Accuracy and Cost-effectiveness of the INDDEX24 Dietary Assessment Platform in Viet Nam.

Technology-enabled approaches to conducting 24-hour dietary recalls (24HR) may reduce dietary assessment bottlenecks in low-resource settings. However, few studies have assessed their performance relative to conventional pen-and-paper interview (PAPI) approaches and none have validated performance against a benchmark (e.g. weighed food record [WFR]) in a low and middle-income country (LMIC). This study assessed relative accuracy and cost-effectiveness of INDDEX24, a technology-enabled approach to conducting 24HR, compared to a PAPI approach and against an observer WFR. Women aged 18-49 years from northern Viet Nam (n=234) were randomly assigned to be interviewed using INDDEX24 or PAPI 24HR following a WFR. The two one-sided T-test approach assessed the equivalence of each recall modality to the benchmark. Difference-in-differences analysis compared the recall-benchmark results across modalities. Cost per percentage point of accuracy for INDDEX24 and PAPI was derived from accuracy results and the cost to conduct the 24HRs. The PAPI and INDDEX24 24HRs were statistically equivalent to the WFR for all nutrients except vitamin A. INDDEX24 diverged significantly less than PAPI from the WFR for iron (0.9 v. -1.3 mg) and PAPI diverged less for protein (-3.7 v. 7.9 grams). At the individual level, 26% of PAPI and 32% of INDDEX24 respondents had energy intakes within +/- 10% of the WFR. INDDEX24 cost $111,004 and the PAPI $120,483 (USD 2019), making INDDEX24 more cost-effective across most indicators. INDDEX24 was an accurate and cost-effective method for assessing dietary intake in the study context and represents a preferred alternative to PAPI 24HRs in Viet Nam and other LMICs.