Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2023 Dec 1.
Published in final edited form as: Health Psychol. 2022 Sep 1;41(12):938–945. doi: 10.1037/hea0001224

Behaviors and psychological states associated with transitions from regaining to losing weight

Jacqueline F Hayes 1, Rena R Wing 1, Jessica L Unick 1, Kathryn M Ross 2
PMCID: PMC9793336  NIHMSID: NIHMS1856092  PMID: 36048078

Abstract

Objective.

Little is known about week-to-week recovery from regains following a behavioral weight loss intervention (BWLI). This study examined changes in behaviors, cognitions, and moods associated with transitioning from weight regain to weight loss during the 9-month weight loss maintenance period after a 3-month Internet-based BWLI.

Methods.

Participants (n=68) self-weighed daily via “smart” scales and answered 40 weekly questionnaires about their weight-related behaviors and psychological states. Mixed models were used to 1) determine whether weight gain in a given week predicted changes in weight, behaviors (e.g., self-monitoring), and psychological states (e.g., mood, temptation) the following week and to 2) compare back-to-back weeks when individuals recovered from weight gain (gained in the first week and lost in the second) versus those in which they gained both weeks.

Results.

Weight gain in a given week predicted greater weight gain and greater report of behaviors and psychological states associated with weight gain the following week. Back-to-back weeks when individuals switched from gaining to losing were few (5%) compared to weeks when individuals continued gaining (60%). Switching from gaining to losing was associated with greater physical activity during the initial weight gain week and greater self-reported behaviors/cognitions consistent with weight loss (e.g., greater calorie self-monitoring, greater importance of “staying on track”) during the following week.

Conclusions.

Engagement in more weight loss consistent behaviors and more favorable ratings of key psychological variables were associated with the rare shifts from gaining to losing. Future research should investigate interventions to help individuals quickly recover from weight regain.

Keywords: Weight loss maintenance, weight regain, weight transitions

Introduction

Although behavioral weight loss interventions (BWLIs) consistently produce clinically-meaningful weight losses (Jensen et al., 2014), the vast majority of participants experience later weight regain, on average returning to their baseline weight after 4 years with no additional intervention (Nordmo et al., 2020). A substantial literature base has focused on examining potential causes of weight regain, which have been associated with reductions in weight loss behaviors (e.g., adherence to a reduced calorie diet, regular engagement in physical activity, and regular self-monitoring of weight, dietary intake, and physical activity) coupled with perceptions of weight loss behaviors as being more effortful and boring, increased feelings of hunger, greater temptations to eat high-calorie foods and skip physical activity, greater experiences of negative mood and stress, and prioritization of other life demands above weight management (Klem et al., 2000; Latner et al., 2013; McGuire et al., 1999; Ross et al., 2019). Consistent with the relapse prevention model (Larimer et al., 1999), participants in BWLIs are counseled to return as quickly as possible to weight loss behaviors following gains; however, research has demonstrated that few individuals are able to recover from periods of weight regain (DerSarkissian et al., 2017; Phelan et al., 2003). Rather, experiences of weight gain can lead to further reductions in adherence to behavioral recommendations, perhaps due to a sense of discouragement (Goldstein et al., 2019; Tanenbaum et al., 2016).

Although recoveries from weight regains are rare, a subset of individuals maintain lost weight long-term (Thomas et al., 2014). A key skill for successful maintenance appears to be the ability to quickly respond to and recover from weight regain (McGuire et al., 1999; Phelan et al., 2003). However, only a few studies have examined recovery from weight gain and little is known regarding individual traits, behaviors, or cognitions that may contribute to these recoveries. Phelan and colleagues (2003) found that smaller increases in depression differentiated long-term weight loss maintainers who were able to recover from a >5% weight gain from those who continued to regain, although no differences were found among a number of other behavioral or psychological variables. A more recent analysis examined concurrent behavioral and psychological factors associated with recoveries from weight gain in the context of a weight gain prevention study (Hayes, Tate, et al., 2021). Results showed that the use of a greater number of weight loss strategies was predictive of weight gain recoveries, although the study did not assess which strategies. Both studies had long follow-up periods and few data points, approximately a year apart, to assess recovery from weight gain. An analysis of more acute behavior is needed to identify short-term processes associated with recovery from weight gain, in order to effectively target weight regain during weight loss maintenance.

The current study addressed this gap by examining recovery from weight regain on a week-to-week level during a 9-month maintenance period following the end of an initial 3-month, Internet-based BWLI. Throughout the 9-month maintenance period, participants were asked to weigh themselves daily using study-provided smart scales and to complete a questionnaire each week that assessed weight management behaviors and psychological states hypothesized to be associated with weight regulation. Using these maintenance-period data, the effect of weight changes in a given week on weight, weight management behaviors (e.g., self-monitoring, physical activity), and psychological states (e.g., mood, temptation to stray from weight management behaviors, importance of “staying on track”) the following week was examined. It was hypothesized that greater weight gain during a given week would be associated with greater weight gain and with less use of behaviors and psychological states consistent with weight loss during the following week. For a more fine-grained examination of the variables that contribute to recovery from weight regain, an exploratory aim was to assess weight-related behaviors and psychological changes during back-to-back weeks when participants transitioned from gaining to losing weight compared to back-to-back weeks in which participants gained weight both weeks.

Methods

Transparency and Openness

Data are available upon request. Data were analyzed using R version 3.6.3 (R Core Team, 2013) and code is provided as a supplementary document. The full questionnaire provided to participants has been published previously (Ross et al., 2019). This article has been reported in adherence to the APA Journal Article Reporting Standards. Neither the parent study nor the current analysis were pre-registered.

Participants

Data for this secondary analysis were collected from the maintenance portion (i.e., last 9 months) of a 12-month study of an Internet-based BWLI in a worksite setting, which was approved by The Miriam Hospital Institutional Review Board. All participants provided written informed consent. Full details regarding the parent study, including participant recruitment, inclusion/exclusion criteria, and sample demographics/baseline characteristics have been published elsewhere (Ross & Wing, 2016a). In brief, participants in the parent study (n=75) were employees of a large health care corporation, were required to have a BMI ≥ 25 kg/m2 but weigh <150 kg (due to constraints of study-provided smart scales), and to have access to home internet. Participants of the parent study were included in the current analysis if they responded to at least one weekly questionnaire in the maintenance phase.

Intervention

The intervention consisted of a 3-month, multicomponent Internet-based lifestyle weight management program adapted from the Diabetes Prevention Program (Diabetes Prevention Program Research Group, 2002) and Look AHEAD (Look AHEAD Research Group, 2006). The program began with an introductory in-person group visit and included 12 weekly 15-minute video lessons. The final lesson was focused on strategies for relapse prevention. Participants were given in-home smart scales and asked to monitor their weight daily. Participants were prescribed calorie (1200–1800 kcal/day), dietary fat (<30% of total daily calories), and physical activity goals (progressive goal to 200 minutes of moderate-to-vigorous physical activity per week). Participants were asked to self-monitor progress toward these goals, and automated feedback messages based on these data were provided each week via the study website. Small financial incentives (ranging from $1–10 per week, with amounts selected using a pattern unknown to participants) were used to promote self-report of weekly self-monitoring data to the study team. Participants achieving only minimal weight loss at 1 month were offered three in-person counseling sessions. At the conclusion of the 3-month intervention, participants were asked to continue logging into the study website each Sunday to complete assessment measures, but automatic feedback was no longer provided, and participants no longer had access to any of the intervention website components. Additional details on the intervention can be found in Ross et al., 2016 (Ross & Wing, 2016a). The current study only utilized data from the weight loss maintenance phase (months 4–12).

Measures

Anthropometrics.

Weight and height were measured at baseline and follow-up by trained study staff using a calibrated digital scale and a wall-mounted stadiometer. Participants wore light clothing and with pockets emptied and shoes removed. Measurements were taken to the nearest 0.1 kg and 0.1 cm, respectively. Study “smart” scales (BodyTrace, Inc), which automatically transmitted weight data to a secure server, were used to collect daily weight data throughout the intervention and maintenance periods. Participants were instructed to weigh themselves using these smart scales first thing in the morning with minimal clothing before eating or drinking, but after voiding. The smart scales used in the current study have been demonstrated to have high concordance (correlations of 0.99) with weights measured in-person via calibrated clinical scales (Pebley et al., 2019; Ross & Wing, 2016b).

Weekly Questionnaires.

At the end of each study week, participants were asked to complete a questionnaire on the study website reporting on their experiences over the prior 7 days (the full questionnaire has been published previously (Ross et al., 2019)). Using a 7-point Likert scale, this questionnaire asked participants to self-report positive mood, negative mood, stress, hunger, boredom with weight control efforts, temptation to eat foods not on their plan, temptation to skip planned physical activity, the degree to which eating choices were consistent with weight loss goals, the degree to which physical activity choices were consistent with weight loss goals, the amount of effort it took to stay on track, and the importance of staying on track compared to other life demands. Higher ratings for each item indicated greater endorsement. The questionnaire also asked the number of days participants self-monitored their weight and caloric intake as well as the total number of minutes they participated in physical activity across the week. Physical activity minutes were converted into hours for data analysis.

Statistical Analysis

Duplicate and outlier weight measurements were removed prior to analysis (see (Ross et al., 2018) for additional details). As described previously (Ross et al., 2019), weekly weight change slopes were calculated using LOESS local polynomial regression models using R package locfit (Loader & Liaw, 2020). LOESS models were fit on an individual level, using a decay function (with a bandwidth of 15 days) that prioritized more recent body weight values to account for missing data and smooth “noise” typically observed in daily weight data (e.g., due to hydration status or salt intake) (Stevens et al., 2006). Weight slopes of single-week increments, aligning with questionnaire administration, were isolated and used in the models described below. Weeks with a negative slope indicated weight loss while weeks with a positive slope indicated weight gain (see for additional details). Patterns of missing data were examined; missing data were not found to be related to baseline demographics.

Random-intercept multilevel models, nested within person with an unstructured covariance matrix and restricted maximum likelihood estimation, were created using R package lme4 (Bates et al., 2014). Univariate models were used to estimate the effect of a given week’s (time “t”) weight change (centered within-person), which was the independent variable in the model, on weight change and behaviors/psychological characteristics the following week (time “t+1”), which were the dependent variables. Linear time (measured in weeks) was tested for inclusion as a covariate as it was likely to affect some variables (e.g., self-monitoring) moreso than others (e.g., stress).

To examine behavioral and psychological differences in “transition” weeks, weeks were grouped into four categories based on weight change at “t” and “t+1”: gain-gain, gain-lose, lose-gain, lose-lose. Any amount of weight lost or gained throughout the week, regardless of magnitude, qualified for inclusion in these categories. Furthermore, due to the nature of the data collection and creation of weight change slopes, no values were exactly equal to zero. Week pairs in which one or both weight change slopes were missing were excluded from analyses. Gain-gain weeks and gain-lose week pairs were isolated and compared using multilevel models nested within person. Only gain-gain and gain-lose were selected for subsequent models because the second aim of this analysis was to understand the behavioral and psychological factors associated with recovery from regain, thus participants needed to be gaining weight in week “t”. Week type (i.e., gain-gain vs. gain-lose) was tested as a dichotomous predictor of the dependent variables, which were weight change at time “t” and behavioral/psychological variables at time “t” and “t+1”. Both time points were assessed to understand if any behavioral/psychological variables pre-empt changes in weight trajectory as well as to examine behavioral/psychological variables in the weeks concurrent with weight trajectory changes. The behavioral/psychological variable at time “t” was included as a covariate in models testing the salient variable at “t+1” to assess change in these variables when an individual shifts from gaining to losing. All models were visually inspected for model assumptions (e.g., normality of residuals, collinearity, influential observations) and were found to adhere sufficiently to these assumptions.

Results

Of the 75 participants included in the parent study, 68 completed at least one weekly questionnaire during the maintenance phase and were thus included in the current study. On average (mean ± SD), these participants were 50.84 ± 10.34 years old and weighed 86.42 ± 16.98 kg at baseline. The majority were female (69.1%) and non-Hispanic white (83%). Using weights collected at assessment visits, participants lost −6.84 ± 4.72 kg during the 3-month program and regained an average of 2.97 ± 4.64 kg during the maintenance period. Over the 9-month (40 week) maintenance period, participants responded to an average of 29.62 ± 10.08 weekly questionnaires (61.8% of participants responded to at least 75% of the weekly questionnaires) and had an average of 38.40 ± 4.79 weeks of objectively measured weight change data.

Multilevel models demonstrated that greater weight regain during a given week predicted greater weight regain, reduced adherence to self-monitoring of weight and caloric intake, less engagement in physical activity, and changes in the expected direction for all of the Likert-style questionnaire items during the following week (see Table 1). Specifically, 1 kg of weight gain in a given week predicted an additional 0.91 kg of weight gain the following week, a decrease of 1.78 days of calorie monitoring and 1.01 days of self-weighing, and a decrease of 0.92 hours of physical activity. For Likert scale variables, 1 kg of weight gain in a given week predicted increases in Likert scale points (out of 7) the following week of 1.34 for boredom, 1.32 for effort, 1.05 for hunger, 0.76 for negative mood, 0.67 for stress, 1.44 for temptation to eat foods “not on plan”, 1.17 for temptation to skip planned physical activity, along with reductions in Likert scale points of 0.67 for positive mood, 1.62 and 1.09 for eating and physical activity choices consistent with weight loss goals, respectively, and 1.40 for the importance of “staying on track” compared to other life demands.

Table 1.

Results from Univariate Multilevel Models with Weekly Weight Change Predicting Weight Change and Behavioral/Psychological Variables the Following Week

Fixed Effect Estimate (SE) BP Variance WP Variance p
Weight change (kgs) a 0.91 (0.01) <0.01 0.01 <0.001
Frequency of self-monitoring caloric intake (days)a −1.78 (0.16) 4.10 3.33 <0.001
Physical activity (hours)a −0.92 (0.14) 8.04 1.62 <0.001
Frequency of self-weighing (days)a −1.01 (0.19) 3.84 4.50 <0.001
Eating choices consistent with weight loss goalsa −1.62 (0.12) 0.66 1.31 <0.001
Physical activity consistent with weight loss goals −1.09 (0.14) 1.94 1.77 <0.001
Boredom with weight control effortsa 1.34 (0.11) 1.26 1.11 <0.001
Effort of “staying on track” 1.32 (0.12) 1.10 1.43 <0.001
Importance of “staying on track” vs. other life demandsa −1.40 (0.12) 1.68 1.22 <0.001
Hunger 1.05 (0.11) 1.06 0.99 <0.001
Negative mood 0.76 (0.12) 0.81 1.31 <0.001
Positive mood −0.67 (0.11) 0.79 1.12 <0.001
Stressa 0.67 (0.14) 0.87 0.71 <0.001
Temptation to eat foods “not on plan” 1.44 (0.12) 1.01 1.26 <0.001
Temptation to skip planned physical activity 1.17 (0.14) 2.22 1.89 <0.001
Open in a new tab
a

Model controlled for linear fixed effect of time due to associations with the outcome variable

Fixed effects estimates are unstandardized; Each table row is a separate model; Unless otherwise noted, the variables were measured using a 7-point Likert scale (1–7), with higher numbers indicating greater endorsement of the indicated variable in the first column; BP= between-person; WP= within-person

Participants had an average of 37.40 ± 4.80 out of 39 (96%) possible week pairs. A total of 2,660 two-week pairs were identified; the majority were gain-gain weeks (60%), followed by lose-lose weeks (29.7%), lose-gain weeks (6.1%), and gain-lose weeks (5.2%). On average, participants experienced 22.16 ± 7.15 gain-gain weeks (range: 2–35), 11.06 ± 6.75 lose-lose weeks (range: 0–34), 2.24 ± 1.07 lose-gain weeks (range: 0–4), and 1.94 ± 1.03 gain-lose weeks (range: 0–4).

Mixed model building began with time predicting psychological and behavioral outcomes. Time was significantly related to frequency of self-monitoring weight and caloric intake, reports of eating consistently with weight goals, importance of weight goals, and stress, which all decreased over time. Time was also related to hours of physical activity and reports of boredom, which increased over time. Thus, time was included as a covariate in the models assessing these outcomes. Models in which linear time was not a significant predictor of outcomes did not include time as a covariate.

Fixed-effects estimates from univariate multilevel models with week pair (gain-gain vs. gain-lose) predicting behavioral and psychological outcomes are presented in Table 2. Table 3 shows the estimated marginal means and standard deviations of the behavioral and psychological outcomes by week-pair types from the multilevel models presented in Table 2. Estimated marginal means reflect the change in the indicated variable in the specified week. In gain-gain weeks, participants gained an average of 0.19 kg in both week “t” and week “t+1” whereas in gain-lose weeks, participants gained an average of 0.06 kg in week “t” and lost 0.05 kg in week “t+1”. Results from models comparing gain-lose weeks to gain-gain weeks at week “t” (i.e., the week in which both groups were gaining weight) showed individuals in gain-lose weeks gained less weight and engaged in 20 more minutes of physical activity than individuals in gain-gain weeks. At week “t+1” (i.e., one group was losing and the other group continued to gain), individuals who lost weight reported greater frequency of self-monitoring of calorie intake (approximately 1/3 of a day more), greater eating consistent with weight loss goals, less boredom, less effort, greater importance of “staying on track” in comparison to other life demands, and greater positive mood, than those who continued to gain weight. No other weight-related behaviors or psychological states were found to be significant at either time “t” or “t+1”.

Table 2.

Results of Individual Univariate Multilevel Models Comparing the Effects of Gain-Gain to Gain-Loss Week Pairs on Behavioral and Psychological Variables in Week “t” (“gain” week in both pairs) and Week “t+1” (“gain” week in gain-gain pair and “loss” week in gain-loss pair)

Week “t” Week “t + 1”
Fixed Effects Estimate (SE) p Fixed Effects Estimate (SE) p
Frequency of self-monitoring caloric intake (days)a −0.20 (0.16) 0.214 0.37 (0.15) 0.015*
Physical Activity (hours)a 0.33 (0.14) 0.018* −0.11 (0.15) 0.485
Frequency of self-weighing (days)a −0.13 (0.20) 0.510 0.17 (0.19) 0.376
Eating choices consistent with weight loss goalsa −0.15 (0.13) 0.252 0.34 (0.14) 0.014*
Physical activity consistent with weight loss goals 0.27 (0.15) 0.069 0.16 (0.15) 0.294
Boredom with weight control effortsa 0.20 (0.11) 0.074 −0.32 (0.13) 0.011*
Effort of “staying on track” 0.12 (0.13) 0.332 −0.28 (0.14) 0.038*
Importance of “staying on track” vs. other life demandsa 0.03 (0.12) 0.814 0.27 (0.13) 0.041*
Hunger 0.13 (0.11) 0.223 −0.08 (0.11) 0.453
Negative mood −0.04 (0.12) 0.768 −0.12 (0.13) 0.361
Positive mood −0.14 (0.11) 0.209 0.28 (0.12) 0.020*
Stressa −0.02 (0.14) 0.893 −0.17 (0.15) 0.278
Temptation to eat foods “not on plan” 0.12 (0.12) 0.318 −0.13 (0.13) 0.311
Temptation to skip planned physical activity −0.08 (0.15) 0.619 −0.20 (0.16) 0.218
Open in a new tab
a

Model controlled for the linear fixed effect of time due to associations with the outcome variable

Gain-gain week pairs are the reference group; Fixed effects estimates are unstandardized; Each table row is a separate model; Unless otherwise noted, the variables were measured using a 7-point Likert scale (1–7), with higher numbers indicating greater endorsement of the indicated variable in the first column

Table 3.

Estimated Marginal Means from the Mixed Models Presented in Table 2 that Compare Gain-Gain and Gain-Lose Weeks

Week “t” Week “t+1”
Gain-Gain Gain-Lose Gain-Gain Gain-Lose
Weight (kgs) 0.19 (0.01) 0.06 (0.02) 0.19 (0.01) −0.05 (0.02)
Frequency of self-monitoring caloric intake (days)a 2.5 (0.26) 2.3 (0.30) 2.31 (0.18) 2.68 (0.22)
Physical Activity (hours)a 2.35 (0.33) 2.69 (0.36) 2.41 (0.23) 2.3 (0.27)
Frequency of self-weighing (days)a 4.10 (0.23) 3.97 (0.29) 4.01 (0.18) 4.18 (0.25)
Eating choices consistent with weight loss goalsa 3.68 (0.12) 3.53 (0.17) 3.64 (0.10) 3.97 (0.16)
Physical activity consistent with weight loss goals 3.58 (0.17) 3.84 (0.22) 3.59 (0.13) 3.75 (0.19)
Boredom with weight control effortsa 3.77 (0.15) 3.97 (0.18) 3.83 (0.12) 3.51 (0.16)
Effort of “staying on track” 5.09 (0.14) 5.21 (0.18) 5.18 (0.11) 4.90 (0.17)
Importance of “staying on track” vs. other life demandsa 4.44 (0.18) 4.47 (0.21) 4.44 (0.12) 4.71 (0.17)
Hunger 4.25 (0.14) 4.38 (0.17) 4.29 (0.10) 4.21 (0.14)
Negative mood 3.12 (0.12) 3.08 (0.17) 3.08 (0.09) 2.95 (0.15)
Positive mood 4.76 (0.12) 4.62 (0.16) 4.78 (0.09) 5.07 (0.14)
Stressa 4.10 (0.12) 4.08 (0.18) 4.12 (0.09) 3.95 (0.17)
Temptation to eat foods “not on plan” 5.07 (0.13) 5.19 (0.17) 5.11 (0.11) 4.98 (0.16)
Temptation to skip planned physical activity 4.62 (0.18) 4.55 (0.23) 4.65 (0.12) 4.46 (0.19)
Open in a new tab
a

Model controlled for linear fixed effect of time

Bolded values indicate significant differences, p<0.05; Unless otherwise noted, the variables were measured using a 7-point Likert scale (1–7), with higher numbers indicating greater endorsement of the indicated variable in the first column

Discussion

The current study demonstrated that greater weight gain on a given week was predictive of greater weight gain the following week; moreover, weight gain on a given week also predicted behaviors and psychological states consistent with continued gains the following week. Weight gains had a greater effect on psychological variables related to weight (e.g., effort of “staying on track”, temptation to eat foods “not on plan”) than it did on more general psychological variables (e.g., positive mood, stress). Shifts from gaining weight one week to losing weight the following week were rare (occurring on only 5% of week pairs compared to 60% of week pairs in which participants continued gaining), supporting previous research suggesting that recovery from weight gain is challenging and that additional intervention during maintenance may be necessary for most individuals to effectively maintain weight (MacLean et al., 2015). Compared to week pairs where participants continued gaining on the second week, weeks when participants successfully switched from gaining to losing weight were associated with smaller weight regains and greater engagement in physical activity the first week, along with greater adherence to self-monitoring of dietary intake, eating choices more consistent with weight loss goals, lower ratings of boredom and effort, greater importance of staying on track given competing life demands, and higher ratings of positive mood the second week.

Comparisons between gain-gain and gain-lose weeks provide some indication of where it may be best to target intervention efforts. The shift from gaining to losing was associated with a smaller magnitude of weight gain and increased use of fundamental weight loss behaviors, including physical activity, self-monitoring of caloric intake, and making eating choices consistent with weight loss, all of which likely contributed to the observed decreases in weight. The time course of these variables differed, with smaller magnitude of weight gain and greater physical activity reported in the week preceding the weight loss and greater diet-related weight loss behaviors reported concurrent with the weight loss. Increasing physical activity by a moderate amount (i.e., 20 minutes) may be the first behavior change made when re-engaging with weight loss behaviors. A recent study found that plans to exercise increased in weeks when BWLI participants experienced a weight gain, supporting the possibility that increasing physical activity may be used as a key strategy for coping with setbacks (Hayes, Balantekin, et al., 2021). Greater physical activity may have contributed to the smaller weight gain preceding a loss week, which may in turn have been a motivating factor to return to weight management behaviors. The presence of smaller weight gains coupled with greater physical activity could potentially be used as indicators to identify individuals likely to transition to losing weight within an adaptive intervention framework (Nahum-Shani et al., 2018).

Increases in diet-related behaviors followed increases in physical activity and occurred concurrently with weight loss. Caloric restriction and self-monitoring of calorie intake are two of the most robust behavioral strategies for weight loss (Burke et al., 2011; Yu et al., 2015). Exercise has been shown to prompt improvements in eating behaviors in the context of weight loss programs (Annesi, 2019; Mata et al., 2011); therefore, greater engagement in physical activity may have catalyzed the use of diet-related weight management strategies. Future studies should investigate the utility of increasing physical activity as a strategy to motivate re-engagement with weight loss behaviors following regains in the maintenance period.

Psychological changes were also observed in the weeks where individuals went from gaining to losing, including increases in the importance of staying on track with weight-management behaviors, compared to other life demands. As the process involved in maintaining weight becomes less rewarding over time, priorities may shift to other life domains (MacLean et al., 2015). Shifting priorities back to weight loss likely reflects a renewed motivation, which would positively influence intentions to reach and capability to complete a goal (Ajzen, 1991). What is unclear from the current findings is what triggered this shift, though this would be of interest for future studies. Nevertheless, interventions that focus on harnessing motivational factors, such as motivational interviewing or acceptance and commitment therapy (ACT), may be helpful to address weight regains during maintenance (Armstrong et al., 2011; Forman & Butryn, 2015).

On week pairs that individuals shifted from gaining to losing, they also reported lower ratings of boredom and effort with weight control efforts the following week. Boredom is a commonly cited barrier in long-term weight control (Leahey et al., 2016), but changing eating and physical activity routines to align more closely with weight loss behaviors may help assuage boredom (Bond et al., 2012; Levy et al., 2010; Sylvester et al., 2018). Continued weight loss maintenance is associated with reduced effort over time (Klem et al., 2000), but it was surprising that decreases in effort occurred so acutely. According to the theory on psychological momentum (Iso-Ahola & Dotson, 2016), once initial success with a goal is achieved, confidence and perceived competence increase. Continued success will then seem less effortful, increasing psychological momentum and the likelihood that success will occur. It may be that only small behavioral successes, perhaps an increase in physical activity as mentioned above, are needed to incite psychological momentum to re-engage with weight loss behaviors, although this remains to be tested. Alternatively, this study did not assess environmental context. It may be that participants experienced a change in their environment during these weeks that may have positively impacted the ability to successfully manage weight with lower effort. For instance, a return from vacation would likely make weight loss less effortful given potential reductions in temptation and greater control of the home food environment.

Finally, greater positive mood was reported in the week when weight loss occurred. Similar to discussion around the other variables that occurred concurrent with weight loss, conclusions around the directionality of the relationship are limited. It may be that the experience of losing weight following a gain is reinforcing and contributes to a greater positive mood. On the other hand, being in a positive mood throughout the week may promote one’s ability or capabilities to “get back on track”. For example, following the broaden-and-built theory, positive emotions serve to broaden perspectives and increase psychological resources to be able to accomplish a specific action, such as losing regained weight (Fredrickson, 2013). Similarly, in problem solving therapy, which has been shown effective at enhancing long-term weight loss maintenance (Perri et al., 2001), having a positive problem orientation and believing in personal ability to solve the problem is the first step to working towards a solution (Nezu & D’Zurilla, 2006).

A number of variables also did not differ between the transition weeks and the continued gain weeks. Most notably, greater frequency of self-weighing has consistently been associated with more positive weight outcomes over time (Linde et al., 2005); however, there was not a difference in frequency of self-weighing between weeks that participants continued to gain weight versus weeks where they shifted from gaining to losing weight. The current findings suggest that this strategy may not be used acutely to shift weight, but may follow after individuals start acting more consistently with weight loss goals. Tanenbaum and colleagues have shown that individuals participating in a BWLI are more likely to skip weighing themselves if they have eaten more calories than normal the previous day (Tanenbaum et al., 2016). It is possible participants who reverse weight gain may feel the need to make more consistent behavioral changes prior to returning to more frequent self-weighing. Another possibility is that the transmission of weight data via the “smart” scale may have artificially inflated the frequency with which individuals who continued to gain would have weighed themselves if data was not being submitted.

The current analysis has a number of limitations. First, there were fewer gain-lose weeks than gain-gain weeks and, although the standard deviations around these means were similar, the limited number of gain-lose weeks in comparison to the gain-gain weeks reduces the ability to make firm conclusions. Second, the categorization of week pairs into gain-lose and gain-gain weeks did not account for the magnitude of the weight change each week nor for “runs” or “streaks” in weight over a series of weeks (e.g., how many weeks participants had been gaining weight before transitioning to a loss). Third, corrections for multiple tests were not made as second aim of the study was also considered exploratory. While this leaves open the possibility that some associations occurred due to chance, it reduces the possibility of missing true relationships, which is important in this hypothesis-generating stage (Perneger, 1998). Fourth, behavioral variables were not objectively measured, which may increase likelihood for bias or error (Burke et al., 2008; Prince et al., 2008), and data collection methods did not account for all possible behaviors that could be related to weight (e.g., calories consumed). Fifth, the multilevel models used in the current analyses assume that missing data are missing at random (MAR), although intervention data are often more likely to be missing not at random (MNAR; e.g., as individuals who drop out from the study tend to have worse outcomes) (Molenberghs & Kenward, 2007). Proper specification of MNAR models require assumptions regarding the missing data; however, no previous studies have investigated these variables on this time scale, precluding ability to develop valid assumptions. Thus, given lack of prior research in this area and simulation studies showing that maximum-likelihood approaches (including multilevel models) are fairly robust even to MNAR data (Shin et al., 2017), these models were most appropriate choice for this exploratory study. Finally, while the study design allows for a more acute examination of behavioral and psychological factors that influence weight changes during maintenance, conclusions about the directionality is limited, particularly around the psychological variables, and external factors that may influence this relationship (e.g., travel or other changes in participants’ routines) were not assessed. It will be important to uncover which psychological variables precede or follow weight losses after weight regains and to control for other variables that affect these relationships, particularly when considering how best to develop interventions that may help shift regain trajectories. Relatedly, participants in the current study were using the scale frequently and completing questionnaires each week, which may have made them more aware of their weight changes as well as their behavioral and psychological changes; it is unclear how findings may generalize to populations engaged in less self-monitoring.

Conclusion

The current study examined the directionality of weekly weight changes and their association with proximal behaviors and psychological states. It is also the first to identify behaviors and psychological states that are associated with initiating a weight loss following a regain during weight loss maintenance. The time course of the behavioral variables (i.e., greater physical activity preceded greater caloric monitoring and better dietary choices) as well as the identification of the psychological variables most related to the shift in weight (i.e., comparative importance of weight loss goals, boredom, effort, and positive mood) can be used as potential targets in developing an intervention for individuals struggling with weight regain. Future studies should seek to develop and test interventions that effectively address weight regain.

Supplementary Material

Supplemental Material

Acknowledgments

Financial support for the study was provided by National Institute of Diabetes Digestive and Kidney Diseases (National Institutes of Health) Grant R21 DK109205 and K23 DK7147270 and the National Heart, Lung, and Blood Institute (National Institutes of Health) Grant T32 HL076134, and by the Lifespan Corporation. RRW is on the Scientific Advisory Board of Noom.

Footnotes

The other authors declared no conflict of interest.

Bibliography/References Cited

  1. Ajzen I (1991). The theory of planned behavior. Organizational behavior and human decision processes, 50(2), 179–211. [Google Scholar]
  2. Annesi JJ (2019). Moderation of mood in the transfer of self-regulation from an exercise to an eating context: short-and long-term effects on dietary change and obesity in women. International journal of behavioral medicine, 26(3), 323–328. [DOI] [PubMed] [Google Scholar]
  3. Armstrong M, Mottershead T, Ronksley P, Sigal R, Campbell T, & Hemmelgarn B (2011). Motivational interviewing to improve weight loss in overweight and/or obese patients: a systematic review and meta-analysis of randomized controlled trials. Obesity reviews, 12(9), 709–723. [DOI] [PubMed] [Google Scholar]
  4. Bates D, Maechler M, Bolker B, Walker S, Christensen RHB, Singmann H, Dai B, & Grothendieck G (2014). Package ‘lme4’. R Foundation for Statistical Computing, Vienna, 12. [Google Scholar]
  5. Bond DS, Raynor HA, Phelan S, Steeves J, Daniello R, & Wing RR (2012). The relationship between physical activity variety and objectively measured moderate-to-vigorous physical activity levels in weight loss maintainers and normal-weight individuals. J Obes, 2012. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Burke LE, Sereika SM, Music E, Warziski M, Styn MA, & Stone A (2008). Using instrumented paper diaries to document self-monitoring patterns in weight loss. Contemp Clin Trials, 29(2), 182–193. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Burke LE, Wang J, & Sevick MA (2011). Self-monitoring in weight loss: a systematic review of the literature. Journal of the American Dietetic Association, 111(1), 92–102. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. DerSarkissian M, Bhak RH, Huang J, Buchs S, Vekeman F, Smolarz BG, Brett J, Ganguly R, & Duh MS (2017). Maintenance of weight loss or stability in subjects with obesity: a retrospective longitudinal analysis of a real-world population. Current medical research and opinion, 33(6), 1105–1110. [DOI] [PubMed] [Google Scholar]
  9. Forman EM, & Butryn ML (2015). A new look at the science of weight control: how acceptance and commitment strategies can address the challenge of self-regulation. Appetite, 84, 171–180. 10.1016/j.appet.2014.10.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Fredrickson BL (2013). Positive emotions broaden and build. In Advances in experimental social psychology (Vol. 47, pp. 1–53). Elsevier. [Google Scholar]
  11. Goldstein SP, Goldstein CM, Bond DS, Raynor HA, Wing RR, & Thomas JG (2019). Associations between self-monitoring and weight change in behavioral weight loss interventions. Health Psychology, 38(12), 1128. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Diabetes Prevention Program Research Group. (2002). Reduction in the incidence of type 2 diabetes with lifestyle intervention or metformin. The New England journal of medicine, 346(6), 393. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Look AHEAD Research Group. (2006). The Look AHEAD study: a description of the lifestyle intervention and the evidence supporting it. Obesity, 14(5), 737–752. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Hayes JF, Balantekin KN, Fitzsimmons-Craft EE, Jackson JJ, Ridolfi DR, Boeger HS, Welch RR, & Wilfley DE (2021). Greater Average Meal Planning Frequency Predicts Greater Weight Loss Outcomes in a Worksite-Based Behavioral Weight Loss Program. Annals of Behavioral Medicine, 55(1), 14–23. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Hayes JF, Tate DF, Espeland MA, LaRose JG, Gorin AA, Lewis CE, Jelalian E, Bahnson J, Wing RR, & Group S. o. N. A. t. W. G. P. R. (2021). Patterns of weight change in a weight gain prevention study for young adults. Obesity, 29(11), 1848–1856. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Iso-Ahola SE, & Dotson CO (2016). Psychological momentum—A key to continued success. Frontiers in psychology, 7, 1328. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Jensen MD, Ryan DH, Apovian CM, Ard JD, Comuzzie AG, Donato KA, Hu FB, Hubbard VS, Jakicic JM, & Kushner RF (2014). 2013 AHA/ACC/TOS guideline for the management of overweight and obesity in adults: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines and The Obesity Society. Journal of the American college of cardiology, 63(25 Part B), 2985–3023. [DOI] [PubMed] [Google Scholar]
  18. Klem ML, Wing RR, Lang W, McGuire MT, & Hill JO (2000). Does weight loss maintenance become easier over time? Obesity research, 8(6), 438–444. [DOI] [PubMed] [Google Scholar]
  19. Larimer ME, Palmer RS, & Marlatt GA (1999). Relapse prevention: An overview of Marlatt’s cognitive-behavioral model. Alcohol Research & Health, 23(2), 151. [PMC free article] [PubMed] [Google Scholar]
  20. Latner JD, McLeod G, O’Brien KS, & Johnston L (2013). The role of self-efficacy, coping, and lapses in weight maintenance. Eating and Weight Disorders-Studies on Anorexia, Bulimia and Obesity, 18(4), 359–366. [DOI] [PubMed] [Google Scholar]
  21. Leahey TM, Fava JL, Seiden A, Fernandes D, Doyle C, Kent K, La Rue M, Mitchell M, & Wing RR (2016). A randomized controlled trial testing an Internet delivered cost–benefit approach to weight loss maintenance. Preventive medicine, 92, 51–57. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Levy RL, Jeffery RW, Langer SL, Graham DJ, Welsh EM, Flood AP, Jaeb MA, Laqua PS, Finch EA, & Hotop AM (2010). Maintenance-tailored therapy vs. standard behavior therapy for 30-month maintenance of weight loss. Preventive medicine, 51(6), 457–459. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Linde JA, Jeffery RW, French SA, Pronk NP, & Boyle RG (2005). Self-weighing in weight gain prevention and weight loss trials. Annals of Behavioral Medicine, 30(3), 210–216. [DOI] [PubMed] [Google Scholar]
  24. Loader C, & Liaw MA (2020). Package ‘locfit’. In.
  25. MacLean PS, Wing RR, Davidson T, Epstein L, Goodpaster B, Hall KD, Levin BE, Perri MG, Rolls BJ, & Rosenbaum M (2015). NIH working group report: innovative research to improve maintenance of weight loss. Obesity, 23(1), 7–15. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Mata J, Silva MN, Vieira PN, Carraça EV, Andrade AM, Coutinho SR, Sardinha LB, & Teixeira PJ (2011). Motivational “spill-over” during weight control: Increased self-determination and exercise intrinsic motivation predict eating self-regulation. [DOI] [PubMed]
  27. McGuire MT, Wing RR, Klem ML, Lang W, & Hill JO (1999). What predicts weight regain in a group of successful weight losers? Journal of consulting and clinical psychology, 67(2), 177. http://psycnet.apa.org/journals/ccp/67/2/177/ [DOI] [PubMed] [Google Scholar]
  28. Molenberghs G, & Kenward M (2007). Missing data in clinical studies. John Wiley & Sons. [Google Scholar]
  29. Nahum-Shani I, Smith SN, Spring BJ, Collins LM, Witkiewitz K, Tewari A, & Murphy SA (2018). Just-in-time adaptive interventions (JITAIs) in mobile health: key components and design principles for ongoing health behavior support. Annals of Behavioral Medicine, 52(6), 446–462. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Nezu AM, & D’Zurilla TJ (2006). Problem-solving therapy: A positive approach to clinical intervention. Springer Publishing Company. [Google Scholar]
  31. Nordmo M, Danielsen YS, & Nordmo M (2020). The challenge of keeping it off, a descriptive systematic review of high-quality, follow-up studies of obesity treatments. Obesity reviews, 21(1), e12949. [DOI] [PubMed] [Google Scholar]
  32. Pebley K, Klesges R, Talcott G, Kocak M, & Krukowski R (2019). E-scale measurements in comparison to clinic weight measurements. Obesity, 27(7), 1107–1114. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Perneger TV (1998). What’s wrong with Bonferroni adjustments. BMJ, 316(7139), 1236–1238. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Perri MG, Nezu AM, McKelvey WF, Shermer RL, Renjilian DA, & Viegener BJ (2001). Relapse prevention training and problem-solving therapy in the long-term management of obesity. Journal of consulting and clinical psychology, 69(4), 722. [PubMed] [Google Scholar]
  35. Phelan S, Hill JO, Lang W, Dibello JR, & Wing RR (2003). Recovery from relapse among successful weight maintainers. The American journal of clinical nutrition, 78(6), 1079–1084. [DOI] [PubMed] [Google Scholar]
  36. Prince SA, Adamo KB, Hamel ME, Hardt J, Gorber SC, & Tremblay M (2008). A comparison of direct versus self-report measures for assessing physical activity in adults: a systematic review. International Journal of Behavioral Nutrition and Physical Activity, 5(1), 1–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Ross KM, Qiu P, You L, & Wing RR (2018). Characterizing the pattern of weight loss and regain in adults enrolled in a 12-week internet-based weight management program. Obesity, 26(2), 318–323. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Ross KM, Qiu P, You L, & Wing RR (2019). Week-to-week predictors of weight loss and regain. Health Psychology, 38(12), 1150. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Ross KM, & Wing RR (2016a). Implementation of an internet weight loss program in a worksite setting. J Obes, 2016. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Ross KM, & Wing RR (2016b). Concordance of in-home ‘smart’scale measurement with body weight measured in-person. Obesity science & practice, 2(2), 224–228. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Shin T, Davison ML, & Long JD (2017). Maximum likelihood versus multiple imputation for missing data in small longitudinal samples with nonnormality. Psychological methods, 22(3), 426. [DOI] [PubMed] [Google Scholar]
  42. Stevens J, Truesdale KP, McClain JE, & Cai J (2006). The definition of weight maintenance. International Journal of Obesity, 30(3), 391–399. [DOI] [PubMed] [Google Scholar]
  43. Sylvester BD, Jackson B, & Beauchamp MR (2018). The effects of variety and novelty on physical activity and healthy nutritional behaviors. In Advances in motivation science (Vol. 5, pp. 169–202). Elsevier. [Google Scholar]
  44. Tanenbaum ML, Ross KM, & Wing RR (2016). Overeat today, skip the scale tomorrow: An examination of caloric intake predicting nonadherence to daily self-weighing. Obesity, 24(11), 2341–2343. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. R Core Team. (2013). R: A language and environment for statistical computing. In R Foundation for Statistical Computing. http://www.R-project.org/ [Google Scholar]
  46. Thomas JG, Bond DS, Phelan S, Hill JO, & Wing RR (2014). Weight-loss maintenance for 10 years in the National Weight Control Registry. American journal of preventive medicine, 46(1), 17–23. [DOI] [PubMed] [Google Scholar]
  47. Yu Z, Sealey-Potts C, & Rodriguez J (2015). Dietary self-monitoring in weight management: current evidence on efficacy and adherence. Journal of the Academy of Nutrition and Dietetics, 115(12), 1931–1933. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplemental Material
NIHMS1856092-supplement-Supplemental_Material.pdf (53.2KB, pdf)

RESOURCES