Greater Average Meal Planning Frequency Predicts Greater Weight Loss Outcomes in a Worksite-Based Behavioral Weight Loss Program

Patients with neurofibromatoses (NF1, NF2, and schwannomatosis) who participated in a live video resiliency program sustained improvements in quality of life domains at one year follow-up.

Abstract

Background

Planning in behavioral weight loss (BWL) programs helps participants enact changes in eating and exercise, although the direct impact on weight loss is unclear.

Purpose

To examine how meal and exercise planning frequencies change in a BWL program and their relations to weight loss outcomes.

Methods

Participants (N = 139) in a 40 week worksite-based BWL program completed a questionnaire regarding meal and exercise planning frequency at Weeks 0, 10, 20, 30, and 40 and were weighed weekly. Growth curve models were used to determine trajectories in meal and exercise planning frequency and to assess the role of an individual’s average meal and exercise planning (between-person effect) and individual variation in planning (within-person effect) on body mass index (BMI).

Results

The best-fitting model, a linear random effect with a quadratic fixed-effect model, demonstrated that meal and exercise planning frequency increased over the course of the program with slowing growth rates. Between participants, higher average meal planning frequency (B = −0.029, t = −3.60), but not exercise planning frequency, was associated with greater weight loss. Within participants, exercise planning, but not meal planning, predicted a higher than expected BMI (B = 3.17, t = 4.21).

Conclusions

Frequent meal planning should be emphasized as a continued, as opposed to intermittent, goal in BWL programs to enhance weight loss. Average exercise planning frequency does not impact weight loss in BWL programs; however, acute increases in exercise planning frequency may be a popular coping strategy during a weight loss setback or, alternatively, may lead to increased calorie consumption and weight gain.

Introduction

Obesity affects over one third of adults in the USA, leading to increased obesity-related diseases and costing over $147 billion a year [1, 2]. While nearly one half of females and one third of males in the USA report intentions to lose weight [3], obesity rates continue to rise [1]. This may signal a failure to translate intentions into behavior, frequently deemed the intention–behavior gap, which is a well-documented phenomenon in studies of health behavior [4]. Action planning, or specifying when, where, and how one would enact behaviors, is critical in transitioning intentions to behavior [5]. Indeed, laboratory studies support planning as a mediator between intentions and behavior [6] and a meta-analysis shows planning strategies can have a medium-to-large effect (d = 0.65) on goal attainment [7]. With relation to individual energy-balance behaviors, planning strategies are effective in both supporting healthy changes to the diet [8] and increasing physical activity [9].

Multicomponent behavioral weight loss (BWL) programs, recommended as the most effective treatment for obesity [10], aim to improve diet and increase physical activity through behavior change strategies for weight loss. Planning ahead for weight loss behaviors (e.g., meals and exercise) is one of the key components to treatment; however, only a few studies have attempted to isolate the potency of its effects on weight loss outcomes. One study by Luszczynska et al. assigned a group of individuals participating in a commercial weight loss program to create and receive feedback on plans for diet and exercise for the following week at the beginning of their participation in the program [11]. After 2 months, participants in the planning group had lost twice as much weight as individuals who participated in the program alone. Follow-up mediation analyses indicated that the effect by condition was explained by the increased planning frequency in those who participated in the planning group. Another study by Benyamini et al. had a group of participants in a weight loss program make a detailed plan about two specific strategies they would like to focus on, one to improve diet (e.g., eat healthy snacks) and the other to improve exercise (e.g., use stairs whenever possible), at both the beginning and end of a 10 week treatment. Compared to individuals who just participated in the program, individuals in the group who made plans lost 40% more weight posttreatment, although differential effects did not persist at the 3 or 12 month follow-up, suggesting that planning effects may not endure over time [12]. A secondary analysis of this study indicated that dietary plans, rather than exercise plans, may be particularly beneficial [13]. Thus, studies indicate that planning in advance for weight loss behaviors may be a potent treatment component within multicomponent BWL programs.

One major drawback of these two existing studies is that planning activities were only completed once or twice across the program, whereas, in practice, participants in BWL programs are ideally utilizing planning skills at many, if not all, treatment sessions, as well as at home. With continued practice, skill level with planning may develop over time. Indeed, studies of other strategies taught within BWL programs, such as self-monitoring, show that the ability to use BWL strategies and the frequency with which an individual uses BWL strategies may change across the course of a program [14]. It is quite possible that people increase their planning if they have setbacks or if they experience unexpected gains, for example. Furthermore, existing planning studies examine differences between people (i.e., do individuals who plan more frequently lose more weight than those who plan less often); however, given that planning frequency may be changing at the individual level, it is worthwhile to know the impact of planning week to week on weight loss. This information may help to identify how effective planning is as a weight loss strategy and inform how to use time most effectively within BWL sessions. As such, the current study seeks to (a) assess how planning frequency of weight change behaviors (e.g., meal planning and exercise planning) evolves across a BWL program, (b) examine how individual differences in planning frequency between participants influence variations in weight loss across the program, and (c) explore how within-person changes across the program in planning frequency influence weight loss across the program.

Method

Overview/Study Design

This study used data collected as part of the MyWay to Health program, a 40 week worksite-based weight loss program adapted from an evidence-based BWL program [15]. The program was offered through an academic medical center within a regional health care system that employs more than 31,000 individuals. Height and weight measurements were completed at baseline and weight measurements were subsequently taken at each session and at the conclusion of the program. Participants also completed a planning questionnaire every 10 weeks, starting at baseline. Data are limited to participants who were ≥18 years and agreed to participate in research related to the program via written informed consent upon initiation of the program (95% of program participants). The study was approved by the affiliated university’s institutional review board.

Participants

Participants were hospital employees (n = 126) or employee family members (n = 13) participating in a multicomponent BWL program (total N = 139), of which 81.2% (n = 113) completed the program. Participants were aged ≥18 (M ± standard deviation [SD] = 49.09 ± 10.10); the majority were female (n = 124) and met criteria for overweight or obesity (body mass index [BMI] M ± SD = 39.11±7.36; 88.4% with obesity). Participants were recruited for the program via an in-house newspaper for employees and word of mouth. To participate, individuals were required to have a BMI of 25 or higher, to be an employee or directly related to an employee (i.e., spouse/partner, child), to be on the hospital health care plan, and to have physician clearance to participate in the program. There were no other exclusion criteria.

The MyWay to Health Worksite-Based BWL Program

The MyWay to Health program was a multicomponent BWL program that targeted diet, physical activity, and behavioral modification to support weight loss. Employees could participate alone or involve a significant other and/or other family member(s). Behavioral strategies included goal setting, planning, self-monitoring, stimulus control, skills training, contingency management, and relapse prevention, all of which are commonly found in effective BWL interventions [16]. The program consisted of 40 1 hr weekly individual sessions held in person, although make-up sessions were provided by phone if a participant was unable to attend. Each participant received their own treatment manual. All participants followed the same general flow of content (Supplementary Table 2), starting with dietary and physical activity changes, although the content was modularized to be able to tailor the content and meet participant needs as they arose.

A particular emphasis was placed on planning meals and exercise, and participants worked with their interventionist to create a detailed meal and exercise plan for the upcoming week. Interventionists worked with participants to select food choices for the daily meal plans to ensure that a participant met their daily calorie goal (i.e., final daily total was at or slightly below goal) beginning in the second week of the program. Exercise plans were created with the goal of 150 min or more moderate-intensity activity each week, although, given participant ability at baseline, this duration of activity was shaped for some participants (i.e., shorter goals but increasing over time to 150 min). Exercise planning followed meal planning and was introduced once interventionists believed that participants had developed meal planning skills; therefore, the timeline differed across participants. Meal and exercise planning were explicitly discussed in-session for at least a month following their introduction. A focus on planning continued across the program, although, given the programmatic nature of the intervention (as opposed to a research-driven protocol), the portion of each session spent discussing planning was based on participant needs. If participants were meeting calorie goals and exercising, particularly if they were making plans for this on their own, it was not discussed in as much detail. On the other hand, if participants were having trouble meeting calorie and exercise goals, planning, as well as addressing barriers to motivation for planning, and other behavioral changes continued to be a large component of the sessions. A greater emphasis was placed on meal planning above exercise planning as calorie reduction ties more strongly to weight loss than exercise [17].

Interventionist Training and Supervision

Interventionists came from multidisciplinary backgrounds and included three PhD level interventionists (two clinical psychologists and one exercise physiologist) and three Master’s level interventionists (two dietitians and one marriage and family counselor). All had prior experience in BWL intervention. Interventionists participated in a 2 day training prior to treatment intervention provided by a clinical psychologist with 20 years of experience in BWL and were provided a detailed treatment manual. This psychologist also provided weekly, 2 hr group supervision for the entire treatment team. Treatment fidelity was not explicitly assessed, although all sessions were audiotaped and selected sessions were reviewed during supervision.

Measures

Anthropometrics

At baseline, height and weight were measured in triplicate using an electronic scale and wall-mounted stadiometer by research staff following a detailed protocol. Participants wore light clothing and removed shoes for measurements. Throughout treatment, weight measurements were taken using the same scale and weight measurement protocol at each treatment session that was attended.

Planning questionnaire

A 16-item questionnaire was developed for this study given the lack of a previously developed questionnaire, which asked participants to report on their experiences with and thoughts about planning meals and exercise. The questions on these subscales were adapted from those used in the study by Luszczynska et al. [11]. Given that this was the first use of this questionnaire, an exploratory factor analysis was used to determine the underlying factor structure. The five components with Eigenvalues above 1 were selected for extraction and a promax rotation specifying five factors was used. The planning questionnaire was then organized into subscales based on factor loadings. Items with loadings greater than .5 were included within each factor, which resulted in two items being dropped. The five factors indicated exercise planning frequency (Factor 1), meal planning frequency (Factor 2), beliefs in planning for weight loss (Factor 3), ability to plan independently (Factor 4) and contingency planning, or making a “Plan B” (Factor 5). Subscales were then created within each factor by averaging responses to the items. The factor loadings of the items and the correlation of the subscales are provided in Supplementary Material 1. To address the study aims, only the exercise planning frequency and the meal planning frequency were included in the current study. Sample items from these subscales are “Each week I make plans regarding which exercise to perform” and “Each week I make plans regarding which meals and snacks to eat,” respectively, which were scored on a Likert scale from 1 to 5.

Statistical Analysis

Multilevel mixed growth models were used to analyze the data. These models assume that repeated observations are nested within persons. Level 1 observations represent the data collected across the program within person (i.e., in-session weights and meal and exercise planning frequency questionnaire scores). Level 2 observations represent between-person individual participants.

For Aim 1, to determine the trajectories of changes in planning subscales of interest and BMI across the course of the program starting at Week 0, both linear and quadratic models were tested. A linear model fit would suggest a steady, linear rate of weight loss, whereas a quadratic model would suggest a rate of weight loss that changes over time following a quadratic curve (e.g., slows over time). Random effects, or effects that account for variability in starting values and change trajectories between participants, were added to each model and were kept if they were determined to improve model fit. The Akaike information criterion and the Bayesian information criterion were compared across models and a χ ² difference test was used to test models against each other to determine a superior model fit.

To assess the role of planning on BMI change across the program (both within participants and between participants; Aims 2 and 3), additional variables were added to the multilevel mixed model predicting BMI change over time. To begin, covariates of sex and age were added at Level 2 given their potential association with weight and weight change [18, 19]. Meal and exercise planning variables were then added at both Level 1 and Level 2. Level 1 effects in the model represent the within-person effects (i.e., how variability within an individual at a certain time point across the program accounts for variability in BMI at that time point). Level 2 effects test between-person differences. The inclusion of meal and exercise planning variables at Level 2 alone test the influence of an individual’s planning frequency on average throughout the program on their BMI at baseline. In order to test the aim that an individual’s average meal and exercise planning frequency scores influence BMI trajectory, the interaction between Level 2 meal and exercise planning variables and time were included. Level 1 planning variables were person centered and the Level 2 planning variables were grand-mean centered. Meal and exercise planning models were tested separately (i.e., only meal planning and covariates and only exercise planning and covariates) and together with all variables in the model to examine independent and dependent effects. For all analyses, t-values of 1.96 or above indicate statistical significance at p < .05.

Results

Growth Models Predicting Meal Planning, Exercise Planning, and BMI Over Time

Models used all available data. Participants had an average of 28.1 ± 9.17 BMI measurements (out of 40 possible) and an average of 3.7 ± 1.38 and 3.7 ± 1.35 measurements for meal and exercise planning, respectively (out of 5 possible). Average scores on the meal planning questionnaire were 2.95 ± 0.95 at baseline to 3.9 ± 0.63 at program conclusion and average scores on the exercise planning questionnaire were 2.58 ± 1.07 at baseline to 3.68 ± 0.92 at program conclusion (highest score possible was 5). In the longitudinal growth models, including a quadratic term for both meal and exercise planning frequency models was superior to simple linear models of change. The quadratic models demonstrate that planning scores for meal and exercise improved over the program but that improvements lessened as the program continued (e.g., individuals increased in their frequency of making plans for what to eat for meals or do for exercise and where and when to do it more rapidly at the beginning of the program, but this growth slowed over time). For meal planning, the inclusion of a linear random effect improved model fit, whereas, for exercise planning, it did not. The random effect for meal planning indicates that people differ on their trajectory of meal planning, with some staying the same and others improving. In contrast, for exercise planning, there were few differences in trajectory among participants; they tended to all improve in planning at a similar rate.

The trajectory of BMI across the program was also assessed. Including a quadratic term was superior to having only a linear term, showing that individuals lost a significant amount of relative weight over the program but that the rate of weight loss decreased over time, with most changes occurring earlier in the program. The inclusion of the random linear effect of time further improved model fit. See Table 1 for model comparison results, Table 2 for final model statistics, and Fig. 1 for growth trajectories of meal planning frequency, exercise planning frequency, and BMI change.

Table 1.

Model comparisons for changes in meal planning, exercise planning, and BMI among participants (N = 139) across the MyWay to Health program^a

		Meal planning				Exercise planning				BMI change
	df	AIC	BIC	Chi-square	p-value	AIC	BIC	Chi-square	p-value	AIC	BIC	Chi-square	p-value
Null	3	1,203	1,215			1,388	1,400			15,463	15,482
Linear	4	1,138	1,154	67.04	>.001	1,297	1,314	92.60	>.001	12,118	12,143	3,346.9	>.001
Quadratic	5	1,083	1,104	57.05	>.001	1,272	1,293	26.79	>.001	11,682	11,713	4,38.28	>.001
Quadratic with random linear effect	7	1,071	1,100	15.82	>.001	1,276	1,305	0.12	.94	7,156	7,200	4,529	>.001

AIC Akaike information criteria; BIC Bayesian information criteria; df degrees of freedom.

^aModels selected for best model fit are in bold.

Table 2.

Growth models estimating changes in planning and BMI change in participants (N = 139) across the MyWay to Health program

	Exercise planning		Meal planning		BMI
Effects	Estimate (SE)	t-value	Estimate (SE)	t-value	Estimate (SE)	t-value
Fixed
Intercept	2.608 (0.088)	29.71	3.042 (0.072)	42.12	0.391 (0.623)	62.80
Linear slope	0.066 (0.008)	8.132	0.070 (0.007)	10.51	−0.206 (0.007)	−28.34
Quadratic slope	−0.001 (0.0,001)	−5.27	−0.001 (0.0001)	−8.16	0.003 (0.000)	49.19
Random
Variance of intercept	0.447	–	0.210	–	53.908	–
Variance of slope	NA	–	0.0001	–	0.006	–
Residual variance	0.534	–	0.389	–	0.231	–

Fixed effects estimates refer to regression coefficients in mixed models, while random effects estimates are estimates of variance.

BMI body mass index; SE standard error.

Fig. 1.

Quadratic models of meal planning frequency (a), exercise planning frequency (b), and body mass index (c) change over time in participants (N = 139) from the MyWay to Health program.

Growth Models With Planning Predicting BMI

Next, the influence of meal and exercise planning on the trajectory of BMI across the program was examined. An initial model, Model 1, was built utilizing the best model fit for BMI (linear random effect and quadratic fixed effect) accounting for sex and age covariates (Table 3). Results showed that BMI decreased across the program, with the rate of BMI loss decreasing over time. A significant interaction effect of time and age was observed demonstrating that individuals who were older had a significantly steeper decline in BMI than younger individuals.

Table 3.

Model results estimating the effects of meal and exercise planning frequency on BMI change in participants (N = 139) across the MyWay to Health program

	Model 1		Model 2
Effects	Estimate (SE)	t-value	Estimate (SE)	t-value
Fixed
Intercept	36.77 (1.881)	19.55*	37.869 (1.870)	20.248*
Slope	−0.227 (0.002)	−11.12*	−0.228 (0.023)	−10.02*
Curvature	0.003 (0.000)	49.21*	0.003 (0.000)	10.618*
Sex	2.643 (1.992)	1.33	1.467 (1.982)	0.740
Age	−0.108 (0.061)	−1.76	−0.058 (0.064)	−0.917
Time × Sex	0.024 (0.026)	1.12	0.032 (0.022)	1.455
Time × Age	−0.002 (0.000)	−2.56*	−0.001 (0.001)	−1.246
Meal L1			−0.530 (1.005)	−0.528
Exercise L1			3.170 (0.753)	4.212*
Meal L2			0.639 (1.014)	0.630
Exercise L2			−3.389 (0.757)	−4.475*
Time × Meal L2			−0.029 (0.008)	−3.604*
Time × Exercise L2			0.004 (0.005)	0.949
Random
Variance of intercept	52.621	–	47.994	–
Variance of slope	0.006	–	0.005	–
Residual variance	0.231	–	0.334	–

Fixed effects estimates refer to regression coefficients in mixed models while random effects estimates are estimates of variance.

L1 Level 1; L2 Level 2; SE standard error.

*p < .05.

To assess the hypothesis that meal and exercise planning frequency between person and within person are associated with weight loss across the program, a second model (Model 2) was built by adding the meal and exercise planning variables both at Level 1 (within person) and at Level 2 (between person), as well as the interaction between Level 2 variables and time (Table 3). Results showed that both the linear and quadratic fixed effects continued to significantly predict weight loss in the same pattern as Model 1. Exercise planning at Level 1 was found to positively predict BMI, indicating that when weekly exercise planning was higher than a person’s average, individuals had greater BMIs than predicted by their trajectories in that week. Variability in meal planning at Level 1 was not significantly related to BMI, indicating that there was no relationship between week-to-week variation in meal planning and changes in BMI across the program. Of note, separate models were tested with just meal planning and just exercise planning both at Level 1 and Level 2, but results were no different from the combined model and are, therefore, not presented.

Between-person results (corresponding with Level 2) demonstrated that average levels of exercise planning across the study period, but not average levels of meal planning, were negatively associated with BMI at baseline, indicating that individuals who had a higher average exercise planning score weighed less at baseline. The interactions of Level 2 planning variables and time evidenced a negative interaction between time and meal planning, but not exercise planning, indicating that individuals who had a higher average meal planning score had a steeper slope (i.e., lost more BMI units) than individuals with a lower average meal planning score (meal planning interaction shown in Fig. 2).

Fig. 2.

Interaction of average meal planning frequency and week of participation on body mass index change in participants (N = 139) of the MyWay to Health program.

Discussion

This assessment of meal and exercise planning across a 40 week BWL program demonstrated that meal and exercise planning frequencies follow a quadratic growth pattern, with the frequency of planning increasing more rapidly at the beginning of the program and slowing in growth as the program went on. Moreover, meal and exercise planning showed differential effects on weight loss at the between- and within-person level. Between-person, higher average meal planning frequency across the program predicted greater weight loss during the program, whereas within-person, exercise planning greater than a person’s typical exercise planning within a given week was associated with greater than expected weight that week. Results did not differ in models that only included meal or exercise planning variables; thus, it appears that these effects are relatively independent of one another.

Generally, results from this study are in line with experimental studies supporting the importance of planning ahead for energy-balance behaviors in weight loss [11, 12]. Moreover, other studies that have measured planning frequency pretreatment and posttreatment indicate that, when participants are taught planning at the beginning of a program, planning frequency increases [11]. This study contributes to the current literature by indicating that both meal and exercise planning frequency increase rapidly at the beginning of a BWL program and level off as the program continues. This result likely reflects the emphasis interventionists placed on preplanning meals and exercise at the start of the MyWay to Health program and suggests that these early gains in planning frequency are maintained throughout the course of the program. Furthermore, meal planning was best modeled using a random linear growth effect, while exercise planning was not. This pattern suggests that there is significant individual variability in meal planning frequency across the course of the program, whereas, with exercise planning frequency, individuals tend to follow a similar trajectory.

The current study found that greater meal planning frequency, on average, across the course of the program was related to greater weight loss. Both Luszczynska et al. and Benyamini et al. found that having participants make plans at the initiation of weight loss program facilitated weight loss compared to participants who did not receive any formal training in planning [11, 12]. Furthermore, greater planning frequency-measured pretreatment and posttreatment has been associated with greater weight loss [11]. The use of a multilevel growth model in the current study allowed for the measurement of both between-person differences and within-person differences in meal planning frequency. The findings of the current study did not support a within-person effect, suggesting individual fluctuations in meal planning frequency, while controlling for average meal planning frequency, did not impact individual fluctuations in weight. Thus, if participants are engaging in consistent and frequent meal planning, variation in meal planning frequency may have less of an impact on weight loss. Conversely, if participants are generally not meal planning, a time period of increased meal planning may not be very beneficial. Meal planning, thus, can be thought of as an important long-term goal rather than an effective short-term goal. This fits with the goal of the current intervention, as well as BWL interventions, more generally [20, 21], which is to help participants develop sustainable habits that will continue following intervention completion and lead to improved health outcomes. Meal planning frequency increased early in the intervention as was intended by the intervention and, while not explicitly tested, it follows that being able to establish this skill earlier lead to increased usage and, therefore, increased benefit for weight loss, for a longer duration of the program.

A different pattern of results was observed for exercise planning frequency. Average exercise planning frequency across the program (between person) was not related to weight loss over time. Studies of weight loss programs indicate that increasing exercise has small effects on weight loss compared to changing diet or changing both diet and exercise [17, 22]. Therefore, even if participants are excellent at consistently planning exercise and following through on their plans, engagement in physical activity may not be sufficient to lead to significant weight loss. Only one previous study has compared the differential effects of meal planning and exercise planning on weight loss outcomes and it found similar results: that meal plans, but not exercise plans, were beneficial for weight loss, specifically in individuals with high initial weight loss goals [13]. Importantly, exercise has a large number of beneficial effects on physical and mental health outcomes that were not measured in the current study [23], so the current findings should not be taken as a suggestion that exercise, and planning for exercise, is unimportant.

Compared to between-person differences in exercise planning, within-person fluctuations in exercise planning frequency were related to weight loss such that, when BMI was higher than predicted by the growth curve, exercise planning frequency was greater. Despite participant counseling around the particular importance of diet change for weight loss, as opposed to exercise change, this positive association between exercise planning frequency and BMI may reflect a tendency for individuals to utilize exercise planning more frequently than usual when they have experienced a weight loss setback in an attempt to “get back on track.” Alternatively, it may mean that, if individuals are making and following exercise plans more often, they may also be consuming more kilocalories, above and beyond what they are using for exercise. It could be that participants are hungrier due to increased exercise or they are overcompensating for perceived energy expenditure. Reports of these effects in the literature are mixed and have high individual variability [24]; thus, future work should replicate and assess the directionality of this relationship. Regardless, clinicians should be weary of acute increases in exercise planning and associated participant expectations at all points in the program so that participants are not discouraged if exercise planning does not lead to intended results.

One other significant result in the model that only included covariates in predicting weight change showed age as a significant predictor of weight change over time, specifically, that older participants had a steeper weight loss trajectory than younger participants. This finding is in line with the general literature suggesting that younger participants have more difficulty losing weight in programs, potentially due to less routinized and more hectic lifestyles [25]. This effect did become nonsignificant in the final model, suggesting that some of the age effect is explained by differences in meal planning and exercise planning frequency by age.

This study was the first to assess meal and exercise planning frequency multiple times during a BWL program, providing the opportunity to examine how the frequency of planning strategies may change. Another strength of the study was the assessment of the role of meal and exercise planning frequency on weight loss growth curve modeling. Growth curve modeling allows for the examination of effects both at the between-person and within-person level, providing more detailed insight into the effects of planning frequency. Limitations of the current study include the use of a planning questionnaire developed in house; however, similar questions have been used in previous studies and the items also showed strong factor loadings on the variables of interest in the factor analysis. Additionally, the questionnaire does not provide an indication of what participants are planning for meals and exercise. It is possible that planned meals and exercise were not in line with program recommendations, although, given participants created their plans with the supervision of their interventionist, this is unlikely. Moreover, it is possible that, at least for some participants, once meal and exercise planning skills were solidified, they transitioned to the use of more informal planning. That is, perhaps individuals were no longer writing down individual meal components and calories or dates and activities for exercise but were planning in more broad strokes in their minds; however, this cannot be gleaned from the current data. Finally, the intervention was programmatic in nature and designed to be tailored to individual participant needs. Therefore, following the initial focus placed on meal and exercise planning for the first month, emphasis on planning was not standardized, thus making it difficult to tease apart whether planning later in the program was initiated by participants or interventionists.

Conclusion

The current study demonstrated that meal and exercise planning frequency increased more steeply at the beginning of a BWL program, in line with the goals of the intervention, and leveled off as the program continued. Participants who more frequently engaged in meal planning, but not exercise planning, across the course of the program, lost more weight. Clinicians should consider promoting frequent meal planning for their patients early on and throughout the course of participation in a BWL program and fostering the maintenance of meal planning as a habit over time. Conversely, individual variations in exercise planning, but not meal planning, were related to weight and indicated that, when weight was higher than expected, exercise planning frequency was higher than average. This possibly highlights the popularity of using exercise planning as a coping strategy for weight loss setbacks or increases in energy consumption. Regardless, it may be more worthwhile for interventionists to spend time in sessions emphasizing meal planning instead of exercise planning.

Funding Program funding came from the sponsoring healthcare system. Work was additionally supported by the National Institutes of Health (T32 HL076134, T32 HL130357, F31DK113700).

Compliance with Ethical Standards

Authors’ Statement of Conflict of Interest and Adherence to Ethical Standards The authors declare that they have no conflict of interest.

Authors’ Contributions J.F.H. conceived of the study; J.F.H., D.R.R., H.S.B., R.R.W.., and D.E.W. contributed to the study design; J.F.H., D.R.R., and H.S.B. aided in data collection; J.F.H., K.N.B., E.E.F.C., and J.J.J. analyzed the data and aided in interpretation of the findings. All authors contributed to and approved the final manuscript.

Ethical Approval 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.

Informed Consent Informed consent was obtained from all individual participants included in the study.