Making large changes or small changes to prevent weight gain in young adulthood: which is preferred and by whom?

Abstract

Knowledge of participant treatment preferences can inform decision-making regarding treatment dissemination and future participant adoption. To compare participant perceptions of two evidence-based approaches for weight gain prevention in young adults to identify the intervention with the greatest likelihood of adoption. As part of a randomized trial (Study of Novel Approaches to Weight Gain Prevention [SNAP]; n = 599) testing weight gain prevention interventions in young adults (18–35 years), individuals assigned to self-regulation interventions using either large changes or small changes reported on perceived personal effectiveness and difficulty of treatment over 3 years. Treatment satisfaction at 2-year follow-up was also reported. Pre-randomization, participants believed the large change intervention would be more personally effective than the small change intervention, although they also considered it more complex. Older age, lower body mass index (p = 0.056), and desire to maintain versus lose weight predicted greater perceived effectiveness of the small change relative to large change intervention. Over follow-up, the large change intervention was no longer perceived as more effective, but perceived effectiveness aligned with assigned treatment. The small change intervention was rated as less complex than the large change intervention at 4 months, but not at other follow-ups. At study conclusion, participants were largely satisfied with both treatments; however, in the small change intervention, individuals who were not successful at preventing weight gain were less satisfied than individuals who were successful. The large and small change interventions are both appropriate for dissemination with no clear advantages based on the participant perceptions.

Young adult participants perceived a large and a small change intervention for weight gain prevention to have different advantages, although both are good candidates for dissemination.

Implications.

Researchers: Given neither the large change nor the small change weight gain prevention intervention was superior for participant adoption, future studies may assess the comparative likelihood of widespread adoption of the large and small change treatment at the organizational level (i.e., clinical and community settings).

Practitioners: For clinicians recommending a weight gain intervention for young adults, both a large change and small change intervention is acceptable, although at the outset a large change weight gain prevention intervention is generally perceived as more compatible while a small change approach is viewed as less difficult.

Policymakers: The large and small change interventions are both viable treatments for dissemination with no clear advantages to adoption based on the participant perceptions that were assessed.

Introduction

Young adults gain weight rapidly and struggle in traditional weight control interventions that do not account for their unique developmental stage [1, 2]. The Study of Novel Approaches to Weight Gain Prevention (SNAP) study tested two approaches to weight gain prevention for young adults: large changes and small changes, both combined with a self-regulatory approach to weight management [3, 4]. The large change approach was designed to buffer against weight gain by having participants lose 5–10 pounds initially. Participants engaged initially in standard behavioral weight loss strategies, including dietary and weight self-monitoring, energy reduction of 500–1,000 kcal per day, and gradual increases in moderate and vigorous physical activity to 250 min per week. Once participants completed the initial weight loss portion, they were encouraged to continue healthy eating and activity behaviors to maintain weight and to return to weight loss behaviors (e.g., dietary self-monitoring) if they began to regain the weight. The small change approach, which encourages participants to make small, discrete changes to diet and activity every day (~100 cal each), was designed to combat a small caloric excess over time that leads to weight gain. Both the large and small change approaches were successful in preventing weight gain over an average follow-up of 3 years compared to a self-guided control condition [4]. Having demonstrated efficacy, it is important to consider factors that may influence the translation of the interventions into clinical practice, including factors that may promote differential uptake among potential participants.

The Rogers’ Diffusion of Innovation Theory highlights characteristics of innovations, or novel ideas, practices or products, that facilitate their adoption through social systems [5]. One component is compatibility or the degree to which an innovation fits with the personal needs and values of the user. A second is the complexity of an innovation or the ease of use. A third is relative advantage or the degree to which one innovation is perceived as better or more satisfactory than an alternative. Innovations that have a high degree of compatibility and relative advantage and a low degree of complexity are theorized to have greater and more rapid rates of adoption [5]. The diffusion of innovation theory can be applied at the service system level (e.g., hospital systems or community centers that would provide the intervention); however, indicators of potential participant adoption may first provide insight into which of the effective interventions to invest in and offer more broadly to young adults.

The current study aimed to compare participant perceptions of the SNAP large change and small change interventions on compatibility, complexity, and relative advantage. These concepts are quantified within the study as perceived personal effectiveness (e.g., which intervention is believed by the participant to be best for themself), difficulty, and satisfaction, respectively, which approximate, though do not precisely measure, Rogers’ identified theory components (e.g., perceived personal effectiveness does not measure how the intervention would fit with a participant’s values per se) (Fig. 1). Perceived personal effectiveness and perceived difficulty were assessed prior to participation (pre-randomization) to understand initial participant impressions. Demographic predictors of these perceptions were examined to inform future targeting or advertising of the large change or small change intervention to certain groups. Next, participant perceptions were compared over the course of treatment and follow-up to ascertain any differences in perceptions of the interventions. Actual intervention effectiveness, or success in meeting treatment goals, can influence perceptions [6]. Therefore, whether participants stayed at or below their baseline weight, which was a primary goal of both treatments, was examined as a moderator of perceptions of intervention attributes.

Fig 1.

Conceptual model of the diffusion of innovation theory, the characteristics assessed in the current study, and theorized effect on participant adoption.

Methods

Study design

The SNAP was a three-arm randomized controlled trial that tested the relative efficacy of two self-regulation interventions for weight gain prevention—a large change and a small change approach—and a self-guided control condition. In-person group sessions in the large and small change interventions were provided across 4 months with 8 weekly and 2 monthly sessions and the self-guided control condition consisted of one in-person group session in the first study week. The primary outcome, defined as the average weight change over the full 3 years of the study, differed significantly among all three arms of the trial—large change: −2.37, small change: −0.56, and control: 0.26. However, using a secondary outcome of weight difference from baseline at 2 years, the large (−1.5 kg) and small change (−0.77 kg) were not significantly different, but both experimental groups still performed better than control (+0.54 kg) [4]. The control condition is not assessed in the current analysis as it was inferior to both the large and small change interventions and would not be a candidate for dissemination [4].

Participants

Participants in the full trial (n = 599) were 18–35 year olds with a body mass index (BMI) of 21.0 to 30.9. As indicated above, only participants in the large change and small change intervention are included in the current secondary analysis (n = 397). Recruitment occurred at two clinical sites in Rhode Island and North Carolina. Recruitment materials advertised a weight gain prevention intervention and noted the possibility of modest weight loss. Inclusion criteria required participants to be English speaking with Internet access and to be able to participate fully with intervention recommendations (e.g., engage in physical activity). Exclusion criteria included a recent weight loss of >10 pounds in the prior 6 months, bariatric surgery or history of anorexia or bulimia, hospitalization for or history of severe psychiatric disorders, and currently or planning to become pregnant. Retention of this sub-sample at 3 years was 81.9% and did not differ by treatment arm [4]. Attendance rates did not differ by treatment arm [4].

Intervention

The large change intervention was designed to create a 5–10 pound weight loss buffer during the first 8 weeks by creating a 500–1000 kcal daily deficit in consumption and to increase activity to 250 min per week using behavioral skills training (e.g., self-monitoring). The small change intervention was designed to counteract small caloric excesses over time by encouraging two 100-calorie changes each day—one to diet and one to physical activity. For safety, participants who experienced a significant weight loss (>20% from baseline, BMI ≤ 18 kg/m², or >15-pound loss in 30 days) were identified and counseled per study protocol. Following the in-person sessions, participants were asked to continue daily weighing and to interpret their weight within a self-regulation framework, using a traffic light zone system, to avoid weight gain over time. Weight ranges for the green and yellow zones varied by the amount of weight loss a participant experienced during the in-person sessions. The green zone was tailored to keep participants around or below their final weight following the initial 8 weeks of in-person sessions and the yellow was any weight in between the green and red zones. The red zone was always defined as any weight greater than their baseline weight.

Participants were given condition-specific behavioral strategies for each zone. In the large change intervention, participants in the green zone were instructed to continue with healthy eating and high levels of physical activity and, if they entered the yellow zone, to problem-solve around why they were gaining weight. If they entered the red zone, participants were asked to return to detailed self-monitoring of diet and exercise and to return to calorie reduction. In the small change intervention, participants in the green zone were asked to continue making two small changes each day and to start monitoring these efforts if they entered the yellow zone. In the red zone, small change participants were asked to continue monitoring their small changes and to increase the number of small changes if they did not experience a weight reduction. Additional details on study protocol, attrition (including a CONSORT diagram), and outcomes can be found in Wing and colleagues [3, 4].

Measures

Masked study staff administered each assessment indicated below at baseline, 4 months, and 1, 2, and 3 years unless an exception is noted in the measure description.

Anthropometrics

Weight was measured in kilograms using a digital scale. Height was measured in centimeters using a wall-mounted stadiometer. The average of two measurements was used. Participants wore light clothing and removed footwear. The “red zone” was defined as any weight higher than baseline weight and the non-red zone was baseline weight or below.

Demographics

Basic demographic data (e.g., age, gender, and race/ethnicity), history of weight control efforts, and data regarding the desired program outcome (lose, maintain, or gain weight) were collected at baseline. No participants indicated they wanted to gain weight so this item-response was dropped from the analysis.

Baseline perception of the program

Participants were given a short description of the large and small change interventions and were asked how difficult it would be for them to adhere to the intervention and how confident they were that the intervention would help them to (a) prevent weight gain and (b) lose weight. The response scale was from 1 to 8, with higher scores indicating greater difficulty and greater confidence. These questions were only assessed at baseline.

Perceived personal effectiveness (compatibility)

Participants were asked “Which approach would be most effective for you in controlling your weight?” to which they could select the large change or small change intervention description.

Engagement with and difficulty of the program (complexity)

This questionnaire assessing difficulty was a branching questionnaire. First, participants reported if they had done anything to manage their weight in the past 4 months with yes/no responses. If participants responded affirmatively, they were asked if they made eating changes (yes/no) and exercise changes (yes/no). If participants responded affirmatively, they were then asked to rate the difficulty (range: 1–8) of the eating or exercise changes they made. This questionnaire was given at each follow-up time point, but not at baseline.

Treatment and treatment outcome satisfaction (relative advantage)

Participants rated satisfaction with their assigned weight management program from 1 (very dissatisfied) to 4 (very satisfied) and indicated whether or not they would recommend their assigned program to others (1: definitely not to 4: definitely would). Due to very low numbers of participants indicating “definitely not,” responses of “definitely not” and “probably not” were combined. To assess satisfaction with outcomes, participants were asked “Given the effort you put into following the SNAP program over the past 2 years, how satisfied are you overall with your progress on (a) maintaining your weight, (b) changing your dietary habits, and (c) changing your physical activity habits” and responded on a Likert scale from 1 (very dissatisfied) to 9 (very satisfied). This questionnaire was given only at year 2.

Statistical analysis

Paired-sample t-tests were used to compare pre-randomization ratings of difficulty and effectiveness of treatments within participants. Logistic regression was used to identify predictors of baseline perceived personal effectiveness. A series of longitudinal regression models implemented with generalized estimating equations with robust standard errors were used to assess changes over time and treatment effects on perceived effectiveness (large change vs. small change), weight, diet, and activity (made changes vs. did not make changes in previous 4 months), and difficulty with assigned treatment over time (range from 1 to 8). Models were built in three steps to assess: (a) main effects of time and treatment, (b) interaction between time and treatment, and (c) a three-way interaction between time, treatment, and red zone status. Finally, chi-square tests for categorical outcomes and independent samples t-tests for continuous outcomes were used to assess differences in treatment satisfaction between the groups at the 2-year time point (when the questionnaire was given). Participants were additionally compared within treatment by 2-year red zone status on their treatment satisfaction to determine whether satisfaction with treatment was influenced by treatment target achievement.

Results

Perceived effectiveness and personal effectiveness of treatment

Pre-randomization, participants were, on average, more confident that the large change intervention would be more effective than the small change intervention for both weight gain prevention, t(1,396) = 10.95, p < 0.001, and weight loss, t(1,395) = 15.26, p < 0.001. When asked which treatment would be most effective specifically for themselves, a larger percentage selected the large change intervention (57%) than the small change intervention (40%). Responses to these questions were related. Individuals who had a personal preference for the large change intervention indicated greater confidence in the large change intervention to prevent weight gain than those who had a personal preference for the small change treatment t(1,387) = 4.00, p < 0.001, and vice versa t(1,387) = −6.71, p < 0.001. Models indicate that predictors of perceived personal effectiveness of treatment were age and desire to lose or maintain weight, and there was a trend for an association with BMI. Older age (i.e., closer to 35 years old; OR: 1.05, 95% CI: 1.001–1.097) and a desire to maintain (vs. lose) weight (OR: 3.53, 95% CI: 1.402–8.860) predicted a preference for the small change intervention above the large change intervention (Table 1). Gender, race, income, student status, and history of previous weight control efforts did not predict perceived effectiveness (ps > 0.05).

Table 1 .

Predictors of perceived personal effectiveness of treatment prior to randomization

Variables	N	B	SE	OR	95% CI	p values
Gender
Male	86	–	–	–	–	–
Female	311	0.32	0.25	1.38	0.848–2.257	0.194
Age	397	0.047	0.02	1.05	1.001–1.097	0.044
Race/ethnicity
White	283	–	–	–	–	–
Black	46	−0.46	0.34	0.63	0.322–1.232	0.177
Hispanic	31	0.43	0.38	1.54	0.730–3.229	0.258
Other	29	−0.60	0.43	0.55	0.235–1.281	0.165
BMI	397	−0.08	0.04	0.92	0.850–1.002	0.056
Income
<$25,000	127	–	–	–	–	–
$25,000–$49,999	135	0.18	0.25	1.20	0.732–1.965	0.471
$50,000–$74,999	74	−0.09	0.30	0.91	0.505–1.655	0.768
>$75,000	37	−0.03	0.39	0.97	0.456–2.060	0.935
Student status
Nonstudent	246	–	–	–	–	–
Student	143	−0.36	0.22	0.70	0.456–1.074	0.102
History of weight control efforts
Yes	317	–	–	–	–	–
No	68	0.44	0.27	1.55	0.914–2.622	0.104
Desired weight outcome
Lose	365	–	–	–	–	–
Maintain	22	1.26	0.47	3.53	1.402–8.860	0.007

While the large change intervention was preferred pre-randomization, a time effect showed that there was an increase in perceived effectiveness for the small change intervention across follow-up. This time effect was modified by assigned treatment, with the individuals in the small change intervention showing a greater preference for the small change intervention and individuals in the large change intervention showing a greater preference for the large change intervention. In other words, the majority of participants believed the condition to which they were assigned was most effective at all follow-up time points, Wald χ² (4, n = 397) = 81.15, p < 0.001 (Fig. 2). Adding red zone status into the model, the three-way interaction was not significant, but a treatment condition by red zone effect emerged, Wald χ² (1, n = 397) = 4.88, p = 0.027. The two-way interaction showed that perceived efficacy did not differ by red zone status for participants in the large change intervention (69% of individuals in the red zone and 70% of individuals in the non-red zone believed the large change intervention to be most efficacious), but in the small change intervention, the percentage of individuals who believed the small change intervention was more effective was related to how successful the participant had been at maintaining their weight. Among those who had been successful (i.e., not in the red zone), 79% believed that small changes were more effective, whereas among those in the red zone, only 66% believed the small change intervention to be more effective.

Fig 2.

Percent of participants indicating assigned condition is the most effective approach for themselves.

Self-reported engagement with and difficulty of treatment

The following analyses only include participants who reported making any changes to manage their weight at each time point (98% of the sample at 4 months which decreased to 73% of the sample at 3 years). Statistical analysis results can be found in Tables 2. The proportion of individuals who reported making changes in their eating and/or exercise decreased significantly from 4 months to 3 years (96% to 86% and 82% to 72% for eating and exercise respectively). A main effect of treatment was found with regard to exercise changes with the small changes intervention group 1.68 times more likely to report making exercise changes than the large change intervention group; however, neither eating nor exercise changes showed a time by treatment interaction (p > 0.05), indicating no differences in the rate of decrease over time for eating and exercise changes. There were no three-way interactions found between time, treatment, and red zone status in exercise changes, although a trend-level three-way interaction was found for eating changes, Wald χ² (3, n = 397) = 7.58, p = 0.055 (Fig. 3). The finding suggests that at 4 months, the number of individuals in the large change group who reported making eating changes was greater in those who were not in the red zone (N = 173) compared to those who were in the red zone (N = 14), but no other time points were significant. No differences were found over time based on red zone status in the small change intervention.

Table 2 .

Parameter estimates for effects of time and treatment on changes to eating and exercise and difficulty of eating and exercise changes

	Changes to eating		Difficulty of eating changes			Changes to exercise		Difficulty in exercise changes
	OR	95% CI	B	SE	95% CI	OR	95% CI	B	SE	95% CI
Main effects
Time
4 mo	–	–	–	–	–	–	–	–	–	–
1 yr	7.15	3.023; 16.903	−0.29	0.122	−0.529; −0.050	3.02	2.120; 4.289	−0.25	0.132	−0.511; 0.006
2 yr	6.02	2.656; 13.638	−0.10	0.126	−0.342; 0.153	3.55	2.484; 5.079	−0.06	0.142	−0.343; 0.215
3 yr	8.95	3.860; 20.730	−0.19	0.142	−0.472; 0.086	2.60	1.769; 3.808	0.08	0.149	−0.212; 0.373
Treatment
Large changes	–	–	–	–	–	–	–	–	–	–
Small changes	0.91	0.555; 1.492	−1.11	0.138	−1.381; −0.840	0.60	0.433; 0.821	−0.70	0.151	−0.996; −0.405
Interaction
Time
4 mo	–	–	–	–	–	–	–	–	–	–
1 yr	6.97	2.398; 20.237	−0.88	0.170	−1.220; −0.542	3.14	1.983; 4.976	−0.41	0.205	−0.815; −0.010
2 yr	4.37	1.574; 12.105	−0.60	0.180	−0.949; −0.245	3.19	1.977; 5.140	−0.38	0.203	−0.780; 0.017
3 yr	5.43	1.844; 15.980	−0.73	0.198	−1.117; −0.341	2.04	1.262; 3.307	−0.27	0.224	−0.703; 0.173
Treatment
Large changes	–	–	–	–	–	–	–	–	–	–
Small changes	0.50	0.092; 2.694	−1.77	0.170	−2.100; −1.431	0.50	0.268; 0.915	−1.01	0.211	−1.427; −0.601
Time × treatment
4 mo × LG	–	–	–	–	–	–	–	–	–	–
4 mo × SM	–	–	–	–	–	–	–	–	–	–
1 yr ×LG	–	–	–	–	–	–	–	–	–	–
1 yr × SM	1.07	0.173; 6.581	1.18	0.235	0.719; 1.640	0.92	0.444; 1.905	0.30	0.267	−0.222; 0.824
2 yr × LG	–	–	–	–	–	–	–	–	–	–
2 yr × SM	2.12	0.380; 11.852	1.00	0.247	0.513; 1.482	1.30	0.626; 2.677	0.60	0.283	0.049; 1.160
3 yr × LG	–	–	–	–	–	–	–	–	–	–
3 yr × SM	2.95	0.504; 17.256	1.07	0.278	0.524; 1.612	1.73	0.785; 3.794	0.67	0.299	0.082; 1.253

LG large change intervention; SM small change intervention; bolded results indicated p < 0.05; mo months; yr year.

Fig 3.

Percent of participants reporting changes to eating by intervention assignment over time.

Prior to randomization, when asked to rate the perceived difficulty of each intervention, participants believed that adherence to the small change intervention would be less difficult (2.92 ± 1.62) than the large change intervention (4.35 ± 1.74), t(1,396) = 14.47, p < 0.001. Over follow-up, when participants reported on the difficulty experienced with their assigned intervention and these were compared between groups, the large change intervention showed greater difficulty with eating (M ± SE: 4.47 ± 0.10) and exercise (M ± SE: 4.7 ± 0.11) than the small change condition (M ± SE: 3.36 ± 0.10 for eating and 4.0 ± 0.10 for exercise). Moreover, a time-by-treatment effect showed difficulty in the large change condition for both eating and exercise changes decreased while difficulty in the small change condition increased (Fig. 4). This effect was significant for eating difficulty (p < 0.001) and trend-level for exercise difficulty (p = 0.088). There were no three-way interactions between time, treatment, and red zone, indicating time by treatment effects were not moderated differentially by red zone status. All results of generalized estimating equations testing time and treatment are reported in Table 2.

Fig 4.

Estimated marginal means of difficulty ratings for participants making eating and exercise changes by treatment over time. (A) Eating. (B) Exercise (trend-level interaction).

Treatment and treatment outcome satisfaction

At 2 years, differences in satisfaction for large and small changes did not differ significantly, χ² (2, n = 317) = 4.62, p = 0.099, with 62% and 53% of participants in the large change and the small change interventions respectively reporting high satisfaction. However, a larger percentage of large change participants (65%) than small change participants (52%) reported that they would “definitely” recommend their assigned treatment to others, χ² (2, n = 321) = 6.04, p = 0.049. When split by red zone status, there were no differences in satisfaction or likelihood of recommendation to a friend within individuals in the large change intervention (ps > 0.05), but within the small change intervention, individuals were more likely to report they were “highly satisfied,” χ² (2, n = 156) = 16.4, p < 0.001, and would “definitely” recommend to others, χ ² (2, n = 156) = 7.35, p = 0.025, if they were below their red zone weight than above their red zone weight. There were no treatment differences in satisfaction with outcomes of the treatment (i.e., weight, diet, or physical activity change) by treatment condition (ps > 0.05) and there were no interaction effects found between treatment and red zone status (ps > 0.05).

Discussion

The current study sought to determine how participants perceived the large and small change self-regulation interventions for weight gain prevention in young adults. We examined this using attributes from the diffusion of innovation theory that influence adoption in real-world settings, namely perceived personal effectiveness, difficulty, and satisfaction. These constructs are considered to be the most important characteristics for adoption and have been demonstrated in other behavioral health studies to predict end-user adoption and/or intentions to adopt [7–10].

Prior to beginning intervention in the current study, participants perceived that the large change intervention would have greater personal effectiveness, although the small change intervention was considered to be less complex. Once participants were engaging in intervention, participants perceived their assigned condition to be the most personally effective. The small change intervention rated making eating and exercise changes as less difficult than the large change intervention at 4 months, though the difference between treatments decreased over time. Participants in both groups were equally satisfied with their assigned intervention and their weight, diet, and physical activity outcomes at year 2, although the large change group was more likely to indicate they would recommend their treatment to others. When red zone status was considered, no difference was found for difficulty ratings by treatment. However, for perceived personal effectiveness and satisfaction, small change participants had poorer perceptions of the small change treatment if they were in the red zone than if they were not in the red zone.

Results regarding perceived personal effectiveness showed a larger number of participants believed large changes would be more personally effective prior to randomization. Nearly all (94%) of participants joined the study wanting to lose, versus maintain, weight. Given that the large change intervention was designed to induce a 5–10 pound weight loss, this may have made the large change intervention seem more personally effective at baseline. Nevertheless, the small changes intervention was perceived as more personally effective by young adults who were relatively older (e.g., closer to 35) and who wanted to maintain, not lose weight, suggesting it appealed to certain sub-populations of young adults. Once participating in the intervention, individuals were likely to perceive their randomly assigned intervention as the most personally effective. As individuals engage in treatment activities of their less preferred treatment, they may find the treatment to be more personally effective than initially believed. Thus, offering the large change intervention in a clinical care setting may generate greater initial interest and uptake given perceived effectiveness; however, once individuals are participating in the intervention, both the large and small changes are considered personally effective.

A shift in preferences may be influenced by experiencing success with their less preferred intervention. This appeared true in the small change intervention as individuals in the red zone were less likely to believe the small change approach was effective for them compared to individuals not in the red zone, although this was not the case for the large change approach. The large change approach is a more traditional, well-known method of weight management, while the small change approach is arguably more novel. It may be there is greater trust in the effectiveness of the large change intervention regardless of how successful one is with it, whereas individuals not succeeding with the small change intervention may be more likely to be critical of intervention characteristics. Another possibility is individuals in the large change intervention had seen some success by losing weight in the initial stages of the study and were thus more likely to believe it was effective, even if they entered the red zone in the later stages. Literature on treatment matching is mixed and meta-analyses have indicated a small effect of receiving a preferred behavioral treatment above a non-preferred treatment on retention and outcomes [11, 12]. Nevertheless, this study and others suggest that participant preferences and perceptions are malleable [13].

Complexity is frequently examined in the context of uptake of technology-based interventions or practices and has shown to impact adoption [7, 9]. In the current study, the small change intervention was seen as easier to use than the large change intervention at baseline, and would likely be most appealing to individuals who value simplicity or who may have limited time. The difference in perceived difficulty level may be the product of both the time involved in making large changes, including self-monitoring calories [14], as well as the greater self-regulatory effort involved in significantly reducing calories [15] in comparison to making small changes. The small change intervention was also rated as easier overall, although over time, difficulty ratings became more similar between treatments. The converging difficulty ratings may be due to the instructions given in the follow-up period by condition. In the large change condition, individuals were asked to maintain physical activity levels and daily self-weighing but were only instructed to return to their detailed self-monitoring and adherence to weight loss calorie goals if they entered the red zone whereas the small change intervention instructions were to continue to make two small changes to eating and exercise each day, even if participants were in the green zone. Thus, the large change intervention components related to diet were only used intermittently, if at all, while adherence to the small change diet components were ongoing, perhaps adding to their difficulty. Limited work has been done to compare the difficulty of the long-term use of an intermittent, intensive weight control treatment to a continuous treatment at a lower intensity for weight management, although it is well-documented that adherence to weight management interventions declines over time, in part due to increasing perceived burden [16].

Continued engagement with making treatment-related changes was also assessed. Participants in both treatments reported less engagement over time, with no difference between arms. Decreasing engagement may reflect a general difficulty with sustaining behavioral change [16] or may be due to the fact that a large number of participants are successful in preventing weight gain and may not need behavior changes. Overall, small change participants were more likely to report making exercise changes compared to the large change participants. Exercise changes were perceived as easier in the small change group compared to the large change group, which may have facilitated their greater implementation. Indeed, other studies have found greater preference for step compared to minute-based physical activity goals [17]. When engagement by treatment condition included an interaction for red zone status, the only effect was seen for eating changes. A trend-level finding (p = 0.055) showed individuals who were in the red zone were making fewer eating changes at 4 months compared to individuals who were not in the red zone in the large change condition, although differences diminished over follow-up. No effects were found in the small change condition. Results reinforce that caloric restriction is helpful for achieving a lower body weight in the short term. Over time, individuals in the large change group not in the red zone may reduce the use of caloric restriction given success at maintaining a lower weight while those in the red zone may continue to use caloric restriction, at least periodically, to attempt to lower weight. This would explain why reported eating changes converge, although caution should be taken in interpreting trend-level findings, particularly given the small number of individuals in the red zone from the large change condition at 4 months (n = 14). Of note, the percent of participants reporting eating changes was generally high (>80%) for both conditions and in both the non-red and red zones across follow-up.

There were no differences in intervention satisfaction by treatment, although individuals from the large change intervention compared to the small change intervention were more likely to “definitely” recommend their treatment to others. Relative advantage, conceptualized as satisfaction in the current study, has been found to be the most salient indicator of intentions to adopt other technology-based behavioral health interventions [7] and may be particularly important for participant adoption. Similar to the perceived effectiveness ratings, intervention satisfaction and recommendation to others did not differ by red zone status in the large change intervention, but in the small change intervention, individuals in the red zone were less satisfied with their intervention and less likely to recommend to others than individuals, not in the red zone. Thus, it may be worthwhile to consider offering individuals who are unsuccessful with the small change treatment the opportunity to shift to the large change treatment. Satisfaction with outcomes did not vary by treatment, no matter whether participants were in the red zone or not. At 2 years, when the questionnaire was given, weight change was relatively similar among the large change and small change interventions, which may explain similar ratings [4]. Moreover, the question regarding satisfaction with weight, diet, and physical activity outcomes was worded such that participants were asked to account for the effort expended. This may account for participants in the red zone who put in less effort and were satisfied with poorer outcomes [6].

A primary limitation of this study, which is framed within the context of the diffusion of innovation theory, is that the measures used were not specifically designed to address compatibility, complexity, and relative advantage and were not validated. As such, some constructs fit very well while others may not fully encompass Rogers’ intended attribute. For example, difficulty ratings are representative of complexity whereas perceived personal effectiveness does not encompass all aspects of compatibility, which is defined as how well the innovation fits with the values, experiences, and needs of the potential adopters. Relatedly, engagement, which is currently discussed as part of complexity due to the organization of the measure of difficulty ratings, is likely also influenced by other factors, including compatibility and satisfaction. The engagement was also self-reported, not objectively measured, and did not specify whether these were maintained changes over time or new changes since the last assessment, which may introduce error. Another limitation of measurement is that difficulty ratings do not include individuals who did not report making eating or exercise changes. These individuals may have chosen not to make eating or exercise changes because they perceived them to be too difficult, which may have biased results. However, this choice was made because collecting difficulty ratings on individuals who may not be trying to make changes to their eating and exercise could also provide inaccurate data.

Conclusion

Neither the large change nor the small change intervention was superior for dissemination when accounting for all three attributes assessed. The large change intervention was generally considered more compatible at the outset, although this shifted once participants began their assigned intervention. The small change intervention was considered to be less complex at baseline and throughout the study, although the difference in the degree of difficulty between the two treatments reduced over time. Finally, the large change intervention had a slight advantage over the small change intervention for treatment satisfaction, which seems primarily related to individuals who had gained weight in the small change intervention having a lower satisfaction. Given no clear “winner,” treatment offerings may best be determined based on the characteristics and values of the participants. For instance, the large change intervention may have broader appeal, but young adults who value simplicity may be most likely to participate in the small change intervention. Future studies may assess attributes of the diffusion of innovation theory among clinical or community settings to determine the comparative likelihood of widespread adoption of the large and small change treatment at the organizational level.

Compliance with Ethical Standards

Conflicts of Interest: DFT serves on the Scientific Advisory board and receives grant funding from WW. JGL receives grant funding from WW. RRW is on the Scientific Advisory Board of Noom. The other authors declared no conflict of interest.

Human Rights: All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. The Institutional Review Board at each institutional study site approved the study (Lifespan—The Miriam Hospital IRB; UNC Chapel Hill Non-Biomedical IRB).

Informed Consent: Informed consent was obtained from all individual participants included in the study. This article does not contain any studies with animals performed by any of the authors.

Transparency Statements: The study was pre-registered at clinicaltrials.gov. The analysis plan was not formally pre-registered. De-identified data from this study will soon be available in a public archive. De-identified data from this study can also be made available by emailing the corresponding author. Analytic code used to conduct the analyses presented in this study are not available in a public archive. They may be available by emailing the corresponding author. Materials used to conduct the study are not publically available.