Identifying children with neurological impairment and disability in resource-poor countries.

BACKGROUND: The burden of neurological impairment (NI) in children living in resource-poor countries (RPCs) is unknown. This lack of data is caused by inappropriate case detection techniques. In RPCs, the most appropriate method should be inexpensive, simple, rapid and accurate. This article reviews methods used to identify children with NI and disability in RPCs, evaluating their effectiveness and suitability. METHODS: A search of relevant articles was performed using the National Library of Medicine via PubMed and Medline search engines. In addition, bibliographies of reviews were also browsed to identify additional articles, particularly those from World Health Organization and United Nations sources and from government and unpublished reports. Key phrases used included impairment, disability or handicap and the following terms: identification, screening, prevalence and developing countries. Studies included were those that fulfilled the following criteria; performed in RPCs, presented data in detail to allow reanalysis and provided data on cost and validity of the methods. RESULTS: Use of the national census, key informants and methods using rapid rural appraisal have low sensitivity and are not able to provide adequate information on diagnostic categories or risk factors. House-to-house surveys using questionnaires have high sensitivities (63-100%) in the detection of impairment, but this approach remains relatively expensive and cannot be applied to an entire population (e.g. a region or country) and is thus less useful for assessing the needs of disability. Furthermore, the sensitivity is decreased in the detection of some domains, e.g. cognition. CONCLUSIONS: Most of the approaches used for identifying individuals with NI or disability suffer from inadequacies, the main ones being low sensitivity and underreporting. To assess the burden, nationwide censuses combined with surveys in selected areas of the country may be useful. These systems, however, require validation to establish their suitability