(P02-021-23) Food Patterns and High Proportion of Obesity in Tuvalu: A Latent Class Analysis in a Pacific Island Country

Objectives: To investigate the association between food patterns and the proportion of obesity in a low-middle-income Pacific island country. It was hypothesized that the obesity trend in Tuvalu, currently 73% obesity and 22% morbid obesity among our adult study population in 2022, is positively associated with an imported food pattern compared to a mixed consumption of culturally conventional foods.

Methods: Our study population in 2022 included 985 adults with complete data. The COMmunity-based Behavior and Attitude survey in Tuvalu (COMBAT) survey collected the frequency of consumption of 25 common foods. Each food item response had five categories, which were recoded into 3 levels: everyday/≥1 per week, ≥1 per month, and < 1 per month/never. A latent class analysis determined four food patterns. A Bayesian multilevel logistic regression estimated the association between food patterns and the proportion of obesity (BMI ≥30kg/m²) and morbid obesity (BMI ≥40 kg/m²), adjusting for age, gender, physical activity, smoking status, chronic disease prevalence, and education level, accounting for clustering by region.

Results: The entropy of the four food classes was 0.94 with an average posterior probability of class assignment for each individual of 0.97. These foods patterns were described as Local: local-moderate diversity (proportion of individuals=28%); Mixed: local and imported-moderate diversity (17%); Limited: local-limited diversity (29%) and Imported: imported-highly diverse (26%). Compared to the Mixed pattern, those classified with an Imported food pattern had 2.52 times greater odds of obesity [95% credible interval (CrI): 1.59 to 3.99]; Limited pattern odds ratio (OR)=1.89 [95%CrI:1.59-3.99]), and Local pattern OR=1.54 [95%CrI:0.94 to 2.50]. Furthermore, using the same reference, the odds of morbid obesity in the Imported pattern were 2.65 times greater [95%CrI: 1.53 to 4.77]; Limited pattern OR=1.77 [95%CrI: 0.99 to 3.24]), and Local pattern OR=2.06 [95%CrI:1.16 to 3.79].

Conclusions: This COMBAT study identified four food patterns and found a positive association of a predominantly imported food pattern with obesity and morbid obesity, compared to a mixed food pattern. Our findings provide insight for future nutrition and food system strategies to combat obesity.

Funding Sources: The Taiwan ICDF and the Tuvalu government