Social Protection and Health Division, Inter-American Development Bank, Washington, D.C., United States of America.
Faculty of Population Health Sciences, Institute of Child Health, University College London, London, United Kingdom.
PLoS One. 2020 Apr 29;15(4):e0231317. doi: 10.1371/journal.pone.0231317. eCollection 2020.
There is increasing global commitment to establish early childhood interventions that promote the development of the millions of disadvantaged children in low- and middle-income countries not reaching their developmental potential. However, progress is hindered by the lack of valid developmental tests feasible for use at large scale. Consequently, there is an urgent need for such tests. Whilst screeners and single-domain tests ('short tests') are used as alternatives, their predictive validity in these circumstances is unknown. A longitudinal study in Bogota, Colombia began in 2011 when 1,311 children ages 6-42 months were given the Bayley Scales of Infant and Toddler Development (Bayley-III) by psychologists and randomized to receive one of two batteries of short tests under survey conditions. Concurrent validity of the short tests with the Bayley-III ('gold standard') was reported. In 2016, at 6-8 years, 940 of these children were given tests of IQ (Wechsler Intelligence Scale for Children, WISC-V) and school achievement (arithmetic, reading, and vocabulary) by psychologists. We compared the ability of the short tests, the Family Care Indicators (FCI), height-for-age, stunting (median height-for-age <-2 SD), and the Bayley-III to predict IQ and achievement in middle childhood. Predictive validity increased with age for all tests, and cognition and language were usually the highest scales. At 6-18 months, all tests had trivial predictive ability. Thereafter, the Bayley-III had the highest predictive validity, but the Denver Developmental Screening Test was the most feasible and valid short test and could be used with little validity loss compared with the Bayley-III. The MacArthur-Bates Communicative Development Inventory at 19-30 months and the FCI under 31 months predicted IQ and school achievement as well as the Bayley-III. The FCI had higher predictive validity than stunting and height-for-age, and could be added to stunting for use as a population-based indicator of child development.
全球越来越重视开展儿童早期干预，以促进为数百万未能充分发挥自身发展潜力的中低收入国家弱势儿童的发展。然而，由于缺乏可行的大规模使用的有效发育测试，进展受到阻碍。因此，迫切需要这类测试。虽然筛查测试和单领域测试（“短测试”）被用作替代方法，但在这种情况下，它们的预测效度尚不清楚。哥伦比亚波哥大的一项纵向研究于 2011 年开始，当时对 1311 名 6-42 个月大的儿童进行了贝利婴幼儿发育量表（Bayley-III）测试，心理学家将他们随机分配接受两种短测试中的一种，在调查条件下进行。报告了短测试与 Bayley-III（“金标准”）的同时效度。2016 年，在 6-8 岁时，其中 940 名儿童由心理学家进行了智商测试（韦氏儿童智力量表，WISC-V）和学校成就（算术、阅读和词汇）测试。我们比较了短测试、家庭护理指标（FCI）、身高年龄别（身高年龄别中位数<-2SD）和 Bayley-III 预测儿童中期智商和学业成绩的能力。所有测试的预测效度都随年龄增长而提高，认知和语言通常是最高的量表。在 6-18 个月时，所有测试的预测能力都微不足道。此后，Bayley-III 具有最高的预测效度，但丹佛发育筛查测试是最可行和最有效的短测试，与 Bayley-III 相比，其有效性损失较小，可以使用。19-30 个月时的麦克阿瑟- Bates 交际发展量表和 31 个月以下的 FCI 可以预测智商和学业成绩，与 Bayley-III 一样。FCI 的预测效度高于发育迟缓的身高年龄别和身高年龄别，可与发育迟缓结合用于作为基于人群的儿童发展指标。
