Field interviewer–administered dietary recalls in participants' homes: A feasibility study using the US Department of Agriculture’s Automated Multiple-Pass Method

Authors: DIGRANDE, LAURA - Rti International, Usa; PEDRAZZANI, SUE - Rti International, Usa; KINYARA, ELIZABETH - Rti International, Usa; HYMES, MELANIE - Food And Nutrition Service, USDA; KARNS, SHAWN - Rti International, Usa; Rhodes, Donna; Moshfegh, Alanna

Submitted to: RTI Press
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 4/16/2021
Publication Date: 5/17/2021
Citation: DiGrande, L., Pedrazzani, S., Kinyara, E., Hymes, M., Karns, S., Rhodes, D.G., Moshfegh, A.J. 2021. Field interviewer–administered dietary recalls in participants' homes: A feasibility study using the US Department of Agriculture’s Automated Multiple-Pass Method. RTI Press. https://doi.org/10.3768/rtipress.2021.mr.0045.2105.
DOI: https://doi.org/10.3768/rtipress.2021.mr.0045.2105

Interpretive Summary: Since 2002, USDA’s Automated Multiple-Pass Method (AMPM) has been the instrument used to collect 24-hour dietary recalls during the National Health and Nutrition Examination Survey (NHANES). The AMPM interviews are conducted when participants visit NHANES mobile examination centers (MECs) to complete physical exams. Although the MEC is a well-controlled environment, the National Center for Health Statistics (NCHS) has expressed interest in methods to mitigate declines in response rates and increases in operational costs. One option for the AMPM is to transition administration to NHANES home interviews which occur before MEC appointments and consist of self-report questionnaires administered by lay interviewers. The objective of this study was to test the feasibility of conducting AMPM dietary recalls by lay interviewers in the homes of 133 adult study participants residing in central North Carolina. The design included computer-assisted personal interviews led by a nutritionist or lay interviewer. Portion estimators tested were three-dimensional food models, a two-dimensional USDA Food Model Booklet, or a tablet with digital images rendered via augmented reality. Mean interview time was 40 minutes (range 13 to 90) with no difference by interviewer type or portion estimator, though timing for nutritionist-led interviews declined significantly over the study period. Forty-five percent of participants used home objects to facilitate recall. Values for the number of food items consumed and nutrients were within reasonably expected ranges regardless of interviewer type or portion estimator used. Average energy intake was 3,011 kcal for men and 2,105 kcal for women. In conclusion, the AMPM dietary recalls conducted in the home are feasible. Those administered by lay interviewers using a food model booklet produce nutrition data within expected limits. Home interviews for nutrition intake studies may be a sensible choice when response rates and cost are concerns. For large-scale studies and public surveillance programs, the home environment offers the potential of reduced cost and respondent burden, as well as increased efficiency of data collection and response rates.

Technical Abstract: The Automated Multiple-Pass Method (AMPM) is a widely used tool for collecting 24-hour dietary recalls. The five-step method, administered in an examination room during the National Health and Nutrition Examination Survey, provides data on food consumption patterns for What We Eat in America. The purpose of this study was to assess the feasibility of administering the AMPM in participants’ homes by lay interviewers. The design included computer-assisted personal interviews led by a nutritionist or lay interviewer. Portion estimators tested were three-dimensional food models, a two-dimensional food model booklet, or a tablet with digital images rendered via augmented reality. Among 133 interviews in residences of adult participants (55 males and 78 females) residing in central North Carolina, most took place in living rooms (52%) or kitchens (22%). The sample was mostly middle aged, non-Hispanic black, urban, with at least a high school diploma with no statistically significant differences were observed in age, race-ethnicity, education, or neighborhood density when men and women were compared. Mean interview time was 40 minutes (range 13 to 90) with no difference by interviewer type (1/3 by nutritionists, 2/3 by lay interviewers) or portion estimator, though timing for nutritionist-led interviews declined significantly over the study period. Forty-five percent of participants used home objects to facilitate recall. Values for the number of food items consumed and nutrients were within reasonably expected ranges regardless of interviewer type or portion estimator used. Average energy intake was 3,011 kcal for men and 2,105 kcal for women. In conclusion, the AMPM dietary recalls conducted in the home are feasible. Those administered by lay interviewers using a food model booklet produce nutrition data within expected limits. Home interviews for nutrition intake studies may be a sensible choice when response rates and cost are concerns.