Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2020 Dec 1.
Published in final edited form as: Health Psychol. 2019 Sep 26;38(12):1128–1136. doi: 10.1037/hea0000800

Associations Between Self-Monitoring and Weight Change in Behavioral Weight Loss Interventions

Stephanie P Goldstein 1, Carly M Goldstein 1, Dale S Bond 1, Hollie A Raynor 2, Rena R Wing 1, J Graham Thomas 1
PMCID: PMC6861632  NIHMSID: NIHMS1051522  PMID: 31556659

Abstract

Objective:

The current study is a secondary analysis of the Live SMART trial, a randomized controlled trial comparing behavioral weight loss (BWL) delivered via smartphone (SMART) to group-based BWL (GROUP) and a control condition (CONTROL). Given the established importance of self-monitoring for weight loss, the aims were to evaluate bidirectional associations between adherence to self-monitoring and weight change and to examine the moderating effect of treatment condition on these associations.

Methods:

Adults with overweight/obesity (n=276; 83% women; 92.8% White; Mage= 55.1 years; MBMI=35.2 kg/m2) were instructed to self-monitor dietary intake, daily weight, and physical activity minutes via paper diaries in GROUP and CONTROL, and via a smartphone application in SMART. All participants were weighed monthly at the research center. Adherence to self-monitoring was assessed via examination of self-monitoring records.

Results:

Generalized linear mixed models revealed that adherence to self-monitoring of dietary intake, self-weighing, and physical activity for each month was associated with weight change throughout that month, such that increased frequency of self-monitoring led to greater weight loss (ps<.001). For the GROUP condition only, poorer weight losses in one month were prospectively associated with poor adherence to self-monitoring the following month (ps≤.01).

Conclusions:

Results provide evidence of a bidirectional association between self-monitoring and weight change. Better self-monitoring was consistently associated with better weight loss across intervention and tracking modalities. Poorer weight loss was prospectively associated with poorer self-monitoring in group treatment, suggesting that social influences could drive adherence in this form of treatment.

TRIAL REGISTRATION:

ClinicalTrials.gov identifier:

Keywords: weight loss, self-monitoring, mobile health, adherence

Background

Obesity remains a major public health concern and behavioral weight loss (BWL) is the gold-standard non-surgical treatment for weight loss (Hales, Fryar, Carroll, Freedman, & Ogden, 2018; Wing et al., 2011). A cornerstone strategy of BWL intervention is self-monitoring of daily dietary intake, physical activity, and weight to track adherence to BWL recommendations and associated progress (Foster, Makris, & Bailer, 2005). Self-monitoring is thought to facilitate the necessary diet and activity changes for weight loss through increased accountability, and enhanced awareness of one’s own actions and how these actions align with a longer-term goal (e.g., weight loss; Bandura, 1998; Khaylis, Yiaslas, Bergstrom, & Gore-Felton, 2010).

For self-monitoring to impact the target behavior, one must record the relevant information in a timely, truthful, and consistent manner (Bandura, 1998). Adherence to self-monitoring weight-related behaviors (i.e., exercise, weight, diet), as defined by the frequency of recommended recording, is associated with greater weight losses in BWL interventions (Burke, Wang, & Sevick, 2011; Madigan, Daley, Lewis, Aveyard, & Jolly, 2015). Unfortunately, research among BWL participants revealed that burden and boredom are barriers to accurate and consistent self-monitoring (Burke, Swigart, Warziski Turk, Derro, & Ewing, 2009). Therefore, tools and techniques that facilitate improved self-monitoring of dietary intake, weight, and physical activity are necessary to improve weight loss and weight maintenance outcomes (Borg, Fogelholm, & Kukkonen-Harjula, 2004; Burke, Styn, et al., 2009; O’Neil, 2001).

One way to improve self-monitoring is to make it less burdensome via technology with access to large food databases (including brand names and restaurant items) and physical activities (Burke, Wang, et al., 2011). These programs, accessible online and via mobile application (“app”), quickly calculate calories and macronutrients consumed and track progress towards a specific calorie/exercise goal. Many programs allow for saved meals, snacks, and exercise routines to increase ease of monitoring, which enhances long-term use and adherence (Peng, Kanthawala, Yuan, & Hussain, 2016).

Electronic self-monitoring tools, as compared to paper-and-pencil tracking, have produced increased adherence to self-monitoring within BWL (Carter, Burley, Nykjaer, & Cade, 2013; Spring et al., 2013; Thomas et al., 2017) and this should presumably lead to greater weight losses (Turk et al., 2013; Wang et al., 2012). If so, electronic self-monitoring could enhance the effectiveness of technology-based BWL interventions to match that of traditional in-person treatments (Tate, Wing, & Winett, 2001). However virtually no studies have directly compared primarily technology-based approaches (i.e., online BWL interventions with electronic self-monitoring) to gold-standard group-based BWL with paper-and-pencil records. As such, comparing how self-monitoring behavior is related to weight change across treatment modalities is critical to understanding the role of mobile health in enhancing obesity treatment.

The recent randomized controlled Live SMART trial (Thomas, Bond, Raynor, Papandonatos, & Wing, 2019), was one of the first investigations to directly assess if smartphone-based BWL intervention could produce weight losses at least as good as gold-standard group treatment, and compare both treatments to a control condition. The trial included 276 adults with overweight/obesity who participated in one of three 18-month BWL treatments (group-based BWL, smartphone-based, BWL, and control). Group-based treatment consisted of meetings (6 months weekly followed by 6 months of biweekly and then 6 months of monthly treatment) and paper-and-pencil self-monitoring with written feedback; smartphone-based treatment consisted of online lessons and electronic self-monitoring with automated feedback and a monthly clinic weigh-in; the control condition consisted of self-monitoring via paper diaries with written feedback and a monthly clinic weigh-in.

Equivalence testing revealed no significant differences in 18-month weight loss across smartphone-based BWL (5.5 kg; 95% CI [3.9, 7.0]), in-person BWL (5.9 kg; 95% CI [4.5, 7.4]), and the control (6.4 kg; 95% CI [3.7, 9.2]), indicating that low-cost technology-based programs can produce equivalent results to intensive in-person programs. Control outperformed expectations, reinforcing the conclusion that self-monitoring is a powerful tool for weight loss, even in the absence of other intervention components. Participants in smartphone-based BWL using electronic self-monitoring more frequently tracked their dietary intake (37.9% days monitored; 95% CI [32.6, 43.2]) than group-based BWL with paper-and-pencil self-monitoring (27.5%; 95% CI [23.6, 31.4]); neither differed significantly from the control condition (32.0%; 95% CI [24.0, 40.0]). Self-weighing rates were also lower in the group-based BWL (21.2%; 95% CI [17.9, 25.9]) compared to smartphone-based BWL (30.7%; 95% CI [26.2, 37.2] and control [29.7%; 95% CI [21.7, 37.7]. Self-monitored physical activity was not significantly different across groups. On average, the smartphone-based BWL groupengaged in more self-monitoring, but between-groups weight losses did not differ.

Importantly, the direct association between self-monitoring and weight change, and the moderating effect of treatment condition, has not been evaluated in Live SMART. Moreover, prior research has largely assumed a unidirectional relationship between self-monitoring and weight change, but a full understanding of self-monitoring behavior within BWL may require studying the bidirectional prospective associations (Carels et al., 2005). For example, short-term weight change could impact subsequent adherence to self-monitoring via emotions and self-perceptions (Burke, Swigart, et al., 2009; Tanenbaum, Ross, & Wing, 2016), and/or effects on motivation (Webber, Tate, Ward, & Bowling, 2010). Ultimately, the literature would benefit from evaluating the directionality of the month-to-month association between weight change and self-monitoring, understanding this association across self-monitoring methods within different treatment modalities, and examining self-monitoring of multiple weight control behaviors. As such, examining the bidirectional temporal associations between weight change and self-monitoring across conditions in Live SMART is critical to understanding the potential mechanisms by which different types of BWL interventions (i.e., smartphone or in-person) impact weight change.

Aims

Given the importance of self-monitoring for success in BWL interventions and the paucity of literature on the potential bidirectional association between weight change and self-monitoring, the current study involves a secondary analysis of data from the Live SMART trial to extend previous group-level analyses by examining temporal patterns within subjects. The current study has the following aims: Aim 1) Evaluate the longitudinal association between monthly self-monitoring adherence and percent weight loss (%WL); Aim 2) Evaluate the prospective association between monthly %WL and the subsequent month of self-monitoring adherence; and Aim 3) Evaluate if the associations in Aims 1 and 2 are moderated by treatment condition. These aims were examined across adherence to self-monitoring of daily dietary intake, weighing, and frequency of physical activity. Based on prior literature (Burke, Wang, et al., 2011), fewer days of self-monitoring during the month was hypothesized to predict poorer %WL. Since negative self-perceptions have been linked with poor adherence to self-monitoring in self-directed weight loss attempts (Burke, Swigart, et al., 2009), poorer %WL was hypothesized to predict poor self-monitoring adherence the following month; this effect was anticipated to be especially pronounced within the control group (which is most akin to a self-directed weight loss attempt).

Methods

Participants

Participants were 276 adults (aged 18–70) with overweight/obesity (body mass index of 25–45 kg/m2) who were willing to attend in-person treatment sessions and use electronic resources for weight loss. Exclusion criteria were: currently enrolled in another weight loss program; taking medication for weight loss; self-reporting weight loss of ≥ 5% of body weight during the past 6 months; currently pregnant, lactating, < 6 months postpartum, or plans to become pregnant over study (18 months); self-reported heart condition, chest pain during physical activity or rest, or loss of consciousness; self-reported medical condition that may affect the safety of participating in unsupervised physical activity (Thomas, Reading, & Shephard, 1992); inability to walk 2 blocks without stopping; and report of conditions that may impact ability to follow protocol (e.g., terminal illness, plans to relocate out of state, history of substance abuse, or other uncontrolled or untreated psychiatric problem).

Procedures

An in-depth description of the parent trial can be found in the Live SMART primary outcomes paper (Thomas et al., 2019). Recruitment and enrollment for Live SMART occurred in four waves from January 2013 to January 2015. Interested individuals were screened via telephone for initial eligibility, and those who appeared eligible were invited to attend an in-person orientation session. The orientation session served as a means for confirming eligibility, completing informed consent procedures, and beginning the baseline assessment. Participants were randomized via computer algorithm utilizing blocks of five and stratified by gender and cohort. The algorithm allocated to three study arms in a 2:2:1 ratio: 1) group-based BWL; 2) smartphone-based BWL; and 3) a control condition. Outcomes of interest for the current secondary data analysis were clinic visit weights and frequency of self-monitoring daily weight, dietary intake, and physical activity. The study protocol and informed consent procedures were approved by The Miriam Hospital Institutional Review Board (IRB). Live SMART was conducted at the Weight Control and Diabetes Research Center of The Miriam Hospital and Brown University in Providence, RI, USA.

Interventions

Participants were prescribed a weight loss goal of 10% of their current body weight (Wadden & Foster, 2000). They were instructed to adhere to a 1200–1800 kcal/day diet (depending on initial body weight) and gradually worked towards a goal of 200 minutes of moderate-to-vigorous physical activity weekly. They were instructed to self-monitor daily dietary intake (i.e., calorie and fat content of items consumed), physical activity (i.e., minutes when activity was performed), and body weight. Self-monitoring recommendations were consistent throughout the 18-month intervention. Study conditions are described below: group-based BWL with self-monitoring and feedback using paper diaries (GROUP), BWL delivered via smartphone with electronic self-monitoring and feedback (SMART), and a control condition with self-monitoring and feedback using paper diaries (CONTROL).

Participants assigned to group-based BWL (GROUP) attended in-person 60-minute sessions, which met weekly for 6 months, bi-weekly for the following 6 months, and monthly for the final 6 months. Format and content of the group meetings closely followed other empirically-supported standard behavioral obesity treatments (Diabetes Prevention Program Research Group, 2002; Look AHEAD Research Group, 2006). In addition to daily self-monitoring of diet, physical activity, and weight, participants were taught behavioral skills such as stimulus control, meal planning, and problem solving to facilitate adherence and to address common barriers to weight loss (Alamuddin & Wadden, 2016). Participants self-monitored daily dietary intake, physical activity, and weight using paper diaries and a nutritional reference book for calorie and fat content of foods consumed. Participants submitted their paper diaries at their group sessions and personalized written feedback was provided the following session. If a group session was missed, then participants were encouraged to submit all paper diaries at the subsequent session.

The smartphone-based BWL condition (SMART) received similar content to those in group-based treatment, delivered via 5-minute skill training videos. The videos were made available 3 times weekly for 6 months, biweekly for 6 months, and weekly for the remaining 6 months. Videos were delivered via an app for smartphones and all released videos were available for the trial duration. Participants were asked to use MyFitnessPal, a free commercially-available app, to self-monitor daily dietary intake, physical activity, and weight. These records were retrieved remotely by the treatment team. Participants received personalized feedback within the study app. This feedback was kept consistent with the GROUP feedback in frequency, length, and content. Participants were weighed monthly at the clinic by study staff. Participants assigned to SMART who did not own a smartphone were lent one for the trial period.

Participants in the control condition (CONTROL) attended monthly weigh-ins and were provided printed information regarding the benefits of weight loss, healthy eating, and regular physical activity. Participants self-monitored daily weight, dietary intake, and physical activity using paper diaries and nutritional reference books (as in the group-based condition). Participants mailed in paper diaries or submitted diaries in at their monthly weigh-ins and received personalized written feedback via mail. Feedback was similar to GROUP and SMART feedback regarding frequency, length, and content.

Measures

Weight was measured to the nearest 0.1 pounds (lbs.) on a calibrated scale in light clothing, without shoes. Weights used in the current study were measured at each clinic visit by study staff. Frequency of self-monitoring dietary intake, weight, and physical activity was automatically recorded in the MyFitnessPal app for those assigned to SMART and recorded by study staff in the GROUP and CONTROL conditions. Adherence to self-monitoring of dietary intake was assessed as the number of days per week that participants recorded at least 50% of the daily calorie goal or recorded at least 3 separate eating occasions (Burke et al., 2012; Thomas & Wing, 2013). Adherence to self-weighing was assessed as the number of days per week that a participant recorded a weight. Frequency of self-monitoring physical activity was measured as the number of days per week in which participants reported any physical activity minutes. Participants who did not attend clinic visits or did not turn in a record were assumed to have zero days of self-monitoring. Participant characteristics were collected via questionnaire at baseline, including gender, age, race, ethnicity, education, and income category.

Statistical Approach

Data were analyzed with R studio version 3.5.1 (RStudio Team, 2015). Descriptive statistics included means and standard deviations of self-monitoring behavior and weight across months of participation. Monthly adherence to self-monitoring daily dietary intake, self-weighing, and frequency of physical activity was calculated by summing the adherent days per week of each month (defined as previous 30-day period). Monthly clinic visit weights were used for analysis in the current study because this was the smallest unit that was measured uniformly across study conditions (participants in the CONTROL and SMART conditions were not instructed to visit the clinic for weekly weights as in the first 6 months of GROUP condition). Missing weights were not imputed because the proposed analyses accommodate multiple missing data points by including a random effect of subject (Edwards, 2000; Krueger & Tian, 2004). All models controlled for body mass index, age, sex, income, education, and race/ethnicity, as these variables have been correlated with weight change in previous investigations. The Live SMART trial was powered to examine proposed primary outcomes for between-groups weight change (Thomas et al., 2019).

Aim 1.

To evaluate the association between self-monitoring and monthly weight loss, linear mixed models with random effects of time (month) and participant were employed using the “lme4” package. Independent variables were condition, self-monitoring adherence, and the interaction between condition and adherence. Models were first calculated for each separate self-monitoring behavior (i.e., dietary intake, weight, and physical activity frequency). To account for underlying characteristics that may govern the general tendency to adhere to self-monitoring recommendations (i.e., motivation), we then controlled for the other self-monitoring behaviors (e.g., the effect of adherence to self-monitoring dietary intake controlling for the effect of self-monitored physical activity and weight). The dependent variable was calculated as %WL during each month of self-monitoring, using baseline weight as a reference point (i.e., (WeightMonth 3 – WeightMonth 2)/WeightBaseline). This procedure ensured that the within-person criterion for comparing the magnitude of % WL stayed constant over time.

Aim 2.

To evaluate the prospective association between monthly weight loss and adherence to self-monitoring, generalized linear mixed models with random effects of time (month) and participant were employed using the “glmmTMB” package. Independent variables were treatment condition, %WL from the previous month (e.g., %WL at the end of month 2 predicted self-monitoring during month 3), and the interaction between condition and %WL. Separate models were run for dependent variables representing each self-monitoring behavior (i.e., dietary intake, self-weighing, and frequency of physical activity). For adherence to self-monitoring dietary intake and weight, a negative binomial distribution with logit link function was used because these count data were positively skewed, and the variance often exceeded the mean (Atkins, Baldwin, Zheng, Gallop, & Neighbors, 2013). For frequency of self-monitoring physical activity, a zero-inflated negative binomial distribution with logit link function was used to account for multiple factors contributing to zero values (i.e., no physical activity v. non-adherence to monitoring physical activity; Atkins et al., 2013).

Aim 3.

To evaluate whether the bidirectional associations between self-monitoring and weight change tested in Aim 1 and Aim 2 differed by treatment condition, the models were reevaluated with treatment condition added as a moderator.

Results

Descriptive Information

Participants were randomized to GROUP (n=106), SMART (n=114) and CONTROL (n=56). The full CONSORT diagram, participant characteristics, and average weight losses over time can be found in the primary outcomes paper (Thomas et al., 2019). Participants were mostly female (83%), middle-aged (M=55.1, SD=9.9), and white Non-Hispanic (92.8%). At baseline, participants’ average body mass index was 35.2 kg/m2 (SD=5.0). Figure 1 illustrates the proportion of participants with available %WL and self-monitoring data across study months. Table 1 depicts the pattern of monthly adherence to self-monitoring dietary intake, weight, and physical activity frequency, as well as average % WL.

Figure 1.

Figure 1.

Open in a new tab

Percentage of Missing Data: Self-monitoring and Percent Weight Loss

Notes: Missing self-monitoring data are representative of records with no days of monitoring or records not turned in; Percent weight loss is representative of attendance at clinic visits (i.e., a value is missing if the participant did not attend the current and/or previous clinic visit for that particular month); Adherence to self-monitoring across conditions can be found in the primary outcomes paper (Thomas et al., 2019).

Table 1.

Days of completed self-monitoring and percent weight change across study months

CONTROL SMART GROUP
M SD M SD M SD
Diet Month 1 19.8 9.9 20.4 9.4 22.0 8.1
Month 6 11.0 11.9 13.4 11.7 10.1 10.2
Month 12 4.7 10.0 6.6 10.1 2.4 7.2
Month 18 2.4 7.3 1.8 6.2 0.9 4.7
Weight Month 1 17.4 10.5 18.1 9.8 18.1 9.9
Month 6 9.9 12 10.7 11.5 7.4 9.9
Month 12 4.5 9.6 5.8 9.7 2.1 7.1
Month 18 2.8 7.9 2.3 7.2 0.9 4.8
Physical
Activity 		Month 1 9.4 8.9 6.1 6.9 7.8 8.1
Month 6 5.2 7.5 7.3 8.7 5.9 7.5
Month 12 2.5 6.3 3.9 7.4 1.7 5.6
Month 18 1.8 5.7 1.5 5.0 0.8 4.3
Percent Weight Loss (%WL) Month 1 2.8 2.2 3.7 1.9 3.2 1.7
Month 6 7.9 5.4 9.1 5.4 10.7 5.6
Month 12 10.3 7.1 9.4 6.2 10.4 6.4
Month 18 8.6 8.5 7.9 6.3 8.8 6.8
Open in a new tab

Note: Values of percent weight loss are drawn from the monthly clinic weights used in the analyses reported in this paper, not the assessment weights used in the primary outcomes analysis reported in Thomas et al. (2019)

Self-monitoring Association with Weight Change

A main effect of adherence to self-monitoring dietary intake on %WL was observed, such that more days of self-monitoring corresponded to greater monthly weight losses during that month. However, when controlling for frequency of self-monitoring physical activity and adherence to self-weighing, the effect of self-monitoring dietary intake became non-significant. There was no interaction effect of condition × frequency of self-monitoring (p=.29) and, as such, the interaction term was removed from the final model reported in Table 2.

Table 2.

Adherence to self-monitoring of dietary intake and self-weighing predicting Percent Weight Loss (%WL)

B SE t-value p-value
Dietary Intake
Diet 0.02 0.003 6.49 <001***
SMART+ 0.06 0.09 0.62 .54
GROUP+ 0.20 0.10 2.05 .04*
Self-Weighing
Weight 0.02 0.003 6.82 <001***
SMART+ 0.08 0.09 0.83 .41
GROUP+ 0.23 0.10 2.30 .02*
Controlling for All Self-Monitoring Behaviors
Diet 0.003 0.01 0.49 .62
Weight 0.01 0.01 1.96 .05*
PA 0.01 0.01 2.72 .007**
SMART+ 0.07 0.09 0.76 .45
GROUP+ 0.21 0.10 2.15 .03*
Open in a new tab

Notes.

+

CONTROL is reference;

*

p≤.05,

**

p≤.01,

***

p≤.001

A main effect of adherence to self-monitoring weight on %WL was observed: additional days of self-weighing corresponded to more monthly weight loss. This remained statistically significant when controlling for adherence to self-monitoring dietary intake and frequency of monitoring physical activity. There was no interaction effect of condition × frequency of self-monitoring (p=.26) and, as such, the interaction term was removed from the final model. See Table 2 for final model results for self-monitoring of dietary intake and self-weighing.

An interaction effect was observed on the frequency of self-monitoring physical activity on %WL, such that more days of self-monitored physical activity was associated with greater %WL in SMART and GROUP conditions, compared to the CONTROL condition. These findings were consistent when controlling for adherence to self-monitoring dietary intake and weight. See Table 3 for final model results.

Table 3.

Adherence to self-monitoring of physical activity predicting Percent Weight Loss (%WL)

B SE t-value p-value
Physical Activity
PA 0.01 0.01 1.10 .27
SMART+ −0.09 0.11 −0.78 .44
GROUP+ −0.01 0.12 −0.11 .91
PA:SMART+ 0.02 0.01 2.14 .03*
PA:GROUP+ 0.03 0.01 2.58 .01*
Controlling for All Self-Monitoring Behaviors
PA −0.002 0.01 −0.27 .78
SMART+ −0.08 0.11 −0.70 .48
GROUP+ 0.02 0.12 0.18 .86
PA:SMART+ 0.02 0.01 2.10 .04*
PA:GROUP+ 0.03 0.01 2.57 .01*
Diet 0.002 0.01 0.42 .67
Weight 0.01 0.01 1.99 .05*
Open in a new tab

Notes.

+

CONTROL is reference;

*

p≤.05,

**

p≤.01,

***

p≤.001

Weight Change Association with Prospective Self-monitoring

When examining the prospective association between %WL and adherence to self-monitoring in the subsequent month, several interaction effects were observed. There were statistically significant interaction effects of condition and %WL on prospective adherence to self-monitoring dietary intake, B=0.16, SE=0.05, z=3.49, p<.001, adherence to self-weighing, B=0.13, SE=0.05, z=2.67, p=.01, and frequency of monitored physical activity, B=0.15, SE=0.05, z=2.83, p=.005 As seen in Figure 2, greater %WL predicted more days of self-monitoring in the subsequent month in GROUP, while there was no association in CONTROL and SMART.

Figure 2.

Figure 2.

Open in a new tab

Prospective association between adherence to self-monitoring and %WL across treatment conditions. In (a), adherence to self-monitoring dietary intake is pictured, followed by adherence to self-monitoring weight (b) and physical activity (c). Note: Greater %WL is indicative of more weight loss

Discussion

This study is one of the first to test bidirectional associations between self-monitoring and weight change in BWL. Effects were compared for BWL delivered in group treatment with self-monitoring using paper diaries (GROUP), BWL delivered via smartphone with electronic self-monitoring (SMART), and a control condition with self-monitoring using paper diaries (CONTROL). This secondary analysis of the Live SMART trial found that greater adherence to self-monitoring key weight control behaviors (i.e., dietary intake, self-weighing, and frequency of physical activity) was associated with greater %WL across treatment conditions. Conversely, greater weight loss was associated prospectively with increased adherence to self-monitoring in the next month, albeit only in the GROUP condition.

The first aim of the current study was to evaluate the association between monthly self-monitoring adherence and monthly weight change. The primary outcomes of Live SMART demonstrated that SMART engaged in more self-monitoring but did not demonstrate greater weight losses. Therefore, one might not expect an association between self-monitoring and weight loss in this sample, despite the prior literature in this area (Burke, Wang, et al., 2011). However, a fine-grained approach using intensive longitudinal modeling revealed that greater frequency of self-monitoring all three key weight control behaviors was associated with weight change from the start to end of the month. As such, this study replicates and extends the literature by examining the month-to-month association between self-monitoring and weight change over the full weight loss attempt. These novel findings illustrate that the association between self-monitoring and weight change is granular (i.e., meaningful over shorter time scales like one month) when examined over an entire long-term BWL (Riley, 2017). Thus, self-monitoring adherence may be an important target for timely intervention throughout BWL. Also, adherence to self-monitoring could be used to identify individuals who need additional intervention to prevent weight regain (Laitner, Minski, & Perri, 2016).

Per Aim 3, the current study also tested whether the magnitude and/or direction of the effect of self-monitoring on %WL differed by condition. One such difference was found; greater physical activity self-monitoring frequency was associated with more %WL in the GROUP and SMART conditions relative to CONTROL. However, the frequency of physical activity self-monitoring was confounded with performance of physical activity (since participants were not asked to record an entry of 0 physical activity minutes on days that none was performed). GROUP and SMART received active BWL, with increasing exercise goals over time. CONTROL did not receive such a physical activity intervention. Thus, this finding is consistent with more GROUP and SMART participants engaging in physical activity, which contributed to increased weight loss for those participants (Jeffery, Wing, Sherwood, & Tate, 2003).

Since self-monitoring behaviors are jointly prescribed in BWL, this study examined the effect of self-monitoring each behavior (e.g., diet) while controlling for adherence to self-monitoring the other behaviors (e.g., self-weighing, physical activity). While the pattern of findings remained the same for self-monitoring weight and physical activity, the effect of self-monitoring dietary intake on monthly weight losses became non-significant while controlling for the frequency of monitoring weight and physical activity. This result is counterintuitive given the established impact of self-monitoring eating behavior on weight change (Burke, Wang, et al., 2011) and is representative of the limitations inherent in assessing adherence to self-monitoring health behaviors that are closely related. For example, it is likely that adherence behaviors have shared variance and clustering given that they are behavioral representations of similar constructs such as motivation (Webber et al., 2010).

The second aim tested the hypotheses that greater %WL in the preceding month would predict higher frequency of self-monitoring in the month that followed; per aim 3, treatment condition was also tested as a moderator of this association. Contrary to hypotheses, this association was found only in the GROUP condition, though it was present for all three behaviors. Compared to the other two conditions, GROUP was characterized by more frequent in-person sessions, which were more intensive than the SMART and CONTROL weigh-ins. This difference in intervention frequency and intensity likely contributed to the different pattern of association in GROUP versus SMART and CONTROL. For example, the GROUP format may have led patients with good weight loss outcomes to receive more frequent or intense praise or other reinforcement for self-monitoring adherence (Bandura, 2011). Conversely, GROUP participants with poor weight loss might have experienced more frequent and/or intense feelings of shame due to concerns about fellow participants’ and interventionists’ perceptions. This may have led participants to avoid self-monitoring, which might remind them of poor outcomes and associated unpleasant emotions (Tanenbaum et al., 2016). Although the current finding requires replication, it suggests that interventionists delivering in-person, group-based treatments should be aware that minor (month-to-month) weight loss failures can have problematic consequences (Schumacher et al., 2016). Future research may evaluate evidence-based strategies specific to coping with weight loss failure during BWL treatment. Further, low-contact interventions may consider attempting to capitalize on positive social supports to facilitate continued self-monitoring after successful weight loss (Payne, Lister, West, & Bernhardt, 2015).

The current study benefited from repeated measurement of weight and adherence to self-monitoring, which allowed for a granular examination of the association between self-monitoring and weight change. This is the first study to provide evidence that the association between self-monitoring and weight change could be bidirectional, particularly in intensive group-based treatment. Lastly, examining the differences in patterns of associations across treatment and tracking modalities is a major strength. The results of the current study can inform further refinement of both in-person and smartphone-based BWL interventions.

There were also numerous limitations in the current study. Consistent with previous trials (Burke et al., 2012), adherence to self-monitoring and attendance at clinic weigh-ins decreased during the maintenance phase. Importantly, the chosen analytic approach, linear mixed models, is robust to missing data. Nonetheless, conclusions should be interpreted with caution, as the missing data are likely confounded with an overall pattern of disengagement (Scherer, Ben-Zeev, Li, & Kane, 2017) that could also be related to weight regain observed during the maintenance period (Thomas et al., 2019). Further, assuming that participants who did not submit records conducted no self-monitoring could have introduced some degree of bias, especially for the GROUP and CONTROL conditions who were required to submit their records in-person or via mail, as opposed to the records of SMART, which were retrieved electronically and thus required no effort to submit. These same research questions should be reexamined in future studies that focus on promoting adherence to self-monitoring and clinic weigh-ins during maintenance. The sample was mostly white, middle-aged, and female, but obesity affects people of all backgrounds (Wang & Beydoun, 2007). Finally, adherence to self-monitoring physical activity was confounded with actual performance of physical activity, which represents an interesting distinction between self-monitoring used for clinical purposes as opposed to an outcome of interest within research. While zero-inflated models can be used to account for behaviors that are only monitored when they occur (as opposed to also noting when they are not occurring), future research on self-monitoring adherence within a clinical context may provide participants with more specific monitoring instructions (e.g., to also monitor when they are not engaging in a behavior).

Future studies should consider replicating findings of the current project given that it was a secondary analysis of data from a larger trial. Future research would benefit from exploring factors that promote adherence to self-monitoring given its relationship with weight loss. For example, it would be valuable to investigate ways various interventions and tracking modalities can be used to capitalize on motivation to engage in self-monitoring and other key weight control behaviors. One method could be matching individuals to preferred tracking tools or intervention modalities (Turner-McGrievy et al., 2013), and another could be using passive sensing technologies to ease the burden of self-monitoring. Given previous studies showing that merely adding tracking tools to BWL may be of no benefit, or even have iatrogenic effects, use of these tools would have to be accompanied by a careful intervention strategy to guide participants in summarizing and interpreting their data to inform behavior change (Jakicic et al., 2016; Thomas et al., 2017). This study demonstrated that treatment and tracking modality did not result in differential patterns of association between self-monitoring and weight change, but future research could explore factors that may impact the strength of the association. Such studies could identify individuals who may need additional supports to enhance the impact of self-monitoring on weight loss. Overall, results of the current study indicate that continuing to assess and intervene on adherence to self-monitoring would be a useful endeavor for enhancing weight loss and weight maintenance in several types of BWL programs.

Funding:

National Institute of Diabetes and Digestive and Kidney Diseases

Footnotes

Disclosure: The authors declare no conflict of interest.

References

  1. Alamuddin N, & Wadden TA (2016). Behavioral treatment of the patient with obesity. Endocrinology and Metabolism Clinics of North America, 45(3), 565–580. doi: 10.1016/j.ecl.2016.04.008 [DOI] [PubMed] [Google Scholar]
  2. Atkins DC, Baldwin SA, Zheng C, Gallop RJ, & Neighbors C (2013). A tutorial on count regression and zero-altered count models for longitudinal substance use data. Psychology of Addictive Behaviors, 27(1), 166. doi: 10.1037/a0029508 [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Bandura A (1998). Health promotion from the perspective of social cognitive theory. Psychology and Health, 13(4), 623–649. doi: 10.1080/08870449808407422 [DOI] [Google Scholar]
  4. Bandura A (2011). Social Cognitive Theory (Vol. 2012). [Google Scholar]
  5. Borg P, Fogelholm M, & Kukkonen-Harjula K (2004). Food selection and eating behaviour during weight maintenance intervention and 2-y follow-up in obese men. International Journal of Obesity, 28(12), 1548. doi: 10.1038/sj.ijo.0802790 [DOI] [PubMed] [Google Scholar]
  6. Burke LE, Styn MA, Glanz K, Ewing LJ, Elci OU, Conroy MB,… Keating AL (2009). SMART trial: A randomized clinical trial of self-monitoring in behavioral weight management-design and baseline findings. Contemporary Clinical Trials, 30(6), 540–551. doi: 10.1016/j.cct.2009.07.003 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Burke LE, Styn MA, Sereika SM, Conroy MB, Ye L, Glanz K, Ewing LJ (2012). Using mHealth technology to enhance self-monitoring for weight loss: a randomized trial. American Journal of Preventive Medicine, 43(1), 20–26. doi: 10.1016/j.amepre.2012.03.016 [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Burke LE, Swigart V, Warziski Turk M, Derro N, & Ewing LJ (2009). Experiences of self-monitoring: successes and struggles during treatment for weight loss. Qualitative Health Research, 19(6), 815–828. doi: 10.1177/1049732309335395 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. 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: 10.1016/j.jada.2010.10.008 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Carels RA, Darby LA, Rydin S, Douglass OM, Cacciapaglia HM, & O’brien WH (2005). The relationship between self-monitoring, outcome expectancies, difficulties with eating and exercise, and physical activity and weight loss treatment outcomes. Annals of Behavioral Medicine, 30(3), 182–190. doi: 10.1207/s15324796abm3003_2 [DOI] [PubMed] [Google Scholar]
  11. Carter MC, Burley VJ, Nykjaer C, & Cade JE (2013). Adherence to a smartphone application for weight loss compared to website and paper diary: pilot randomized controlled trial. Journal of Medical Internet Research, 15(4), e32. doi: 10.2196/jmir.2283 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Diabetes Prevention Program Research Group. (2002). The Diabetes Prevention Program (DPP): description of lifestyle intervention. Diabetes Care, 25(12), 2165–2171. doi: 10.2337/diacare.25.12.2165 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Edwards LJ (2000). Modern statistical techniques for the analysis of longitudinal data in biomedical research. Pediatric Pulmonology, 30(4), 330–344. doi: [DOI] [PubMed] [Google Scholar]
  14. Foster GD, Makris AP, & Bailer BA (2005). Behavioral treatment of obesity. The American Journal of Clinical Nutrition, 82(1), 230S–235S. doi: 10.1093/ajcn.82.1.230S [DOI] [PubMed] [Google Scholar]
  15. Hales CM, Fryar CD, Carroll MD, Freedman DS, & Ogden CL (2018). Trends in obesity and severe obesity prevalence in US youth and adults by sex and age, 2007–2008 to 2015–2016. Journal of the American Medical Association, 319(16), 1723–1725. doi: 10.1001/jama.2018.3060 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Jakicic JM, Davis KK, Rogers RJ, King WC, Marcus MD, Helsel D,… Belle SH (2016). Effect of wearable technology combined with a lifestyle intervention on long-term weight loss: the IDEA randomized clinical trial. Journal of the American Medical Association, 316(11), 1161–1171. doi: 10.1097/01.ogx.0000511086.69375.3b [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Jeffery RW, Wing RR, Sherwood NE, & Tate DF (2003). Physical activity and weight loss: does prescribing higher physical activity goals improve outcome? The American Journal of Clinical Nutrition, 78(4), 684–689. doi: 10.1093/ajcn/78.4.684 [DOI] [PubMed] [Google Scholar]
  18. Khaylis A, Yiaslas T, Bergstrom J, & Gore-Felton C (2010). A review of efficacious technology-based weight-loss interventions: five key components. Telemedicine and e-Health, 16(9), 931–938. doi: 10.1089/tmj.2010.0065 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Krueger C, & Tian L (2004). A comparison of the general linear mixed model and repeated measures ANOVA using a dataset with multiple missing data points. Biological Research for Nursing, 6(2), 151–157. doi: 10.1177/1099800404267682 [DOI] [PubMed] [Google Scholar]
  20. Laitner MH, Minski SA, & Perri MG (2016). The role of self-monitoring in the maintenance of weight loss success. Eating Behaviors, 21, 193–197. doi: 10.1016/j.eatbeh.2016.03.005 [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. 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: 10.1038/oby.2006.84 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Madigan CD, Daley AJ, Lewis AL, Aveyard P, & Jolly K (2015). Is self-weighing an effective tool for weight loss: a systematic literature review and meta-analysis. International Journal of Behavioral Nutrition and Physical Activity, 12(1), 104. doi: 10.1186/s12966-015-0267-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. O’Neil PM (2001). Assessing dietary intake in the management of obesity. Obesity Research, 9, 361S–366S. doi: 10.1002/oby.2001.9.s5.361 [DOI] [PubMed] [Google Scholar]
  24. Payne HE, Lister C, West JH, & Bernhardt JM (2015). Behavioral functionality of mobile apps in health interventions: a systematic review of the literature. JMIR mHealth and uHealth, 3(1), e20. doi: 10.2196/mhealth.3335 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Peng W, Kanthawala S, Yuan S, & Hussain SA (2016). A qualitative study of user perceptions of mobile health apps. BMC Public Health, 16(1), 1158. doi: 10.1186/s12889-016-3808-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Riley WT (2017). Behavioral and social sciences at the National Institutes of Health: Methods, measures, and data infrastructures as a scientific priority. Health Psychology, 36(1), 5. doi: 10.1037/hea0000464 [DOI] [PubMed] [Google Scholar]
  27. Team RStudio. (2015). RStudio: Integrated Development for R. Boston, MA: RStudio, Inc; Retrieved from http://www.rstudio.com/ [Google Scholar]
  28. Scherer EA, Ben-Zeev D, Li Z, & Kane JM (2017). Analyzing mHealth engagement: joint models for intensively collected user engagement data. JMIR mHealth and uHealth, 5(1). doi: 10.2196/mhealth.6474 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Schumacher LM, Gaspar M, Remmert JE, Zhang F, Forman EM, & Butryn ML (2016). Small weight gains during obesity treatment: normative or cause for concern? Obesity Science & Practice, 2(4), 366–375. doi: 10.1002/osp4.73 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Spring B, Duncan JM, Janke EA, Kozak AT, McFadden HG, DeMott A,… Pellegrini CA (2013). Integrating technology into standard weight loss treatment: a randomized controlled trial. JAMA Internal Medicine, 173(2), 105–111. doi: 10.1001/jamainternmed.2013.1221 [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. 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: 10.1002/oby.21650 [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Tate DF, Wing RR, & Winett RA (2001). Using Internet technology to deliver a behavioral weight loss program. Journal of the American Medical Association, 285(9), 1172–1177. doi: 10.1001/jama.285.9.1172 [DOI] [PubMed] [Google Scholar]
  33. Thomas JG, Raynor H, Bond D, Luke A, Cardoso C, Wojtanowski A,… Foster G (2017). Weight loss and frequency of body-weight self-monitoring in an online commercial weight management program with and without a cellular-connected ‘smart’scale: a randomized pilot study. Obesity Science & Practice, 3(4), 365–372. doi: 10.1002/osp4.132 [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Thomas JG, Bond D, Raynor H, Papandonatos G, & Wing RR (2019). Comparison of Smartphone-based Behavioral Obesity Treatment to Gold Standard Group Treatment and Control: A Randomized Trial. Obesity, 27(4), 572–580. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Thomas JG, & Wing RR (2013). Health-e-call, a smartphone-assisted behavioral obesity treatment: pilot study. JMIR mHealth and uHealth, 1(1), e3. doi: 10.2196/mhealth.2164 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Thomas S, Reading J, & Shephard RJ (1992). Revision of the physical activity readiness questionnaire (PAR-Q). Canadian Journal of Sport Sciences, 17(4), 338–345. [PubMed] [Google Scholar]
  37. Turk MW, Elci OU, Wang J, Sereika SM, Ewing LJ, Acharya SD,… Burke LE (2013). Self-monitoring as a mediator of weight loss in the SMART randomized clinical trial. International Journal of Behavioral Medicine, 20(4), 556–561. doi: 10.1007/s12529-012-9259-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Turner-McGrievy GM, Beets MW, Moore JB, Kaczynski AT, Barr-Anderson DJ, & Tate DF (2013). Comparison of traditional versus mobile app self-monitoring of physical activity and dietary intake among overweight adults participating in an mHealth weight loss program. Journal of the American Medical Informatics Association, 20(3), 513–518. doi: 10.1136/amiajnl-2012-001510 [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Wadden TA, & Foster GD (2000). Behavioral treatment of obesity. Medical Clinics of North America, 84(2), 441–461. doi: 10.1016/S0025-7125(05)70230-3 [DOI] [PubMed] [Google Scholar]
  40. Wang J, Sereika SM, Chasens ER, Ewing LJ, Matthews JT, & Burke LE (2012). Effect of adherence to self-monitoring of diet and physical activity on weight loss in a technology-supported behavioral intervention. Patient Preference and Adherence, 6, 221. doi: 10.2147/ppa.s28889 [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Wang Y, & Beydoun MA (2007). The obesity epidemic in the United States—gender, age, socioeconomic, racial/ethnic, and geographic characteristics: a systematic review and meta-regression analysis. Epidemiologic Reviews, 29(1), 6–28. doi: 10.1093/epirev/mxm007 [DOI] [PubMed] [Google Scholar]
  42. Webber KH, Tate DF, Ward DS, & Bowling JM (2010). Motivation and its relationship to adherence to self-monitoring and weight loss in a 16-week Internet behavioral weight loss intervention. Journal of Nutrition Education and Behavior, 42(3), 161–167. doi: 10.1016/j.jneb.2009.03.001 [DOI] [PubMed] [Google Scholar]
  43. Wing RR, Lang W, Wadden TA, Safford M, Knowler WC, Bertoni AG,… Wagenknecht L (2011). Benefits of modest weight loss in improving cardiovascular risk factors in overweight and obese individuals with type 2 diabetes. Diabetes Care, 34(7), 1481–1486. doi: 10.2337/dc10-2415 [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES