Community health workers: to train or to restrain? A longitudinal survey to assess the impact of training community health workers in the Bolama Region, Guinea-Bissau

BACKGROUND The shortage in human resources for health affects most dramatically developing countries which frequently use community health workers (CHW) as the basis for health programmes and services. The traditional definition refers CHWs as members of the community who are recruited and trained in health prevention and promotion to provide services within their community. In Guinea-Bissau, CHWs play a fundamental role in the diagnosis and treatment of childhood diarrheal diseases - one of the main health problems in the country. METHODS This study is based on 22 CHW, 79% of the total number of CHW in the Sanitary Region of Bolama. The main goal was to assess how training CHW on diarrheal diseases impacted the accuracy of the diagnosis and treatment of these diseases in children under the age of 5 years. Three evaluations were made throughout time - one evaluation before the training and two follow-up evaluations.An observation grid was developed to evaluate the identified signs, symptoms, diagnosis and treatments prescribed by the CHW in consultations to children with a suspicion of diarrhoeal disease. A similar grid was filled by a medical doctor who took the role of the external validation standard.Friedman's variance analysis and Cochran's Q test were performed to compare the accuracy depicted by CHWs in identifying items throughout time. A logistic regression model was also used to check the possible influence of socio-demographic characteristics of CHWs on the accuracy of the diagnosis and treatment prescribed by the CHW. RESULTS The results show that CHWs improve significantly their performance in identifying the correct diagnosis in the first follow-up moment after the training (P = 0.001, n = 22) but, 3 months later, the effectiveness decreases. No statistical evidence was found for the logistic regression models applied. This progressive loss of performance after training may occur because CHWs fail to apply treatment algorithms and guidelines over time. A limited set of socio-demographic characteristics of the CHWs can influence their performance and should not be disregarded when selecting CHW candidates. CONCLUSION The selection, supervision, support and continuous training of CHW are as important as the training provided.