Background: Despite emerging evidence that a whole food plant-based diet (WFPBD) can reduce risk for CVD morbidity and mortality, adoption rates in the US remain low (~4%). Little is known about why this disconnect exists, and understanding this is critical to designing and implementing interventions intended to increase WFPBD consumption in the US. However, no validated survey instruments intended for use in US adults has been tested. Objective: The purpose of this study is to validate the WFPBD survey instrument using confirmatory factor analysis (CFA). Methods: Survey development was based on previous research in Australia and Europe, and items include “pros” and “cons” of consuming a WFPBD using constructs from the Social Cognitive Theory and the Transtheoretical Model. Sixty-four statements (33 “con” and 31 “pro”) were presented with five-item Likert scale answer options (strongly disagree to strongly agree). The WFPBD survey was administered to 412 English speaking adults in the US via MTurk. Summary statistics and CFA were used to describe outcomes and reduce the number of survey items while optimizing explained variance. Examination of item variance and covariance was used to iteratively fit a second order twelve factor model, a first order eight factor model, and a first order six factor model. Results: The mean age of participants was 34.9 years, and the sample was 12% (48) African American, 6% (25) Asian, 6% (25) Hispanic, 1% (5) Other, and 75% (309) White. The majority (85.7% (349)) had at least a Bachelor’s degree, and 33% (136) made >$65K. After adjustment for low factor loadings and high cross loadings, the first order six factor model remained significant ( X 2 = 491.87, p < 0.001). This model resulted in the best fit, as indicated by low RMSEA = 0.051 and SRMR = 0.048. The six factors were split between “cons” (social support, instrumental support, self-efficacy, health), and “pros” (instrumental support, health). The largest factor loadings for the latent variable WFPBD “cons” were “I wouldn't get enough energy or strength.” (0.85) and “I am not confident that I know how to prepare plant-based meals.” (0.81), while for the latent variable WFPBD “pros”, the highest factor loadings were for “Have a better quality of life.“ (0.77) and “Be fit.” (0.77). Attitudes towards animals and environmental concerns were not highly implicated and were excluded from the final model. Conclusion: The 25-item WFPBD survey produced internally consistent measures of the latent variables WFPBD “pros” and WFPBD “cons” in this sample of US adults. The reduced survey can be used to assess attitudes and beliefs towards a WFPBD and results can guide behavioral interventions development to promote adoption of a WFPBD. Results suggest the importance of reducing negative perceptions of a WFPBD related to energy and strength, increasing self-efficacy for preparing plant-based meals, and emphasizing health benefits.
Read full abstract