In Brazil, the Ministry of Health (MoH) adopted IMCI guidelines produced by WHO/PAHO since 1996, and these have been adapted to the national and sub-national epidemiological conditions. Implementation of IMCI began in 1997 in several states, particularly in the North-East (Ceará, Pernambuco, Paraíba, Bahia and Sergipe) and North (Pará) Regions, which have the worst socioeconomic and health indicators. By 2002 all states had begun implementation, in the context of the Family Health Programme (FHP), supported by the World Bank and the MoH. Low-income municipalities (or neighbourhoods in larger cities) are provided with family health teams, which include a family doctor, a registered nurse, two health auxiliaries, and four to six community health workers (CHW). The teams are based in first-level government facilities, known as Family Health Programme facilities.

Each FHP team covers from 600 to 1000 families; each CHW visits at least once a month 100 to 200 families. CHW training guidelines have been reviewed for compatibility with IMCI. Wide CHW coverage ensures the vigorous implementation of the community component of IMCI. The team is expected to provide comprehensive health care to all families in the area, offering health promotion, prevention, early diagnosis, treatment and rehabilitation, and referral to specialists as necessary. By January 2001, up to 5139 FHP teams had been assigned to 1933 municipalities, and 111,564 CHWs were employed in 4073 municipalities. The programmes are aimed primarily at increasing access to health care services by the poorest segment of the population. The degree of compatibility between messages delivered by these programmes and IMCI key messages is under review.

To join the FHP, a municipality applies to the federal government and makes a financial contribution. Coverage is patchy, therefore. Some municipalities have one or more teams; others have none. The rate of introduction of teams is high, however; by October 2001, more than 20% of municipalities in some North-Eastern states had at least one team.

IMCI training is provided mainly for team members. Most IMCI municipalities, therefore, also have FHP. However, IMCI training coverage is increasing much more slowly than FHP coverage; many FHP municipalities, consequently, are still without IMCI.

The general objective of the Brazil MCE was to evaluate the impact of the IMCI strategy on child health, including child mortality, child nutrition, and family behaviour, as well as its associated costs, by comparing IMCI and non-IMCI municipalities.

The specific objectives were:

• To document IMCI implementation at the health-facility and community levels in a selected number of municipalities with strong IMCI programmes.

• Using existing data, to assess baseline health, socioeconomic, demographic and environmental characteristics in similar IMCI and non-IMCI municipalities.

• To compare child health care in first-level health facilities in municipalities with and without IMCI, as regards health worker performance, health systems support structure and intensity of community activities.

• To compare health-related behaviour, intervention coverage and impact indicators (mortality and nutrition) in municipalities with and without IMCI.

• To examine whether contextual factors (e.g., socioeconomic, environmental) could account for possible differences in impact indicators between IMCI and non-IMCI municipalities.

• To assess the costs associated with IMCI implementation and measure cost-effectiveness.

The originally proposed design was modified in the light of the stage of IMCI implementation. The actual design is mixed retrospective-prospective, since IMCI is already well implemented in many municipalities.

A municipal database was created to summarize existing data from censuses, MoH statistics, and indicators from UNICEF and UNDP. It has three main purposes:

• to allow the matching of IMCI and comparison municipalities according to baseline health and socioeconomic characteristics;

• to measure contextual factors (confounders) that may affect results interpretation;

• to use prospectively the MoH databases on utilization and hospital admissions to monitor intermediate outcomes of IMCI implementation.

A health facility survey was carried out for 2002 in IMCI and comparison municipalities, with the following objectives:

• to compare child health care in first-level health facilities in municipalities with and without IMCI, as regards health worker performance, health systems support structure, and intensity of community activities;

• to obtain information necessary for certifying that implementation had been adequate in the IMCI municipalities;

• to identify implementation weaknesses that might be corrected during the survey.

The survey included 58 municipalities with FHP (28 IMCI and 30 non-IMCI) and 48 facilities with IMCI and 48 without have been evaluated. Children studied numbered 653 (295 with IMCI and 358 without). All children aged two months to five years who were seen at the health facilities for the first time during an illness episode were included in the study. The survey used the standard MCE tool and provided information on the key and supplementary indicators, as well as on six integrated indices referring to child assessment, treatment and counselling, and readiness and capacity of facilities to deliver IMCI, manage severe illness and avoid unnecessary/harmful practices. Training in IMCI was associated with improved performance in many health worker tasks. A scientific paper summarizing these results has been published. 

Information on costs was collected at local, state and national level, as part of the Health Facility Survey. A Time and Motion Study was carried out in a sub-sample of the municipalities to provide essential information on personnel costs. These studies found that IMCI implementation was not associated with increased costs. 

High rates of staff turnover, along with a marked reduction in training activities due to reduced interest in IMCI at the MoH level, led to the finding that only 23 municipalities, out of over 400 identified in the states included in the evaluation, had more than 60% of their health workers trained in IMCI consistently over a three-year period. An analysis of mortality trends, using secondary data, did not show a significant reduction in mortality in municipalities that implemented IMCI, compared to similar municipalities without IMCI. 

Three scientific papers have been published and a final report was produced. Feedback is being regularly provided to national, state, and municipal health managers, and a home page in Portuguese is available with the main study findings. See publications for a complete list of reports, papers and a policy brief.