Global Database on Child Growth and Malnutrition


Methods and standardized data presentation

The information included in the WHO Global Database on Child Growth and Malnutrition complies with the following standardized format:

  • systematic use of the NCHS/WHO international reference population (1);
  • display of growth retardation prevalence for under-5-year-olds, as measured by the proportion of weight-for-age, height-for-age and weight-for-height below -2 and -3 standard deviations (SDs) (Z-scores);
  • display of the prevalence of overweight, as measured by the proportion of children with weight-for-height above +2 Z-scores;
  • display of Z-score means and SDs for the three indices; and
  • stratification of the results according to age, sex, region, and rural/urban strata.

The required criteria for entering surveys in the database are:

  • A clearly defined population-based sampling frame, permitting inferences to be drawn about an entire population;
  • A probabilistic sampling procedure involving at least 400 children (allowing for an estimation of prevalence with a random error of £5% at a confidence level of 95%);
  • Use of appropriate equipment and standard measurement techniques (1);
  • Presentation of results as Z-scores in relation to the NCHS/WHO reference population.
  • For those surveys where results are presented using a different classification system, reference population, or prevalence cut-offs, the principal investigators are contacted and encouraged to re-analyze their data sets following WHO standardized presentation or, otherwise, to provide the raw data to the WHO Programme of Nutrition for re-analysis. Survey results are systematically checked for inconsistencies and these are brought to the attention of the investigators, with a request for clarification. A hard copy of the survey documentation, together with any corrigendum or additional item of information received from the authors is filed under the survey reference number. The aim is to keep the database as fully documented and comprehensive as possible, so that queries concerning compiled data can be answered quickly.