A Positive Psychology-Motivational Interviewing Program to Promote Physical Activity in Type 2 Diabetes: The BEHOLD-16 Randomized Trial

Abstract

Background:

The majority of persons with type 2 diabetes (T2D) do not meet recommended levels of physical activity, despite clear links between physical activity and superior medical outcomes in this population. The objective of this trial was to assess the feasibility and impact of a novel 16-week combined positive psychology-motivational interviewing (PP-MI) program to promote physical activity among inactive persons with T2D.

Methods:

This pilot randomized trial compared the 16-week, phone-delivered PP-MI intervention to an attention-matched diabetes counseling condition among 70 persons with T2D and low levels of baseline moderate to vigorous physical activity (MVPA; <150 minutes/week). The primary study outcomes were feasibility (assessed via rates of session completion) and acceptability (assessed via mean participant ratings [0–10] of the ease and utility of weekly sessions). Key secondary outcomes included between-group differences in improvement in positive affect, other psychological outcomes, and accelerometer-measured physical activity, assessed using mixed effects regression models, at 16 and 24 weeks.

Results:

Participants completed a mean 11.0 (SD 4.4; 79%) of 14 PP-MI phone sessions, and composite mean ratings of ease/utility were 8.6/10, above our a priori benchmarks for feasibility/acceptability (70% session completion; 7.0/10 mean ratings). PP-MI participants had small to medium effect size (ES) difference improvements in MVPA (ES difference=.34) and steps/day (ES difference=.76) at 16 weeks, with sustained but smaller effects at 24 weeks (ES difference=.22–.33).

Conclusions:

Next-step studies of this PP-MI program in T2D patients can more rigorously explore the intervention’s effects on physical activity and clinical outcomes.

Introduction

Among persons with type 2 diabetes (T2D), physical activity is associated with better glucose control, lower rates of T2D complications, and reduced mortality, independent of other health behaviors.^1,2 Despite the benefits of physical activity in T2D, the majority of T2D patients fail to meet ADA physical activity recommendations.³ Multicomponent programs for physical activity and other health behaviors do exist and have had positive results in research studies,⁴ but these programs have been difficult to implement in clinical settings,^5,6 likely due to their complexity and intensity. For this reason, new physical activity interventions in T2D are needed that are effective enough to precipitate change and feasible enough to be delivered in real-world clinical settings.

Motivational interviewing (MI) could serve as the foundation of an effective and pragmatic physical activity intervention in T2D. MI is an evidence-based style of intervention communication that works to enhance intrinsic motivation for change through a structured, patient-centered approach,⁷ and it has been used to promote physical activity and other health behaviors in numerous settings.⁸ MI is more straightforward than multicomponent interventions and can be delivered remotely with good fidelity.⁹ However, among persons with T2D, MI alone has had somewhat limited effects (effect size approximately 0.2) on physical activity,⁴ and therefore additional intervention components that can amplify the effectiveness of MI are needed.

Promoting psychological well-being could be an effective additional approach to increasing activity in T2D. Higher levels of positive psychological attributes (e.g., positive affect, optimism) predict greater increases in subsequent physical activity, controlling for demographics, medical comorbidities, and the adverse effects of anxiety and depression.^4,10 To promote these aspects of psychological well-being, positive psychology (PP) interventions utilize straightforward well-being activities (e.g., recalling positive life events, using personal strengths) in a systematic manner.¹¹ Importantly, PP interventions are simple for patients to complete and can be delivered by a wide variety of disciplines.^12,13 In randomized trials, PP interventions have led to increases in physical activity and other health behaviors in patients with medical illness.^14,15 Given that many T2D patients have diminished psychological well-being,¹⁶ PP could add a simple, effective component to an intervention for inactive persons with T2D.

A combined PP-MI program could be a powerful, scalable approach to promoting physical activity in T2D, based on a published theoretical framework.¹⁷ The program could promote physical activity though several distinct routes. The PP intervention component could directly promote constructs (positive affect and optimism) linked with increased activity.¹⁸ The MI component could act to reduce ambivalence and increase motivation to perform physical activity,¹⁹ and the PP content can further promote these MI-associated changes by boosting psychological factors that increase engagement in MI and related interventions. Specifically, PP exercises lead to increased optimism and outcome expectancy,^20,21 confidence and self-efficacy,^22,23 and greater perceived social support and interpersonal connectedness.^24,25 These individual and social-level constructs, in turn, are associated with greater efficacy of MI and other health behavior interventions.^26–29 These combined and complementary effects of PP and MI may be much more powerful in this vulnerable population than either approach alone.³⁰ A single-arm proof-of-concept trial found a phone-delivered PP-MI program in T2D to be well-accepted and associated with improvements in physical activity.³¹ However, PP-MI had not been studied in a randomized trial among T2D patients with low physical activity.

Accordingly, we conducted the Boosting Emotional well-being and Happiness in Outpatients Living with Diabetes 16-week (BEHOLD-16) pilot randomized trial to assess the feasibility, acceptability, and impact of a 16-week PP-MI phone-delivered intervention to promote physical activity in T2D patients, compared to an attention-matched diabetes counseling condition. We hypothesized that the PP-MI intervention would exceed pre-determined benchmarks for feasibility and acceptability, and that it would lead to small to medium effect size (ES) greater improvements in psychological outcomes and accelerometer-measured physical activity, compared to the control condition.

Methods

Overview

Participants in the two-arm BEHOLD-16 randomized pilot trial comparing PP-MI and an attention-matched diabetes health counseling condition were primary care patients at an academic medical center with a diagnosis of T2D, enrolled June 2017-May 2019. The study was approved by the hospital system’s Institutional Review Board (IRB) and was pre-registered on ClinicalTrials.gov (registration #: NCT03001999). All participants provided written informed consent.

Participants

Study criteria

Inclusion criteria:

1. T2D. Eligible patients met current ADA criteria³² for T2D (e.g., hemoglobin A1c [A1C] >6.5%, fasting glucose >126 mg/dl), with diagnosis confirmed by their primary medical provider. Patients with adequate blood glucose control were still eligible if they had low physical activity because such inactivity still placed them at risk for diabetes complications and adverse medical events.³³

2. Low baseline physical activity. This criterion was assessed at initial screening using the well-validated International Physical Activity Questionnaire (IPAQ).³⁴ Low physical activity was defined as a total of <150 minutes/week of moderate to vigorous physical activity (MVPA), based on ADA recommendations.³⁵

Exclusion criteria:

(1) cognitive impairment, assessed via a six-item screen,³⁶ (2) a medical condition likely to lead to death within 6 months, as noted by clinical providers or the medical record, (3) lack of fluency in English, (4) inability to participate in MVPA (e.g., due to a comorbid condition, such as severe arthritis), and (5) lack of telephone access.

Recruitment/enrollment

T2D patients in the primary care practices of an urban academic medical center were identified using the electronic medical record. Primary care clinicians reviewed potential participants for appropriateness, and study staff conducted phone screening for study criteria in those who were approved for contact by providers. Following screening, eligible patients had an initial in-person study visit at the medical center. During this initial visit, participants provided informed consent, completed self-report outcome measures, and had vital signs measured and blood drawn. Blood pressure was measured using the protocol of our institution’s Translational and Clinical Research Center by trained research nurses. Blood pressure was measured using automated blood pressure devices calibrated to industry standards, from a seating position following rest of at least five minutes.

Participants also received a hip-worn accelerometer to wear for 1 week to assess baseline levels of physical activity, then returned with the accelerometer for a second baseline study visit.

At the second baseline visit, following confirmation of adequate accelerometer wear time, participants were randomized to a study condition and met with a psychologist study interventionist to begin their corresponding treatment program. If participants did not have adequate wear time, they rewore the accelerometer, then returned again for the remainder of the visit. The interventionist provided and reviewed the condition-specific (PP-MI/diabetes counseling) treatment manual and assigned an initial exercise/activity to be completed independently by the participant for the following week; the remainder of the intervention sessions occurred by phone. All participants also received an Omron pedometer to further support activity.

Interventions

PP-MI (Table 1).

Table 1.

PP-MI intervention content (14 sessions over 16 weeks^*)

Session	PP component**	MI component***
1 (in person)	Gratitude for positive events Participants identify and reflect on three positive events that occurred in the past week.	Moving for better health/activity tracking Participants report their current activity level, set an overall activity goal, discuss the importance and confidence in making the change, and consider the pros/cons of changing their activity.
2	Gratitude letter Participants write a letter of gratitude thanking a person for their kindness.	Setting a SMART physical activity goal Participants learn about and set a SMART (specific, measurable, attainable, relevant, and time-based) activity goal.
3	Capitalizing on positive events Participants identify a positive event in daily life and then capitalize on it by telling another person or celebrating the event in some other way.	Barriers and problem solving Participants consider barriers and facilitators to being more physically active.
4	Gratitude skills application Participants select a useful activity from prior weeks, consider how to adapt it to daily life, and create a plan to utilize this skill regularly.	Reviewing and reflecting on physical activity Trainers assist participants with reviewing graphs of their weekly step counts and reflect on results.
5	Recalling past success Participants recall a successful past life event and then write about the event, their personal contribution to the success, and positive feelings elicited by recalling it.	Finding new routes Participants explore their neighborhood using an audit tool that helps them identify new places to walk.
6	Using personal strengths Participants find a specific new way to use one of their ‘signature strengths’ in the next 7 days.	Using neighborhood resources Participants consider neighborhood resources that can help them be active.
7	Using personal strengths (part 2) Participants select a second strength and find a specific new way to use it in the next 7 days.	Using equipment resources Participants consider equipment they have (or need) that could promote activity.
8	Using perseverance Participants plan and then use perseverance to work towards a specific goal.	Using social resources Participants consider social supports and resources that could help them be active.
9	Using strengths in daily life This session focuses on implementing strengths-based interventions and skills into daily life.	Reviewing and reflecting on physical activity Trainers assist participants with reviewing graphs of their weekly step counts and reflect on results.
10	Enjoyable and meaningful activities Participants complete an enjoyable activity alone, an enjoyable activity with another person, and an activity that is more deeply meaningful to them.	Managing Slips Participants learn tips about how to manage ‘slips’ to lower levels of physical activity.
11	Acts of kindness Participants plan and then complete three acts of kindness toward others within a single day.	Reducing sedentary time Participants assess the time they spend sitting and discuss strategies for reducing their sitting time.
12	The “Good Life” Participants write about what their “good life” would look like over the next year in one or more life domains.	Standing breaks Participants discuss regular standing breaks as a strategy to reduce sitting time.
13	Focusing on Meaning in Life This session focuses on implementing enjoyment- and meaning-based skills in daily life.	Increasing your strength through exercise Participants learn simple strength-training exercises and set a strength goal in addition to their regular weekly goal
14	Skills application + future planning Participants make a plan for continuing to use their PP-based skills in the future.	Reviewing progress/considering the future Trainers assist participants with reviewing their accomplishments and help them to create a plan for physical activity for the near future.

^*Bonus sessions for those completing 14 sessions prior to 16 weeks consisted of repeating a favorite PP activity, with different content or focus, and continuing the MI-based approach to physical activity, including considering barriers/facilitators to activity and setting a weekly goal.

^**Each week, the study interventionist assists the participant in considering how to apply skills from the PP exercises in daily life.

^***Each week, the interventionists also use the 5A’s model to: a) ask participants about their activity goals, (b) advise them about current activity guidelines, (c) assess readiness to set an activity goal, (d) assist participants in clarifying goals and problem-solving barriers to reaching those goals, and (e) arrange for the next session by summarizing the participant’s plan for the activity goal and scheduling the next session.

PP-MI participants independently performed weekly intervention activities as assigned in the treatment manual and completed phone sessions (30–45 minutes) each week with their study interventionist. The PP intervention component focused on the completion of PP-based activities (e.g., performing kind acts) and integrating related skills into daily life. The MI component used MI principles and goal-setting to specifically target physical activity (full manual available from authors).

The PP exercises were derived from published PP interventions and the team’s PP-MI studies.^12,31 Each weekly PP session involved a specific weekly topic and exercise (e.g., writing a gratitude letter, capitalizing on positive events; see Table 1) adapted to T2D. Fourteen weekly intervention sessions were divided into three modules focused on gratitude, personal strengths, and meaning/purpose. The sessions focused on specific weekly activities, with the final session in each module devoted to consolidating those activities and skills and using them in daily life.

We chose 14 sessions, delivered over 16 weeks, given our experience with phone-based interventions in medically ill persons^12,37 finding that this was the most feasible quantity of sessions given the need for rescheduling, vacations, illness, and other issues that led to occasional skipping of weeks even in highly motivated persons. Additional ‘bonus’ sessions (see Table 1) were created with supplemental PP and MI content for those who completed all 14 sessions before 16 weeks.

Interventionists explained each PP exercise via a guided review of the treatment manual; participants then completed an assigned exercise independently during the week and wrote about the exercise and its effects. At weekly phone sessions, the interventionist reviewed the exercise with the participant and introduced a new PP exercise for the subsequent week.

For the MI intervention component (Table 1), across all weeks, interventionists used an overall “5A’s” strategy (Ask, Advise, Assess, Assist, Arrange) from behavior change theory⁷ to assess and promote internal motivation to change. In addition, a specific MI-related weekly topic (e.g., identifying importance of [and confidence in] increasing activity, managing ‘slips’ to less activity) was discussed with participants. Participants also set and reviewed a weekly physical activity goal. As with the PP component, every fourth session was devoted to integrating program skills into daily life.

All sessions were recorded. At weekly interventionist meetings, intervention supervisors (CC, RM) listened to sessions with interventionists and provided feedback. Supervisors also rated sessions for fidelity to the PP and MI intervention components using a 14-point team-developed scale used in our prior work.³⁷

Diabetes counseling condition (Table 2).

Table 2.

Diabetes Counseling intervention

Session	Health Behavior
Part One: Self-monitoring and Healthy Behaviors
1 (in-person)	Diabetes and Risk Factors Participants learn about diabetes complications, such as cardiovascular, neurologic, and visual problems. They also learn about modifiable and non-modifiable risk factors for diabetes complications.
2	The Importance of Self-Monitoring Participants learn how and why they should monitor blood sugar, blood pressure, and cholesterol.
3	Monitoring your Blood Sugar Greater detail is provided regarding when to monitor blood sugar and how to interpret and track results.
4	Foot Care and Eye Exams Participants learn about potential foot and eye complications from diabetes and how to prevent these problems.
Part Two: Medication Adherence
5	The Importance of Medications in Diabetes Participants review medication options for managing diabetes, including insulin, other injectable medications, and oral medications, and discuss the importance of medication adherence.
6	Keeping Track of Medications Participants are introduced to methods to track their medications and are prompted to consider which may be most feasible for them.
7	Barriers to and Resources for Medication Adherence Participants are provided with practical tips for managing their medication use, after considering personal barriers to adherence and instrumental resources that can assist with adherence.
Part Three: Healthy Eating
8	The Importance of Healthy Eating in Diabetes Participants are introduced to the importance of eating healthy foods, such as fruits and vegetables and lean meats are important for good health among people with type 2 diabetes.
9	A Healthy Eating Plan Participants are provided specific strategies to make their diets healthier.
10	Barriers to and Resources for Healthy Eating Participants discuss barriers to healthy eating, along with dietary resources such as family, support groups, and their treatment teams.
Part Four: Physical activity
11	The Importance of Physical Activity in Diabetes Participants are introduced to the numerous mental and physical health benefits of being physically active.
12	How You Can Become More Physically Active Participants are prompted to consider how they can become more physically active and given strategies and practical tips to reach physical activity goals.
13	Barriers to and Resources for Physical Activity Participants are prompted to identify personal barriers to becoming active and resources they can use to overcome those barriers.
14	Planning for the Future This session serves as an overarching review of the healthy behaviors discussed during the program. Participants discuss how they can continue to make healthy choices and make a specific plan with the interventionist for the near future.

This time- and attention-matched condition was used to provide a relevant, T2D-focused control condition. It had a parallel structure to the PP-MI intervention, with a treatment manual, weekly exercises, weekly phone sessions, and pedometer provision. The diabetes counseling condition assisted participants in understanding important cardiometabolic health behaviors and their relevance to prognosis in T2D. The focus of the intervention was on multiple health behaviors pertinent to T2D, including physical activity. At weekly phone sessions, participants discussed diabetes self-care and health behaviors, and they answered questions related to their own self-care and health behavior adherence (manual available from authors). The intervention was divided into four modules over 14 weeks: overall diabetes self-care, medication adherence, diet, and physical activity (Table 2).

The same interventionists delivered both conditions to avoid interventionist-specific effects on outcomes. Diabetes counseling sessions were recorded (as in PP-MI), and study supervisors and interventionists listened to diabetes counseling sessions at weekly meetings to ensure intervention fidelity and lack of cross-condition contamination. Sessions were rated for fidelity (and lack of contamination) using a scale created for the study.

Outcome assessments and study measures

Baseline participant characteristics (sociodemographic and medical variables) were obtained at the initial baseline study visit, supplemented via the medical record. Study outcome measures were collected at the initial baseline visit and repeated at a 16-week post-intervention in-person follow-up conducted by blinded study staff. Finally, a 24-week follow-up assessment (conducted by phone to reduce participant burden) collected self-report and physical activity data; activity was assessed via accelerometers that were mailed to patients, worn for 7 days, and returned by mail. This final assessment did not assess health metrics in this trial that primarily focused on feasibility, psychological outcomes, and physical activity.

Primary outcomes: feasibility and acceptability of PP-MI

The primary aim of BEHOLD-16 was to assess the feasibility and acceptability of the PP-MI intervention. For feasibility, study interventionists captured whether subjects completed weekly study sessions (i.e., completed the week’s PP-based activity and participated in the week’s phone session [with PP discussion and MI-based goal-setting]). For acceptability, during weekly phone sessions, participants provided separate ratings of the ease and utility (‘helpfulness’) for both the PP activity and the MI topic discussed during the prior week on a 0–10 Likert scale (0=very difficult, not at all helpful; 10=very easy, very helpful) for a total of four ratings each week.

Secondary outcomes: between-group outcomes

A key secondary aim was to explore group differences in the intervention’s impact on psychological outcomes (e.g., positive affect) and accelerometer-measured physical activity at 16 and 24 weeks. Physical activity was measured via the hip-worn GT3X+ accelerometer (Actigraph, Pensacola, FL), worn by participants for 1 week at baseline, week 16, and week 24. Participants were required to have 8 hours of wear time for 4+ days, as in prior GT3X+-based studies. We assessed both MVPA (measured in mean minutes/day) and daily steps. We focused on MVPA given its links to fewer diabetes complications, reduced risk of heart disease, and lower mortality,^38,39 with the accelerometer cutoff for MVPA set at 1952 counts/minute. Total steps were also a key outcome given data that even lighter intensity activity is associated with superior medical prognosis.⁴⁰

Specific psychological measures included positive affect (main psychological target given its links to health outcomes⁴¹), measured via the Positive and Negative Affect Schedule (PANAS⁴²; internal consistency [α] in this sample=.91), and optimism, measured via the Life Orientation Test—Revised⁴³ (α=.84). We also assessed depression and anxiety (Hospital Anxiety and Depression Scale⁴⁴ (α=.84 [depression subscale] and α=.88 [anxiety subscale]), and resilience (Brief Resilience Scale⁴⁵; α=.82). Additional outcome measures included physical function (PROMIS-PF 20-item scale⁴⁶; α=.89) and self-efficacy for exercise (Self-Efficacy for Exercise scale⁴⁷; α=.87).

We selected positive affect, optimism, and resilience as main psychological outcomes because they were both intervention targets and have been associated with superior health outcomes. The PP-MI intervention focuses on improving positive affect, optimism, and resilience through its modular approach to increasing well-being. Positive affect, as measured by the PANAS, has been prospectively associated with superior health outcomes, including lower mortality.⁴⁸ Optimism measured by the LOT-R is linked with lower rates of developing heart disease, lower rates of cardiac mortality, and lower rates of all-cause mortality.^30,49 Furthermore, resilience has been linked with superior health outcomes in diabetes, including blood sugar control.⁵⁰

Finally, for the exploratory outcomes related to blood pressure, body mass index (height/weight), and A1C, these values were obtained by trained study nurses. Blood pressure was measured using the protocol of our institution’s Translational and Clinical Research Center by trained research nurses. Blood pressure was measured using automated blood pressure devices calibrated to industry standards, from a seating position following rest of at least 5 minutes. A1C was measured via LabCorp (Middleborough, MA) and our hospital’s Diabetes Research Center, an IFCC A1C reference lab.

Statistical Analysis

Descriptive statistics (means, standard deviations [SDs], and proportions) were used to summarize baseline characteristics of enrolled participants, with chi-square analyses, Fisher’s exact tests, and independent samples t-tests used to identify between-group differences. For feasibility, we recorded the mean number of completed sessions (out of 14 possible core sessions; bonus sessions were not included in the calculation). For acceptability, we calculated separate means and SDs of the ease/utility scores for the PP and MI intervention components, for a total of 4 separate mean ratings. Based on prior work using similar assessments of feasibility and acceptability,^12,37 we denoted 70% completion of assigned sessions and mean ratings of 7.0 on 0–10 ease/utility scales as our a priori benchmarks for feasibility and acceptability. We assessed whether the intervention was numerically superior to these benchmarks, rather than reaching statistical significance in surpassing these benchmarks, as per prior work using such benchmarks.^12,37

For our secondary aim examining group differences in study outcomes immediately post-intervention and 8 weeks later, we used mixed effects regression models with an unstructured covariance matrix to examine between-group differences in change from baseline on each outcome at 16 and 24 weeks, using an intent-to-treat model. In these models, we controlled for age, gender, medical comorbidity [measured via the Charlson Comorbidity index⁵¹], and baseline MVPA, given that these factors may be related to psychological and activity-based measures, to reduce variance in this small trial and allow comparison with other PP-MI work in T2D.⁵² Mixed effects models allow inclusion of all participants in study analyses, even those with some missing data. We used p=.05 (two-tailed) to denote statistical significance. Because this study (with a primary focus on feasibility and acceptability) was not designed to detect significant between-group differences, we also calculated effect size (ES) differences between groups by dividing the coefficient (estimated mean difference [EMD]) by the estimated SD of the residual of the mixed model at baseline. All analyses were performed using Stata 15.2 (StataCorp: College Station, TX).

The sample size of N=70 was selected based on similar studies of PP-based interventions in cardiac patients that examined feasibility and acceptability in the same manner. In the most recent such study, 84% of sessions were completed;¹² based on this rate and a planned 35 participants assigned to PP-MI, the BEHOLD-16 trial was powered at >80% to detect a true proportion of 70%+ sessions completed, assuming a moderate SD of 15%. For acceptability, based on prior ratings for the ease and utility of PP-related sessions (mean 8.1 [SD 2.3]),¹² this trial had >95% power to detect mean ratings of 7.0 or greater for these scales. This initial study was not designed to detect statistically significant between-group differences in study measures at 16 and 24 weeks (secondary/exploratory aim) with N=70.

Results

Seventy eligible patients (n=35 per group) were enrolled and randomized (see flow diagram, Figure 1). Participant characteristics are provided in Table 3. Participants’ mean age was 63.7 (SD 10.4) years, 32 (46%) were women, and 54 (77%) were non-Hispanic White. Follow-up data were obtained from 57 (81%) of participants at one or both time points (n=57 at 16 weeks; n=52 at 24 weeks). There were four serious adverse events (medical hospitalizations) in the PP-MI group and none in the control group; rates of events did not significantly differ between groups (Fisher’s exact test [two-tailed]: p=.11).

Figure 1.

Study flow diagram

Table 3.

Baseline participant characteristics

Characteristic	Overall	PP-MI	MI alone	Test Statistic	p
Sociodemographic characteristics
Age (M [SD])	63.7 (10.4)	64.6 (10.2)	62.9 (10.7)	t= −0.66	.51
Female gender (N [%])	32 (45.7)	16 (45.7)	16 (45.7)	X2= 0.00	1.0
Non-Hispanic White (N [%])	54 (77.1)	24 (68.6)	30 (85.7)	*	.37
Medical characteristics
Neuropathy (N [%])	12 (17.1)	6 (17.1)	6 (17.1)	X2= 0.00	1.0
Nephropathy (N [%])	11 (15.7)	5 (14.3)	6 (17.1)	X2= 0.11	.74
Hyperlipidemia (N [%])	67 (95.7)	34 (97.1)	33 (94.3)	*	1.0
Hypertension (N [%])	53 (75.7)	28 (80.0)	25 (71.4)	X2= 0.70	.40
Coronary artery disease (N/%)	16 (22.9)	9 (25.7)	7 (20.0)	X2= 0.32	.57
Charlson Comorbidity Index (age-adjusted; M[SD])	3.9 (1.7)	4.0 (1.8)	3.8 (1.6)	t= −0.63	.53
Medications at enrollment
Metformin (N [%])	68 (97.1)	35 (100)	33 (94.3)	*	.49
Insulin (N [%])	3 (4.3)	2 (5.7)	1 (2.9)	*	1.0
Sulfonylureas/meglitinides	27 (38.6)	13 (37.1)	14 (40.0)	X2= 0.06	.81
Antidepressant (N [%])	16 (22.9)	6 (17.1)	10 (28.6)	X2=1.30	.26
Anxiolytic (N [%])	8 (11.4)	3 (8.6)	5 (14.3)	*	.71

^*Fisher’s Exact Test used; no test statistic

Feasibility and acceptability.

PP-MI participants completed a mean of 11.0/14 (SD 4.4) sessions (79% of all possible sessions), and 28 participants (80%) completed a majority of PP-MI phone sessions (8+). Of note, 79% (mean 11.1/14 sessions [SD 3.9]) of diabetes counseling intervention sessions were also completed. For acceptability, participants’ mean ratings of PP-MI activity ease and utility were 8.2 (SD 2.0) and 8.8 (SD 1.5), respectively, for the PP component, and 8.4 (SD 1.8) and 9.0 (SD 1.4) for MI.

Psychological and functional outcomes (Table 4).

Table 4.

Secondary outcomes: between-group differences in change from baseline on study outcome measures

Measure	16 weeks					24 weeks
	EMD	SE	z	p	ES**	EMD	SE	z	p	ES**
Positive affect
Positive affect (PANAS)	1.16	1.52	0.76	.45	.16	1.81	1.83	0.99	.32	.24
Physical activity
MVPA (minutes/day)	4.05	3.04	1.33	.18	.34	2.66	3.26	0.82	.41	.22
Steps (per day)	1278	369	3.46	.001*	.76	546	406	1.35	.18	.33
Additional psychological measures
Optimism (LOT-R)	1.85	1.09	1.70	.09†	.35	1.37	1.26	1.09	.28	.26
Resilience (BRS)	−0.55	0.85	−0.65	.51	−.27	−0.41	0.78	−0.53	.60	−.20
Depression (HADS-D)	0.62	0.74	0.84	.40	.17	−0.43	0.80	−0.54	.59	−.12
Anxiety (HADS-A)	0.18	0.90	0.20	.84	.04	0.64	0.91	0.70	.48	.16
Additional functional/behavioral measures
Exercise self-efficacy (SEE)	21.48	6.74	3.19	.001*	1.01	19.10	6.24	3.06	.002*	.90
Physical function (PF-20)	−0.14	1.49	−0.09	.93	−.02	1.22	1.59	0.76	.44	.14
Medical outcomes (16 weeks only)
Hemoglobin A1c (%)	−0.14	0.25	−0.54	.59	−.09	-	-	-	-	-
Systolic BP (mm Hg)	−6.57	4.58	−1.44	.15	−.54	-	-	-	-	-
Diastolic BP (mm Hg)	−3.29	2.78	−1.18	.24	−.37	-	-	-	-	-
Body mass index (kg/m2)	−0.33	0.29	−1.13	.26	−.06	-	-	-	-	-

^*p<.05,

^†p<.10

^**Effect size (ES) differences between groups calculated by dividing the coefficient (estimated mean difference) by the estimated SD of the residual of the mixed model at baseline

BP=blood pressure; BRS= Brief Resilience Scale; EMD=Estimated mean difference; ES=Effect size; HADS=Hospital Anxiety and Depression Scale (HADS-A=Anxiety Subscale, HADS-D=Depression); LOT-R=Life Orientation Test, Revised; MVPA=Moderate to vigorous physical activity; PANAS=Positive Affect Negative Affect Schedule; PF-20=PROMIS Physical Function 20-item measure; SEE=Self Efficacy for Exercise scale.

PP-MI was associated with small, non-significant ES difference improvements in change from baseline PANAS (positive affect) score at both 16 and 24 weeks (16 weeks: estimated mean difference [EMD] 1.16 [SE 1.52], p=.45, ES=.16; 24 weeks: EMD 1.81 [SE 1.83], p=.32, ES=.24) and LOT-R (optimism) score (16 weeks: EMD 1.85 [SE 1.09], p=.09, ES=.35; 24 weeks: EMD 1.37 [SE 1.26], p=.28, ES=.26), compared to the attention-matched control condition. There were more mixed effects on depression, anxiety, and physical function.

Regarding physical activity, compared to the diabetes counseling condition, PP-MI was associated with greater improvements in MVPA (16 weeks: EMD 4.05 [SE 3.04], p=.18, ES=.34; 24 weeks: EMD 2.66 [SE 3.26], p=.41, ES=.22) and daily steps, with differences in steps reaching statistical significance at 16 weeks (16 weeks: EMD 1278 [SE 369], p=.001, ES=.76; 24 weeks: EMD 546 [SE 406], p=.18, ES=.33). PP-MI also led to significantly greater improvements in exercise self-efficacy (16 weeks: EMD 21.48 [SE 6.74]; p=.001; ES=1.01; 24 weeks: EMD 19.10 (SE 6.24); p=.002; ES=.90).

Health outcomes (vital signs and A1C; exploratory).

PP-MI was associated with small to moderate ES difference improvements in blood pressure (systolic: EMD −6.57 [SE 4.58]; p=.15; ES=−0.54; diastolic: EMD −3.29 [SE 2.78]; p=.24; ES=−0.37) and small ES difference improvements in body mass index and A1C (Table 4); these were not statistically significant.

Discussion

In this randomized trial, we found that a 16-week, phone-delivered PP-MI program to promote physical activity among T2D patients exceeded a priori thresholds for feasibility and acceptability, with over three-quarters of all possible phone sessions completed by participants and mean ratings of intervention ease and utility exceeding 8.0 out of 10. These outcomes are consistent with such metrics for related behavioral interventions in medical settings,^12,53 and suggest that a phone-delivered PP-MI intervention may be a well-accepted approach to promote physical activity among inactive T2D patients, who are at risk for adverse medical outcomes.³²

In addition, compared to the time- and attention-matched diabetes counseling control condition, PP-MI led to greater effects on positive affect and optimism (small ES differences) at the 16 and 24-week time points, though these effects were not statistically significant in this randomized pilot trial. These effects mirror those in a recent study of PP-MI among acute coronary syndrome (ACS) patients,¹² though they are lesser than in such a prior study. It may be the case that the longer intervention (16 weeks) was less successful overall than the 12-week ACS intervention given both lower session completion rates and smaller effects on positive affect, or that the chronic nature of T2D (as opposed to a more acute event) makes it more difficult to modify such affect. Such findings may be important given that these psychological constructs are independently associated with superior medical outcomes in T2D and cardiovascular disease.^30,54

Furthermore, PP-MI led to greater improvements from baseline in MVPA and total steps at 16 and 24 weeks, with small (to medium) effect size differences (ES difference=.22–.34) in MVPA at follow-up and somewhat larger effects (ES difference=.33–.76) on steps, with only the effects on steps at 16 weeks reaching significance. These improvements were concurrent with significantly greater improvements in exercise self-efficacy among those in the PP-MI condition. These findings regarding the magnitude of effect on activity match observational studies finding relationships between psychological well-being and physical activity in healthy persons, medical populations, and T2D patients,³⁰ and the prior randomized trial of PP-MI in post-ACS patients observed similar effects on MVPA and steps.¹² It is notable that lighter activity, measured via steps, appeared to improve more than MVPA, and it may be that the intervention could be better-suited to promoting light activity, an increasingly important outcome that is associated with superior health outcomes.^40,55

Though PP-MI numerically surpassed feasibility and acceptability thresholds and appeared to have effects on positive psychological well-being and physical activity in this initial randomized trial, several additional outcomes are notable. Rates of session completion were somewhat lower than in prior work, including a 12-week PP-MI intervention in cardiac patients.¹² This suggests that a 16-week intervention requiring weekly PP and MI activities and weekly phone sessions may be too intensive for some patients. In addition, the intervention’s effects on other measures that were not directly targeted (e.g., depression, anxiety, physical function, and medical outcomes) were mixed, suggesting that elements of the intervention may need to be further optimized.

Larger studies of an optimized PP-MI program, with longer follow-up, are needed to better understand the intervention’s efficacy. However, PP-MI could have substantial promise as a practical, effective physical activity intervention in T2D. PP activities are simple, require minimal provider training, and can be efficiently delivered during office visits or remotely. Likewise, MI training is broadly available, and a wide range of clinical providers can deliver MI-based programs with good fidelity.⁷

This trial had strengths and limitations. Study strengths included use of a relevant, diabetes-focused, and attention-matched control condition, rigorous and ongoing assessments of intervention fidelity, and objective measurement of physical activity via research-grade accelerometry. Limitations included the conduct of the trial in a single academic medical center and intervention delivery by clinical psychologists, though in prior work PP and PP-MI interventions have been delivered by social workers and bachelor’s level clinicians with high fidelity and to good effect.^12,13 This study was powered to measure feasibility and acceptability, but it was not powered to detect significant differences in clinical outcomes and a larger trial is needed to examine such outcomes.

Finally, an important aspect of this trial is that while the control condition was substantial, and matched for time and a focus on health behaviors, the control condition—in contrast to the experimental condition—was not specifically focused solely on physical activity. This suggests that the differences on physical activity outcomes and exercise self-efficacy may have been driven by the PP-MI intervention’s focus on physical activity rather than a specific novel element of the program. Future studies could use a specific MI-based physical activity program as a control to assess the incremental benefits of PP-MI.

In conclusion, a phone-delivered PP-MI program to promote psychological well-being and physical activity among T2D patients was feasible, well-accepted, and led to increases in physical activity compared to an attention-matched, diabetes-specific counseling condition. Next-step studies can more definitively assess the sustained effects of PP-MI on physical activity, glycemic control, and cardiometabolic risk markers. If effective, such an intervention has the potential to improve diabetes and cardiovascular health outcomes in this high-risk population.