Sociodemographic and Health Indicators of Diet Quality in Pre-Frail Older Adults in New Zealand.

Abstract

We collected detailed dietary intake from pre-frail (FRAIL scale 1-2) older adults using NZ Intake24, an online version of 24 h multiple pass dietary recall. Diet quality was ascertained with the Diet Quality Index-International (DQI-I). We used regression generalized linear models to determine predictors of diet quality as well as classification and regression tree (CART) analysis to examine the complex relationships between predictors and identified profiles of sub-groups of older adults that predict diet quality. The median age in this sample (n = 468) was 80.0 years (77.0-84.0). Living with others, a high deprivation index and a higher BMI were independent predictors of poorer diet quality. With CART analysis, we found that those with a BMI > 29 kg/m2, living with others and younger than 80 years were likely to have a lower diet quality. We found that BMI, living arrangement and socioeconomic status were independent predictors of diet quality in pre-frail older adults, with BMI being the most important variable in this sample when the interaction of these variables was considered. Future research is needed to determine the similarities and/or differences in the profile of subgroups of older adults with poorer diet quality.