Abstract
Objective
To examine whether the addition of online MI chats to a web-based, group behavioral obesity treatment program augments weight loss outcomes relative to the web-based online weight control program alone.
Methods
Healthy individuals (N=398, 24% minority) with overweight/obesity were randomized to a 36-session group internet behavioral weight control treatment (BT) or the same group internet treatment plus 6 individual MI chat sessions (BT+MI). Both conditions received weekly synchronous online chat group sessions for 6 months followed by 12 monthly group chats. Participants in both groups received identical behavioral lessons and individualized therapist feedback on progress toward meeting exercise and calorie goals. BT+MI also received 6 individual MI sessions delivered by a separate MI counselor via web chat. Weight loss was measured at 6- and 18-months.
Results
There were no significant differences in weight loss between BT (−5.5±6.0 kg) and BT+MI (−5.1±6.3 kg) at 6-months or at 18-months (−3.3±7.1 kg vs −3.5±7.7 for BT and BT+MI, respectively). Attendance at group chats did not differ between groups, nor did self–monitoring patterns, suggesting comparable engagement in the weight control program in both conditions.
Conclusions
Online MI chat sessions are not a viable strategy to enhance web-based weight control treatment outcomes.
Keywords: Internet, Motivation, Adult, Weight Management
Introduction
Obesity presents a vexing public health challenge. Effective behavioral weight control treatments are available1 and projections show these interventions could have a significant impact on overall population health if the interventions were more broadly available.2 Internet delivery of behavioral lifestyle interventions has potential for broad reach and is thus an attractive option,3 although average weight losses tend to be lower than in a comparable behavioral program delivered in-person.4 Nonetheless, internet-delivered synchronous behavioral weight control programs can produce an average 5% weight loss4 and may be more cost effective than in-person delivery of the same program.5 However, it would be beneficial to identify methods to enhance outcomes produced by online delivery of weight control to maximize utility of this potentially broad-reaching approach.
Motivational Interviewing (MI) is a collaborative counseling style that focuses on eliciting personally-relevant reasons for change as a strategy to develop and strengthen motivation to change behavior, MI further seeks to reinforce autonomy and self efficacy for behavior change using directive reflective listening.6 MI has been shown to be effective in promoting weight loss among adults,7 with randomized trials demonstrating significantly greater weight loss when group-based behavioral weight control is augmented with individual MI sessions than achieved with the group program alone.8 MI has also been shown to enhance the outcomes of self-help weight loss interventions,9 and bolster outcomes of individuals who struggle to achieve minimum weight goals during a group lifestyle program when compared to others who also struggle to achieve the weight loss goals but do not receive MI.10 Meta-analyses suggest the addition of MI to behavioral weight control interventions can increase weight losses by approximately 2 kgs.7
These interventions incorporating MI for weight loss have been delivered in-person; however, early online use of MI with adults for weight management suggests it may also be effective. For example, a small pilot study which intertwined some motivational interviewing constructs during online text-based group chats, reported promising weight losses at 2 months.11 Unfortunately, the study did not include a control group to allow determination of whether MI enhanced the online group weight loss program. Indeed, Webber et al.11 may have underestimated the potential impact of online MI since it was administered in a group setting and there are indications that group delivery of MI may attenuate impact compared with individually-delivered MI.12 Another study examined a self-directed, online adaptation of an evidence-based lifestyle weight loss program with a single in-person session delivered in MI-congruent style (“Minimal” condition) compared with a condition that received these elements plus weekly online group chats sessions featuring MI elements focused on building confidence in and importance of the specified behavior changes, and a decisional balance about behavior change (“Enhanced” condition).13 Weight losses at 16 weeks in the two conditions were not significantly different from one another. Again, there was no control group without elements of motivational interviewing, therefore it is impossible to isolate the impact of this approach in conjunction with an online weight control program or allocate weight losses achieved in the self-directed behavioral program versus those attributable to the initial MI-influenced face-to-face meeting. Additionally, in the Enhanced condition, engagement in weekly group MI sessions was modest, with only half of the sessions attended on average. Nonetheless, there were indications that greater weight loss was associated with greater attendance at the weekly chats. Collectively, these studies suggest a role for MI in online text-based weight control, but specific contributions of MI to online behavioral lifestyle programs cannot be determined from these preliminary studies.
Therefore, the current study examined whether the addition of individual MI web chats to a group-based, Internet-delivered behavioral weight control program confers any benefit in weight loss outcomes or treatment engagement over the evidence-based group behavioral program alone.
METHODS
Study Design
An overview of the study design has been reported.14 Briefly, it was a randomized, controlled trial conducted at two clinical sites (Arkansas and Vermont). Participants were randomly allocated to either (1) a 36-session, 18-month online adaptation of group-based behavioral lifestyle programs,8,15 which had been demonstrated effective in a previous study,4 (BT) or to (2) the same online group program augmented with six individual online MI chat sessions (BT+MI). Data were collected at baseline, 6-months and 18-months. The study was approved by the Committee on Human Research in the Behavioral Sciences at the University of Vermont and the Institutional Review Board at the University of Arkansas for Medical Sciences.
Participants
Volunteers were recruited through community-based efforts (newspaper, flyers, health fairs, etc.), targeted emails using available distribution lists (e.g. worksites, professional organizations, and sororities) and word of mouth. Interested individuals applied via a study recruitment website that provided a study overview and obtained initial screening data; a phone interview then determined likely eligibility. To be eligible, individuals had to be in generally good health, at least 18 years old and have a Body Mass Index (BMI; kg/m2) between 25 and 50. Individuals were ineligible if they took medications that might affect weight loss, reported substantial recent weight loss, had a history of bariatric surgery, were enrolled in another weight reduction program, or had a condition for which weight loss was contraindicated. Access to a computer (at home or work) and the Internet were required. Participants were required to successfully complete a behavioral run-in consisting of logging into the group chat webpage and completing an online food diary for 7-day and they agreed to be randomized. Participants indicated availability for pre-determined group times; these intact groups were stratified by baseline BMI percentile and then randomized using a biased coin approach.
Group behavioral weight control intervention
The 18-month manualized intervention focused on changing dietary and physical activity patterns using self-management skills and other behavioral strategies. One hour online synchronous chat sessions of 12 to 19 participants were moderated by experienced behavioral weight control counselors. Chats were offered weekly during the first 6 months and monthly for twelve additional months and combined participants from both clinical sites. Participants had access to a secure, password-protected dynamic website with behavioral lessons posted to accompany each chat session, a bulletin board for group communications, educational resources, regularly updated weight loss tips and healthy recipes, and notices of local physical activity events.16 A self-monitoring tool with a personalized dietary monitoring feature and a weight graphing feature and a compendium of physical activities with associated caloric expenditure information were also available on the website. Participants were instructed to record dietary intake, minutes of physical activity and weight daily in the online journal. Group counselors provided a weekly email with tailored feedback to participants based on this online journaling. A calorie-restricted diet and dietary fat goal corresponding to ≤ 25% of calories from fat were prescribed, and graded exercise goals that progressed to 200 min/week of moderate to vigorous exercise were provided. Pedometers were given to assist in self-monitoring steps and a goal of 10,000 steps/day was provided. Behavioral strategies to assist in making habit changes included self-monitoring, goal setting, problem solving, and relapse prevention. Weekly homework corresponding to the lesson topic and facilitating enactment of the featured behavioral strategy was assigned. The same group-based, goal-directed intervention was provided to both conditions.
Online MI Intervention
The online MI intervention has been described in detail previously.14 Briefly, six individual MI sessions using an interactive, synchronous form of private chat integrated within the same website as the group weight loss program were offered to participants randomized to the BT+MI condition. The text-based chat was selected because many participants lived in rural areas without consistent access to the technology required to support alternatives like video chat. Chats were designed to last approximately 30 minutes and followed a protocol which allowed flexibility in session flow and provided content tailored to the participant's treatment engagement and weight change experience, while standardizing the session across individuals. The first MI session was conducted before the group program started. The second MI session was after session 5 of the weekly group program, when early indications that individuals may be struggling with behavior change efforts or weight loss can emerge.17,18 The next five MI chats were offered at 3-month intervals. There are no empirical data examining different patterns of MI delivery in obesity treatment to guide the number or timing of MI sessions; therefore, we mirrored the approach taken in successful in-person programs.8
MI counselors were clinical psychologists, who had delivered MI for weight management in previous studies and/or in clinical practice. All MI counselors received training and ongoing supervision in MI from the first author, a motivational interviewing network trainer and a clinical psychologist who has conducted MI in conjunction with behavioral weight control for two decades.19 MI chat transcripts were reviewed and constructive feedback provided to refine therapist skills. Group telephonic coaching was provided weekly with a focus on maintaining an MI spirit6, adhering to the protocol and role-playing around difficulties encountered during MI chats.
MI sessions were synchronous and conducted in a private chat room on the study website and focused on the four processes of MI,6 with initial emphasis on engaging the participant and establishing rapport. Due to the nature of the study enrollment screening process, the desired outcome of weight loss had been established prior to the MI chats; therefore, initial chats emphasized personalized reasons underlying this desire and related behavior changes. Eliciting and elaborating change talk and collaboratively identifying behavior change strategies which the individual recognized as helpful were key elements of early MI sessions. Reflective statements and summaries were used to clarify, reinforce and promote further elaboration. Each chat concluded with a collaboratively-identified goal, if appropriate. The goals could be a behavior change strategy previously recommended within the group-based program, such as engaging in self-monitoring using the online journal, or something else the participant identified as likely to be effective for him/her. Participants were asked to rate their confidence in their ability to accomplish the goal and how important they believed the self-selected short-term goal to be in relation to their overall weight loss goals. Counselors reflected both importance and confidence when a participant's confidence was high and focused on reflecting importance of the goal when confidence was low.6 A semi-structured interview guided each MI chat.
Outcome measures
All measures were obtained at baseline, 6- and 18-months unless otherwise indicated.
Body weight
Weight change was the primary dependent measure. Weight was measured in street clothes, without shoes, on a calibrated digital scale. Height was measured using a wall-mounted stadiometer. BMI was calculated as weight (kg) / height (m)2 and obesity was defined as BMI ≥ 30. The proportion who achieved clinically significant weight losses of ≥5% and ≥10% from baseline were also examined.
Sociodemographic characteristics
Self-reported sociodemographic characteristics were collected at baseline by online questionnaire.
Process data
Group facilitators recorded participant attendance at online group chat sessions, as well as self-monitoring journal submissions, using an online adherence monitoring tool. MI counselors recorded attendance at individual MI chats.
Statistical Analysis
The study was designed to detect a 1.72 kg group weight loss difference with a standard deviation of 5.2 kg, a 5% Type I error rate and 80% power. Group comparisons used chi-square tests for contingency tables and Wilcoxon Rank Sum Tests for continuous variables. Group-specific multiple imputations (k=6) were implemented for missing weight data using baseline covariates, including sex, race, education, age, and obesity status (BMI ≥ 30). Comparison of imputed and complete case data averages at each time point and longitudinally appeared consistent. All analyses of weight changes at each time point were conducted with imputed data using a mixed linear model with repeated measures nested within participants and participants nested within randomized clusters (online chat groups). All models included treatment and clinical site as fixed effects. To gain insight into the potential joint impact of group behavioral weight control and individual MI chats on weight loss, a mediational analysis of MI engagement was conducted for the BT+MI treatment arm.20 Linear regression models between measures characterized direct effects and 95% confidence intervals (CIs) for slopes. Mediation effects were considered directional. Interpretation of regression-based mediation effects was assisted using median splits for attendance at group chats and MI sessions to define four engagement subgroups and their weight loss estimates. All analyses were conducted using SAS Version 9.4 (SAS Institute: Cary, NC). Across all tests, statistical significance was defined as p < 0.05 (2-tailed).
RESULTS
Participants
A total of 398 participants were randomized (n = 197 from AR; n = 201 from VT). Participants were predominantly female, obese and well-educated, with 24% African Americans. There were no significant differences between conditions on socio-demographic characteristics, weight or proportion of participants providing outcome data (Table 1). Significantly more participants recruited in Vermont were randomized to the BT condition and therefore study site was controlled in all analyses.
Table 1.
Baseline sample characteristics and study retention
| All | BT | BT+MI | p | |
|---|---|---|---|---|
| n = 398 | n = 199 | n = 199 | ||
| Age at baseline (years) (M±SD) | 48.4 ± 10.1 | 48.9 ± 10.7 | 47.9 ± 9.5 | 0.333 |
| Female (N [%]) | 357 (89.7%) | 179 (90.0%) | 178 (89.5%) | 0.869 |
| African American | 96 (24.1%) | 43 (21.6%) | 53 (26.6%) | 0.241 |
| Clinical site (N [%]) | ||||
| VT | 201 (50.5%) | 112 (56.3%) | 89 (44.7%) | 0.021 |
| AR | 197 (49.5%) | 87 (43.7%) | 110 (55.3%) | |
| Weight (kg) (M±SD) | 98.3 ± 18.6 | 98.2 ± 18.4 | 98.4 ± 19.0 | 0.894 |
| BMI (M±SD) | 36.0 ± 6.0 | 36.1 ± 6.1 | 35.9 ± 6.0 | 0.692 |
| Obese (≥ 30BMI) (N [%]) | 326 (81.9%) | 163 (81.9%) | 163 (81.9%) | 1.000 |
| Education (N [%]) | ||||
| High school/Vocational Training | 31 (7.8%) | 14 (7%) | 17 (8.5%) | 0.274 |
| Some college | 73 (18.3%) | 44 (22.1%) | 29 (14.6%) | |
| College degree | 155 (38.9%) | 75 (37.7%) | 80 (40.2%) | |
| Graduate/Professional | 139 (34.9%) | 66 (33.2%) | 73 (36.7%) | |
| Retained at 6 months (N [%]) | 357 (89.7%) | 181 (91.0%) | 176 (88.4%) | 0.410 |
| Retained at 12 months (N [%]) | 332 (83.4%) | 166 (83.4%) | 166 (83.4%) | 1.000 |
| Retained at 18 months(N [%]) | 323 (81.2%) | 160 (80.4%) | 163 (81.9%) | 0.701 |
Note. Proportions were tested using the chi-square goodness-of-fit test. Continuous variables were tested using the independent samples t-test.
Follow-up data were provided by 90% of randomized participants at 6-months and 81% at 18-months, with no difference between conditions in retention rates. However, those retained at 6-months were significantly older at baseline than those lost to follow-up (48.8±10.2 years vs. 45.3±9.4 years, respectively, F(1,396)=4.22, p=0.040). Similarly, participants who attended 18-month assessments were older than those not providing long-term follow-up data (48.9±10.2 vs.46.1±9.6 years, respectively, F(1,396)=4.81, p=0.029).
Weight Change
Weight losses were comparable at each point with no differences between conditions (Table 2). Participants in BT+MI lost 5.1 kg at 6-months, compared with 5.5 kg for participants in BT. Among those in BT+MI, 46.7% lost at least 5% by 6-months, compared with 49.8% of those in BT. A similar pattern was evident at 18-months (−3.5 kg for BT+MI vs −3.3 kg for BT). The proportion losing at least 5% of initial weight at 18-months was comparable between conditions.
Table 2.
Treatment Outcomes and Engagement
| BT | BT+MI | p | |
|---|---|---|---|
| n = 199 | n = 199 | ||
| Weight loss (kg) (M±SD) | |||
| Baseline to 6 months | 5.5 ± 6.0 | 5.1 ± 6.3 | 0.212 |
| Baseline to 18 months | 3.3 ± 7.1 | 3.5 ± 7.7 | 0.883 |
| Proportion (%) of participants meeting weight loss goal of ≥ 5% | |||
| Baseline to 6 months | 49.8 | 46.7 | 0.547 |
| Baseline to 18 months | 29.7 | 33.2 | 0.450 |
| Proportion (%) of participants meeting weight loss goal of ≥ 10% | |||
| Baseline to 6 months | 24.6 | 24.6 | 1.000 |
| Baseline to 18 months | 17.1 | 19.6 | 0.517 |
| Attendance | |||
| Behavioral weight control group chats (M±SD) | |||
| 0–6 months (out of 24) | 15.8 ± 6.5 | 15.2 ± 6.7 | 0.349 |
| 6–18 months (out of 12) | 3.3 ± 3.9 | 3.7 ± 3.9 | 0.373 |
| 0–18 months (total) (out of 36) | 19.1 ± 9.5 | 18.8 ± 9.8 | 0.843 |
| Ml chat sessions (M±SD) | |||
| 0–6 months (out of 3) | — | 2.5 ± 0.8 | — |
| 6–18 months (out of 3) | — | 1.4 ± 1.3 | — |
| 0–18 months (total) (out of 6) | — | 3.8 ± 1.9 | — |
| Self-monitoring frequency | |||
| Body weight (days) (M±SD) | |||
| 0–6 months (out of 168) | 77.7 ± 58.8 | 69.0 ± 59.2 | 0.126 |
| 6–18 months (out of 356) | 30.4 ± 63.7 | 25.9 ± 56.2 | 0.744 |
| 0–18 months (total) | 108.1 ± 109.9 | 94.9 ± 104.3 | 0.147 |
| Dietary intake (days) (M±SD) | |||
| 0–6 months (out of 168) | 104.0 ± 59.9 | 96.2 ± 60.3 | 0.221 |
| 6–18 months | 35.6 ± 72.8 | 30.9 ± 64.7 | 0.829 |
| 0–18 months (total) | 139.5 ± 117.0 | 127.0 ± 111.3 | 0.240 |
| Physical Activity (days) | |||
| 0–6 months | 68.7 ± 50.9 | 58.6 ± 47.5 | 0.062 |
| 6–18 months | 29.2 ± 63.0 | 21.5 ± 49.6 | 0.575 |
| 0–18 months (total) | 97.9 ± 101.8 | 80.1 ± 87.3 | 0.093 |
| Physical activity minutes/day (M±SD) | |||
| 0–6 months | 42.2 ± 22.8 | 39.3 ± 18.1 | 0.341 |
| 6–18 months | 37.0 ± 171.2 | 24.4 ± 27.0 | 0.934 |
| 0–18 months (total) | 45.0 ± 38.7 | 39.9 ± 18.4 | 0.349 |
Note. Proportions were tested using the chi-square goodness-of-fit test. Continuous variables were tested using the Wilcoxon Rank Sum Test. For intent-to-treat analyses, missing values for weight change used imputed data; missing attendance and self-monitoring variables were replaced with “0”.
Group Program Treatment Engagement
Attendance at online group chats was comparable between groups, with no significant difference between conditions. BT+MI averaged 15.2 group chats during the initial 6-months or 63% of available weekly sessions, which was reduced to 31% or an average of 3.7 sessions during the monthly maintenance phase. Participants in BT attended an average of 15.8 weekly sessions during the first 6-months (66%) and 3.3 of the monthly chats. Website log-ins to self-monitor key behaviors was also comparable between conditions, with no significant differences in recording dietary intake, physical activity or body weight (Table 2).
MI Session Engagement
Attendance at MI sessions during the initial 6-months was high. An average of 2.5 (83%) of the 3 sessions offered were attended. Engagement declined later in the program, with an average of 1.4 out of 3 sessions (47%) attended during weight maintenance. In total, an average of 3.8 (63%) of the available MI sessions were attended. All 6 MI sessions were attended by 57 participants (28.6%).
MI Session Mediation
Positive direct effects on weight loss at 6-months were seen for both MI session engagement (p< 0.0001; 95%CI [2.65, 4.35]) and number of group chat sessions (p < 0.0001; CI[0.41, 0.61]). Similar direct effects were seen for 18-month weight loss with MI session attendance (p < 0.0001; 95%CI [1.04, 2.20]) and number of group chats (p < 0.0001; 95% CI [0.26, 0.41]). Significant direct positive relationships were also observed between group chats and MI engagement at 6-month and 18-months (both p < 0.0001) with a less pronounced relationship at 6-months (95% CI [0.08,0.10]) compared to 18-months (95%CI [0.14, 0.16]). Examination of the indirect effect of group chat sessions via the mediational effect of MI engagement indicated that MI engagement had an expected positive mediating impact on group chats at 6-month (p = 0.086; 95%CI [−0.03,017]) and at 18-months (p = 0.066; 95%CI [−0.03,0.21]).
This mediation effect of MI engagement on weight loss can be seen descriptively with a median breakdown on the number of group chats and the number of MI sessions at both 6- and 18-months (Table 3). Those participants who were high on both intervention components lost more weight (8.36 kg at 6-months and 7.35 kg at 18-months) compared to those who were high on chats but low on MI engagement (5.16 kg at 6-months and 5.64 kg at 18-months), so there was some benefit from MI above and beyond the group intervention sessions among those who chose to engage in MI. Those who were low on group chats but high on MI lost more at 6-months than those low on group chats and low on MI as well (4.49 vs 0.47 kg) but they differed only minimally at 18-months (0.99 vs 0.10 kg).
Table 3.
Baseline Demographic Characteristics of BT+MI Group by Median Splits of MI Session Engagement Crossed with Group Chat Attendance*
| All | HiChat - HiMI | HiChat - LoMI | LoChat - HiMI | LoChat - LoMI | p | |
| At 6-Months | ||||||
| n = 199 | n = 93 (47%) | n = 10 (5%) | n = 36 (18%) | n = 60 (30%) | ||
| MI Session Attendance (M ± SD)1 | 2.5 ± 0.8 | 3.0 ± 0 | 1.9 ± 0.3 | 3.0 ± 0 | 1.5 ± 0.7 | <0.001 |
| Group Chat Attendance (M ± SD)2 | 15.1 ± 6.7 | 20.6 ± 2.0 | 20.5 ± 1.4 | 13.3 ± 2.8 | 7.0 ± 4.3 | <0.001 |
| Weight Loss (kg) (M; 95% Confidence Interval) | 5.1 (4.2, 6.0) | 8.4 (7.1, 9.6) | 5.2 (1.6, 8.7) | 4.5 (2.2, 6.7) | 0.5 (−0.2, 1.2) | <0.001 |
| Age at baseline (years) (M ± SD) | 47.9 ± 9.5 | 48.3 ± 9.4 | 56.9 ± 8.5 | 49.0 ± 8.4 | 45.1 ± 9.4 | 0.002 |
| Female | 178 (89.4%) | 81 (87.1%) | 10 (100%) | 34 (94.4%) | 53 (88.3%) | 0.575 |
| African American | 53 (26.6%) | 19 (20.4%) | 3 (30%) | 10 (27.8%) | 21 (35.0%) | 0.238 |
| Weight at baseline (kg) (M ± SD) | 98.4 ± 19.0 | 99.3 ± 19.5 | 93.1 ± 16.7 | 94.0 ± 19.9 | 100.7 ± 17.8 | 0.164 |
| BMI at baseline (M± SD) | 35.9 ± 6.0 | 36.0 ± 6.3 | 35.2 ± 5.8 | 34.6 ± 5.9 | 36.5 ± 5.6 | 0.384 |
| Obese (≥ 30 BMI) at baseline | 163 (81.9%) | 78 (83.9%) | 8 (80%) | 26 (72.2%) | 51 (85.0%) | 0.406 |
| Education | ||||||
| High school/vocational training | 17 (8.5%) | 8 (8.6%) | 1 (10%) | 2 (5.6%) | 6 (10.0%) | 0.267 |
| Some college | 29 (14.6%) | 8 (8.6%) | 1 (10%) | 5 (13.9%) | 15 (25.0%) | |
| College degree | 80 (40.2%) | 42 (45.2%) | 4 (40%) | 12 (33.3%) | 22 (36.7%) | |
| Graduate/Professional | 73 (36.7%) | 35 (37.6%) | 4 (40%) | 17 (47.2%) | 17 (28.3%) | |
| Retained at 6 months | 176 (88.4%) | 92 (98.9%) | 10 (100%) | 35 (97.2%) | 39 (65.0%) | <0.001 |
| Retained at 18 months | 163 (81.9%) | 87 (93.5%) | 9 (90%) | 32 (88.9%) | 35 (58.3%) | <0.001 |
| At 18-Months | All | HiChat - HiMI | HiChat - LoMI | LoChat - HiMI | LoChat - LoMI | p |
| n = 90 (45%) | n = 14 (7%) | n = 21 (11%) | n = 74 (37%) | |||
| MI Session Attendance (M ± SD)3 | 3.8 ± 1.9 | 5.4 ± 0.8 | 2.9 ± 0.4 | 4.9 ± 0.8 | 1.8 ± 0.9 | <0.001 |
| Group Chat Attendance (M ± SD)4 | 18.8 ± 9.8 | 27.2 ± 4.1 | 25.6 ± 3.2 | 14.5 ± 3.4 | 8.6 ± 5.2 | <0.001 |
| Weight Loss (kg) (M; 95% Confidence Interval) | 3.5 (2.4, 4.6) | 7.3 (5.6, 9.1) | 3.0 (−1.1, 7.0) | 1.0 (−2.0, 3.9) | −0.4 (−1.5, 0.8) | <0.001 |
| Age at baseline (years) (M ± SD) | 49.4 ± 9.1 | 53.9 ± 9.5 | 47.9 ± 9.4 | 45.0 ± 9.3 | 0.002 | |
| Female | 79 (87.8%) | 13 (92.9%) | 20 (95.2%) | 66 (89.2%) | 0.864 | |
| African American | 21 (23.3%) | 3 (21.4%) | 5 (23.8%) | 24 (32.4%) | 0.591 | |
| Weight at baseline (kg) (M ± SD) | 98.5 ± 19.3 | 94.7 ± 15.0 | 98.3 ± 21.7 | 99.1 ± 18.8 | 0.922 | |
| BMI at baseline (M± SD) | 35.8 ± 6.1 | 35.3 ± 6.0 | 36.2 ± 7.3 | 36.0 ± 5.6 | 0.973 | |
| Obese (≥ 30 BMI) at baseline | 75 (83.3%) | 12 (85.7%) | 16 (76.2%) | 60 (81.1%) | 0.868 | |
| Education | ||||||
| High school/vocational training | 8 (8.9%) | 1 (7.1%) | 1 (4.8%) | 7 (9.5%) | 0.534 | |
| Some college | 8 (8.9%) | 1 (7.1%) | 5 (23.8%) | 15 (20.3%) | ||
| College degree | 38 (42.2%) | 8 (57.1%) | 7 (33.3%) | 27 (36.5%) | ||
| Graduate/Professional | 36 (40.0%) | 4 (28.6%) | 8 (38.1%) | 25 (33.8%) | ||
| Retained at 6 months | 90 (100%) | 14 (100%) | 21 (100%) | 51 (68.9%) | <0.001 | |
| Retained at 18 months | 88 (97.8%) | 12 (85.7%) | 19 (90.5%) | 44 (59.5%) | <0.001 |
Out of 3 possible.
Out of 24 possible.
Out of 6 possible.
Out of 36 possible.
Note. Linear regression models between measures characterized direct effects and 95% confidence intervals (CIs) for slopes in a mediational analysis of MI engagement conducted for the BT+MI treatment arm. Interpretation of mediation effects was assisted using median splits for attendance at group chats and MI sessions to define four engagement subgroups and their weight loss estimates. Proportions were tested using Fisher's Exact Test. Continuous variables were tested using the Kruskal-Wallis Test.
DISCUSSION
Incorporating individual MI chat sessions in an online group-based behavioral weight control program conferred no weight loss advantage relative to the online group program alone. Further, the addition of individual MI chat sessions with the online group intervention did not produce greater attendance at group sessions or self-monitoring. This contrasts with previous studies of MI delivered in-person in which both attendance and self-monitoring were increased.8
Weight change outcomes with the online behavioral weight control program featuring synchronous delivery of group lifestyle sessions are comparable to those achieved previously,4 suggesting the robustness of this approach. Average weight losses exceeded 5%, suggesting clinically significant weight loss21 can be achieved with online delivery, and almost a quarter of participants lost at least 10% of their body weight at 6-months via the online program. Although impressive, these weight losses are not comparable to those achieved with similar lifestyle programs delivered face-to-face, underscoring the need to identify other approaches to improve online group weight control before technology-delivered approaches can realize maximal weight losses.
These data contrast with meta analyses suggesting that MI delivered face-to-face provided small but significant increases in weight loss.7,22 However, the current findings are consistent with some in-person programs showing no benefit with the addition of MI to a behavioral obesity intervention in African American women23 or in a stepped-care protocol.10 In addition, other studies combining MI with online behavioral weight control suggest no incremental advantage to augmentation with MI, although these interventions were short and had graduate students administering the MI.11,13
Early engagement in the MI sessions was good in the current study, with 83% of chat sessions completed in the first 6 months. There was a marked decrease in the proportion of MI chat sessions completed in the later part of the program, which paralleled a decrease in attendance at group chats after transitioning to monthly sessions. Others have noted similar challenges in online weight control program engagement24 and a decrease in attendance is observed when in-person programs transition to less frequent meetings as well.25 Overall, engagement in MI sessions was lower with online delivery than has been reported with face-to-face delivery in previous weight loss studies; 84% of available MI sessions were attended (4.2 out of 5) in person8 compared with 63% (3.8 out of 6) in the current study. Individuals who engaged in MI chats lost more weight than those who did not, particularly if they also attended more group chats, although the superior 6-month weight loss among those with high engagement in MI sessions but relatively lower group session attendance dissipated by 18-months. However, it is unclear whether individuals who were losing weight were more motivated to engage in MI and group sessions or those who engaged in MI and group sessions lost more weight. Nonetheless, there may have been a salubrious effect of MI for some individuals which did not emerge in intent-to-treat analyses of average response, perhaps reflecting the fairly bimodal distribution of MI engagement in BT+MI; 47% of participants in BT+MI were relatively higher in both MI session and group chat attendance and 30% were in the lower half of both treatment components. It appears that most individuals self-selected either higher or lower engagement in both of the treatment components and few were engaged in one aspect of treatment but not the other.
These results have implications for other online behavior change interventions utilizing interactive text-based MI. Loss of non-verbal cues because of the chat format proved problematic for both counselor and participant, as we observed anecdotally. As reported earlier, emails and other rapidly-written text formats can often be misconstrued without the context of facial and voice nuances.26 Furthermore, reflective statements, which are the heart of MI6, are distinguished from questions by voice deflection at the end of the statement and this is lost when MI is text-based. Finally, poor typing facility can impede the natural flow of the consultation. Despite the counselors reassurances that spelling was not important, their modeling less than perfect spelling and encouragement to type as one might text, participants may have felt uncomfortable doing so with their MI counselors. Newer video technologies which were unavailable at the start of this study may allow web-based but non-text-based MI, which may confer benefits similar to those observed with face-to-face MI in weight control.
CONCLUSION
The addition of online MI chats to an internet-delivered, synchronous behavioral weight control program failed to enhance weight losses. However, this replication of our previous implementation of an online behavioral lifestyle program confirms the potential of the approach, and merits continued exploration to identify methods to enhance weight outcomes. Incorporating MI in a non-text based fashion or optimization with other strategies may provide viable avenues to explore.
What is already known about this subject?
Online, synchronous group behavioral weight control produces clinically-significant weight losses which are, on average, lower than those of the same program delivered in-person.
Motivational interviewing can improve weight loss outcomes when added to face-to-face lifestyle treatment programs.
What does your study add?
Examination of whether online motivational interviewing increases the weight loss outcomes achieved with a proven synchronous, web-based group behavioral program in a large randomized trial
Motivational interviewing delivered via online chat in the context of an evidence-based online group weight control programs did not enhance weight losses
Other behavioral health promotion programs considering online motivational interviewing chats may wish to consider other avenues of delivery
Acknowledgments
Funding: This project was supported in part by NIDDK R01DK056746 to Drs. West and Harvey
Trial registration NCT01232699
References
- 1.Look AHEAD Research Group Long term effects of lifestyle intervention on weight and cardiovascular risk factors in individuals with type 2 diabetes: four year results of the Look AHEAD trial. Arch Intern Med. 2010;170(17):1566–1575. doi: 10.1001/archinternmed.2010.334. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Albright AL, Gregg EW. Preventing Type 2 Diabetes in Communities Across the U.S.: The National Diabetes Prevention Program. American Journal of Preventive Medicine. 2013;44(4, Supplement 4):S346–S351. doi: 10.1016/j.amepre.2012.12.009. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Krukowski RA, West DS, Harvey-Berino J. Recent Advances in Internet-Delivered, Evidence-Based Weight Control Programs for Adults. Journal of Diabetes Science and Technology. 2009;3(1):184–189. doi: 10.1177/193229680900300122. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Harvey-Berino J, West D, Krukowski R, et al. Internet delivered behavioral obesity treatment. Preventive Medicine. 2010 Aug;51(2):123–128. doi: 10.1016/j.ypmed.2010.04.018. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Krukowski RA, Tilford JM, Harvey-Berino J, West DS. Comparing Behavioral Weight Loss Modalities: Incremental Cost-Effectiveness of an Internet-Based Versus an In-Person Condition. Obesity. 2011 Aug;19(8):1629–1635. doi: 10.1038/oby.2010.341. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Miller WR, Rollnick S. Motivational interviewing: Preparing people for change. 3rd ed Guilford Press; New York: 2013. [Google Scholar]
- 7.Armstrong M, Mottershead T, Ronksley P, Sigal R, TS C, Hemmelgarn B. Motivational interviewing to improve weight loss in overweight and/or obese patients: a systematic review and meta-analysis of randomized controlled trials. Obesity Reviews. 2011 doi: 10.1111/j.1467-789X.2011.00892.x. [DOI] [PubMed] [Google Scholar]
- 8.West DS, Gore SA, DiLillo V, Greene PG, Bursac Z. Motivational interviewing improves weight loss in women with Type 2 Diabetes. Diabetes Care. 2007;30(5):1081–1087. doi: 10.2337/dc06-1966. [DOI] [PubMed] [Google Scholar]
- 9.DiMarco I, Klein D, Clark V, Wilson G. The use of motivational interviewing techniques to enhance the efficacy of guided self-help behavioral weight loss treatment. Eat Behav. 2009;10:134–136. doi: 10.1016/j.eatbeh.2009.02.001. [DOI] [PubMed] [Google Scholar]
- 10.Carels RA, Darby L, Cacciapaglia HM, et al. Using motivational interviewing as a supplement to obesity treatment: a stepped-care approach. Health Psychol. 2007;26(369–374) doi: 10.1037/0278-6133.26.3.369. [DOI] [PubMed] [Google Scholar]
- 11.Webber KH, Tate DF, Quintiliani LM. Motivational interviewing in Internet groups: A pilot study for weight loss. Journal of the American Dietetic Association. 2008 Jun;108(6):1029–1032. doi: 10.1016/j.jada.2008.03.005. [DOI] [PubMed] [Google Scholar]
- 12.Lundahl BW, Kunz C, Brownell C, Tollefson D, Burke BL. A Meta-Analysis of Motivational Interviewing: Twenty-Five Years of Empirical Studies. Res Social Work Prac. 2010 Mar;20(2):137–160. [Google Scholar]
- 13.Webber KH, Tate DF, Bowling JM. A randomized comparison of two motivationally enhanced Internet behavioral weight loss programs. Behav Res Ther. 2008 Sep;46(9):1090–1095. doi: 10.1016/j.brat.2008.06.008. [DOI] [PubMed] [Google Scholar]
- 14.Krukowski RA, DiLillo V, Ingle K, Harvey JR, West DS. Design and Methods of a Synchronous Online Motivational Interviewing Intervention for Weight Management. JMIR Res Protoc. 2016;5 doi: 10.2196/resprot.5382. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.The Look AHEAD Research Group The Look AHEAD Study: a description of the lifestyle intervention and the evidence supporting it. Obesity. 2006;14(5):737–752. doi: 10.1038/oby.2006.84. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Krukowski RA, Harvey-Berino J, Ashikaga T, Thomas CS, Micco N. Internet-Based Weight Control: The Relationship Between Web Features and Weight Loss. Telemed J E-Health. 2008 Oct;14(8):775–782. doi: 10.1089/tmj.2007.0132. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Krukowski RA, Harvey-Berino J, Bursac Z, Ashikaga T, West DS. Patterns of Success: Online Self-Monitoring in a Web-Based Behavioral Weight Control Program. Health Psychology. 2013;32:164–170. doi: 10.1037/a0028135. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Unick JL, Hogan P, Neiberg R, et al. Evaluation of early weight loss thresholds for identifying nonresponders to an intensive lifestyle intervention. Obesity. 2014;22(1608–1616) doi: 10.1002/oby.20777. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Smith DE, Heckemeyer CM, Kratt PP, Mason DA. Motivational interviewing to improve adherence to a behavioral weight control program for older obese women with NIDDM: A pilot study. Diabetes Care. 1997;20(1):52–54. doi: 10.2337/diacare.20.1.52. [DOI] [PubMed] [Google Scholar]
- 20.MacKinnon D, Dwyer J. Estimating mediated effects in prevention studies. Evaluation Review. 1993:144–158. [Google Scholar]
- 21.Williamson DA, Bray GA, Ryan DH. Is 5% Weight Loss a Satisfactory Criterion to Define Clinically Significant Weight Loss? Obesity. 2015;23:2319–2320. doi: 10.1002/oby.21358. [DOI] [PubMed] [Google Scholar]
- 22.Burke B, Arkowitz H, Menchola M. The efficacy of motivational interviewing: A meta-analysis of controlled clinical trials. Journal of Consulting & Clinical Psychology. 2003;71(5):843–861. doi: 10.1037/0022-006X.71.5.843. [DOI] [PubMed] [Google Scholar]
- 23.Befort C, Nollen N, Ellerbeck E, Sullivan DK, Thomas J, Ahluwalia JS. Motivational interviewing fails to improve outcomes of a behavioral weight loss program for obese African-American women: a pilot randomized trial. J Behav Med. 2008;31:367–377. doi: 10.1007/s10865-008-9161-8. [DOI] [PubMed] [Google Scholar]
- 24.Wing RR, Tate DF, Gorin AA, Raynor HA, Fava JL. A self-regulation program for maintenance of weight loss. New Engl J Med. 2006 Oct 12;355(15):1563–1571. doi: 10.1056/NEJMoa061883. [DOI] [PubMed] [Google Scholar]
- 25.Bartfield JK, Stevens VJ, Jerome GJ, et al. Behavioral Transitions and Weight Change Patterns Within the PREMIER Trial. Obesity. 2011;19(8):1609–1615. doi: 10.1038/oby.2011.56. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Byron K. Academy of Management Review. 2008;33:309–327. [Google Scholar]