Feasibility Trial of a Mind–Body Activity Pain Management Program for Older Adults With Cognitive Decline

Abstract

Background and Objectives

The relationship between chronic pain (CP) and cognitive decline (CD) is bidirectional among older adults. The CP–CD comorbidity can progressively worsen cognitive, physical, emotional, and social functioning with aging. We explored the feasibility and outcomes associated with 2 mind–body activity programs for CP and CD that focus on increasing walking using time goals (Active Brains) or step-count reinforced via Fitbit (Active Brains–Fitbit).

Research Design and Methods

Older adults with CP and CD participated in a nonrandomized open pilot of Active Brains (n = 6) and Active Brains–Fitbit (n = 6) followed by exit interviews. Quantitative analysis explored feasibility markers and signals of improvement on physical, cognitive, and emotional function, as well as additional program targets. Qualitative analyses were predominantly deductive and applied the Framework Method to enhance the programs and methodology.

Results

Both programs met a priori feasibility benchmarks. We found within-group improvements for pain intensity, pain-specific coping, physical function, and cognitive function in both programs. Exit interviews confirmed high satisfaction with both programs.

Discussion and Implications

Our mixed-methods data provide preliminary evidence of feasibility, showed promise for improving outcomes, and yielded critical information to further enhance the programs. We discuss “lessons learned” and future directions.

Chronic pain (CP) is common, costly, and challenging to treat (Glajchen, 2001). The prevalence of pain increases with age, 25%–50% among older adults in the community (Patel et al., 2013) and over 80% among those living in nursing homes (Won et al., 2004), and is associated with substantial decline in emotional, physical, and social function (Dueñas et al., 2016). Older adults with CP are two times more likely to also report cognitive decline (CD; Cravello et al., 2019).

Objective (confirmed by cognitive testing) or subjective CD (self-report only), defined as decreases in cognitive performance that exceed normal aging (Patel et al., 2013), is also common among older adults and associated with decline in emotional, physical, and social function (Ma, 2020). The CD–CP relationship is bidirectional among older adults. CP accelerates CD and increases risk for dementia (Whitlock et al., 2017), while neurodegeneration associated with CD alters pain perceptions (van Kooten et al., 2016), increasing risk for CP. Consequently, older adults with CP and CD can become caught in a “disability spiral” whereby cognitive, physical, emotional, and social functioning progressively worsen over time (Gagliese et al., 2018; Mace et al., 2020).

There are no evidence-based treatments that comprehensively address the CP–CD comorbidity among older adults (Cravello et al., 2019). First-line pharmacological approaches treat only CP or CD and are limited by minimal efficacy (Raina et al., 2008), adverse events (Shorr et al., 1992), and cognitive dysfunction side effects (Wright et al., 2009). Nonpharmacological interventions that teach adaptive coping skills are relatively efficacious in improving physical, emotional, and social functioning in CP (Wetherell et al., 2011), but do not address CD and are not tailored to older patients who have CD. Increasing pain self-efficacy has the potential to preserve functioning despite coping with CP (Cheng et al., 2018) and reduce catastrophic beliefs about pain linked to anxiety and depression (Wood et al., 2016).

Physical activity programs focused on walking are safe for older adults with CP (Wood et al., 2016) and have shown benefit in improving both emotional and physical function (Terrier et al., 2019). Physical activity has also documented benefits for improved cognition. Quota-based walking (i.e., noncontingent on pain level or cognitive ability), reinforced by a digital monitoring device (e.g., Fitbit), can address engagement and adherence, although problems due to misconceptions about pain and fear avoidance remain (O’Connor et al., 2015).

Our team developed Active Brains and Active Brains–Fitbit through qualitative focus groups with older adults who have CP and CD (Mace et al., 2020). Here, we report on a nonrandomized open pilot of the two programs followed by qualitative exit interviews. Our methodology and hypotheses follow the Obesity-Related Behavioral Intervention Trials (ORBIT) and NCCIH models (Czajkowski et al., 2015; National Center for Complementary and Integrative Health [NCCIH], 2020) of behavioral intervention development. In the initial steps, both frameworks emphasize defining the methodology and respective treatment targets, assessing feasibility, demonstrating signal of improvement, and refining prior to efficacy testing. Accordingly, we hypothesize that both programs would meet a priori set feasibility benchmarks (Greenberg et al., 2019; Hypothesis 1), and that both programs will be associated with signals of improvement in physical, cognitive, and emotional outcomes, as well as additional program targets (Hypothesis 2). We also expect that individual exit interviews will assist in further optimizing the programs and methodology prior to conducting a clinical trial with a larger and more diverse sample.

Method

Participants

We recruited participants through referrals from the Pain Clinic at the Massachusetts General Hospital, and Institutional Review Board (IRB)-approved flyers. A trained research coordinator conducted phone screens and a licensed clinical health psychologist reviewed all cases prior to enrollment. See Supplementary Figure 1 for the participant flow.

Inclusion criteria were: (a) age ≥60, (b) nonmalignant musculoskeletal pain for over 3 months, (c) self-reported CD, (d) ability to perform a 6-min walk test (6MWT), (e) access to a smartphone, (f) on a stable dose of pain and/or psychotropic medication, and (g) willingness to complete the program and study assessments.

Exclusion criteria were: (a) medical illness expected to worsen in the next 6 months, (b) serious untreated psychiatric disorders, (c) current substance abuse, (d) current suicidal ideation, (e) regular use of a digital monitoring device (e.g., Fitbit) in the last 3 months, (f) routine intensive physical exercise daily for >30 min, (g) practice of meditation or yoga for >45 min once a week in the last 3 months, and (h) inability to walk without any assistance (e.g., cane, walker).

Table 1 presents demographics and clinical characteristics for the final sample. We included older adults with CP and patients with either objective or subjective CD given that both: (a) are associated with CP (Cravello et al., 2019), (b) are common in pain clinics (Binnekade et al., 2018), (c) can indicate preclinical dementia (Jessen, 2014), and (d) do not necessitate different approaches to pain coping skills (Mace et al., 2020). These eligibility criteria are consistent with similar mind–body trials with CP (Greenberg et al., 2019; Wetherell et al., 2011).

Table 1.

Demographic and Baseline Clinical Characteristics by Open Pilot Group

Variable	Open pilot: Active Brains, n = 5		Open pilot: Active Brains–Fitbit, n = 6
	M (SD)	N (%)	M (SD)	N (%)
Age	73.6 (4.99)		67.1 (5.91)
Gender
Male		1 (20.0)		1 (16.7)
Female		4 (80.0)		5 (83.3)
Race
Asian		0 (0.0)		1 (16.7)
Black/African American		2 (40.0)		1 (16.7)
White		3 (60.0)		4 (66.7)
Marital status
Single, never married		1 (20.0)		2 (33.3)
Married		2 (40.0)		3 (50.0)
Separated or divorced		1 (20.0)		0 (0.0)
Widowed		1 (20.0)		1 (16.7)
Income
<$10,000		1 (20.0)		2 (33.3)
$10,000 to <$35,000		2 (40.0)		2 (33.3)
$35,000 to <$75,000		2 (40.0)		0 (0.0)
≥$75,000		0 (0.0)		1 (16.7)
Choose not to answer		0 (0.0)		1 (16.7)
Education
Less than high school (<12 years)		0 (0.0)		1 (16.7)
Some college/associates degree (<16 years)		2 (40.0)		2 (33.3)
Completed college (16 years)		3 (60.0)		1 (16.7)
Graduate/professional degree (>16 years)		0 (0.0)		2 (33.3)
Pain medications
Opioids		1 (20.0)		2 (33.3)
Other analgesics		3 (60.0)		2 (33.3)
None		1 (20.0)		2 (33.3)
Reported pain conditions
Fibromyalgia		3 (60.0)		2 (33.3)
Migraine		1 (20.0)		3 (50.0)
Spinal pain		2 (40.0)		2 (33.3)
Neck pain		1 (20.0)		2 (33.3)
Osteoarthritis		2 (40.0)		1 (16.7)
Other		4 (80.0)		3 (60.0)
Pain duration
≤5 years		1 (20.0)		3 (50.0)
6–10 years		2 (40.0)		2 (33.3)
≥11 years		2 (40.0)		1 (16.7)

Note: The five most common pain conditions reported by patients are listed individually.

Procedures

Our IRB (#2018P002152) approved this study. The first six eligible participants were assigned to Active Brains (Group 1, September 2019) and the next six were assigned to Active Brains–Fitbit (Group 2, January 2020). The clinician reviewed the consent form with patients in a group setting. Participants completed the quantitative assessments and received a wGT3X-BT ActiGraph accelerometer. Participants were instructed to wear the accelerometer every day while awake (except when showering or swimming) for 1 week until the first intervention session (see Supplemental Materials for additional details). Subsequently, participants completed 8 weeks of in-person Active Brains or Active Brains–Fitbit (90 min each). Procedures for the quantitative assessments and accelerometer were repeated postintervention. All program completers elected to participate in an in-person (Group 1) or phone (Group 2 due to coronavirus disease 2019 [COVID-19] precautions) exit interview (60 min). Participants earned $30–$50 for each of the two assessments, $10 for transportation to each of the eight sessions, and $30 for the optional exit interview ($190 maximum).

Active Brains and Active Brains–Fitbit programs

Our iterative plan for intervention development, consistent with the goals of early feasibility studies, is illustrated in Supplementary Figure 2. Supplementary Table 1 outlines the topics and skills for each session of Active Brains and Active Brains–Fitbit (Mace et al., 2020). Both programs teach (a) walking skills to gradually increase step-count through SMART goal setting, individualized nonpain contingent quota-based pacing (e.g., walk for 30 min or meet a step goal of 5,000 steps), and engagement in meaningful activities; (b) mind–body skills to reduce reactivity and catastrophizing to pain or fear of CD through diaphragmatic breathing, body scanning, and mindfulness exercises; (c) pain–cognition awareness skills to correct misconceptions about CP and CD that may impede participation and understand the disability spiral (e.g., how sedentariness perpetuates CP and CD); (d) cognitive functioning skills to develop cognitive compensatory strategies and increase intellectual stimulation; and (e) social and emotional functioning skills to manage negative reactions from others and cope with stress or walking setbacks (positivity, self-compassion, and gratitude). The two programs are identical except for the Fitbit and quota-based pacing described below.

Quota-based walking in Active Brains–Fitbit (Group 2).

Initial walking goals were determined by accelerometer step-count data from the baseline assessment extracted via ActiLife (ActiLife, 2013) software. Participants received a Fitbit (Fitbit Alta HR, 2018) on Session 1 and were instructed to charge, sync, and wear it daily on their wrist during waking hours (except showering). The research coordinator programmed each participant’s Fitbit walking goal each week. At the end of each week, participants who achieved their step goal and adhered to the 5-day valid wear time criterion (≥7 hr) increased their step goal by 10%–20% according to guidelines for quota-based pacing (Greenberg et al., 2019, 2020; Patel et al., 2015). Nonwear days (<7 hr) were excluded from goal calculations (Schrack et al., 2016). Participants in the Active Brains group created weekly SMART goals without the Fitbit (i.e., self-assessment of walking progress).

Measures

Quantitative measures

We selected quantitative measures informed by on our prior mixed-methods development study (Mace et al., 2020). We assessed the following: (a) physical function: wGT3X-BT ActiGraph accelerometer (Actigraph, Pensacola, FL), 6MWT (Jamison et al., 2017), PROMIS Physical Function (Stone et al., 2016), World Health Organization Disability Assessment Schedule (WHODAS; Üstün et al., 2010), and Godin Leisure-Time Exercise Questionnaire (Amireault & Godin, 2015); (b) cognitive function: Montreal Cognitive Assessment (MoCA; Nasreddine et al., 2005) and 12-item Everyday Cognition Scale (eCog-12; Tomaszewski Farias et al., 2011); (c) emotional function: the PROMIS Anxiety (v1.08a) and PROMIS Depression (v1.08b; Pilkonis et al., 2011); (d) social function: PROMIS Emotional Support (4a; Hahn et al., 2014) and 8-item UCLA Loneliness Scale (Russell et al., 1978); (e) pain intensity: Numerical Rating Scale (Farrar et al., 2001); (f) pain-specific coping: Pain Catastrophizing Scale (Sullivan et al., 1995) and Pain Self-Efficacy Questionnaire (Nicholas, 2007); and (g) general coping: Measure of Current Status—Part A (MOCS-A; Carver, 2006), Cognitive and Affective Mindfulness Scale— Revised (Feldman et al., 2007), and 6-item Gratitude Questionnaire (GQ-6; McCullough et al., 2002). The Supplementary Materials contain additional details on the quantitative measures.

Feasibility measures

All benchmarks were determined a priori based on intervention development guidelines (Czajkowski et al., 2015) and our previous feasibility pilot studies (Greenberg et al., 2019, 2020; Vranceanu et al., 2018). The Supplementary Materials contain additional details on the feasibility measures.

Exit Interviews

A semistructured exit interview guide was used to elicit a priori themes: (a) Active Brains skills for enhancing physical, cognitive, and emotional functioning (e.g., “What were your impressions of the Body Scan?” “How can we better teach this skill?”); (b) barriers/facilitators to home practice, Fitbit, and assessments (e.g., “In what ways did the Fitbit help or hinder your walking?”); and (c) the program structure and in-group experiences (e.g., “What made it hard to participate in the program?”). The Supplementary Materials contain additional exit interview details.

Statistical Methods

Quantitative analysis

The goal of our quantitative analysis was to preliminarily assess the extent to which the groups were responsive to the population-specific needs identified in our prior work (Mace et al., 2020), and if our measures were sensitive to changes in outcomes (Leon et al., 2011). We used R (R Development Core Team, 2019) and RStudio (RStudio Team, 2019) for quantitative analyses. The sample size was appropriate for exploring feasibility and outcomes for future trials (Leon et al., 2011). We calculated proportion of patients with improvement, paired t tests, and Cohen’s d effect sizes (Sullivan & Feinn, 2012) for each quantitative measure. We used nonparametric tests when appropriate and refrained from between-group analyses (Leon et al., 2011). We prioritized interpretations of pre–post differences in this pilot to small (Cohen’s d = 0.20), medium (d = 0.50), and large (d = 0.80) categorizations based on Cohen (1988).

Qualitative analysis

The goal of our qualitative analysis was to further enhance the programs and study methodology for future trials based on direct participant feedback. Qualitative and quantitative data were integrated to corroborate the feasibility of the programs, contextualize the findings at the group and individual participant levels, and understand why changes in the outcomes may have occurred (O’Cathain et al., 2010). Qualitative analysis was primarily deductive (Braun & Clarke, 2006) using the Framework Method based on our prior work (Mace et al., 2020), which allowed for some inductive flexibility to explore unexpected needs and preferences of participants. Briefly, all exit interview recordings were manually transcribed, de-identified, and checked for accuracy, and qualitative data were organized using NVivo 12 (Richards, 2018). The research coordinators and lead investigators individually read all transcripts, discussed an initial thematic framework, and created the codebook. The research assistants independently conducted line-by-line deductive coding (Corbin & Strauss, 2008) of critical participant statements on one practice transcript. We incorporated gaps identified by the independent raters (initially coded as “other”; Gale et al., 2013), incorporated divergent viewpoints, and agreed on the final codebook for remaining transcripts (interrater reliability, κ = 0.93). Disagreements were resolved through consensus. We charted and visualized the qualitative data using a spreadsheet matrix, balancing data reduction while maintaining the original context and sentiment, to inform improvements in programs and methodology (Gale et al., 2013). The Qualitative and Mixed Methods Research Unit at MGH provided consultation on all Framework Method analytical steps.

Results

Participant Characteristics

Out of the 34 inquiring patients, 27 (79.4%) agreed to complete screening Supplementary Figure 1) and 12 adults with CP and CD enrolled in the two open pilot groups (Group 1: Active Brains, n = 6; Group 2: Active Brains–Fitbit, n = 6). One participant dropped from Active Brains due to financial and transportation concerns before Session 1 for an overall completion rate of 91.7%. All but one (who underwent an elective spinal surgery) of the 11 participants completed the open pilot and assessments and an exit interview (90.9%).

As presented in Table 1, the overall sample was majority female (81.8%) and White (63.6%). The sample was more evenly distributed based on marital status, income, and education level. All reported multiple pain conditions, the duration of which widely varied between ≤5 years (36.4%), 6–10 years (36.4%), and ≥11 years (27.2%). Nearly a third (27.2%) were prescribed opioids while the remainder took other analgesics (45.6%) or did not take pain medications (27.2%).

At baseline, participants in both groups (four out of five, Active Brains; five out of six, Active Brains–Fitbit) were sedentary (<5,000 steps) at baseline (Tudor-Locke & Bassett, 2004). Participants reported moderate-to-severe pain intensity at rest (Active Brains: M = 7.2, SD = 1.9; Active Brains–Fitbit: M = 6.8, SD = 1.6) and with activity (Active Brains: M = 8.4, SD = 1.1; Active Brains–Fitbit: M = 6.5, SD = 1.4) that exceed population estimates (Nicholas et al., 2008) and prior work (Mace et al., 2020). MoCA scores (<26) indicated that, on average, both groups had clinically significant cognitive impairment (Active Brains: M = 24.2, SD = 4.8; Active Brains–Fitbit: M = 23.7, SD = 6.7). eCog-12 scores (Active Brains: M = 2.0, SD = 0.8; Active Brains–Fitbit: M = 2.7, SD = 0.7) were comparable to mild cognitive impairment samples, suggesting that participants’ daily functioning was impaired by CD (Farias et al., 2011). Active Brains–Fitbit had higher anxiety and depression PROMIS scores at baseline than Active Brains (+1 SD).

Feasibility Markers

Table 2 presents the feasibility and acceptability results for both groups (Table 2). Active Brains and Active Brains–Fitbit met criteria for “excellent” on most of the a priori benchmarks. Adherence to the ActiGraph accelerometer was “good” in both groups. Credibility, expectancy, and satisfaction were high in both groups. No adverse events were reported.

Table 2.

Feasibility and Acceptability of the Open Pilot Groups

Outcome	Active Brains	Active Brains–Fitbit
Feasibility of recruitment	27 participants out of 34 successfully contacted agreed to complete screening (excellent)
Credibility	5 out of 5 participants (100%) scored over midpoint (excellent)	5 out of 6 participants (83.3%) scored over midpoint (excellent)
Expectancy	4 out of 5 participants (80%) scored over midpoint (excellent)	5 out of 6 participants (83.3%) scored over midpoint (excellent)
Client satisfaction	5 out of 5 participants (100%) scored over midpoint (excellent)	5 out of 6 participants (83.3%) scored over midpoint (excellent)
Acceptability of treatment	4 out of 5 participants (80%) attended ≥6 out of 8 sessions (excellent)	6 out of 6 participants (100%) attended ≥6 out of 8 sessions (excellent)
Adherence to ActiGraph and Fitbit	3 out of 5 participants at baseline and 4 out of 5 at posttest recorded ≥5 days (of ActiGraph data [70% total; good])	4 out of 6 participants at baseline and 5 out of 6 participants at posttest recorded ≥5 days (of ActiGraph data [75% total; good])
		5 out of 6 participants (83.3%) recorded at least 5 out of 7 days of Fitbit data every week (excellent)
		Participants met 66.7% of their step-count goals
Adherence to homework	4 out of 5 participants (80%) completed ≥5 out of 7 weeks of homework (excellent)	5 out of 6 participants (83.3%) completed ≥5 out of 7 weeks of homework (excellent)
Therapist adherence to sessions	100% adherence (excellent)	100% adherence (excellent)
Feasibility of quantitative measures	99.1% of questionnaires were partially completed (excellent)	98.5% of questionnaires were partially completed (excellent)
Patients’ perception of improvement	4 out of 5 participants (80%) reported overall improvement (excellent)	5 out of 6 participants (83.3%) reported overall improvement (excellent)
Adverse events	None	None

Quantitative Outcomes

Descriptive and change statistics for each outcome by group are presented in Table 3 (physical function, cognitive function, pain intensity) and Supplementary Table 3 (emotional functioning, social functioning, pain-specific coping, general coping).

Table 3.

Objective, Performance-Based, and Self-reported Physical, Pain, and Cognitive Functioning Outcomes

Physical function assessments	Active Brains						Active Brains–Fitbit
	Baseline	Posttest	M difference (post–pre)	t value	p value	Cohen’s D	Baseline	Posttest	M difference (post–pre)	t value	p value	Cohen’s D
	M (SD)	M (SD)					M (SD)	M (SD)
ActiGraph average steps	2,604.0 (2,806.3)	3,134.2 (2,983.3)	530.2	0.3	.8	0.2	4062.9 (2,276.3)	5,176.9 (3,176.6)	1,114.0	0.7	.5	0.5
6MWT (m)	284.0 (99.3)	337.5 (70.4)	53.5	0.9	.4	0.7	471.7 (106.7)	415.3 (136.5)	−56.4	0.7	.5	0.5
PROMIS Phys	37.0 (2.0)	38.4 (8.8)	1.4	0.4	.7	0.3	37.2 (5.4)	42.8 (9.7)	5.6	1.2	.2	0.8
WHODAS	24.6 (11.9)	15.9 (8.3)	–8.7	1.3	.2	1.0	31.4 (7.5)	28.6 (10.7)	−2.8	0.5	.6	0.3
GLQ (total)	17.4 (21.3)	40.0 (32.0)	22.6	1.3	.2	0.9	22.0 (17.4)	35.7 (27.8)	13.7	1.0	.3	0.6
Pain at rest	7.2 (1.9)	4.4 (1.7)	–2.8	2.5	.04	1.7	6.8 (1.6)	4.7 (3.4)	−2.1	1.4	.2	0.9
Pain with activity	8.4 (1.1)	5.0 (2.2)	–3.4	3.0	.02	2.1	6.5 (1.4)	3.4 (3.1)	−3.1	2.2	.1	1.5
MoCA	24.2 (4.8)	25.3 (3.5)	1.1	0.4	.7	0.3	23.7 (6.7)	26.0 (3.7)	2.3	0.6	.6	0.5
eCog-12	2.0 (0.8)	1.7 (0.7)	–0.3	0.7	.5	0.5	2.7 (0.7)	2.3 (0.7)	−0.4	0.8	.4	0.5

Note: 6MWT = 6-min walk test; eCog-12 = Everyday Cognition Scale; GLQ = Godin Leisure-Time Questionnaire; MoCA = Montreal Cognitive Assessment; PROMIS Phys = PROMIS physical function; WHODAS = World Health Organization Disability Assessment Schedule.

Physical function

Both groups showed comparable improvements in physical function (Supplementary Figure 3). Average improvement in step-count (Active Brains = 530.2; Active Brains–Fitbit = +1114.0) were small–medium. Both groups reported medium–large improvements in physical function on the PROMIS (10/11 improved; Active Brains = +1.4; Active Brains–Fitbit = +5.6) and WHODAS (7/11 improved; Active Brains = −8.7; Active Brains–Fitbit = −2.8). The 6MWT results were mixed, with increased performance in Active Brains (+53.5) but decreased in Active Brains–Fitbit (−56.4). The perceived increase in step-count was higher for the Active Brains–Fitbit group (75%) compared to the Active Brains group (43%), which aligned with the accelerometer data.

Cognitive function, pain intensity, and pain-specific coping

Both groups showed comparable improvements in cognitive function and pain (Supplementary Figure 4). Decreases in pain intensity were large (10/11 improved) and clinically significant (improvement > −2.0; Salaffi et al., 2004) for Active Brains–Fitbit (at rest: −2.1, p = .20, d = 0.9; with activity: −3.1, p = .10, d = 1.5) and Active Brains (at rest: −2.8, p = .04, d = 1.7; with activity: −3.4, p = .02, d = 2.1). Both groups reported large decreases in pain catastrophizing (7/11 improved; Active Brains = −9.7; Active Brains–Fitbit = −10.6) and small–medium increases in pain self-efficacy (8/11 improved; Active Brains = +4.2; Active Brains–Fitbit = +4.5). Both groups showed small–medium improvements in objective cognition (5/8 improved [3 unable to complete the MoCA due to COVID-19]; Active Brains = +1.1; Active Brains–Fitbit = +2.3) and decreased severity of subjective CD (6/11 improved; Active Brains = −0.3; Active Brains–Fitbit = −0.4).

Emotional functioning, social functioning, and general coping

Within-group changes were mixed for the measures of emotional and social functioning, as well as general coping skills (Supplementary Figure 5). Participants in Active Brains reported medium–large improvements (three out of five improved across measures) in depression (PROMIS = −4.0), anxiety (PROMIS = −7.2), emotional support (PROMIS = +6.2), and coping (MOCS-A = +12.2) while these changes in Active Brains–Fitbit were negligible. Additionally, gratitude increased (three out of five improved; GQ-6 = 1.4) in Active Brains but decreased in Active Brains–Fitbit (one out of six improved; GQ-6 = −3.8). Mindfulness and loneliness were negligible for both groups.

Exit Interviews

Three main themes emerged across semistructured exit interviews across the 10 program completers (Table 4). We report overall findings from each theme below and propose improvements to the program based on participant critiques in Supplementary Table 2.

Table 4.

Themes and Subthemes From Exit Interviews

Theme	Subtheme	Quotes
Active Brains skills	Cognitive functioning skills	“I found that I was more cognitively aware of things and able to do things and think more clearly after my walks.”
	Pain–cognition awareness skills	“And what really changed it for me too was learning about chronic pain that hurt doesn’t always mean harm, doesn’t always mean danger. That was just really, really powerful for me too.”
	Walking skills	“Although, the best thing on the face of the earth that I learned from your program is to walk … it’s valuable and it makes me rethink everything about what I thought was good and not good for me.”
		“Because I remember the times when I was walking at a brisk pace pace. And then I said to myself, slow down. I don’t have to walk this far, this fast.”
		“And with pacing it’s like you set the tone, you set the pace.”
	Mind–body skills	“You know, the difficulty was I’d never done it before. But I persisted and I try it. And then I remember the first time I felt a change, it was awesome, in the body, how my body relaxed.”
		“I found that I could ease up the pain, when mowing the lawn or when not paying attention to it, you know, you can scan it and practically that pain takes off.”
	Social and emotional functioning skills	“The thing is you heard of what you know, we know about self-care, self-compassion to others. But being compassionate to yourself, I thought it was so, so important.”
Barriers and facilitators	Accelerometers	“So, so that’s the only thing that I worry about the sensitivity of the actigraphy, as it related to the person’s physical activity.”
	Home practice	“I think the homework is a good idea because it keeps your mind on it. You don’t forget that day to day that you are doing this. The homework is a good thing.”
		“I just found it very hard to give myself the space to do it, but that’s not the program’s fault. That’s my own issue. I don’t know if there’s something the program could do about that.”
	In-person participation	“But let me tell you that a lot of times, I didn’t want to come in. Getting that ride and going home late and come home in the afternoon, sometimes six o’clock at night, worn out.”
	Audio recordings	“Yeah, and here’s what I loved about your audios. People brought more of a person, kind of their own little story in it. I loved the one about walking and it motivated me more than ever to walk.”
	Using the Fitbit	“Certainly knowing my steps and how those changed over the week was helpful in terms of acknowledging the progress that I was making.”
Group format and process	Clinician factors	“I thought you did a remarkable job of being sensitive to people’s concerns as they come up around the group and I commend you [Author] for that.”
		“I think drawing out the whole spectrum of responses was very important and only you in that group could’ve done that, so that was good.”
	Group dynamics	“I enjoyed it because I liked the diversity of the group.”
		“And that, that to some degree was helpful in the group to hear other people try to struggle through experience to move forward.”
	Manual structure	“I liked everything about the program, you know, very plain and simple. It was easy to follow. Just like, like I said earlier, I think the most important part of it is coming together and in sharing of information.”
	Participant factors	“Like I had towards the end of this, like the lightest flare up and I was feeling really down, and I mean physically. Mentally start listening to some of the negatives more than you would otherwise.”
		“I got pain too. But I don’t let it affect me. I don’t let it stop me going out there.”
	Recruitment	“And I think we need to have those hoops up. So you can check the person out and so on. You know, so you find out if they will fit the program or not, or find out if that’s what they want or they don’t want.”

Theme 1: perceptions of program skills

Participants spoke highly of the program skills and provided several examples of how practicing them improved physical, emotional, and cognitive functioning. Participants emphasized the importance of SMART goals and quota-based pacing for keeping them motivated and focused on increasing walking. For example, one participant said: “Although, the best thing on the face of the earth that I learned from your program is to walk … it’s valuable and it makes me rethink everything about what I thought was good and not good for me.” Learning to distinguish harmless pain while walking from injury further enhanced participants’ coping and decreased their kinesiophobia (“And what really changed it for me too was learning about chronic pain that hurt doesn’t always mean harm, doesn’t always mean danger.”). After integrating mind–body skills into their daily routine, participants found them helpful for stress and pain management (“I remember the first time I felt a change, it was awesome, in the body, how my body relaxed.”). Participants benefitted from skills that targeted emotional and social functioning, including empathy and social support (“… being compassionate to yourself, I thought it was so, so important.”). Participants recalled feeling challenged by the group to self-reflect on how CP and CD impacted their relationships. Participants agreed on the need for education on CD and several reported that the cognitive skills enhanced their organization and daily functioning. One participant connected the cognitive benefits to walking: “I found that I was more cognitively aware of things and able to do things and think more clearly after my walks.”

Theme 2: barriers and facilitators

Participants highlighted multiple barriers and facilitators that influenced their program experience. The Fitbit group found the device helpful for quantifying their progress, enhancing motivation, and maintaining accountability toward goals. According to one participant: “Certainly knowing my steps and how those changed over the week was helpful in terms of acknowledging the progress that I was making.” A few participants expressed skepticism about the Fitbit’s accuracy and felt discouraged, frustrated, or guilty when they did not meet their step goal. Some participants found it difficult to remember to wear and sync the Fitbit device when beginning the program. Similarly, some participants were concerned about the accelerometer’s accuracy, had difficulty remembering to wear the device, or found it inconvenient. Most participants spoke highly of the weekly homework and found that it helped them learn and apply program skills. Reminders from study staff (calls, texts, or emails based on preference) were viewed as supportive and helpful. Participants enjoyed audio recordings (“I loved the one about walking and it motivated me more than ever to walk.”). External factors, including bad weather, transportation costs, home commitments, and COVID-19 (end of Group 2) were identified as barriers to participation. Participants expressed interest in participating via live video to overcome barriers to the in-person group.

Theme 3: group format and process

Participants reported high satisfaction with the program format and structure, specifically highlighting the appropriateness of the program’s content, length, and pace. All noted that the social interaction and opportunity to learn from others was one of the most important aspects of the program. Specifically, learning that the other participants faced similar problems with CP and CD was reassuring to most participants (“And that, that to some degree was helpful in the group to hear other people try to struggle through experience to move forward.”). Participants enjoyed practicing the mind–body skills in-session and suggested allotting more time for this in future groups. Participants were encouraged by other members’ strategies for walking and skill practice despite initial pain. Other participants noticed a tendency to compare their perceived progress with other group members, which discouraged engagement. Participants agreed that facilitating group discussions, understanding population-specific needs, and encouraging shared goals were key qualities in the group leader.

Discussion

Consistent with our first hypothesis, both programs met all feasibility benchmarks (Greenberg et al., 2019, 2020; Vranceanu et al., 2018). We demonstrated the feasibility of recruiting and retaining older adults with CP–CD into both programs. Retention rates exceeded those reported in pilot studies in adults with CP (Greenberg et al., 2020; Veehof et al., 2016) and were comparable to those reported in studies with older adults (Morone et al., 2009). Credibility, expectancy, satisfaction, and session attendance were high for both groups, underscoring their acceptability for this population. We also observed that participants were highly adherent to homework, accelerometer, and Fitbit. Furthermore, accelerometer adherence improved in Group 2 after we revised our procedure (individualized wear plan and reinforcement system). These findings are important given concerns about challenges with technology among older adults in general (Lee, 2014), and those with cognitive challenges in particular (Hedman et al., 2016).

The overall within-group improvements in outcomes for both groups provided support for our second hypothesis. The strongest signal of improvement was on measures of pain intensity and pain-specific coping. The major emphasis on pain management skills in both programs may explain the clinically significant improvement in pain intensity found for both groups. Both groups were also associated with small to large improvements in step-count and physical function. The observed improvement in physical function, and step-count, in particular, support combining mind–body with activity skills and compliment research on these interventions for CD or CP (Ambrose & Golightly, 2015; Morone et al., 2009; Wetherell et al., 2011). We observed small–medium improvements for objective global cognition and severity of subjective CD. This suggests that the cognitive benefits associated with mind–body (Farhang et al., 2019) and physical activity (Falck et al., 2019) interventions may apply to older adults with CP and objective or subjective CD. Our study also shows that both performance-based and self-report measures provide sensitive and complementary assessments of changes in cognitive functioning.

Emotional functioning, social functioning, and general coping results were mixed, with improvements mostly displayed by Active Brains. This may be explained by the higher baseline severity of anxiety and depression in Active Brains–Fitbit at baseline (1 SD more than Active Brains), reflecting emotional functioning needs that exceeded the group. Nevertheless, signals of overall improvement indicate that both groups fit the target population and that the measures were sensitive to changes in outcomes (Leon et al., 2011). These findings are notable given the high needs of participants, who were sedentary (<5,000 steps), exhibited CD (MoCA <26), and reported moderate-to-severe pain (intensity between 6 and 8) on average at baseline.

The qualitative exit interviews confirmed high feasibility and suggested how to further improve uptake. Overall positive views of the program skills aligned with previous research on walking interventions for both management and increasing physical function (Terrier et al., 2019), as well as mind–body skills to enhance physical and emotional functioning, such as reduced stress and pain reactivity (Ball et al., 2017). In future trials, we will encourage participants to tie quota-based pacing to personal values and long-term goals to further address common barriers to starting activity, such as soreness, ambivalence, and kinesiophobia (Greenberg et al., 2019). We will also allocate more time in-session to explain and practice the mind–body skills to address misconceptions about mindfulness and its potential benefits for CD. Improved perceptions of CD and daily functioning may relate to the development of cognitive compensatory strategies and engagement in intellectually stimulating activities (Krell-Roesch et al., 2017). Participants recommend that we increase the emphasis on cognition in the programs—its greatest differentiator from available treatments—by introducing these skills earlier, adding resources on lifestyle factors of brain health, and enhancing awareness of the CP–CD comorbidity. Finally, participants reported that the social and emotional functioning skills encouraged self-compassion and self-reflection on the impact of CP and CD on relationships.

Exit interviews also highlighted the potential for technology to enhance adherence and program engagement. Feedback from the Fitbit group aligns with recent research on the device among older adults for augmenting behavior change (Patel et al., 2015) and high motivation to learn technology (Martínez-Alcalá et al., 2018). We will follow guidelines for technology adoption (e.g., Hickman et al., 2007) and discuss the purpose and limitations of activity trackers. In addition to the external barriers identified by participants (e.g., weather, transportation), the Active Brains–Fitbit group endorsed fears of COVID-19 at post-testing, which may have negatively impacted their emotional functioning outcomes. We will explore adaptations for live video delivery which have been feasibly used for similar mind–body programs (Vranceanu et al., 2018), to increase patient safety and accessibility.

Strengths of the study include: (a) applying ORBIT and NCCIH models to iteratively adapt programs and methodology, (b) emphasizing feasibility benchmarks prior to efficacy testing, (c) collecting multimodal physical function data following IMMPACT criteria (Gewandter et al., 2019), (d) exploring within-group analyses to estimate signal of improvement rather than between-group tests (Leon et al., 2011), and (e) collecting exit interviews to further enhance the programs (Supplementary Table 2).

There are also limitations. Given the small sample size and heterogeneity of the patient population, interpretations of our quantitative results are limited to guiding the iterative development of both programs and emphasize the need for larger trials. Two outliers disproportionally influenced group-level quantitative results (Supplementary Figures 3 and 4), but provided invaluable qualitative lessons to accommodate patients with the greatest treatment needs. Multimodal and longitudinal assessment in future trials can help rule out confounders, including COVID-19 distress and weather. Our sample was also mostly White, female, and urban. Older adults who are motivated to participate in a mind–body activity program may have self-selected for this study. This highlights the need to increase multicultural representation in future feasibility and acceptability testing.

To our knowledge, this the first study to integrate mind–body and activity skills to address the CP–CD comorbidity among older adults. Open pilot and exit interview data provided evidence of feasibility and showed promise for improving outcomes. Our mixed-methods approach yielded critical information to further enhance the program outlined in Supplementary Table 2. The next phase of intervention development will evaluate these program enhancements through an efficacy trial against an attention educational placebo control in larger and more diverse samples.

Funding

This work was supported by a supplement from the National Institute on Aging (3R34AT009356-02S1) to an R34 grant funded by the National Institute of Complementary and Integrative Health (1R34AT009356-01A1).

Conflict of Interest

None declared.