Differences in dietary acceptability, restraint, disinhibition, and hunger among African American participants randomized to either a vegan or omnivorous soul food diet

Abstract

The Nutritious Eating with Soul study was a 24-month, randomized behavioral nutrition intervention among African American adults. This present study, which is a secondary analysis of the NEW Soul study, examined changes in dietary acceptability, restraint, disinhibition, and hunger. Participants (n=159; 79% female, 74% with ≥ college degree, mean age 48.4 y) were randomized to either a soul food vegan (n=77) or soul food omnivorous (n=82) diet and participated in a two-year behavioral nutrition intervention. Questionnaires assessing dietary acceptability (Food Acceptability Questionnaire; FAQ) and dietary restraint, disinhibition, and hunger (Three-Factor Eating Questionnaire; TFEQ) were completed at baseline, 3, 6, 12, and 24 months. Mixed models were specified with main effects (group and time) and interaction effects (group by time) to estimate mean differences in FAQ and TFEQ scores using intent-to-treat analysis. After adjusting for employment, education, food security status, sex, and age, there were no differences in any of the FAQ items, total FAQ score, dietary restraint, disinhibition, and hunger at any timepoint except for one item of the FAQ at 12 months. Participants in the vegan group reported a greater increase in satisfaction after eating a meal than the omnivorous group (mean difference 0.80±0.32, 95% CI 0.18, 1.42; P=0.01). This is one of the first studies to examine differences in dietary acceptability, hunger, and other eating factors among African American adults randomized to either a vegan or omnivorous soul food diet. The findings highlight that plant-based eating styles are equally acceptable to omnivorous eating patterns and have similar changes in hunger, restraint, and disinhibition. These results suggest that plant-based eating styles can be an acceptable dietary pattern to recommend for cardiovascular disease prevention and may result in greater post-meal satisfaction.

1. Introduction

Little research has investigated the acceptability of therapeutic diets in the context of a randomized nutrition intervention among African American adults. This is particularly true when examining the use of plant-based diets for reducing cardiovascular disease risk factors. Several studies have highlighted important reasons for not adopting plant-based or vegan diets that can be addressed. These reasons include perceived stigma from friends and family for not eating meat (Markowski & Roxburgh, 2019), a perception that a vegan diet will not be acceptable in terms of taste and convenience (Fehér et al., 2020), enjoyment of meat and difficulties in preparation of plant-based meals (Pohjolainen Pasi et al., 2015), and lack of self-efficacy for adopting a healthy diet (Okpara et al., 2022).

Previous randomized trials have found similar dietary acceptability among participants randomized to a vegan diet or other therapeutic diets (Barnard et al., 2000; Barnard, Gloede, et al., 2009; Katcher et al., 2010); however, these studies have been mostly conducted in white populations. There has been limited work examining the acceptability of therapeutic diets that aim to reduce cardiovascular risk factors among African American adults. One study that examined ways to improve dietary acceptability for plant-based diets among African American adults found that acceptability can be increased through partnering with local plant-based soul food restaurants (Crimarco et al., 2020). Adherence and acceptability of therapeutic diets may differ among racial/ethnic groups due to differences in several factors, such as cultural preferences (Satia, 2009), food availability (Bennett et al., 2022), or taste preferences (James, 2004).

In addition to dietary acceptability, few studies have examined if there are differences in changes in dietary restraint, disinhibition, and hunger among African American participants randomized to either vegan or other therapeutic diets. Previous randomized studies have found no difference in dietary restraint, disinhibition, and hunger among participants randomized to vegan or other therapeutic diets (Moore et al., 2015) or have found increases in dietary restraint among a low-fat diet with no change in the vegan group (Barnard et al., 2004), but these studies have mostly been conducted among white populations.

Dietary acceptability, restraint, disinhibition, and hunger are important components to assess during dietary interventions because they can impact the adoption of plant-based and other dietary patterns. For example, hunger during a dietary intervention inhibits dietary adoption (O’Connor et al., 2022) and weight loss maintenance (Pasman et al., 1999). Components that make up dietary acceptability, such as time needed for preparation, cost of foods, or taste, are also commonly cited as inhibiting the adoption of healthy diets (James, 2004; Kearney & McElhone, 1999). Therefore, studying if therapeutic diets differentially affect dietary adoption is important to inform the development and recommendations of therapeutic diets for cardiovascular disease prevention.

Several factors can influence the adherence and acceptability of plant-based diets. People who follow a vegan diet have indicated that having an ethical, animal welfare, or health reason for following the diet as motivational, whereas feeling the diet is inconvenient or has a lack of available options as barriers (Cruwys et al., 2020; North et al., 2021; Williams et al., 2023). Barriers to other types of healthy diets can be similar, such as lack of cooking skills, social pressure to eat differently, lack of time, and wanting to get the most value for food for the money (Loor et al., 2023; Macdiarmid et al., 2013). While inconvenience or limited availability of food options are often cited as barriers, hunger, feeling too restrained when eating, or feeling out of control when eating are not mentioned frequently in the literature around following a plant-based diet.

The Nutritious Eating with Soul (NEW Soul) study was a 24-month, randomized behavioral nutrition intervention among African American adults (G. M. Turner-McGrievy et al., 2023). The NEW Soul study compared a vegan and omnivorous diet, both emphasizing soul food cuisine (Crimarco et al., 2019; Miller, 2013). The goal of the main trial was to examine the differential impacts of adopting each diet on risk factors for cardiovascular disease.

The goal of the present study was to conduct a secondary analysis of the NEW Soul study to examine if changes in dietary acceptability, restraint, disinhibition, and hunger differed among African American participants randomized to a vegan vs. omnivorous soul food diet in the NEW Soul study. This research is important because there is a belief among consumers that plant-based diets are less satisfying than omnivorous meals (Vural et al., 2023) and health care providers may be reluctant to recommend pant-based diets due to perceived difficulties with acceptability (Tuso et al., 2013). Examining long-term dietary acceptability, restraint, disinhibition, and hunger among African American adults in the context of a two-year randomized behavioral nutrition intervention can help provide insight into how barriers to dietary acceptability are similar or different from other therapeutic dietary approaches and is an understudied area. We hypothesized that there would be no differences between the groups and that both groups would see increases in dietary acceptability.

2. Methods

The NEW Soul intervention, inclusion criteria, measures, recruitment strategies, and main outcomes have been described elsewhere (G. Turner-McGrievy, 2023; G. Turner-McGrievy et al., 2020; G. M. Turner-McGrievy et al., 2021, 2023). Briefly, NEW Soul was a randomized two-year intervention among African American participants with overweight or obesity (18–65 years, BMI between 25–49.9 kg/m²). Participants were recruited in two cohorts separated by one year starting in 2018. In relation to the present analysis, participants completed questionnaire items assessing dietary acceptability, restraint, disinhibition, and hunger at baseline and at 3, 6, 12, and 24 months.

2.1. Measures

Dietary acceptability was assessed using the Food Acceptability Questionnaire (FAQ) (Barnard et al., 2000). The FAQ is a 10-item questionnaire assessing acceptability with consuming a diet, assessing items like ease of preparation, liking of foods eaten, and satisfaction after eating. Items are scored from 1 to 7. A composite score was calculated so the total FAQ score ranged from 10 (lowest acceptability) to 70 (highest acceptability). The FAQ has been used in prior studies (Barnard, Cohen, et al., 2009; Katcher et al., 2010) and has demonstrated test-retest reliability of 0.70 to 1.00 (Barnard et al., 2000).

Dietary restraint, disinhibition, and hunger were assessed using the Three-Factor Eating Questionnaire (TFEQ). The TFEQ consists of 36 items (scored either as 1 for present or 0 as absent) with three factors that measure dietary restraint (scoring from 0–21), disinhibition (scoring from 0–16), and hunger (scoring from 0–14) (Stunkard & Messick, 1985). Dietary restraint measures the ability or self-control to restrict food and adhere to a diet, disinhibition measures the lack of restraint or uncontrollable eating, and the hunger factor measures the feeling of being or staying hungry. A higher score in each factor indicates a higher level of susceptibility in that category (Stunkard & Messick, 1985). The TFEQ has demonstrated validity in prior studies (Karlsson et al., 2000). In addition, the TFEQ has demonstrated to work equally well among both white and Black respondents (Atlas et al., 2002) and those respondents with a normal, overweight, or obese weight status (Bohrer et al., 2015).

2.2. Dietary intervention

Participants in the two-year study were randomized to follow a vegan or low-fat omnivorous diet, both focusing on soul food cuisine. Participants attended weekly classes for six months, bi-weekly classes for six months, and monthly classes for the last year of the study. Classes consisted of a nutrition topic, cooking demonstration, and time for participants to share successes and challenges and set SMART goals related to following their diet, managing stress and engaging in physical activity. Due to the COVID-19 pandemic, content after year one for cohort 1 was delivered remotely via Zoom and content after six months was delivered remotely for cohort 2.

The dietary recommendations provided to each group have been described elsewhere (G. Turner-McGrievy et al., 2020). Participants in the vegan group were asked to consume whole plant foods (fruits, vegetables, legumes, grains, nuts, and seeds) and avoid all animal products. Dietary recommendations for the omnivorous group were based on the Therapeutic Lifestyle Recommendations (NHLBI. Your Guide to Lowering Cholesterol with Therapeutic Lifestyle Changes (TLC). NIH Publication No. 06–5235., 2005). Participants were asked to limit red meat and other sources of saturated fat, and favor fruits, vegetables, whole grains, legumes, poultry, fish, egg whites, and low-fat dairy. Neither group was given an energy limit or asked to self-monitor their intake. Participants could eat until they were satiated; however, the omnivorous group did have limitations on some foods (no more than five ounces of lean meat per day and no more than two egg yolks per week). To ensure cultural relevancy of the diets, the study partnered with local soul food chefs and restaurant owners to provide cooking demonstrations. Participants also suggested entrees they wanted to have demonstrated in future classes. Sample one-day menus for each diet group, showing typical entrees recommended and demonstrated in the NEW Soul classes, are provided in Table 1. The study was approved by a university Institutional Review Board, and all participants provided written consent.

Table 1:

Sample one-day menus for the vegan and omnivorous groups of the NEW Soul study showing example meals that were demonstrated in class and recommended for consumption

Meal	Vegan group	Omnivorous group
Breakfast	Whole-wheat toast with peanut butter	Whole-wheat toast with light butter
	Scrambled tofu with nutritional yeast	Scrambled egg whites with reduce fat cheddar cheese
	Sliced bananas	Sliced bananas
	Coffee or tea	Coffee or tea
Lunch	Veggie chili	Turkey chili
	Cornbread	Cornbread
	Salad	Salad
Dinner	Hoppin’ John with brown rice, peppers, and onions	Cajun catfish with brown rice, okra, and tomatoes
	Collard greens	Collard greens
	Nice cream (ice cream made from frozen bananas)	Low-fat frozen yogurt
Snack	Apple slices and cashews	Apple slices and string cheese

2.3. Statistical analysis

Sample size calculations have been described elsewhere (G. Turner-McGrievy et al., 2020) and were conducted to detect differences in weight loss and LDL cholesterol between groups at 12 months. The present study represents a secondary analysis of the NEW Soul main trial. Changes in FAQ and TFEQ scores were compared between diet groups as well as across time points within each group. Mixed models were used to obtain least-square means adjusted for covariates. We specified the mixed models with main effects (group and time) and interaction effect (group by time) controlling for employment, education, food security status, sex, and age. Next, we compared differences in least-square means for FAQ and TFEQ within groups as well as between diet groups. The change values from baseline to each assessment time point were tested within groups. These change values were contrasted across diet groups at each time point. To test the hypothesis whether the differences in least-square means were different from zero, we conducted independent t-tests with Satterthwaite approximation. PROC MIXED procedure in SAS 9.4 (SAS Institute) analyzed the models through maximum likelihood estimation. In addition, the analytic plan was pre-specified and all analyses are clearly identified and have been discussed appropriately.

3. Results

Results, including demographics and dietary adherence, have been reported elsewhere (G. M. Turner-McGrievy et al., 2023). There were 159 participants in the NEW Soul study who were randomized to either a vegan diet (n=77) or omnivorous diet (n=82). Participants were mostly female (79%), educated (74% with a college degree or greater), employed (74%), middle-aged (mean 48.4 y), and food secure (86%). The FAQ and TFEQ questionnaires were completed by all participants at baseline, by 142 (89%) at 3 months, 122 (77%) at 6 months, 109 (69%) at 12 months, and 93 (58%) participants at 24 months. Dietary adherence (assessed via three 24-hour recalls at each timepoint) was 41% (n=32) vegan and 24% (n=20) omnivorous participants at three months, 30% (n=23) vegan and 22% (n=18) omnivorous at six months, 35% (n=27) vegan and 26% (n=21) omnivorous at 12 months, and 29% (n=22) vegan and 24% (n=20) omni at 24 months (G. M. Turner-McGrievy et al., 2023).

Analyses of main outcomes examined group by time interactions for dietary acceptability, restraint, disinhibition, and hunger (Table 2). As hypothesized, there were no differences in changes in total dietary acceptability (least squares mean change in total FAQ score at 24 months: vegan 1.81±1.67 (−1.47, 5.09), omnivorous −1.25±1.59 (−4.38, 1.88); P=0.19) or in the TFEQ factors of dietary restraint (least squares mean change at 24 months: vegan 0.85±1.00 (−1.12, 2.81), omnivorous 2.09±0.95 (0.22, 3.96), P=0.37), disinhibition (least squares mean change at 24 months: vegan −1.70±0.48 (−2.63, −0.77), omnivorous −1.69±0.45 (−2.58, −0.80); P=0.99) or hunger (least squares mean change at 24 months: vegan −1.87±0.48 (−2.81, −0.93), omnivorous −1.98±0.46 (−2.88, −1.08; P=0.87) across groups. The pattern of change for overall dietary acceptability and dietary restraint, disinhibition, and hunger are shown in Figures 1 and 2 respectively. There were also no differences over time in changes in any of the individual items of the FAQ except satisfaction after eating at 12 months (Table 2, Figure 2). Vegan participants reported greater increase in satisfaction after eating meals than the omnivorous group (least squares mean difference of 0.80±0.32, 95%CI 0.18, 1.42; P=0.01).

Table 2:

Baseline values and changes from baseline in all the Food Acceptability questions and overall score and in the three domains of the Three-Factor Eating Questionnaire presented as adjusted means ±standard error (Confidence Intervals) ^a

	Vegan N=77	Omnivorous N=82	Difference between groups and P value
Food Acceptability Questionnaire
Overall Total Food Acceptability Score range 10 for low acceptability to 70 high acceptability)
Baseline	47.39±1.65 (44.14, 50.64)	49.15±1.65 (45.90, 52.40)
3-mo	−1.31±1.45 (−4.15, 1.53)	−2.77±1.38 (−5.48, −0.05) b	1.46±2.00 (−2.47, 5.39) P=0.47
6-mo	−0.34±1.50 (−3.29, 2.61)	−3.53±1.47 (−6.42, −0.63) b	3.18±2.10(−0.94, 7.31) P=0.13
12-mo	0.0008±1.55 (−3.04, 3.04)	−2.72±1.54 (−5.75, 0.31)	2.72±2.18 (−1.57, 7.01) P=0.21
24-mo	1.81±1.67 (−1.47, 5.09)	−1.25±1.59 (−4.38, 1.88)	3.06±2.307(−1.47, 7.59) P=0.19
Individual items of the Food Acceptability Questionnaire (items 1–10)
How well do you like these foods? (1 not at all, 7 extremely)
Baseline	5.01±0.21 (4.6, 5.42)	5.19±0.21 (4.78, 5.60)
3-mo	−0.36±0.20(−0.75, 0.032)	−0.42±0.19 (−0.79, −0.038) b	0.05±0.28 (−0.49, 0.59) P=0.85
6-mo	0.09±0.21 (−0.32, 0.49)	−0.43±0.20 (−0.83, −0.04) b	0.52±0.29 (−0.05, 1.08) P=0.07
12-mo	0.09±0.21 (−0.33, 0.51)	−0.21±0.22 (−0.63, 0.21)	0.30±0.30 (−0.29, 0.89) P=0.32
24-mo	0.23±0.23 (−0.22, 0.68)	−0.19±0.22 (−0.62, 0.24)	0.43±0.32 (−0.19, 1.05) P=0.18
How well do you like the taste of these foods? (1 not at all, 7 extremely)
Baseline	5.14±0.21 (4.72, 5.56)	5.31±0.21 (4.89, 5.73)
3-mo	−0.58±0.19 (−0.97, −0.20) b	−0.62±0.19 (−0.99, −0.26) b	0.036±0.27 (−0.49, 0.56) P=0.89
6-mo	−0.23±0.20 (−0.63, 0.17)	−0.62±0.20 (−1.01, −0.23) b	0.39±0.28 (−0.17, 0.94) P=0.17
12-mo	−0.20±0.21 (−0.63, 0.21)	−0.46±0.21 (−0.87, −0.05) b	0.26±0.29 (−0.32, 0.83) P=0.38
24-mo	0.002±0.22 (−0.44, 0.44)	−0.48±0.21 (−0.89, −0.05) b	0.48±0.31 (−0.13, 1.08) P=0.13
How appealing or unappealing do you find the appearance of these foods? (1 extremely unappealing, 7 extremely appealing)
Baseline	5.20±0.22 (4.77, 5.63)	5.15±0.22 (4.72, 5.58)
3-mo	−0.65±0.2 (−1.05, −0.25) b	−0.48±0.19 (−0.87, −0.10) b	−0.16±0.28 (−0.71, 0.39) P=0.56
6-mo	−0.29±0.21 (−0.7, 0.13)	−0.57±0.21 (−0.98, −0.17) b	0.29±0.30 (−0.29, 0.86) P=0.33
12-mo	−0.17±0.22 (−0.60, 0.25)	−0.28±0.22 (−0.70, 0.15)	0.11±0.31 (−0.50, 0.71) P=0.73
24-mo	0.07±0.23 (−0.39, 0.53)	−0.19±0.22 (−0.63, 0.25)	0.26±0.32 (−0.38, 0.89) P=0.43
How boring are these foods? (Item is reverse scored) (1 not at all, 7 extremely)
Baseline	4.90±0.24 (4.42, 5.36)	5.08±0.24 (4.62, 5.55)
3-mo	−0.23±0.26 (−0.74, 0.28)	−0.12±0.25 (−0.60, 0.37)	−0.11±0.36 (−0.81, 0.59) P=0.75
6-mo	−0.31±0.27 (−0.83, 0.21)	−0.37±0.26 (−0.88, 0.14)	0.06±0.37 (−0.67, 0.79) P=0.87
12-mo	−0.38±0.27 (−0.92, 0.15)	−0.44±0.27 (−0.98, 0.09)	0.060±0.39 (−0.7, 0.82) P=0.88
24-mo	−0.06±0.30 (−0.64, 0.52)	−0.37±0.28 (−0.92, 0.19)	0.31±0.41 (−0.49, 1.11) P=0.45
How easy or difficult has it been for you to prepare these foods? (1 extremely difficult, 7 extremely easy)
Baseline	4.73±0.25 (4.24, 5.21)	4.79±0.25 (4.30, 5.28)
3-mo	−0.52±0.22 (−0.96, −0.09) b	−0.75±0.21 (−1.16, −0.33) b	0.22±0.30 (−0.37, 0.82) P=0.46
6-mo	−0.47±0.23 (−0.92, −0.02) b	−0.89±0.224(−1.33, −0.45) b	0.42±0.32 (−0.20, 1.05) P=0.19
12-mo	−0.56±0.24 (−1.02, −0.10) b	−0.54±0.23 (−1.00, −0.081) b	−0.02±0.33 (−0.67, 0.63) P=0.96
24-mo	−0.28±0.25 (−0.78, 0.22)	−0.19±0.24 (−0.66, 0.29)	−0.10±0.35 (−0.79, 0.59) P=0.78
How easy or difficult has it been for you to purchase these foods? (1 extremely difficult, 7 extremely easy)
Baseline	5.04±0.24 (4.57, 5.51)	5.10±0.24 (4.63, 5.57)
3-mo	−0.38±0.21 (−0.79, 0.034)	−0.47±0.20 (−0.86, −0.07) b	0.09±0.29 (−0.49, 0.66) P=0.76
6-mo	−0.33±0.22 (−0.76, 0.098)	−0.75±0.21 (−1.17, −0.33) b	0.42±0.30 (−0.18, 1.02) P=0.17
12-mo	−0.20±0.23 (−0.64, 0.25)	−0.48±0.22 (−0.92, −0.04) b	0.29±0.32 (−0.34, 0.91) P=0.37
24-mo	0.17±0.24 (−0.31, 0.65)	0.02±0.23 (−0.44, 0.48)	0.15±0.34 (−0.51, 0.81) P=0.65
How easy or difficult has it been for you to maintain your current diet at restaurants? (1 extremely difficult, 7 extremely easy)
Baseline	4.17±0.25 (3.67, 4.67)	4.30±0.25 (3.80, 4.80)
3-mo	−0.02±0.25 (−0.51, 0.47)	−0.34±0.24 (−0.81, 0.13)	0.32±0.34 (−0.35, 0.99) P=0.35
6-mo	−0.13±0.26 (−0.63, 0.37)	−0.01±0.25 (−0.50, 0.47)	−0.12±0.36 (−0.83, 0.58) P=0.74
12-mo	−0.14±0.26 (−0.66, 0.38)	−0.07±0.26 (−0.59, 0.44)	−0.07±0.37 (−0.80, 0.67) P=0.86
24-mo	0.20±0.29 (−0.36, 0.76)	−0.04±0.27 (−0.57, 0.50)	0.24±0.39 (−0.54, 1.01) P=0.55
How much effort does it take for you to stay on this diet? (1 more effort than is possible, 7 no effort at all)
Baseline	4.46±0.23 (4.01, 4.91)	4.62±0.23(4.17, 5.07)
3-mo	−0.06±0.24 (−0.53, 0.40)	−0.51±0.26 (−0.96, −0.07) b	0.45±0.33 (−0.19, 1.09) P=0.17
6-mo	−0.12±0.24 (−0.60, 0.36)	−0.52±0.24 (−0.99, −0.05) b	0.40±0.34 (−0.27, 1.07) P=0.24
12-mo	−0.15±0.25 (−0.65, 0.34)	−0.58±0.25 (−1.08, −0.09) b	0.43±0.35 (−0.27, 1.13) P=0.23
24-mo	−0.002±0.27 (−0.53, 0.53)	−0.7±0.26 (−1.21, −0.19) b	0.70±0.37 (−0.04, 1.43) P=0.06
How satisfied or dissatisfied do you feel after eating a meal on this diet? (1 extremely dissatisfied, 7 extremely satisfied)
Baseline	4.56±0.22 (4.13, 4.99)	5.04±0.22 (4.60, 5.47)
3-mo	0.58±0.21 (0.17, 0.99) b	0.08±0.21 (−0.31, 0.48)	0.50±0.29 (−0.07, 1.07) P=0.09
6-mo	0.48±0.22 (0.05, 0.91) b	−0.005±0.21 (−0.42, 0.41)	0.48±0.30 (−0.12, 1.08) P=0.11
12-mo	0.72±0.22 (0.28, 1.16) b	−0.08±0.22 (−0.52, 0.36)	0.80±0.32 (0.18, 1.42) P=0.01
24-mo	0.56±0.24 (0.09, 1.03) b	0.15±0.23 (−0.31, 0.60)	0.41±0.33 (−0.24, 1.07) P=0.22
Overall, how satisfied or dissatisfied are you with the diet? (1 extremely dissatisfied, 7 extremely satisfied)
Baseline	4.19±0.23 (3.73, 4.65)	4.62±0.23 (4.16, 5.08)
3-mo	0.92±0.21 (0.50, 1.34) b	0.86±0.21 (0.46, 1.26) b	0.06±0.30 (−0.52, 0.64) P=0.84
6-mo	0.97±0.22 (0.53, 1.40) b	0.71±0.22 (0.28, 1.14) b	0.26±0.31 (−0.35, 0.87) P=0.40
12-mo	1.01±0.23 (0.55, 1.46) b	0.47±0.23 (0.01, 0.91) b	0.54±0.32 (−0.10, 1.18) P=0.0958
24-mo	0.96±0.25 (0.47, 1.45) b	0.76±0.24 (0.29, 1.22) b	0.20±0.34 (−0.47, 0.87) P=0.56
TFEQ: Restraint, Disinhibition, and Hunger
Dietary Restraint (range 0 for low restraint to 21 for high restraint)
Baseline	8.99±0.85 (7.31, 10.66)	8.33±0.85 (6.66, 10.00)
3-mo	12.01±0.87 (10.30, 13.72) b	12.77±0.83 (11.14, 14.40) b	−0.76±1.20 (−3.12, 1.61) P=0.53
6-mo	0.25±0.90 (−1.52, 2.02)	1.95±0.88 (0.21, 3.68) b	−1.7±1.26 (−4.17, 0.78) P=0.18
12-mo	1.97±0.93 (0.15, 3.79) b	2.74±0.92 (0.92, 4.55) b	−0.77±1.31 (−3.34, 1.79) P=0.55
24-mo	0.85±1.00 (−1.12, 2.81)	2.09±0.95 (0.22, 3.96) b	−1.24±1.38 (−3.95, 1.47) P=0.37
Disinhibition (range from 0 for low disinhibition to 16 for high disinhibition)
Baseline	5.33±0.54 (4.27, 6.39)	5.54±0.54 (4.48, 6.61)
3-mo	1.90±0.41 (1.09, 2.70) b	1.85±0.39 (1.07, 2.62) b	0.05±0.57 (−1.07, 1.17) P=0.93
6-mo	−1.21±0.43 (−2.05, −0.37) b	−1.54±0.42 (−2.36, −0.72) b	0.33±0.60 (−0.84, 1.50) P=0.58
12-mo	−1.41±0.44 (−2.28, −0.55) b	−1.14±0.44 (−2.00, −0.28) b	−0.27±0.62 (−1.49, 0.95) P=0.66
24-mo	−1.70±0.48 (−2.63, −0.77) b	−1.69±0.45 (−2.58, −0.80) b	−0.01±0.66 (−1.30, 1.29) P=0.99
Hunger (range 0 for low hunger to 14 for high hunger)
Baseline	5.20±0.47 (4.28, 6.12)	5.12±0.47 (4.20, 6.04)
3-mo	2.16±0.42 (1.34, 2.98) b	2.58±0.40 (1.79, 3.36) b	−0.42±0.57 (−1.54, 0.71) P=0.47
6-mo	−1.61±0.43 (−2.45, −0.76) b	−1.68±0.42 (−2.52, −0.85) b	0.08±0.60 (−1.11, 1.26) P=0.90
12-mo	−1.79±0.44 (−2.66, 0.91) b	−1.31±0.44 (−2.18, −0.44) b	−0.48±0.63 (−1.71, 0.75) P=0.45
24-mo	−1.87±0.48 (−2.81, −0.93) b	−1.98±0.46 (−2.88, −1.08) b	0.11±0.66 (−1.19, 1.41) P=0.87

^aAll models were adjusted for baseline socioeconomic status (education and employment), food security status, sex, and age. Within-group baseline data presented as baseline adjusted means ± standard error (Confidence Intervals). Within-group 3-, 6-, 12-, and 24-month data presented as adjusted mean change from baseline ± standard error (Confidence Intervals).

^bP<0.05 for difference in within-group change from baseline.

Fig 1.

Change in total food acceptability score and satisfaction after eating a meal over 24 months between vegen and omnivorous participants.

Fig 2.

Change in dietary restraint, disinhibition, and hunger over 24 months between vegan and omnivorous participants.

Although the primary aim was to examine group by time changes in outcomes, within-group changes were also examined to help elucidate patterns in change within each group and help to inform future interventions. Examining within-group changes, the omnivorous group reported within-group decreases in how much they liked the foods they were eating at three and six months. Vegan participants indicated a within-group decrease in liking the taste and the appearance of the foods, but only at three months, whereas the omnivorous group reported a decrease in liking the taste at all timepoints and the appearance at three and six months. Both groups reported within-group increases in difficulty preparing the foods on their diet pattern at 3, 6, and 12 months, whereas only the omnivorous group reported within-group increases in difficulty purchasing foods for their diet pattern at the same timepoints. At all timepoints, only the omnivorous group reported within-group increases in effort to maintain their diet and only the vegan group reported within-group increases in satisfaction after eating a meal. Overall, both groups reported within-group increases in satisfaction with the diet they were assigned to at all timepoints. There were no changes in overall dietary acceptability (total FAQ score) among vegan participants with within-group decreases in total FAQ score at 3 and 6 months among omnivorous participants. There were no changes in either group for boredom with eating the foods or difficulty when dining out. There were within-group changes for all three TFEQ constructs at all timepoints for both groups with the exception of six-month within-group changes in dietary restraint among vegan participants. Overall directions of TFEQ factors point to increases in scores for dietary restraint, disinhibition, and hunger at 3 months and lowering of scores after that.

4. Discussion

This secondary analysis of the NEW Soul study found no group by time differences in any of the FAQ items, total FAQ score, dietary restraint, disinhibition, and hunger at any timepoint apart from a greater increase in satisfaction after eating a meal in the vegan as compared to the omnivorous group. This is one of the first studies to examine differences in acceptability, hunger, and other eating factors among African American adults randomized to either a vegan or omnivorous soul food diet to reduce cardiovascular disease risk factors. The findings highlight that plant-based eating styles have similar levels of dietary acceptability as compared to omnivorous eating patterns and also have similar changes in hunger, restraint, and disinhibition. These results suggest that plant-based eating styles can be a potential dietary pattern to recommend for cardiovascular disease prevention and may lead to greater satisfaction after a meal is consumed.

Although not different between groups, the direction of the findings was the opposite of the hypothesized direction that food acceptability would increase across all timepoints in both groups. Total dietary acceptability (FAQ score) decreased within the omnivorous group at three and six months and the vegan group was the only group that had higher scores than baseline at 24 months. In addition, the pattern of the findings examining dietary restraint, disinhibition, and hunger (TFEQ) indicate that both groups had greater challenges with their assigned diet early on at three months and then adjusted to the diets as time went on. Changes in dietary restraint, disinhibition, and hunger were all the highest at three months compared to the later timepoints.

Several previous studies have demonstrated equal dietary acceptability between participants randomized to follow a vegan or other therapeutic diet. For example, a study examined differences in dietary acceptability among participants randomized to a vegan diet, one of three plant-based diets (vegetarian, pesco-vegetarian, or semi-vegetarian), or an omnivorous diet. That study, which also used the FAQ, found no differences in dietary acceptability among the five diet patterns (Moore et al., 2015). Other studies have found similar acceptability between vegan and other therapeutic diets except for food preparation. Previous studies have found that participants in the vegan groups felt their diets were less easy to prepare than participants randomized to either a low-fat diet (Barnard et al., 2004), a diet focusing on the American Diabetes Association guidelines (Barnard, Gloede, et al., 2009), or compared to their usual intake (Barnard et al., 2000). These studies, however, have mostly been conducted among white populations. One study that examined a short-term plant-based diet intervention among African American participants randomized participants to either a vegan diet or vegan diet plus vouchers to a local vegan soul food restaurant. While both groups in that study followed a vegan diet, the findings point to the positive affect that partnering with a local restaurant can have on dietary acceptability, with an increase in perception of the availability of plant-based foods in the restaurant group (Crimarco et al., 2020). The present study did not find any differences between groups in difficulty with meal preparation. The NEW Soul study is one of the first long-term behavioral intervention studies to find differences between vegan and omnivorous participants in satisfaction after a meal. Higher satisfaction post-meal among vegan participants may have been due to differences in protein sources between the groups. Previous work has shown that meals containing plant-based proteins (e.g., legumes) are more satiating than those containing animal-based proteins (Klementova et al., 2019; Kristensen et al., 2016; Muhlhausler et al., 2022). In addition, people consuming vegan diets tend to consume more fiber than individuals consuming diets containing animal products (Davey et al., 2003) and high fiber diets tend to be more satiating than low fiber diets (Slavin, 2005). Perception of the satiating and satisfying effects of plant-based meals by consumers, however, is often in the opposite direction with individuals thinking that plant-based meals would be less filling and pleasant than animal-based meals (Vural et al., 2023).

Other studies have also examined dietary restraint, disinhibition, and hunger between participants randomized to vegan or other therapeutic diets. These studies have found conflicting results from the TFEQ with one study finding no differences in dietary restraint, disinhibition, and hunger between participants randomized to a vegan or other therapeutic diets (Moore et al., 2015) and others finding no differences in disinhibition and hunger but finding increases in dietary restraint among participants randomized to a low-fat diet (Barnard et al., 2004) or a diet following American Diabetes Association guidelines (Barnard, Gloede, et al., 2009). The increases in dietary restraint seen with these two studies may be due to the type of comparison diet used in these interventions. While the Moore et al. (2015) study compared five diets with no portion size or energy limits, the participants in the comparison groups of the other studies required some tracking of portion sizes, such as limiting servings of meat or fat intake, which may have led to the increases in dietary restraint (Barnard et al., 2004; Barnard, Gloede, et al., 2009). Participants randomized to the vegan diet groups in all three studies were told to eat an ad libitum plant-based diet but reduce fat intake (Barnard et al., 2004; Barnard, Gloede, et al., 2009; Moore et al., 2015). Our study also utilized a low-fat ad libitum vegan diet as compared to the omnivorous diet that required some restrictions in portion sizes, particularly for meat and egg yolks. We found no differences between groups with any of the TFEQ constructs but demonstrated a pattern of larger changes early on at three months among the TFEQ constructs. The study is one of the first studies to examine these outcomes in the context of adopting a plant-based diet among African American adults.

A focus of the NEW Soul study was to ensure that recipes were culturally relevant for both diet groups by partnering with local soul food chefs and restaurant owners. This cultural tailoring may have been why there were minimal differences between the groups. Previous studies have highlighted the importance of culturally tailoring dietary interventions for reducing chronic disease (Winham, 2009). For example, a dietary intervention found that African American participants were not as adherent to a DASH diet intervention as their white counterparts and the authors highlighted the need to culturally tailor diet patterns (Epstein et al., 2012). Cultural tailoring may have helped to provide higher acceptability among participants in both groups of the NEW Soul study.

The present study has several strengths including the use of validated questionnaires and measures across a two-year intervention. In addition, the study adds to the literature on dietary acceptability, restraint, disinhibition, and hunger among African American adults randomized to follow either a vegan or omnivorous soul food diet. The study’s findings should also be interpreted within the context of the limitations. Participants were mostly female and highly educated, limiting the generalizability of the findings. All items were self-reported and physiological measures for items, such as assessing ghrelin levels for hunger (Matu et al., 2018), were not obtained. In addition, while the TFEQ has been assessed for use in African American populations, the FAQ has not been validated among African American adults. Lastly, as a secondary analysis of the parent study, this study was not powered to detect differences in TFEQ or FAQ. Therefore, it is possible the study was underpowered to detect differences.

5. Conclusion

The NEW Soul study demonstrates that a plant-based soul food diet had equal dietary acceptability, restraint, disinhibition, and hunger as compared to an omnivorous soul food diet among African American adults participating in a two-year, behavioral nutrition intervention. In addition, the pattern of changes highlights the importance of longer-term interventions beyond three months to ensure participants become accustomed to their new diet pattern. The findings suggest that a culturally tailored vegan diet is well accepted and has the potential to be utilized as a therapeutic dietary pattern for cardiovascular disease risk reduction among African American adults.

Abbreviations

NEW Soul

Nutritious Eating with Soul

BMI

Body Mass Index

FAQ

Food Acceptability Questionnaire

TFEQ

Three-Factor Eating Questionnaire