NEUROLOGICAL EXAMINATION FOR CHILDREN - RELIABILITY AND UTILITY IN STUDIES OF HIV-INFECTION

A valid and reliable measure of the neurologic signs associated with H IV infection in children is needed to quantitate disease progression a nd to evaluate efficacy of treatments. We are developing a structured Neurological Examination for Children (NEC) for this application. This paper reports the results of a test-retest reliability study of items for children under 2 years. The NEC included 95 items to measure head and fontanel size, vision and eye movements, tone, reflexes, power, c erebellar function, symmetry, and adventitious movements. The sample c omprised 31 children aged 6-47 months; 22 were 24 months or younger, a nd 13 were infected with HIV. Twenty-eight items had good-to-excellent reliability (Kappa or intraclass correlation coefficients > 0.60). Me dian reliability coefficients per domain were: head size and fontanel, 0.93; tone, 0.66; reflexes, 0.58; and power, 0.81. Items with reliabi lity coefficients less than 0.60 (n = 15) were revised (n = 6), delete d (n = 6) or retained (n = 3) for further evaluation. For the remainin g 39 items, we could not estimate reliability because either all child ren exhibited a single response and agreement between examiners was pe rfect or disagreement involved only 1 or 2 cases. Items to assess seve ral of the signs most characteristic of HIV infection had good-to-exce llent reliability. We anticipate that, with training and practice, sim ilar results can be obtained in other settings.