Studies of children in developing countries. How soon can we prevent neurodisability in childhood?

Abstract

Neurodevelopmental disability is a functional limitation due to a neurological disorder with an onset early in life and occurs across a range of domains: cognition, movement, seizure disorders, vision, hearing, and behaviour. Disability can occur within a single domain or involve more than one domain in the case of multiple disabilities. Disabilities can be caused by a variety of etiological insults, many of which are preventable at a primary, secondary, or tertiary level. Neurodevelopmental disabilities are an important but largely unaddressed problem in low-income countries (Durkin et al. 1991). For this paper, the term 'developing countries' includes low-income countries with few professional and institutional health resources and inadequate infrastructure to deliver known effective preventive or rehabilitative mea-sures nationally. It is understood that higher-income coun-tries may also fail to address the challenges faced in preventing disability or in effective rehabilitation. Though many of the causes of neurodevelopmental dis-abilities in children in developing countries are identical to those in developed countries, there are causes which have an impact solely in developing countries such as cerebral malar-ia or trachoma. The prevalence of a disabilitywill increase or decrease in any given population depending on a variety of factors: for example, age distribution of the population, prevalence of the causal agents, availability of preventive measures or of initial treatment, and the longevity of those at risk. As a result, the epidemiology of neurodevelopmental disabilities show a markedly different pattern in developing countries than in the developed world. Estimating the prevalence of neurodisability It is not easy to obtain good estimates of the prevalence of disability in either developed or developing countries. For most disabilities, it is believed the rates found in epidemio-logical studies are higher in less developed countries (Institute of Medicine [IOM] 2001), though it is difficult to be sure because the study design and nature of the study sam-ples vary widely in the small number of epidemiological studies from developing countries that have been published. Many studies of prevalence of neurodisability in devel-oped countries depend on counting cases already recog-nized, since routine surveillance of development in children frequently occurs either through the health service or the educational system. This approach is not rigorous, but prag-matically it can provide rapid cost efficient prevalence esti-mates. Because medical services are often non-existent in developing countries, this approach is not feasible there. Disabilities may go unrecognized in developing countries. Therefore population-based studies are required. For instance, congenital rubella syndrome had not been report-ed in Ghana, but when research surveillance was put in place, a series of 18 cases were identified in one central city (Lawn et al. 2000). Many disabilities are age dependent, so that if studies do not clearly identify the age range of the sam-ple, the prevalence is difficult to interpret. For instance, in a study of blindness in Ethiopia, the prevalence of blindness was less than 1 per 1000 in those under 20 years of age and 175 per 1000 in those aged 70 or over (Zerihun and Mabey 1997). There is not universal agreement on how to identdy children with disabilities even in population-based surveys. The effectiveness of using the 'key informant' method to identify children with disabilities has been debated. For instance, a study of epilepsy in West Bengal found that a household survey identified fewer cases than using key infor-mants (Pal et al. 1998) while another study, of disability in Jamaican children (Thorburn et al. 1991), found the oppo-site, concluding that house-to-house surveys was the only reliable approach. A large study of learning disability* in chil-dren in China used multiple sources to identify children, hospital records, household survey, and records of develop-mental tests (Zuo et al. 1986). Since the quality of research in developing countries is quite variable and the methods of obtaining information dif-ferent, it is difficult to rely on the accuracy of prevalence esti-mates found. In a review of epidemiological studies of hearing loss around the world in 2000, Mencher reported country-specific rates of sensorineural hearing loss ranging from 1 to 14 per 100. It is not clear from the review if the dif-ferences result from the differences in definition and study design or are real. High-quality, low-cost comparative research will be critical to evaluating future interventions and to monitor progress towards decreasing the burden of childhood disability. Low-cost methods to get accurate estimates of prevalence and to investigate risk factors have been validated and used in many countries (Davidson 1994). These methods involve two-stage study designs, using a brief questionnaire adminis-tered by a community worker at the first stage and a clinical assessment of all children screening positive and a sample of those screening negative at a second stage (Stein et al. 1992). When adopting this approach and identical study designs with rigorous methodological approaches of tested validity and reliability (Durkin et al. 1994, 1995), differences in prevalence are likely to indicate real differences in the popu-lations studied. For instance, the prevalence of seizures in Bangladesh and Jamaica was markedly lower than that found in Pakistan using identical research methods (Durkin et al. 1992) where we found a history of epilepsy in 16 of 1000 chil-dren aged 2 to 9 years. The lower rate in Bangladesh (6.5 of 1000) may have been due to a higher mortality rate in