Appropriate Use of Linear Growth Measures to Assess Impact of Interventions on Child Development and Catch-Up Growth.

Linear growth is increasingly used as the sole or primary outcome for evaluating interventions, but impact is often not seen. Sometimes there is interest in whether children catch up to where they otherwise would have been had detrimental conditions not occurred, but the literature is confusing because of claims for evidence of catch-up growth based on inappropriate methods. This article examines the use of linear growth measures to evaluate intervention impact and catch-up. Focus on linear growth as a measure of impact has resulted in a lack of demonstrable intervention effects, evaluations that do not use measures responsive to nutrition-sensitive and integrated interventions, insufficient emphasis on adolescent girls and women before conception and children after the first 1000 d, and insufficient investment in developing other measures. Nutrition interventions may benefit children but may not discernibly affect linear growth deficits in immediate or intermediate periods. Interventions intended to affect one domain may have positive or negative impacts on others. Child nutrition and growth are part of early childhood development; the focus should be on improving conditions in which children grow and develop throughout childhood and adolescence because poor conditions in a population affect all children. To assess the impact of nutrition interventions, it is important to use a broad set of measures and indicators of outcomes and immediate and underlying determinants. The 4 criteria for demonstrating catch-up in growth, which are much more stringent than those for demonstrating intervention impact, require demonstration of the following: 1) an inhibiting condition and 2) reduced velocity during 1 period, 3) followed by alleviation of or compensation for the inhibiting condition, and 4) higher-than-normal velocity during a subsequent period. Partial catch-up growth is sometimes possible when constraints are removed, but whether it is beneficial to the child is not clear. Research is needed to develop, refine, and validate measures and indicators for the purposes intended.