Lora Sabin1, Maxton Tsoka, Mohamad I Brooks, Candace Miller. 1. Center for Global Health and Development, Department of International Health, School of Public Health, Boston Unviersity, Boston, Massachusetts 02118, USA. Candace@bu.edu
Abstract
OBJECTIVES: To validate the Child Status Index (CSI) as an instrument that can meaningfully measure the vulnerabilities of orphaned and vulnerable children, including those infected and affected by HIV/AIDS. METHODS: Two age-specific instruments, comprised of previously validated tools and indicators commonly considered best practice, were administered to 102 children aged 5-10 years and 100 children aged 11-17 years in Mchinji, Malawi. Respondents were randomly sampled from a roster of children recently scored with the CSI. For each of the CSI's 12 subdomains, we assessed construct validity using Spearman Rank correlation coefficients. We also calculated cross tabulations to explain the resulting correlation coefficients. Analyses were conducted separately for the 2 age groups. RESULTS: No relationships exceeded the standard for high construct validity (≥0.7). Only 2 were moderate (0.3-0.7), both for the younger age group: food security (0.4) and wellness (0.36). All other relationships were weak or negative. In most subcategories, a substantial proportion of surveyed children indicated distress that was not evident from CSI scores. In the abuse and exploitation subdomain, all children were rated as "good" or "fair" by the CSI, but among surveyed children aged 11-17, 20% or more reported being beaten, kicked, locked out of the house, threatened with abandonment, cursed, and made to feel ashamed. CONCLUSIONS: In this rural Malawi population, we were not able to validate the CSI as a tool for assessing the vulnerabilities of orphaned and vulnerable children. We recommend caution in interpreting CSI scores and revisions to the tool before global scale-up in its use.
OBJECTIVES: To validate the Child Status Index (CSI) as an instrument that can meaningfully measure the vulnerabilities of orphaned and vulnerable children, including those infected and affected by HIV/AIDS. METHODS: Two age-specific instruments, comprised of previously validated tools and indicators commonly considered best practice, were administered to 102 children aged 5-10 years and 100 children aged 11-17 years in Mchinji, Malawi. Respondents were randomly sampled from a roster of children recently scored with the CSI. For each of the CSI's 12 subdomains, we assessed construct validity using Spearman Rank correlation coefficients. We also calculated cross tabulations to explain the resulting correlation coefficients. Analyses were conducted separately for the 2 age groups. RESULTS: No relationships exceeded the standard for high construct validity (≥0.7). Only 2 were moderate (0.3-0.7), both for the younger age group: food security (0.4) and wellness (0.36). All other relationships were weak or negative. In most subcategories, a substantial proportion of surveyed children indicated distress that was not evident from CSI scores. In the abuse and exploitation subdomain, all children were rated as "good" or "fair" by the CSI, but among surveyed children aged 11-17, 20% or more reported being beaten, kicked, locked out of the house, threatened with abandonment, cursed, and made to feel ashamed. CONCLUSIONS: In this rural Malawi population, we were not able to validate the CSI as a tool for assessing the vulnerabilities of orphaned and vulnerable children. We recommend caution in interpreting CSI scores and revisions to the tool before global scale-up in its use.
Authors: L Sherr; A Macedo; L D Cluver; F Meinck; S Skeen; I S Hensels; L T S Sherr; K J Roberts; M Tomlinson Journal: Health Psychol Behav Med Date: 2017-01-30