Premium
Comparing the FGI‐10R Dietary Diversity Scale and Quantitative Food Intake: A Study in Urban Bangladesh
Author(s) -
Waid Jillian L.,
Ali Masum,
Akter Rina,
Thilsted Shakuntala Haraksingh
Publication year - 2016
Publication title -
the faseb journal
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.709
H-Index - 277
eISSN - 1530-6860
pISSN - 0892-6638
DOI - 10.1096/fasebj.30.1_supplement.1153.13
Subject(s) - respondent , statistics , diversity (politics) , demography , sample (material) , environmental health , cohen's kappa , medicine , mathematics , chemistry , chromatography , sociology , political science , anthropology , law
Objectives The FAO has just released a universal indicator (FGI‐10R) for women's individual dietary diversity, which was developed from analysis of secondary quantitative food intake data data from multiple countries. The study validates the measure theoretically by deriving the dietary diversity measure from quantitative dietary assessment and not from dietary diversity modules as they are administered in the field. However it has not yet been verified that the two very different modules will elicit the same response. Methods A dietary diversity questionnaire (DDQ) was added to a sub‐sample of households in quantitative food assessment (QFA) survey administered among a panel of households in poorer neighborhoods of Dhaka City Corporation. The DDQ included an estimate of the amount of the food type eaten (greater or less than 15 grams during the day). Information from the respondent was collected through free ‐recall followed by list based probes and coded into a 21‐item questionnaire by the enumerator. The DDQ was asked to the respondent before the QFA. We use kappa and indicator validation statistics to compare the performance of indicators directly as measured in the DDQ and as derived from the QFA across 139 women included in the first round of dietary assessment. Once the three rounds of data collection are complete in January 2016, we will update results with the annual sample (estimated to be approximately 450 women) and also compare the DDQ performance to the mean probability of adequacy of individuals based on the QFA. Results Agreement between the two types of measures was moderate and nearly the same when both the QFA measure and DDQ measure were restricted to observations of consumption above 15 grams (k=55%) and when the QFA measure was restricted but the DDQ measure was not (k=54%). The F1 score of these two modules is also similar at 76% for the restricted DDQ and 73% when the DDQ measure was not restricted to 15 grams. Among food groups, the greatest agreement was for dark green leafy vegetables (k=80% for both comparisons) and worst for other fruits as the QFA measure indicated that all respondents had consumed more than 15 grams but only around three quarters of individuals were recorded as consuming these items using the DDQ (k=0%). Conclusions This analysis indicates that there is only a limited improvement of the dietary diversity instrument when information on the amount of food consumed is included. This indicates that individuals do not report trace amounts of many items they consume when asked a traditional dietary diversity module. Support or Funding Information Data was collected and analysis undertaken as part of the project “Aquaculture and the Poor: improving fish production, consumption and nutrition linkages”, conducted in partnership with WorldFish and funded by the Deutsche Gesellschaft für Internationale Zusammenarbeit (GIZ).