Methodological Lessons Learned From Conducting a Population-Based Phone Survey on Nutrition Practices in the Kyrgyz Republic

Abstract

ObjectivesTo share lessons from a survey on nutrition practices, where the methodology shifted from in-person to phone interviews due to the COVID-19 pandemic. MethodsThe USAID Advancing Nutrition Project is conducting an impact evaluation in Kyrgyz Republic to determine the effect of project interventions on nutrition behaviors. Municipalities in Batken and Jalalabad regions were randomly assigned to intervention or comparison areas.The stepped wedge design includes baseline, midpoint, and endline surveys, carried out at 12-month intervals. The baseline survey was completed in late 2020. Difference in differences analysis and analysis of project exposure variables will be used to determine impact. Due to COVID-19, all interviews were carried out using computer assisted telephone interviewing (CATI). ResultsWe interviewed 2,091 women with children under two years of age, with phone numbers obtained from health facilities in survey areas. The response rate (11.6% completed interviews) was similar or higher than typical phone surveys. Interviews included complex questions related to 11 nutrition practices, including women’s and children’s dietary recall. Average interview length was 31 minutes. 4% of respondents dropped before completing the call. Advantages of phone interviews included reduced travel costs, improved quality control, and real-time data review. Disadvantages included high non-response, inability to observe items such as handwashing stations, and potential biases related to phone ownership. ConclusionsOur experience suggests that it is possible to carry out a complex, population-based nutrition survey using CATI. Response rates may be improved by building rapport at the start of the call. Interview length can be reduced using block randomization. Baseline levels of most indicators were similar to previous in-person surveys in Kyrgyzstan, suggesting CATI survey results can potentially provide accurate estimates of nutrition behaviors. Funding SourcesUnited States Agency for International Development.