Abstract
Data collection of diet quality is important to estimate global dietary transitions affecting public health. Mobile-phone based tools can provide a low-cost and rapidly deployable modality to complement enumerator-collected data. This study validated self-administration of the Diet Quality Questionnaire using mobile phones, comparing accuracy against enumerator administration, measuring both against an observed benchmark. Quantitative dietary intake data were gathered from 308 participants in northwest Rwanda, using a weighed food record. Intake data were used to calculated 'observed' responses to the questionnaire, half the participants responding to enumerators, and the other half using a mobile-administered version of the questionnaire. After filtering for low quality data, agreement in observed and reported responses, for all questionnaire questions, were statistically compared. Agreement rates (observed-reported) of self-administered and enumerated responses were high (91% vs 95%, p = 0.05), respectively. Agreement was significantly lower for the mobile-administered modality among older and lower income respondents (by about 5 percentage points), with no significant differences by gender or time of response. Mobile-administration cost approximately USD 0.70 per response, versus the marginal cost of USD 0.79 for the enumerator-administered. Our results confirm self-administered reporting by mobile-phone as a valid, low-cost method for collecting dietary data, with only marginal (yet significant for some subgroups) differences in agreement rates, compared to enumerated data. Data collection by mobile phone represents an agile complement to enumerated collection, administrable in the absence of existing survey platforms; and provides a useful option for high-frequency data collection to monitor dietary dynamics in target sub-populations.