Feasibility and Acceptability of a Lifestyle Intervention for Individuals with Bipolar Disorder

Abstract

Individuals with bipolar disorder are at greater risk for cardiovascular disease and are less likely to adhere to lifestyle interventions than the general population. To decrease cardiovascular risk and improve adherence to lifestyle interventions, we developed the Nutrition Exercise and Wellness Treatment (NEW Tx). NEW Tx is an 18 session, 20-week cognitive behavioral therapy (CBT)-based treatment comprising 3 modules: Nutrition, Exercise, and Wellness. To evaluate the feasibility and acceptability of this intervention as well as predictors of treatment satisfaction and expectations, 38 adult outpatients with bipolar disorder were randomized to either NEW Tx or a waitlist control condition. There was no statistically significant difference in drop-out rates between the groups (26.3% in NEW Tx, 31.6% in the control condition). In the NEW Tx condition, participants attended a mean of 66.7% of sessions and reported moderate to high satisfaction. There were no study-related adverse events. We also found that expectations, but not perceived credibility (or believability), of NEW Tx (as measured by the Credibility/ Expectancy Questionnaire) at baseline predicted treatment satisfaction (as measured by the Care Satisfaction Questionnaire) post-treatment. Manic symptoms at baseline predicted treatment satisfaction, and marital status predicted one’s expectations of lifestyle interventions. Data suggest that NEW Tx is a feasible and acceptable intervention for individuals with bipolar disorder and that further research is warranted to explore potential moderators of treatment expectations and credibility in this clinical population.

Individuals with bipolar disorder are more likely to develop obesity and cardiovascular diseases than the general population.^1–3 They also report poorer general health as well as more concurrent physical health problems (eg, substance use disorder, chronic health conditions).^4,5 Paradoxically, many of the medications used to treat bipolar disorder can cause weight gain and contribute to metabolic syndrome,⁶ further increasing cardiovascular risk. These patients also tend to have sedentary lifestyles and poor diets.⁷ Thus, there is a need for adjunctive therapies to help individuals with bipolar disorder improve their physical health and subsequently reduce their risk of mortality.

The challenge facing lifestyle interventions for bipolar disorders lies in overcoming residual depressive symptoms characteristic of bipolar disorder, such as fatigue, lack of motivation and interest, and social isolation, all of which lower adherence to lifestyle interventions. For example, Daumit et al⁸ conducted a lifestyle intervention for individuals with serious mental illness (ie, bipolar disorder, schizophrenia, and major depression) in outpatient psychiatric rehabilitation programs. Although they found that participants lost more weight relative to a control group over 18 months, nearly half (44%) of these participants did not attend all sessions in months 1 to 6 and 91% did not attend all sessions in months 7 to 18. Similarly, another study found that patients experiencing depression after a myocardial infarction were less likely to adhere to recommendations for reducing cardiac risk during recovery, including lifestyle changes (eg, low-fat diet, increased exercise).⁹

To improve adherence rates for individuals with bipolar disorder, we developed the Nutrition Exercise and Wellness Treatment (NEW Tx), a lifestyle intervention specifically designed for these individuals. NEW Tx was first piloted as a group intervention (N = 10) and yielded promising study outcomes and feasibility, with participants completing all study assessments and attending 82% of the sessions.¹⁰ However, participants requested more individual feedback and treatment; thus, the NEW Tx manual was revised to be administered as an individual therapy and piloted again in a small open trial (N = 5).¹¹ These preliminary results yielded promising acceptability and feasibility, as well as positive trends on all physical and mental health outcomes. On the basis of additional feedback from pilot participants, we further lengthened the NEW Tx lifestyle program to 18 sessions over 20 weeks. In this study, we specifically examined the feasibility and acceptability of the revised NEW Tx compared to a waitlist condition. The main outcomes of this randomized trial that addressed overall study aims (ie, physical outcomes, such as weight loss, and psychological outcomes, including changes in mood symptoms) are presented in another publication.¹² We expected that NEW Tx would yield high adherence rates as well as treatment satisfaction and credibility ratings, given that it addresses aspects unique to individuals with bipolar disorder (eg, the intervention is tailored based on mood state, discusses the side effects of medications, and shares data that weight loss can happen despite pharmacotherapy). Finally, we explored whether baseline characteristics of the individual (age, marital status, education, mood symptoms) were associated with treatment satisfaction, expectation and credibility.

METHOD

Participants

Adult outpatients were recruited from the local community using advertisements on public transportation, flyers, internet advertisements, and referrals by community providers. Eligible participants had a primary diagnosis of bipolar disorder (Type I or II). Participants were between the 18 and 65 years of age, currently symptomatic, and overweight or obese [body mass index (BMI) ≥ 25]. Those with a BMI between 25 and 30 were also required to have higher-than-normal body fat percentage (ie, greater than or equal to 32% for women and 25% for men). To ensure patient safety, if patients endorsed any contraindication to physical activity at the baseline visit, they were required to get approval from their physician before enrolling. Individuals were ineligible if they had a diagnosis of anorexia nervosa, bulimia nervosa, or substance dependence in the month before to study entry, were actively suicidal [score > 4 on item 10 of the Montgomery Asberg Depression Rating Scale (MADRS)¹³], were currently pregnant, were exercising regularly (ie, 5 days per week for at least 30 minutes), or had a neurologic disorder or history of head trauma.

Of 58 participants who were screened, 38 met the eligibility requirements. Reasons for ineligibility included being euthymic (CGI-BP ≤ 2, n = 3), exercising regularly (at least 150 minutes per week, n = 2), having a diagnosis of bulimia nervosa (n = 2) or substance dependence (n = 1) within the past month, not having a primary diagnosis of bipolar disorder (n = 5), having contraindications to an exercise or diet intervention (eg, a history of eating pathology, n = 2), not being able to comply with study procedures (eg, not being able to be reached to complete screening visit procedures, n = 2), having a BMI below 25 (n = 1), or having a combination of these exclusion criteria (ie, one participant both did not meet criteria for bipolar disorder and was exercising regularly, and one participant both did not meet criteria for bipolar disorder and had a diagnosis of a substance use disorder, n = 2).

Eligible participants were randomized to either NEW Tx (n = 19) or a waitlist control group (n = 19). Participants in the NEW Tx group were mainly female (n = 13, 68.4%) and white (n = 16, 84.2%) and on average 39.7 years old (SD = 12.5). The control group was also predominantly female (n = 12, 63.2%) and white (n = 16, 84.2%). The average age of participants in the control group was 44.3 years (SD = 11.9). At baseline, NEW Tx participants had an average score of 13.1 (SD = 8.9) on the MADRS and 5.4 (SD = 8.6) on the Young Mania Rating Scale (YMRS).¹⁴ At the post-treatment assessment, the average MADRS and YMRS scores were 8.1 (SD = 6.7) and 2.5 (SD = 3.1), respectively. These scores suggest that participants were moderately depressed and mildly manic at study entry, and that they had mild symptoms of depression and very mild symptoms of mania by the end of the study. Participants randomized to the control group had baseline MADRS and YMRS scores of 12.4 (SD = 9.2) and 6.0 (SD = 5.9), and post-treatment MADRS and YMRS scores of 9.6 (SD = 8.2) and 5.3 (SD = 2.8), respectively. See Table 1 for additional demographic data on the participants.

Table 1.

Participant Demographics: Total Consented (N = 38)

Variable	All Participants Mean ± SD	NEW Tx Mean ± SD	Control Mean ± SD
Age (yr)	42.0 ± 12.3	39.7 ± 12.5	44.3 ± 11.9
	All Participants N (%)	NEW Tx N (%)	Control N (%)
Primary diagnosis
Bipolar type I	32 (84.2)	18 (94.7)	14 (73.7)
Bipolar type II	6 (15.8)	1 (5.3)	5 (26.3)
Gender
Female	25 (65.8)	13 (68.4)	12 (63.2)
Male	13 (34.2)	6 (31.6)	7 (36.8)
Race
Caucasian/white	32 (84.2)	16 (84.2)	16 (84.2)
Black/African American	1 (2.6)	1 (5.3)	0 (0.0)
Asian/Asian American	1 (2.6)	1 (5.3)	0 (0.0)
Other	2 (5.3)	0 (0.0)	2 (10.5)
Not reported	2 (5.3)	1 (5.3)	1 (5.3)
Ethnicity
Hispanic/Latino	5 (13.2)	1 (5.3)	4 (21.1)
Not Hispanic/Latino	31 (81.6)	17 (89.5)	14 (73.7)
Unknown	1 (2.6)	0 (0.0)	1 (5.3)
Not reported	1 (2.6)	1 (5.3)	0 (0.0)
Marital Status
Single	13 (34.2)	7 (36.8)	6 (31.6)
Divorced/separated	6 (15.8)	4 (21.1)	2 (10.5)
Married/living as married	17 (44.7)	7 (36.8)	10 (52.6)
Widowed	1 (2.6)	0 (0.0)	1 (5.3)
Not reported	1 (2.6)	1 (5.3)	0 (0.0)
Education
Less than high school	1 (2.6)	1 (5.3)	0 (0.0)
Some college (at least 1 year)	9 (23.7)	2 (10.5)	7 (36.8)
Technical school/associate degree	3 (7.9)	2 (10.5)	1 (5.3)
College diploma	15 (39.5)	10 (52.6)	5 (26.3)
Graduate/professional degree	9 (23.7)	3 (15.8)	6 (31.6)
Not reported	1 (2.6)	1 (5.3)	0 (0.0)
Income
< $25,000	11 (29.0)	5 (26.3)	6 (31.6)
< $50,000	7 (18.4)	3 (15.8)	4 (21.1)
< $75,000	6 (15.8)	3 (15.8)	3 (15.8)
>$75,000	13 (34.2)	7 (36.8)	6 (31.6)
Not reported	1 (2.6)	1 (5.3)	0 (0.0)

Description of NEW Tx

NEW Tx is an 18 session, 20-week cognitive behavioral therapy (CBT)-based treatment comprising three modules: Nutrition, Exercise, and Wellness. In the Nutrition/Weight Loss Module, clinicians and participants review nutritious foods (eg, foods high in vitamins and minerals and low in fat) in each of the food groups, the concept of a balanced diet, and the importance of vitamins and minerals (especially those beneficial for mood disorders) to inform healthy food choices. Participants monitor their food consumption and caloric intake using food diaries and utilize CBT skills such as cognitive restructuring and behavioral strategies to enhance adherence to the treatment. The goal of the Exercise Module is to achieve a healthy level of weekly exercise—at least 30 minutes of moderate intensity exercise 5 days per week, as recommended by the American College of Sports Medicine.¹⁵ Participants learn the benefits of exercise for enhancing mood and set weekly exercise goals using reward charts; they also identify obstacles to exercising and use CBT strategies to overcome these obstacles (eg, changing negative thoughts, functional analysis, building support). The Wellness Module supports healthy decision-making with the use of CBT skills (eg, functional analysis, relapse prevention plans) and covers other obstacles to creating a healthy lifestyle (eg, poor sleep, substance use, lack of social support).

Therapist Training

NEW Tx therapists were master’s-level clinicians in doctoral psychology programs. All therapists had an understanding of NEW Tx, knowledge of bipolar disorder, and proficiency in the techniques used to achieve NEW Tx treatment goals (ie, motivational interviewing and CBT). They received additional NEW Tx training that included the following components: (1) reviewing the manuals with a trained NEW Tx therapist; (2) listening to portions of selected treatment tapes of trained NEW Tx therapists; (3) role playing NEW Tx skills and modules with a trained therapist; and (4) ongoing weekly supervision with a trained NEW Tx therapist to review audiotapes from their sessions.

Description of Waitlist

Participants in the control group continued their treatment as usual (ie, they did not receive any additional treatment aside from their regular care) for the 20-week duration of NEW Tx. After they completed the waitlist phase, they had the option of receiving the NEW Tx intervention.

Procedure

The Partners HealthCare’s Institutional Review Board approved all study procedures and participants provided informed consent prior to initiation of any study procedure. Participants recruited from the community were directed to contact the study coordinator to determine initial eligibility. Following a brief telephone screening interview, participants were asked to attend an in-person screening visit. Eligible participants were then randomized to either NEW Tx or the waitlist control group.

Participants randomized to the NEW Tx group met individually with a study clinician to complete 18 individual sessions over 20 weeks of treatment. They also completed assessments at mid- (week 10) and post-treatment (week 20) visits. Symptom states were monitored via these assessments, and managed with guideline-informed treatments for bipolar disorder.¹⁶ Participants were also able to discuss concerns with their study therapists each session, and study therapists coordinated patient care with community providers as needed (eg, patients with community providers received monthly booster sessions of therapy, and NEW Tx therapists worked with the community providers for a smooth transition back to regularly scheduled therapy sessions). As per the protocol, participants would be withdrawn from the study if their clinical condition deteriorated substantially or if a higher level of care was warranted. Participants in the control group were asked to complete the same assessment visits as the NEW Tx group (ie, at weeks 0, 10, and 20). Only NEW Tx participants received $20 total in compensation for completing the mid- and post-treatment visits as part of their work with NEW Tx therapists in establishing and achieving treatment-related goals.

Measures

Baseline characteristics

Independent raters assessed clinical diagnoses using the Mini International Neuropsychiatric Interview (MINI Plus)¹⁷ and severity of current symptoms with the Clinical Global Impressions Scale–Bipolar Version (CGI-BP).¹⁸ Participants were considered currently symptomatic and therefore eligible if they scored at least a 3 on the CGI-BP. Risks in starting an exercise program were assessed with the Physical Activity Readiness Questionnaire (PAR-Q).¹⁹ Participants were also asked about their medical history, psychiatric history, and demographic information at the initial visit.

Feasibility and acceptability

The Credibility/Expectancy Questionnaire (CEQ),²⁰ a 6-item, self-report measure that assesses a participant’s perceptions of and expectations for treatment before beginning an intervention, was administered at the pre-treatment visit. Items are rated on either a 9-point scale (from 1 to 9) or an 11-point scale (from 0% to 100%).²⁰ This measure creates two subscales, credibility and expectancy scores, that are separately standardized and summed together with higher scores indicating greater credibility and expectancy on their respective scales.²⁰ The Care Satisfaction Questionnaire (CSQ-8),²¹ an 8-item, self-report measure of acceptability of treatment, was completed at the post-treatment (week 20) visit. This measure assessed participant satisfaction with and perceived quality and tolerability of the intervention. Items were rated from 1 (question-specific negative responses; eg, “Poor”, “No, definitely not”) to 4 (question-specific positive responses; eg, “Excellent”, “Yes, definitely”) and summed together with higher scores indicating greater satisfaction with treatment.²¹ The NEW Tx Scale, a 10-item measure using a 5-point Likert scale ranging from 1 (“Strongly Agree”) to 5 (“Strongly Disagree”), was also administered at the post-treatment visit. This measure assessed whether participants thought that the treatment was helpful at targeting weight loss, eating more nutritiously, and increasing weekly exercise, as well as whether they learned skills to help them live a healthier lifestyle and if they would recommend the treatment to others. At week 20, participants were also asked to provide qualitative feedback on aspects of the treatment that were helpful, not helpful, or could be improved. Attendance and retention data were also collected.

Mood symptoms

Depressive and manic/hypomanic symptoms were assessed using the MADRS¹³ and the YMRS,¹⁴ respectively. These measures were administered at Weeks 0 (pre-treatment/waitlist), 10 (midpoint), and 20 (post-treatment/waitlist). The MADRS is a 10-item clinician-rated measure of depression that assesses the presence and severity of current depressive symptoms, with each item rated on a scale of 0 to 6 and anchors at 2-point intervals. Total scores can range from 0 to 60, with higher scores indicating greater severity of depressive symptoms. The YMRS is an 11-item clinician-rated measure that assesses the presence and severity of manic symptoms. Four of these items are graded on a scale of 0 to 8 and double-weighted; 7 are graded on a scale of 0 to 4, with higher numbers representing greater severity of manic symptoms.

Statistical Analysis

Feasibility

Descriptive statistics regarding attrition, session adherence, and adverse events are reported to examine feasibility of completing the treatment intervention.

Acceptability

To examine post-treatment satisfaction with the intervention, we report the average score on the CSQ-8 at the post-treatment visit for the NEW Tx group. We conducted multiple regressions on continuous variables and analyses of variance (ANOVAs) on categorical variables to determine whether baseline characteristics predicted treatment satisfaction as measured by the CSQ-8. We also conducted linear regression analyses to examine whether participants’ perceptions of the intervention at week 20, as measured with the NEW Tx scale, were associated with initial perceptions of intervention credibility (CEQ) at baseline. Regressions and analyses of covariance (ANCOVAs) were used to examine whether a participant’s perception of the intervention’s credibility (ie, believability of the intervention; CEQ) and expectations (ie, beliefs about anticipated personal achievements during treatment) predicted their treatment satisfaction (CSQ-8). Credibility and expectancy subscale scores from the CEQ were calculated based on the procedure detailed by Devilly and Borkovec.²⁰ Given that the analyses presented in this paper are exploratory, we did not control our level of statistical significance. The post-treatment feedback questionnaire was qualitatively analyzed using conventional content analysis,²² a qualitative method of analysis in which themes are identified directly from the text.

RESULTS

Feasibility

Over the course of the study, there was no significant difference between the groups in drop-out rates (P > 0.05). Five participants did not complete the NEW Tx condition (26.3%) and 6 did not complete the control condition (31.6%). One NEW Tx participant attended all sessions, but did not complete the post-treatment assessments. The 26 completers with baseline and post-treatment assessment data were included in analyses. In the NEW Tx group, reasons for drop out were being lost to follow up (ie, participant was unable to be reached by study staff; n = 3), transportation issues making regular study visits difficult (n = 1), and an unrelated medical issue (n = 1). In the control group, reasons for drop-out were being lost to follow up (n = 3), an unrelated medical issue (n = 2), and taking a medication which made the participant ineligible to participate (n = 1). Participants randomized to NEW Tx attended a mean of 66.7% of sessions. Individuals who completed NEW Tx (n = 14) attended 81.8% of sessions on average (Figure 1).

Figure 1.

NEW Tx Session Adherence (%) by Session for Completers (n = 14)

NEW Tx participants reported a total of 12 adverse events, which could be classified into three categories: (1) physical health or medical issues (eg, breaks, sprains, or infections unrelated to participation in NEW Tx; n = 7); (2) unrelated life events (eg, car accident, work injury; n = 4); and (3) worsening psychiatric illness requiring participation in a partial hospitalization program; n = 1). None of the adverse events was related to the intervention or study procedures. Of note, participants randomized to the waitlist condition were offered NEW Tx following the waitlist period, and 9 (47.4%) of those 19 participants opted to receive the NEW Tx intervention after the waitlist period.

Acceptability

The mean score on the CSQ-8 at post-treatment was 25.8 (SD = 6.9), reflecting a medium level of satisfaction (a score > 27 suggests a high level of satisfaction).²¹ To examine potential predictors of satisfaction, we conducted multiple regressions with continuous variables and ANOVAs with categorical variables of baseline characteristics predicting scores on the CSQ-8. We did not find that marital status, education, age, baseline depression (MADRS), post-treatment depression (MADRS), or post-treatment mania (YMRS) predicted CSQ-8 scores (all Ps > 0.05). However, baseline mania (YMRS) did predict satisfaction with the treatment (CSQ-8; ß = −0.68, t(11) = −4.10, P = 0.002), with participants with higher levels of mania at baseline found to be less satisfied with the treatment.

On the basis of the NEW Tx Scale, participants, on average, endorsed the intervention as helpful across all items, with the highest scores on the following items: “I found this treatment was helpful for eating more nutritiously” (mean = 2.1, SD = 1.4), “I learned how to increase healthy behaviors” (mean = 2.1, SD = 1.4), and “Overall, I found this treatment to be helpful” (mean = 2.2, SD = 1.5; see Table 2).

Table 2.

Average Scores on Individual Items of the NEW Tx Scale at Week 20

NEW Tx Scale Item	Average Mean (SD)
1. I found this treatment was helpful for losing weight	2.8 (1.4)
2. I found this treatment was helpful for eating more nutritiously	2.1 (1.4)
3. I found this treatment was helpful for increasing my weekly exercise	2.5 (1.6)
4. I learned how to increase healthy behaviors	2.1 (1.4)
5. I learned skills to change my unhealthy habits	2.3 (1.4)
6. I learned strategies to help me stick to a diet and exercise program	2.5 (1.3)
7. I will keep using these skills on my own	2.4 (1.3)
8. I found this treatment to be challenging	2.4 (1.7)
9. Overall, I found this treatment to be helpful	2.2 (1.5)
10. I would recommend this treatment for someone else	2.4 (1.7)

Note: Items were rated on a 5-point Likert scale ranging from 1 (“Strongly Agree”) to 5 (“Strongly Disagree”).

Expectancy and the perceived credibility (or believability) of NEW Tx, as measured by the CEQ at baseline, were highly correlated (r(11) = 0.66, P = 0.015); therefore, each of these items was modeled separately. Results from linear regression models found expectancy but not credibility at baseline predicted treatment satisfaction (CSQ-8) at post-treatment (ß = 0.618, t(11) = 2.61, P = 0.024). We also examined the association of scores on the CEQ with scores on the NEW Tx scale at Week 20 and found that credibility ratings were only associated with item 8 on the NEW Tx Scale (“I found this treatment to be challenging”; ß = 0.43, t(11) = 3.0, P = 0.013). For this item, the overall regression model was significant (R² = .443, F(1, 11) = 8.76, P = 0.013).

Marital status had an impact on expectations concerning treatment (F(2,10) = 4.58, P = .039), with divorced participants (mean = 1.3, SD=0.9) having higher expectations for the intervention than single participants (mean = −2.4, SD=1.7). We did not find that age, education, or symptoms (pre- or post-intervention) were associated with expectations concerning the lifestyle intervention (all P > 0.05).

On the post-treatment feedback questionnaire (n = 13), 2 participants (15.4%) indicated that they had no suggestions for how NEW Tx could be improved and 7 participants (53.9%) did not find any aspects of NEW Tx unhelpful. Three participants (23.1%) were dissatisfied with their relationships with their specific therapists (eg, “I don’t think I meshed with the therapist”), and 1 participant (7.7%) reported not enjoying maintaining food and exercise diaries but also stated that the food and exercise diaries were helpful. All of the responses on the post-treatment questionnaire were grouped based on themes identified in the content analysis, as shown in Table 3.

Table 3.

Participant Responses to NEW Tx Post-Treatment Questionnaire

Category	n	Representative Quotes
Helpful aspects of treatment	Total: 15
Homework	3	“Recording food in-take; recording exercise” “Keeping the food journals and sessions” “The thought diaries were helpful to help me see the black and white quality of my thoughts”
Therapist and/or study staff	2	“The support of my therapist” “Study staff”
Aspects of treatment	7	“ALL of it! I wish I had been offered this 10 years ago when I was diagnosed w/ bipolar. I desperately need help in all of the points you stressed.” “The nutrition part. Healthy eating habits.” “The sections on nutrition and exercise were most helpful.” “Gradual change and explanations of changing my diet.” “Validating small goals and breaking them even smaller if needed.” “Recognizing my own cognitive distortions”
Other	3	“Realized I can have a healthy lifestyle always!” “I at first resisted the therapy. I found it difficult to relate to therapist. Then she made me realize I need to give myself rewards and think more positively.” “I learned that I can control myself and get knowledge in different sides of my health.”
Suggestions for improvement	Total: 14
Aspects of treatment	9	“How to figure out/measure the size of food and calories.” “Spend more time on nutrition–explaining how to read nutrition labels.” “More exercise tips.” “Focus more on weight.” “More mention of foods specifically good for managing bipolar.” “Focus more on mood aspects of exercise and nutrition.” “I think we should have groups with people who are involved with the study. I firmly believe in peer interaction.” “Spend more time on negative thoughts problem solving.” “Stress good sleep habits.”
Homework	3	“Digital tracking rather than paper, but I understand that is rather expensive.” “Better comprehensive handouts for mood monitoring.” “Focus on fewer tools/techniques. Too much information can be overwhelming.”
Therapist and/or study staff	1	“Provide a choice of therapists.”
Other	1	“Please allow participants to have make-up sessions for those that they missed.”

DISCUSSION

To our knowledge, this is the first randomized trial to examine a lifestyle intervention specifically tailored for bipolar disorder. We found that completion rates were comparable to or higher than those of other lifestyle interventions for individuals with serious mental illness.^8,23 Overall, most participants found the intervention helpful and indicated that they would recommend it to others. Regarding intervention attendance, Figure 1 highlights a dip in attendance around Weeks 4 to 6 as well as from Weeks 11 to 14. This is consistent with our clinical observations that some participants seemed to lose motivation after the first month of treatment, as they often did not meet their personal goals within this amount of time (despite the therapists’ efforts to make their goals realistic and attainable). We observed a similar drop in motivation approximately three-quarters of the way through the intervention, concurrent with some participants’ stated concerns that they were “running out of time” to accomplish their treatment goals. Further research is warranted to explore these visual and anecdotal observations, especially given the historical challenges with patients adhering to lifestyle interventions.

We did not find that age or education predicted participants’ perception of the intervention, suggesting that these factors may not be moderators of the main treatment outcomes, including weight loss. Unexpectedly, participants who were less manic at study entry had higher treatment satisfaction ratings at post-treatment, and vice versa. It is possible that participants experiencing less manic symptoms at baseline attributed the lifestyle intervention to improving their mania and therefore tended to rate their satisfaction with the treatment higher at post-treatment; however, future studies are needed to further explore this relationship. Regarding marital status, we found that divorced participants had higher expectations for treatment than single participants, suggesting that divorced participants may be more motivated to make lifestyle changes. However, further research is warranted to examine this association.

Feedback from participants also suggested that future lifestyle interventions should incorporate technology, such as weight-loss applications and activity trackers for mobile devices, as several participants did not like the paper and pencil diaries. Although there is limited research on the efficacy of mobile applications for patients with serious mental illness,²⁴ some evidence suggests that that using mobile applications can support lifestyle changes, such as weight loss.²⁵ The comments from the participants on the feedback questionnaire highlight the importance of the relationship between the therapist and the patient. This may suggest that a strong therapeutic relationship is important for lifestyle interventions.

These preliminary findings must be considered within the context of several limitations. The sample was small and relatively homogeneous, raising concerns about the generalizability of these data. Specifically, 84% of NEW Tx participants were Caucasian and 68% were women. We also did not measure motivation to make lifestyle changes during treatment, as motivation to lose weight and improve health tends to vary over time and can affect weight loss outcomes.²⁶

CONCLUSIONS

In sum, we found NEW Tx to be a feasible and acceptable lifestyle intervention for individuals with bipolar disorder. There appears to be a pattern to participants’ attendance rates that merits further research to maximize attendance. We also found that manic symptoms at baseline may predict treatment satisfaction, and that marital status may impact one’s expectations of treatment outcome. These data are preliminary, and further work is needed to examine and optimize lifestyle interventions for bipolar disorder, especially given recent advances in technological tools for diet and exercise tracking.