Print version ISSN 1415-790X
Rev. bras. epidemiol. vol.5 n.1 São Paulo Apr. 2002
Evaluating the impact of healt interventions
Cesar G. Victora
Programa de Pós-Graduação em Epidemiologia Universidade Federal de Pelotas, RS, Brasil
Are our preventive and/or curative actions in health improving the health of the population? The question may seem obvious to many people, but there are many reasons why the answer may be no. An efficacious drug, for example, may be prescribed wrongly, a vaccine may not be stored correctly and may lose its immunogenic properties, a micronutrient may be distributed to individuals who do not need supplementation, or an educational campaign may fail to change risk behavior. In addition to proving the efficacy of a specific intervention through experimental studies performed under ideal conditions, it is also essential to prove that interventions are effective in decreasing morbi-mortality under routine conditions in health services1. For all these reasons, it is essential that epidemiologists and other health professionals continually study whether interventions in health are attaining the expected impact.
Evaluations in the field of health may be divided into two large categories process evaluation and impact evaluation2. The former investigate whether the intervention or service is being offered to the population adequately, if the quality is appropriate, if the population is really using the service, and if the coverage attained is adequate. This kind of evaluation is frequently performed in our country, with many examples of excellent quality. On the other hand, evaluation of the impact of interventions or services on morbi-mortality, or on direct health-related behavior is relatively rare. That is why the authors of the III Master Plan for the Development of Epidemiology in Brazil 2000-2004, who met in Brasília in August 20003, identified the need to organize a national seminar on impact evaluation. The Epidemiology Graduate Program of UFPEL was nominated by ABRASCO to coordinate the activity.
The preparation for the seminar began through systematic research into impact studies performed in Brazil during the past 10 years. The review, performed using the LILACS and MEDLINE database, was supplemented by consulting ABRASCO's Epidemiology Committee, leading to the proposals of seminar participants. The seminar was carried out in Pelotas, from the 6th to the 8th of November, 2001, with the presence of 25 Brazilian researchers and of Prof. Jean Pierre Habicht, from Cornell University (USA), a world authority on the issue.
It was a highly productive seminar. Eighteen impact evaluations and three studies of sources of data for evaluation were presented. Most impact studies evaluated health and infant nutrition programs: six dealt with promotion of breastfeeding, four with nutritional status, three with diarrhea, one was on anemia and another on neonatal mortality. The remaining studies discussed Chagas' Disease, Schistosomiasis and Diabetes Mellitus. The high standard of the methodology of studies became evident, as was their suitability for the characteristics of health programs and available data. Various approaches were used: randomized community assays, case-control studies, cohort studies, cross-sectional, and ecologic studies. Regarding the terminology proposed by Habicht et al.2, the various studies that were presented allowed inferences on adequacy (descriptive studies), plausibility (observational studies) and probability (experimental studies).
On the other hand, few studies were identified during the literature review, especially studies of the major health problems of our population, such as chronic-degenerative conditions and external causes. The lack of studies of the cost-effectiveness of interventions was also noted. Finally, the need to plan evaluation during the initial stages of new programs in order to allow for baseline studies and therefore prospective evaluations was pointed out.
Health service and intervention evaluations are increasingly important and publishing this special number of Revista Brasileira de Epidemiologia aims to join the different methodologies used in the area in order to stimulate Brazilian researchers to undertake new studies.
1. Black N. Why we need observational studies to evaluate the effectiveness of health care. Br Med J 1996; 312: 1215-8.
2. Habicht JP, Victora CG, Vaughan JP. Evaluation designs for adequacy, plausibility and probability of Public Health programme performance and impact. Int J Epidemiol 1999; 28: 10-8.
3. ABRASCO. do III Plano Diretor para o Desenvolvimento da Epidemiologia no Brasil 2000-2004. Rio de Janeiro; 2000.