Dietary adherence (adapted Mediterranean diet score) significantly improved with culturally integrated nutrition counseling but worsened with usual DSME

Description #

In this pilot cluster RCT of Cantonese-speaking Chinese American (CA) diabetes patients (n=18 baseline; 15 at follow-up), dietary adherence measured by an adapted Mediterranean diet score (16 items, range 0–80) improved among participants who received the integrative nutritional counseling curriculum (DSME+INC) but worsened among participants who received usual diabetes self-management education (DSME). At baseline the scores were similar (control 54.3, INC 55.6), but at 6 months the control group declined by 5.5 points while the INC group rose by 4.3 points, a significant group–time interaction (z=4.26, p<0.001) (Table 3). This was the study's only significant behavioral outcome.

"Food frequency, as measured by the adapted Mediterranean Diet score, improved among participants in DSME+INC, but worsened for DSME participants (significant group–time interaction z=4.26, p<0.001)." (Ho, 2020)

"Mediterranean diet*54.3±3.3−5.5±1.855.6±2.64.3±1.4" (Ho, 2020)

"*p<0.01 for within-group changes from baseline to 6 months, and group–time interaction based on linear mixed model." (Ho, 2020)

Methods Context #

What? #ⓘ

The observable: dietary adherence operationalized as a food-frequency score on a culturally adapted Mediterranean diet scale (16 items, possible range 0–80), measured at baseline and 6-month follow-up.

"Participants were asked about dietary adherence through two validated measures of food frequency: Starting the Conversation (8 items, possible range 0–16) and a culturally sensitive adaption of the Mediterranean diet score (16 items, possible range 0–80). … We adapted the scale by including Chinese-specific foods, such as dumplings, fried dim sum, and sweet boba drinks." (Ho, 2020)

How? #ⓘ

Pilot two-arm cluster RCT; DSME classes randomized to usual nutrition curriculum (ADA-based) vs. INC curriculum (ADA + Chinese medicine principles, cotaught by a nurse educator and a licensed acupuncturist). Linear mixed models with random effects for persons nested within clusters and fixed effects of group, time, and group–time interaction tested change in adherence from baseline to 6 months. See ART - Integrative Nutritional Counseling (INC) curriculum.

"Linear mixed models (LMMs) were used to test significance of change in means in dietary self-efficacy, dietary adherence, weight, HbA1c, and other outcomes of interest. LMMs included random effects for persons nested within clusters and fixed effects of treatment group (DSME vs. DSME+INC), time, and group–time interaction." (Ho, 2020)

Who? #ⓘ

18 fluent Cantonese-speaking/reading Chinese Americans ≥21 with type 2 diabetes or prediabetes enrolled in DSME classes at the Chinatown Public Health Center, San Francisco (7 control, 11 intervention); predominantly female (83%), older (mean age 62), low income (100% <$50,000); 15 completed 6-month follow-up (6 control, 9 intervention).

"Participants were Cantonese-speaking patients with diabetes, 83% female, 100% with income < $50,000, with an average age of 62 years, and slight preference for western medicine over Chinese medicine." (Ho, 2020)

Other Notes #

The companion food-frequency measure (Starting the Conversation, 0–16) showed no significant group difference (control −0.6, INC −0.8 at 6 months) (Table 3).

Caveats #

• Very small pilot sample with cluster (not individual) randomization limits power and generalizability of the integrative counseling findings Only 18 patients were enrolled (7 control, 11 intervention) and 15 completed 6-month follow-up, so the trial was a feasibility pilot underpowered to detect most differences. Randomization was by course (cluster) rather than individual, further reducing power. Because participants were recruited blind to assignment, it is unknown whether self-selection differed by interest in Chinese-medicine content. These constraints mean the single significant adherence finding, and the nonsignificant clinical trends, should be treated as preliminary and not generalized beyond this Cantonese-speaking safety-net population.