Development of an instrument to assess nutritional risk factors for children infected with human immunodeficiency virus.

OBJECTIVE: To produce a simple and effective instrument to evaluate and monitor the nutritional risk of children infected with the human immunodeficiency virus (HIV). DESIGN: The test instrument was developed in consultation with 5 physicians, 5 nutritionists, and 5 social workers with expertise in caring for HIV-infected children. Patient information was collected through medical record review for 19 sociodemographic, 10 anthropometric, 4 biochemical, 6 dietary intake, and 19 medical factors. As a part of routine nutrition care, anthropometric data were obtained and the caregiver was asked to complete a 3-day diet record. Also recorded were the most recent CD4+ T-cell numbers and serum HIV p24 antigen and plasma HIV RNA levels. SUBJECTS/SETTING: Thirty-nine HIV-infected children were selected using quota sampling; that is, subjects were stratified by clinical class as defined by the Centers for Disease Control and Prevention. STATISTICAL ANALYSIS: The severity or degree of potential nutritional risk in each section (anthropometric, biochemical, dietary intake, and medical data) was graded (0 to 4, 0 = low risk) and summed. Reliability of internal consistency was determined through covariance matrixes. Validity was determined through Pearson product moment correlation coefficients to measure convergent and divergent validity; predictive validity was determined using analysis of variance. Correlation for validity was compared to 6 selected dependent variables: weight for height, weight growth velocity, lean body mass, serum albumin level, CD4+ T-cell numbers, and quantitative plasma HIV RNA levels. RESULTS: Of the 38 factors that were analyzed for reliability, 11 fell in the strongly reliable range: height for age, weight for age, clinical class, somatic protein stores, mid-arm circumference, weight for height, serum albumin, immunologic status, body mass index, energy intake, and opportunistic infection. CONCLUSIONS: Anthropometric, dietary intake, and medical data were reliable indicators of nutritional risk. The entire instrument was reliable after 8 of the weakest items were removed. The instrument was found to be valid and a good predictor of nutritional risk in HIV-infected children.