Am. J. Trop. Med. Hyg., 64(1, 2)S, 2001, pp. 76–84 Copyright 䉷 2001 by The American Society of Tropical Medicine and Hygiene
TOWARD A FRAMEWORK AND INDICATORS FOR MONITORING ROLL BACK MALARIA J. H. F. REMME, F. BINKA, AND D. NABARRO United Nations Development Programme/World Bank/World Health Organization Special Programme for Research and Training in Tropical Diseases, World Health Organization, Geneva, Switzerland; Roll Back Malaria, World Health Organization, Geneva, Switzerland
Abstract. Roll Back Malaria (RBM) is a new global partnership that aims to halve the malaria burden by the year 2010. A framework and indicators for monitoring the outcomes and impact of RBM have been defined through an extensive consultative process. The framework identifies critical areas for monitoring RBM action relating to 1) the impact on malaria burden, 2) improvements in malaria prevention and treatment, 3) related health sector development, and 4) support for RBM action (including partnerships). A set of RBM indicators has been defined that corresponds to these critical areas but that reflects the major variations in malaria epidemiology and the principal interventions in different parts of the world. Countries would select indicators that are appropriate for their situation. Four global indicators are proposed for use by all countries in which RBM action is under way. Data collection procedures are discussed, and it is indicated how monitoring RBM action can build on existing data-collection mechanisms. south of the Sahara.3 This article describes the framework and indicators for monitoring the outcomes and impact of RBM. General principles in developing a monitoring framework. The development of a framework and indicators for monitoring RBM was guided by the following general principles. Broad consensus. Roll Back Malaria is the collective effort of many partners and multiple disciplines, and it is essential that there is a broad consensus among them about the approach to monitoring. Much work and time have been devoted, therefore, to establishing a consensus on the monitoring methodology while ensuring that it remained technically sound. This work started with the creation of a multidisciplinary, cross-cluster group within WHO that developed a first proposal for the approach to monitoring. The proposal was then reviewed with other groups in WHO and other RBM partners at various occasions (including the conference of the American Society of Tropical Medicine and Hygiene in December 1999), and the monitoring methodology was revised repeatedly to meet the requirements of the different partners. The resulting framework and indicators were presented in February 2000 to the third Global RBM Partners Meeting, which endorsed the proposed approach after making a few final modifications and recommending further work to strengthen monitoring of malaria in pregnancy. Relevance to RBM objective. The framework and indicators should be directly relevant to the RBM objective. They should allow an assessment of the impact of RBM action on malaria burden and enable the monitoring of the principal malaria interventions and related efforts to reinforce the health sector. Standardized but adaptable approaches. The epidemiology of malaria, intervention strategies, and health sector development vary considerably between countries and regions, and this variation needs to be taken into account in the monitoring methodology. It has been attempted, therefore, to develop a general framework that covers all situations and to develop a set of RBM indicators that reflects the major variations in malaria epidemiology and the principal interventions. Countries and regions are encouraged to select from the basic set those indicators that are the most appropriate
INTRODUCTION
Malaria is a major public health problem, endemic in over 100 countries in the world. The World Health Organization (WHO) estimates there are ⬃300 million clinical cases every year, with over a million deaths. Over 90% of the burden occurs in Africa.1 The director-general of WHO, in consultation with African heads of state, agreed to tackle malaria as a priority health disease. A global partnership to roll back malaria in the world was initiated.2 The objective of the Roll Back Malaria (RBM) partnership is to halve the malaria burden in participating countries through interventions that are adapted to local needs and by reinforcement of the health sector. The principal mechanism for achieving this is through intensified national action by country-level partnerships working together toward common goals within the context of health sector development and using agreed strategies and procedures. Country partnerships will be supported by a global partnership, and technical support networks will provide the necessary technical assistance. Roll Back Malaria will also encourage strategic investments in the development of better tools and intervention strategies through focused support for research. RBM was launched in October 1998. During the preparatory phase, which lasted until February 2000, activities concentrated on partnership building, RBM inception at country level, and development of mechanisms to facilitate RBM action. One of these preparatory activities was the development of a framework and indicators for monitoring the outcomes and impact of RBM. An effective system for monitoring progress and outcomes will be critical for the success of RBM. Monitoring is needed to provide local feedback to RBM operations and to monitor the progress and impact of RBM at the country, regional, and global levels. RBM will need to report on progress and lessons learned and identify which general modifications are needed in later phases of RBM. One of the major challenges for effective monitoring and evaluation of the impact of the initiative is the inaccuracy of the information on clinical cases and death due to malaria as a result of irregular reporting and underreporting, especially in areas known to be highly endemic, such as in Africa
76
77
FRAMEWORK AND INDICATORS FOR MONITORING RBM
FIGURE 1.
Critical areas for monitoring Roll Back Malaria.
for their specific epidemiological situation and intervention strategy. This approach ensures standardization of RBM indicators and methods while allowing for adaptation to local variation. Local feedback to control. The first priority for data collection at the community and district level is to provide feedback to malaria control and health care systems, and the monitoring system and the selected indicators should facilitate this process. Minimal data collection. The burden of record-keeping and reporting is high in many health care settings, and much of the information collected and reported is never used. Data collection for RBM is to be kept to a minimum and should only be undertaken if the data are likely to be reliable and useful for decision-making. Furthermore, existing mechanisms for data collection should be used as much as possible. There exist many such mechanisms that are relevant to RBM. FRAMEWORK FOR MONITORING RBM
A monitoring framework has been developed that identifies critical areas for monitoring the impact and outcomes of RBM. The framework is shown in Figure 1, with each box representing a critical area for monitoring and the shaded boxes referring to country-level monitoring. The critical areas relate directly to the objective of RBM and to the monitoring of 1) the impact on malaria burden, 2) improvements in malaria prevention and treatment, 3) related health sector development, and 4) support and partnerships. The framework describes the main components of RBM
action, especially at country level. The ultimate objective of RBM is to halve the burden of malaria, and one of the critical areas to monitor is obviously the impact on disease burden. The reduction in burden will be achieved through interventions that are initiated by the national RBM partnership, and this partnership is another critical area to monitor. The actual interventions will vary according to malaria epidemiology and status of the health sector. However, malariaspecific interventions will always include the critical areas of protection and of early diagnosis and treatment. The delivery of these interventions requires strengthening, and thus monitoring, of the relevant components of the health sector ranging from health policy, health systems management, and service delivery, especially at first-line health facilities. A problem as important and complex as malaria is not just an issue for the health sector; the involvement of other important sectors needs to be monitored also. The national interventions will require international support, and other critical areas to monitor are the resources made available at the national and global level, the technical support provided to countries, and the effectiveness of research and development to develop new tools and control strategies. INDICATORS
It is proposed that the principal monitoring of RBM at country level be based on a small number of indicators that represent the critical areas in the monitoring framework. These indicators should be intervention oriented and should provide information for action at the relevant operational level, especially at the district level.
78
REMME AND OTHERS
Table 1 provides a list of indicators that meets these criteria and that are proposed by RBM. It also provides an operational definition for each of the indicators and an indication of the likely source of information. It is recommended that each region or subregion and country select from this list only those indicators (at least one per critical area) that it considers to be appropriate for the local malaria epidemiology and intervention strategy. An example of such a selection for a typical malaria-endemic country in Africa is given in Table 2. Some general comments on the indicators are given below. Impact indicators. The objective of the RBM partnership is ‘‘to halve the burden of malaria through interventions adapted to local needs and strengthening of the health sector.’’3 Given that malaria mortality is by far the most important contributing factor to the burden of malaria as measured in disability-adjusted life years, it is proposed that malaria-related mortality be the principal impact indicator for RBM. The relevant mortality indicators in endemic areas are the crude death rate and the malaria death rate of children aged 0–59 months by age and sex. In Africa, where 90% of malaria deaths occur, no country has a functioning vital registration system. Hence, malaria mortality has always been an estimate derived from models that are based on the outcome of a few epidemiological studies.4,5 Within this framework, an attempt will be made to improve on the estimates of the determinants of malariarelated morbidity and mortality and the population at risk. Data will be collected from several sources in malaria endemic countries. In Africa, existing demographic sentinel sites will be developed into a platform to provide better estimates of crude death rate and malaria mortality (www.indepth-network.org). Both crude death rate and malaria death rate are recommended for the following reasons: 1) there are circumstances where changes in crude death rate can be reliably measured, but changes in malaria death rate cannot: deaths counted in surveys sometimes cannot accurately be attributed to malaria; 2) interventions against malaria can have indirect as well as direct benefits, reducing deaths partly attributable to other conditions, and it is highly desirable to quantify these additional, indirect benefits; 3) malaria is sometimes a competing risk, in which case fewer malaria deaths are offset by more deaths from other causes; when there is no measurable change in crude death rate, we need to distinguish between 2 possible explanations: the failure of malaria control and compensating mortality; and 4) a number of RBM interventions will not be specific to malaria (e.g., management of anemia in pregnancy). Although RBM’s global target for malaria control has been expressed in terms of mortality (50% reduction in deaths by 2010), many countries will want to set additional impact targets. For many countries outside sub-Saharan Africa and Papua New Guinea, the single best core indicator is the annual parasite incidence (by age, sex, and parasite species), which is measured through routine surveillance. Many countries in Europe, Asia, Oceania, North Africa, and Latin America have shown that they can measure annual parasite incidence by routine surveillance (patients with symptoms contacting health services) and can identify Plas-
modium species by microscopy. A morbidity indicator will suit countries where reducing incidence is the principal goal, and reduced incidence is likely to mean fewer deaths. Furthermore, many countries have high case loads but few malaria deaths—for instance, where Plasmodium vivax predominates and where most cases of Plasmodium falciparum infection are treated. Outcome indicators: malaria prevention and treatment. The indicators for malaria prevention and treatment reflect the most important interventions for reducing the global burden of malaria. One of these is the effective use of insecticide-impregnated nets. Although insecticide-impregnated nets will not be used everywhere, this intervention is considered critical for reducing malaria mortality, especially among children in areas with stable transmission in Africa, which is the group at highest risk of death due to malaria. That this is a good indicator for guiding country action was obvious from the results of pretesting data collection instruments in 4 African countries: though the presence of bed nets varied greatly, not a single bed net was found that had been treated with insecticides. An alternative indicator for epidemic-prone areas, where insecticide-impregnated nets may not be used, is the timely detection and control of epidemics. Intermittent treatment during pregnancy is recommended as an appropriate and effective method for reducing the consequences of malaria in pregnancy in highly endemic areas, especially for first and second pregnancies.6 At present, only few countries in Africa have adopted such intermittent treatment as national policy, but it is expected that more countries will soon follow. Early diagnosis and appropriate treatment is probably the most important action against malaria. The proposed indicator is defined as the percentage of high-risk people with a malaria attack who receive appropriate treatment within 24 hr. Where good diagnostic and clinical services are available and actively used, this indicator may be restricted to parasitologically confirmed cases of malaria. However, in most endemic areas, especially in Africa, such services are not widely available, and many patients are managed at home. In such areas, a more appropriate indicator is the percentage of children with fever who receive appropriate treatment within 24 hr. Preliminary results of pretesting an instrument to measure this at the community level in malaria-endemic areas in 2 African countries indicated that ⬍ 10% of children with fever receive appropriate treatment for malaria. Monitoring the improvement in this indicator will be extremely important for rolling back malaria. Health sector development. The indicators for health sector development are not restricted to malaria but touch on broader health sector issues that need to be tackled to ensure effective malaria control. They measure the presence of the relevant health policies and the implementation of these policies at the operational level (usually the district). It is proposed to monitor whether these operational units have adequate funds and skilled staff to implement the policies and to provide basic services (pretesting showed that some health districts spent ⬎ 95% of their budget on personnel and that they had virtually no funds to provide services). Two indicators deal with malaria-related services: whether first- and second-line antimalarial drugs are present at the facility level and, for areas where this is policy, wheth-
79
FRAMEWORK AND INDICATORS FOR MONITORING RBM
TABLE 1 Operational definitions and possible source of information of proposed indicators for country-level monitoring of Roll Back Malaria (RBM) Indicator
Impact Percentage reduction in mortality of children ⬍5 years old (all causes)
Percentage reduction in malaria mortality (children ⬍5 years old and other age groups)
Percentage reduction in malaria incidence rate
Malaria prevention and treatment Protection Percentage of children ⬍5 years old sleeping under insecticide-treated nets
Percentage of pregnant women on intermittent antimalarial treatment
Malaria epidemics detected within 2 weeks of onset and properly controlled
Operational definition
Estimated percentage reduction in mortality of children ⬍5 years old (⫽ number deaths per year of children 0–59 months/number of live births per year). For countries with high transmission, mainly in Africa south of the Sahara. 1) Estimated percentage reduction in malaria mortality in children ⬍5 years old (⫽ number deaths of children 0–59 months attributable to malaria per year/number of live births per year). 2) Estimated percentage change in MDR for all ages or specific age groups (MDR ⫽ number of deaths per year attributable to malaria/population at risk). By convention, (1) is a proportion, (2) is a rate. Mainly (1) in high-transmission areas, (2) elsewhere. Estimated percentage reduction in annual parasite incidence (⫽ number of parasitologically confirmed malaria cases per year/population at risk). Not recommended for high-transmission areas where specificity is low, nor for areas where health information systems are weak or where many malaria cases are not seen by the health system Proportion of children ⬍5 years old who sleep under insecticide-treated nets (in malaria risk areas). Numerator: number of children ⬍5 years old who slept under an insecticide-treated net† the previous night. Denominator: number of children ⬍5 years old surveyed. Proportion of pregnant women who receive intermittent antimalarial treatment in accordance with national policy (only applicable for countries that have a policy of intermittent antimalarial treatment during pregnancy). Numerator: number of women who report to have delivered within the last 6 months and to have received intermittent antimalarial treatment in accordance with national policy during their pregnancy. Denominator: number of women who report to have delivered within the last 6 months and who would have been eligible to receive intermittent treatment according to national policy. Numerator: number of epidemics detected in a specific geographical area (country, district, etc.) within 2 weeks during the last 12 months and for which appropriate control measures have been initiated within the following 2 weeks. Denominator: number of malaria epidemics recorded during the last 12 months within a specific geographical area. ‘‘Appropriate control measures taken’’ means ‘‘action based on preparedness plan of action where such control measures are defined or, in case they are not locally defined, based on global guidelines prepared by WHO.’’
Source of information
1) DSS; 2) Cross-sectional household surveys, including DHS.
1) Routine data on malaria deaths in health facilities; 2) DSS ⫹ DHS, with cause of death determined by verbal autopsy.
Routine surveillance data (passive or active case finding).
Household surveys.
Household surveys.
Health information systems.
80
REMME AND OTHERS
TABLE 1 Continued Indicator
Early recognition and treatment Percentage of high-risk people with a malaria attack getting appropriate therapy within 24 hr
Health sector development Health policy Policy of coverage for all with basic priority health interventions, including malaria interventions Health sector management Percentage of districts having plans or reports reflecting the national policy of providing priority health interventions to whole population Percentage of government recurrent health expenditure within the district health system spent on providing services Percentage of districts systematically using health information for planning Presence of technical skills (including integrated management of childhood illness) at the required level
Service delivery Percentage of facilities with first- and second-line antimalarial drugs available according to national policy Percentage of facilities with adequate parasite detection services according to national policy
Operational definition
High-transmission areas (especially Africa): proportion of children ⬍5 years old with fever who received an appropriate antimalarial treatment within 24 hr. Numerator: number of children ⬍5 years old who are reported to have had fever in the previous 2 weeks and reported to have received the locally recommended antimalarial treatment within 24 hr of onset of fever. Denominator: number of children ⬍5 years surveyed who are reported to have had fever in the previous 2 weeks. Epidemic-prone areas with good diagnostic facilities: proportion of people with a parasitologically confirmed attack of malaria who received an appropriate antimalarial treatment within 24 hr. Numerator: number of people with a parasitologically confirmed attack of malaria and reported to have received the locally recommended antimalarial treatment within 24 hr of onset of symptoms. Denominator: number of people with a parasitologically confirmed attack of malaria
Source of information
Household surveys.
Health information system.
Availability of an explicit national policy to cover the whole population with priority health interventions, including malaria interventions, or a national strategic plan for the health sector that reflect the same policy.
1) National health policy document; 2) National strategic health plan.
Numerator: number of district health plans consistent with the national policy. Denominator: number of district health plans surveyed. Numerator: government recurrent health expenditure within the district health system, excluding salaries. Denominator: government recurrent health expenditure within the district health system, including salaries. Proportion of districts with plans that show that locally collected data is used for planning (health education, prevention and control, epidemic preparedness, outbreak investigation). Index of case management, using the child as denominator: the proportion of children with fever presenting at first-level health facilities who are managed correctly (assessment, classification, treatment, counseling of the mother).
District health plans.
First- and second-line antimalarial drugs available at the recommended cost in the health facility according to national policy on the day of the audit or survey. Proportion of peripheral health services in malarious areas with functional microscopy or dipsticks or support malaria diagnosis in accordance with national policy. Numerator: number of peripheral health services with functional microscopy or dipsticks to support malaria diagnosis in conformity with national policy. Denominator: number of surveyed peripheral health services that require parasite detection services according to national policy.
Health facility surveys.
1) National health accounts (WHO); 2) Public health expenditure review (World Bank); 3) District health costing surveys.
1) District health reports; 2) Supervision visits; 3) Comprehensive or specific systems assessment reports. Health facility surveys.
Health facility surveys.
81
FRAMEWORK AND INDICATORS FOR MONITORING RBM
TABLE 1 Continued Indicator
Intersectoral linkage Effective collaboration in operational research between national research institutions and the MOH
Malaria prevention and treatmentseeking in primary school curriculum
Environmental malaria risk factors are taken into account in design of development projects
Community action Community groups taking action on malaria issues
Support National partnership Inclusive, multisectoral membership established and maintained
Percentage of total agreed RBM budget for the country met Technical support Perceived effectiveness and timeliness of technical support to countries
Operational definition
Source of information
National RBM partnership has identified priority questions for operational research, and national research institutions and the MOH have jointly undertaken operational research on the top two priority questions during the preceding year. Proportion of primary schools that have introduced malaria prevention and treatmentseeking in the school curriculum. Numerator: number of primary schools which have introduced malaria prevention and therapy in the school curriculum. Denominator: number of primary schools. Numerator: number of development projects with possible environmental implications for malaria transmission, which have taken malaria risk factors into account in project design. Denominator: number of development projects with possible environmental implications for malaria transmission.
National RBM partnership reports and progress reports on operational research.
Reports from ministry of education; school visits.
Development project plans. World Bank environmental assessments of projects.
Percentage of endemic communities where groups of community members take action in IEC on malaria, improved home management and referral, and/or provision and treatment or retreatment of bed nets. Numerator: number of endemic communities where groups of community members (not taking into account members of the formal health system) are reported to take action in IEC on malaria, improved home management and referral, and/or provision and treatment or retreatment of bed nets. Denominator: number of endemic communities surveyed.
Key informant interview.
National RBM partnership has been established and has representatives from MOH, other sectors (including ministry of finance and ministry of education), the principal NGOs working on malaria-related issues in the country, national health research institutions, international agencies including World Bank, UNICEF, UNDP, and WHO, other important partners. Percentage of the total budget for all RBM activities planned, and agreed upon, by the national partnership for the preceding budget year that has been met.
Reports of national RBM partnership; country offices of WHO and other multilateral agencies.
Proportion of MOH requests for technical assistance to which partner organizations responded within 1 week and for which the provided assistance was perceived as effective by the responsible officers in the MOH. Numerator: number of official MOH requests for technical assistance to which the concerned partner organization responded within 1 week and for which the assistance was perceived as effective by the responsible officer in the MOH. Denominator: number of official MOH requests of technical assistance to partner organizations.
WHO registry; country reviews using reports from WHO country offices, regional offices and headquarters, MOH.
RBM partnership reports and reported expenditure by partners.
* DSS ⫽ Demographic Surveillance System (INDEPTH network coordinates 27 DSS sites, 17 in Africa); DHS ⫽ Demographic and Health Surveys (USAID funded; MACRO International is implementing group); MCE ⫽ Multi-Country Evaluation of Integrated Management of Childhood Illness (coordinated by WHO; under way in Tanzania, Uganda, Peru, and Bangladesh); MDR ⫽ malaria death rate; MOH ⫽ ministry of health; WHO ⫽ World Health Organization; NGO ⫽ non-governmental organization; UNDP ⫽ United Nations Development Project. † Insecticide-treated net is defined as a bed net that has been immersed in an insecticide solution or directly sprayed with insecticide.
82
REMME AND OTHERS
TABLE 2 Example of Roll Back Malaria indicators for a malaria-endemic country in Africa Impact
Malaria prevention and treatment Protection
Early recognition and treatment
Health sector development Health policy
Health sector management
Service delivery Intersectoral linkage
Community action Support National partnership
Technical support
Percentage reduction in malaria-related mortality (children ⬍5 years old, other age groups). Percentage of children ⬍5 years old sleeping under insecticide-impregnated nets. Percentage of pregnant women on intermittent antimalarial therapy. Malaria epidemics detected within 2 weeks of onset and properly controlled. Percentage of children ⬍5 years old with fever receiving appropriate therapy in 24 hr. Policy of universal coverage for all with a basic intervention package, including malaria interventions. Percentage of districts with plans that reflect above national policy. Percentage of government recurrent health expenditure spent on services. Percentage of districts using health information for planning. Presence of technical skills (including Integrated Management of Childhood Illness) at the required level. Percentage of facilities with first- and second-line antimalarial drugs available. Malaria prevention and therapy-seeking in primary school curriculum. Effective collaboration in operational research between national research institutions and ministry of health. Community groups taking action on malaria issues. Inclusive, multisectoral membership established and maintained. Percentage of total agreed Roll Back Malaria budget for the country met. Perceived effectiveness and timeliness of technical support to countries.
er there are parasite-detection services. Pretesting of instruments showed the importance of the indicator on antimalarials for RBM: 5 out of 10 health facilities in one district had no antimalarials at all, whereas the remaining 5 facilities combined had some 3,000 chloroquine tablets but no sulfadoxine-pyrimethamine. This compared with over 100,000
chloroquine tablets and 153 tins of sulfadoxine-pyrimethamine in shops in the same villages. Intersectoral collaboration is another important element of the RBM approach. It is proposed to monitor the collaboration in operational research between national research institutions and the Ministry of Health, because collaboration between research and control is often reported to be poor. It will also be monitored whether prevention and treatment seeking for malaria is taught in primary schools in endemic areas and whether environmental risk factors for malaria are taken into account in the planning of development projects. The involvement of the community will be critical in most areas, but the proposed indicator is not very specific, and local adaptations will be required. Support. An effective national partnership is the key to success for RBM. It is proposed to monitor whether the national partnership really brings all potential partners together and whether these partners generate the necessary resources to roll back malaria in the country. With respect to technical support, it will be assessed whether the clients are satisfied with the technical support provided to them. Additional indicators for monitoring RBM at regional and global levels. The monitoring of RBM action at regional and global levels will be mainly a matter of aggregating the results for the country-specific indicators. In addition, there are a few global indicators proposed for research and development and for the global partnership. For research and development, 2 indicators have been selected that are considered most relevant for the short and medium term: the number of antimalarial drugs that have progressed to the level of Phase 3 trials, and the number of cost-effective interventions that have been developed and field tested. The main indicator for monitoring the global partnership is the extent to which the RBM global budget has been met by partners. Global indicators. Although most indicators will vary between countries, there are 4 indicators that are considered so important that they have been selected as global indicators. It is recommended that all RBM countries report on these global indicators wherever they apply. These 4 global indicators are as follows: the percentage reduction in malaria death rate; the percentage of children ⬍ 5 years old sleeping under insecticide-impregnated nets; the percentage of highrisk people having a malaria attack getting appropriate therapy within 24 hr; and the percentage of facilities with firstand second-line antimalarial drugs available. APPROACHES TO DATA COLLECTION
There are 5 main approaches to collecting data for the proposed RBM indicators. These are the regular health information system, demographic surveillance systems, community surveys, health facility surveys, and review of documents. Routine surveillance (health information system). Both the health care system and the health information system are well developed in several countries in which malaria is endemic. In these countries, the health information system can be used to monitor trends in malaria morbidity and mortality at the facility level. Furthermore, for many of these countries, it is believed that a large proportion of patients with
83
FRAMEWORK AND INDICATORS FOR MONITORING RBM
malaria are seen at the health facility and that the reported trends can be used as a good proxy of the national trends in malaria morbidity and mortality. Effective reporting systems are often in place that are adequate for the purpose of RBM. Demographic surveillance systems. Unfortunately, in many countries, especially in Africa, the quality and reliability of information generated by the health information system is poor, and the health information system rarely provides information on the burden of malaria at the community level. Hence, the information provided by the health information system in these countries has limited value for monitoring the impact of RBM, especially in Africa, where ⬎ 90% of the disease burden is found. The only option for monitoring trends in malaria mortality in Africa may be through demographic surveillance systems in sentinel sites. Fortunately, such demographic surveillance systems are now operational in 28 sites in 14 African countries, and these sites have joined together in the INDEPTH network that works toward standardization of the methodology. Because of the need for reliable information on trends in malaria mortality, RBM intends to support the INDEPTH network to include surveillance of malaria mortality in its African sites. RBM will also work with INDEPTH to extend the network to include new sites in the areas not covered to date, especially in Central Africa and Nigeria. Community and household surveys. Community-based information on prevention and treatment practices will be critical for monitoring the effectiveness of related RBM interventions. Such information is especially important for monitoring the outcomes of RBM action in areas where a large proportion of patients are managed at home and where the burden of malaria is usually most severe. This community-based information is not readily available, and it will be necessary to undertake special community surveys for this purpose. Community surveys tend to be time-consuming and relatively costly, and they can therefore only be undertaken in selected sentinel sites in each country and at intervals of 2–3 years. There are several ongoing activities in which information on the proposed community-based indicators is already being collected. The demographic and health surveys, funded by the United States Agency for International Development and executed by MACRO International, have been extended to include a malaria module that allows the collection of community-based information on several of the RBM indicators. The same is true for the multiple indicator cluster survey of UNICEF, which was undertaken in many countries in the year 2000. Finally, RBM has developed a methodology for situation analysis, which includes instruments for community-level assessment of indicators such as the percentage of children ⬍ 5 years old sleeping under insecticideimpregnated nets, provision of intermittent treatment in pregnancy, provision of timely and appropriate treatment of children with fever, and community action against malaria. At least 30 countries intended to undertake a situation analysis during the year 2000 as part of their RBM strategy development, and the information to be generated will provide important baseline data on key RBM indicators. Special surveys are not a sustainable solution to the need for community-based information, and RBM will support the development of alternative approaches that would enable the
health information system to routinely collect communitylevel data, and for the community itself to monitor key RBM indicators. Health facility assessment. The WHO/UNICEF initiative on integrated management of childhood illness has developed and tested a methodology and instruments for a multicountry evaluation of the integrated management of the sick child. The instruments include an assessment of the clinical skills of health care staff and an assessment of the available supplies and equipment at the health facility. The evaluation provides all the information needed for the proposed facility-based RBM indicators. The advantage of the integrated management of childhood illness approach is that the assessment is not limited to skills for the management of malaria only; it also addresses the management of the sick child, including children with malaria. It is therefore proposed to use for the health facility assessment the relevant sections of the integrated management of childhood illness evaluation methodology or to rely on the results of the integrated management of childhood illness evaluation where this is undertaken. The application of the method requires special skills, and it cannot be done in every facility. It is recommended, therefore, to combine the health facility assessment with the community-based surveys and to undertake them in the same districts and at the same interval. In case the indicator on technical skills of health care staff is not selected, the health facility assessment becomes much simpler, consisting only of an assessment of the presence of the required antimalarial drugs and, where this is policy, of the presence of parasite-detection services. This simplified facility assessment can be done as part of routine supervisory visits or be easily combined with the community surveys Review of documents. This is the easiest and cheapest of the data collection methods. The main requirement is that the necessary documents exist and are available; some special effort and travel may be needed to ensure that this is indeed the case. It will also be important to retain copies of the relevant documents so that these can be made available as supporting evidence for the monitoring findings on the selected indicators. DISCUSSION
The consultative preparatory work on monitoring RBM action has resulted in a broad consensus on a global monitoring framework and a set of corresponding standardized indicators. There is also agreement that the selection of indicators will need to be adapted to the local epidemiological situation and status of health sector development. Thus, there is a broad understanding about what needs to be done with respect to monitoring RBM action. The main challenge is now to ensure that the required monitoring systems are implemented and that the necessary data are collected regularly. The main action for rolling back malaria will be undertaken by national RBM partnerships in the malaria endemic countries. Monitoring the outcomes of this action will also need to be the responsibility of these national RBM partnerships. It is recommended that each national partnership reviews
84
REMME AND OTHERS
the proposed framework and indicators, decides to what extent they wish to monitor the outcomes and the impact of their interventions, which of the proposed RBM indicators they wish to adopt, and how and when they will use the monitoring results for decision-making on rolling back malaria. The national partnership needs also to consider the expected costs of implementing a selected monitoring system and to decide how these costs will be covered. The global RBM partnership will contribute resources to specific activities such as the INDEPTH network for monitoring of malaria mortality in Africa, but the other country-level monitoring costs will need to be met by national partnerships. If so requested, WHO and the other global partners will provide the national partnership with technical assistance for monitoring and with data collection instruments for selected indicators. As country activities to roll back malaria are accelerating, it is becoming urgent to get the monitoring systems going and to collect the necessary baseline data. This will enable monitoring of change in the short to medium term on key outcome indicators such as the percentage of children with fever getting adequate treatment and the use of bed nets by high-risk populations. Showing progress on these indicators will be critical to maintain commitment of the national and global RBM partnership. But good baseline data are also needed to show the long-term impact on malaria morbidity and mortality and to document whether RBM is indeed achieving its aim of halving malaria mortality by the year 2010. The latter will require reliable baseline and follow-up data on the main impact indicator, especially for Africa. The monitoring framework as presented in this article forms a core structure which will continue to evolve. WHO will provide regular updates of the framework on the RBM website
at www.RBM.WHO.int, and readers are encouraged to consult the website for the latest official version. Acknowledgments: We acknowledge the contributions of the members of the WHO cross-cluster group on RBM monitoring, the WHO regional advisors on malaria, and many others from different RBM partners who contributed to the development of the monitoring framework. We also thank the research teams from Zambia, Nigeria, Ghana, and Mali who pretested the situation analysis instruments and to whose findings we have made reference in the text. Authors’ addresses: J. H. F. Remme, UNDP/World Bank/WHO Special Programme for Research and Training in Tropical Diseases, World Health Organization, Geneva, Switzerland. F. Binka and D. Nabarro, Roll Back Malaria, World Health Organization, Geneva, Switzerland. Reprint requests: J. H. F. Remme, UNDP/World Bank/WHO Special Programme for Research and Training in Tropical Diseases, World Health Organization, 20 Avenue Appia, Ch-1211 Geneva 27, Switzerland, Telephone (⫹41 22) 7913815, Fax (⫹41 22) 791 4774 (email
[email protected]). REFERENCES
1. World Health Organization, 1999. The World Health Report 1999: Making a Difference. Geneva: World Health Organization, 49–63. 2. Nabarro DN, Tayler EM, 1998. The Roll Back Malaria campaign. Science 280: 2067–2068. 3. World Health Organization, 1999. Weekly epidemiological report. 74: 265–272. 4. Murray CJL, Lopez, 1997. Mortality by cause for eight regions of the world: global burden of disease study. Lancet 349: 1269–1276. 5. Cooper RS, Osotimehin B, Kaufman JS, Forrester T, 1998. Disease burden in sub-Saharan Africa: what should we conclude in the absence of data. Lancet 351: 208–210. 6. WHO Technical Report Series, in press. Report of the 20th WHO Expert Committee on Malaria.
