How many children really have severe acute malnutrition?
Public Health Nutrition Editorial Highlight: ‘Using cross-sectional surveys to estimate the number of severely malnourished children needing to be enrolled in specific treatment programmes’. Authors: Dale, N. M., Myatt, M., Prudhon, C., & Briend, A.
The burden of malnutrition has become a topic of considerable attention since the development and implementation of highly effective and high coverage methods of treating severe acute malnutrition in the early years of this century. International agencies have tried to quantify actual figures of malnourished children, globally, to understand the magnitude of the problem and to estimate the impact of nutritional interventions that have been conducted in attempts to combat this potentially deadly problem.
To plan for nutritional interventions for children with severe acute malnutrition, it is necessary to estimate the number of children that will require treatment. This is essential for accurate planning of personnel as well as logistic supplies and funding.
The use of prevalence surveys in conjunction with population estimates can give an estimate of the number of cases existing at the time of the survey but does not provide information on the subsequent incidence cases. Many health managers use a prevalence to incidence conversion factor to help estimate the number of incident cases.
International agencies and government bodies continue to use the current recommendation of a standard conversion factor of 1.6 that was derived from a study based on data collected more than 20 years ago. There has been very little work done to determine the validity of this value until recently. The questions at stake have been highlighted by UNICEF and are simply: what exactly is the most accurate value for this conversion factor? and is it a constant value or does it vary among regions?
The aim of our study was to estimate the incidence conversion factor empirically in different contexts. The way we did this was by collecting survey and nutrition program data from international agencies who have opened or continued to operate nutrition programs because of prevalence survey results. Collection of data proved to be challenging but we managed to obtain 24 data sets from a variety of different regions throughout the world and attempted to correlate the expected number of children to be treated based on the prevalence calculated from surveys and population estimates, and the number of children treated after the surveys took place. We used a working assumption that if using a constant conversion factor to convert prevalence into caseload is correct then we would see a high correlation among these 2 variables and the slope of the regression line forced to pass through the origin would provide the best estimate of the conversion factor.
Our results showed a statistically significant relationship was found between the 2 variables (Pearson’s r = 0.48 (95% CI 0.12; 0.72)) with a value of the slope being 2.17 (95% CI = 1.33; 3.79). When considering the possible variation that inevitably occurs with coverage of programs, we found a corrected slope range from 2.8-11.21 using 100% and 38% coverage (a recent global average) respectively. Our results did show regional variation however the number of datasets used was too small to test for significance. Other recent studies that have shown similar range of values from 4.3 to 9.5 with significant variation among regions
Our results, viewed in conjunction with others, highlight the complexity of estimating the burden of severe acute malnutrition and the need for further work to be done in this area to fully understand the impact of one of the most devastating conditions facing children in the world today.
The paper,‘Using cross-sectional surveys to estimate the number of severely malnourished children needing to be enrolled in specific treatment programmes‘ is published in the journal Public Health Nutrition and is freely available until 30 February 2017.