanda, Ethiopia, or two districts in Zambia with mass LLIN distribution.

Several factors may be involved in the limited impact. First, there was insufficient funding to conduct nationwide distribution of LLINs to all children <5 years or to all households. Instead, LLINs were distributed to all children <24 months in November 2006 with the limited LLINs that were available. Environmental conditions, increased rainfall, fees for public-sector ACTs, limited data to measure pre-intervention baseline, and higher malaria transmission, alone or in combination, could be responsible for the unexpected finding.


Our investigation revealed that surveillance is a powerful tool for quickly and continuously monitoring interventions with high impact at the health facility, district, and national level. In addition, the "slide positivity rate" (percentage positive out of total patients with laboratory test results) was shown to be an excellent indicator in both Ethiopia and Rwanda. The slide positivity rate declined progressively from 30-60% to near 10% and below in most health facilities in Ethiopia and Rwanda. Surveillance data was not being used as a management tool in most countries and districts that we visited, which is not unexpected since the Roll Back Malaria partnership and The Global Fund have not fully supported use of surveillance data to monitor impact, either locally or at national level, in high-burden African countries. Management information systems monitoring stock-outs at the health facility and district level were also not in place. Going forward, we believe that decentralized monitoring of surveillance and logistics data, and information systems to support analysis and use of data will be key to achieving maximal program performance and effectiveness. ...

Although great progress has been made, much more needs to be done in the four countries to reach malaria mortality reduction levels of >75% in all districts. Many households still do not have 2 LLINs per one person and many children are not consistently sleeping under LLINs each night. The percentage of children that receive an ACT within 24 hours of onset of fever is not optimal.

In summary, this initial data indicates that widespread distribution of sufficient (at least to all children <5 years) LLINs and ACTs in the public sector resulted in widespread dramatic reductions in the burden of severe malaria morbidity and mortality. More limited impact in Zambia (nationwide) and Ghana was associated with lack of nationwide distribution of sufficient LLINs. Surveillance data revealed a potential issue with impact in Ghana. International partners should urgently collaborate with national governments to ensure that all households have at least one LLIN per two persons, and that surveillance and logistics monitoring systems are in place in all high-burden countries to highlight management issues and enable action to resolve them. The magnitude of decline (>50%) found in Rwanda and Ethiopia is similar to that needed to reach Abuja mortality reduction targets for 2010 (>50%). It appears that dramatic reduction in malaria mortality can be achieved quickly and may enable many African countries to make rapid progress towards the child survival Millennium Development Goal.

AfricaFocus Bulletin is an independent electronic publication providing reposted commentary and analysis on African issues, with a particular focus on U.S. and international policies. AfricaFocus Bulletin is edited by William Minter.

AfricaFocus Bulletin can be reached at Please write to this address to subscribe or unsubscribe to the bulletin, or to suggest material for inclusion. For more information about reposted material, please contact directly the original source mentioned. For a full archive and other resources, see

Read more on |Africa Health|

URL for this file: