Understanding goal setting and behavior change mechanics in an older adult sitting reduction intervention

Abstract

Purpose:

We explored intervention fidelity, participant satisfaction, and the goals and reminder strategies participants chose to reduce sitting.

Approach:

Mixed methods approach leveraging data collected during study coaching and fidelity monitoring.

Setting:

A successful 6-month randomized controlled trial of a sedentary behavior (SB)intervention for adults ≥60 years in Washington, USA.

Participants:

N = 283 (140 intervention, 143 attention control); mean age 69, 66% women, 69% Non-Hispanic White

Intervention:

Theory-based SB reduction intervention structured around phone-based health coaching and goal setting. Attention control received equal coaching on non-SB health topics.

Method:

Coaches tracked all participant goals, and 8% of visits were randomly observed and fidelity coded using a structured template. Participants completed a satisfaction questionnaire at study end. Goals data were qualitatively grouped by reminder strategy and topic. Fidelity and satisfaction data were summarized and compared by study arm using two-sided paired t-tests.

Results:

Both participants’ satisfaction (>90% satisfied, between-group p=0.195) and coach fidelity to intervention content and techniques were high (96% sessions set SMART goals, p=0.343) across both arms. Intervention participants primarily set goals leveraging outward (e.g., fitness band prompts) and habit (e.g., adding standing to a daily meal) reminder strategies highly tailored to individual preferences and lifestyle.

Conclusion:

Participants’ SB-related goals varied widely, suggesting tailored intervention approaches are important to change sitting behavior, particularly for older adults with chronic conditions.

Introduction

Sedentary behavior (SB) consists of all activities expending little energy and performed in a sitting or lying posture and is highly prevalent among adults and older adults in the United States.^1,2 High levels of SB and low levels of physical activity are associated with more physical and mental health complications and chronic conditions such as obesity, diabetes, cardiovascular disease, and depression, as well as higher mortality.^3–6 Reducing SB could be a more feasible and acceptable behavioral target than increasing more intense physical activity,⁷ and the design and testing of SB interventions has garnered much interest in the field. To date, many trials are laboratory-based and the effectiveness of SB interventions to change behavior and downstream health outcomes is still inconclusive.^8,9 There is a critical need to translate laboratory and other preliminary findings into actionable behavior change techniques for real-world SB reduction.

Based on Social Cognitive Theory, strategies to foster self-regulation are key to behavior change. Chief among such self-regulation strategies is goal setting,^10,11 which concretely defines a plan or intention to engage in a new or changed behavior. Recent reviews of published SB reduction interventions across age groups have identified strategies like motivational interviewing, which typically includes goal setting and proactive problem-solving of barriers, self-monitoring through real-time feedback, social support, and the use of supportive technologies (e.g., wearable devices, smartphone apps) are effective in facilitating SB change.^12,13 Notably, interventions that combined multiple strategies yielded the largest behavior changes.¹² In contrast, an earlier review focused on physical activity interventions in older adults suggested that motivational interviewing and proactive barrier identification and problem solving were key for this population, but that techniques like goal setting and social support may be less effective.¹⁴ Importantly, this review was focused on physical activity behavior, rather than SB, and to our knowledge, these review findings have not been replicated.

A growing literature has identified common facilitators (e.g., self-monitoring and reminder tools, restructuring the environment, and the ease of incorporating SB reduction into current lifestyle) and barriers (e.g., environmental constraints, opposing social norms, existing health conditions, and enjoyment of sedentary activities) to sedentary behavior change.^15–17 However, there has not yet been any investigation of how barriers, facilitators, and individual needs and preferences translate to the types of SB reduction goals participants set when engaged in a real-world randomized controlled trial (RCT) of sitting reduction. Furthermore, real-world fidelity to planned intervention components, including motivational interviewing, goal setting, and proactive problem-solving, has not been described in a sitting reduction intervention. To support future successful SB interventions, we leverage data from a successful SB intervention in older adults to describe: 1) fidelity to planned intervention components, 2) the types of SB reduction and physical activity goals participants set, and 3) participant satisfaction.

Methods

Trial Overview

Details of the Healthy Aging Resources to Thrive (HART) trail protocol and main outcomes have been published previously.^18,19 Between January 1, 2019 and November 21, 2022, we conducted a parallel two-group RCT with balanced randomization (1:1) of a 6-month SB intervention (called I-STAND) targeting highly sedentary older adults. All study procedures were approved by the Kaiser Permanente Washington (KPWA) Human Subjects Division institutional review board, and the trial was registered with ClinicalTrials.gov (NCT03739762).

Participants & Recruitment

Participants were recruited from the membership panels of KPWA, an integrated healthcare delivery system in Washington State and Idaho, and were oversampled based on race and ethnicity to increase sample diversity. Individuals were assessed for eligibility based on electronic health record (EHR) and self-reported information after a brief phone screening. EHR eligibility criteria included: Washington State residence; age 60–89; body mass index (BMI) of 30–50 kg/m2; continuous KPWA enrollment over prior 12 months; no ICD-10 codes indicating residence in skilled nursing or an intermediate care facility or current receipt of palliative or hospice care; no cancer diagnosis in the last year; no diagnosis of deafness/significant hearing loss, dementia, or serious mental health disorder in the last 24 months. Self-report criteria included: ≥6 hours of sitting time per day; ability to stand from a seated position without assistance; ability to walk one block; fluency in English and ability to read study materials; ability to talk on the phone; no concurrent enrollment in other studies at KPWA; and willingness to complete necessary study procedures. In total, 283 participants screened eligible and agreed to participate; 140 were randomized to the I-STAND intervention arm and 143 to the attention control arm.

Intervention & Attention Control Condition Overview

I-STAND Sitting Reduction Intervention.

Two behavior change theories underpinned the intervention. Social cognitive theory²⁰ emphasizes self-regulation (e.g., goal setting, feedback) and the use of environmental cues to promote self-efficacy for behavior change. Habit formation^21,22 posits that behaviors like sitting become automatic and changing these behaviors necessitates bringing conscious awareness through cues and disruption strategies. The I-STAND program was structured upon ten ~30 minute one-on-one sessions with a health coach over 6 months, each focused on setting goals leveraging 3 different reminder strategies to sit less: 1) inner, which rely on mindfulness and the perception of internal cues that it’s time to stand or move; 2) outward, which leverage external cues, prompts, and environment changes to change behavior; and, 3) habit, which attach standing or moving to daily habits the participant is already doing. The study provided tools to facilitate outward reminder strategies: 1) cue cards to place on chairs where they often sit as a reminder to stand or move, 2) a table-top standing desk, and 3) a wrist-worn fitness tracker that provided frequent reminders to break up sitting. A workbook with didactic content was provided to participants and reviewed during coaching sessions, and paper tracking logs were provided to track goal progress. Participants were provided feedback charts documenting their data from several 1-week wear sessions of the activPAL accelerometer, which tracks sitting, standing, and stepping behaviors.²³ This feedback allowed participants to see progress in their behavior change with objective data and informed subsequent goal setting.

Attention Control Condition.

Attention control participants also received ten ~30-minute one-on-one sessions with a health coach over 6 months. Participants selected topics to review and set relevant goals guided by a workbook covering 15 general health topics for older adults (e.g., bone health, fall prevention, cognitive activity, social activity, depression, stress management, sleep, and healthy diet). Paper tracking logs were provided to participants to track progress. In instances where a participant organically expressed interest in a goal related to physical activity, exercise, or sitting less, coaches were instructed to redirect from the topic if possible. However, participants were allowed to set activity-related goals on their own, outside of their study goals.

Health Coach Training.

All coaches were trained by the study principal investigator, a licensed clinical psychologist (D.R.), to use motivational interviewing strategies (e.g., reflective listening, open-ended questions, affirmations, and summaries) and problem-solving techniques to support behavior change. Coaches were also trained in the use of a custom study database for documentation and tracking. During each coaching session (I-STAND or attention control), coaches recorded notes from the session, the participant’s progress on prior goals, and documented participants’ new or continued goals. Ongoing training and supervision were provided in weekly meetings with the principal investigator.

Fidelity Monitoring.

All coaching sessions were audio recorded with participant consent. Throughout the trial, approximately 10% of the I-STAND and 5% of the control group coaching sessions were retrospectively reviewed on a rolling basis from the audio recording and coded for fidelity by a trained coder with a graduate degree in social work. A structured coding protocol and rubric were used to review and document fidelity for each session reviewed. A copy of the fidelity rubric is available in Appendix 1. To enhance subsequent fidelity, feedback from fidelity reviews was given to coaches in weekly supervision meetings.

Data Collection and Structure

Participant characteristics were collected at baseline through participant self-report (gender, race and ethnicity, retirement status, marital status, education, arthritis), EHR data (age, Elixhauser comorbidity index, diabetes, hypertension), and direct study measurement (BMI, activPAL daily sitting time).

Fidelity Data.

Items assessing coaching techniques (motivational interviewing components, goal setting, etc.) were scored as either 0=not used or 1=used. Interventional topic items were scored using a 0–2 point scale, with 0 indicating the topic was not discussed, 1 indicating a brief mention (no more than 1 exchange between coach and participant), and 2 indicating a full discussion of the topic (two or more exchanges between coach and participant). For ease of interpretation, brief mention and full discussion categories were grouped for comparison across groups. To assess the frequency and pattern of discussions of physical activity in both the I-STAND and control arms, additional items were captured to assess who initiated the discussion of a physical activity topic (coach vs. participant) and the coach’s response to the topic (sustaining the topic vs. redirecting from the topic). Additionally, an overall session rating was assigned based on the fidelity coder’s assessment of how well the session adhered to protocols and addressed expected topics across all domains reviewed (rating scale: 0=Did not follow content; 1=addressed some content but skipped key concepts; 2=mainly followed content with minor changes; 3=followed content as laid out).

Goals Data.

All goals set in each coaching session were individually tracked in the study’s Microsoft Access tracking database using discrete fields (i.e., one field per goal) for all participants in both arms. Additionally, goals set at a prior coaching session were reviewed at the beginning of the subsequent session and could be carried forward for continued work.

Participant Satisfaction Data.

As part of the 6-month study measurement survey, participants were invited to provide study feedback through closed-ended survey items, with additional opportunity to provide open-text comments. Four closed-ended items assessed how helpful participants found study workbooks, coaching sessions, goal setting, and goal-tracking logs and were scored on a 0 to 3-point scale from 0=“Did not use” to 3=“Very helpful.” An additional four items asked participants to rate their overall satisfaction with other components of the study (i.e., wearing the thigh-worn activity monitor, the frequency of study contacts, and the topics of sessions) and the study overall on a scale of 0=“Not satisfied” to 3=“Very satisfied”. A copy of the presented satisfaction survey items are available in Appendix 4. For analysis and presentation, values were binarized by combining the two highest response categories and the two lowest response categories. Two additional questions assessed whether participants noticed any health improvements or worsening health status during the study (yes vs. no).

Data Analysis

Fidelity Analysis.

We generated descriptive summaries of all fidelity items overall and by intervention arm. For some fidelity measures, only a subset of sessions (e.g., only baseline or last sessions) were expected to cover an assessed topic, and, in these cases, frequencies were calculated based on sample sizes restricted to only those reviewed sessions meeting that criterion. Differences in the distributions between I-STAND and attention control groups were assessed using Chi-square tests for categorical measures or two-sample t-tests for continuous measures. P-values p<0.05 were considered statistically significantly different between groups. All analyses were conducted using Stata v17.0 (SataCorp, College Station, TX).

Goals Analysis.

Upon study completion, all goals data (free text) and goal progress data were exported by study arm (I-STAND vs. control) for analysis. Goals text was reviewed by a team member trained in clinical chart review (L.Y.). Data were cleaned to clarify ambiguous goals based on prior data (e.g., carrying forward prior listed goals in place of goal statements like “same goals as prior week”), and thematic groups were inductively defined based on preliminary review. Goals were then iteratively reviewed by a larger analytic team with two additional researchers trained in qualitative methods (M.A.G.H and D.E.R.) and final groupings were assigned based on common themes. Similar goals with differing phrasing (e.g., “stand to do yard work” and “stand to weed garden”) within and between participants were combined to derive a list of unique goals across the I-STAND and control groups separately. Unique goals related to key topics of interest (e.g., TV, standing desk, etc.) among reminder strategy types (i.e., outward, habit, and inner) were identified using key word searches, including common phrases and synonyms, of the goals text data. To tabulate the most frequently chosen outward reminder-focused goal categories, goals text was reviewed for all individuals in the I-STAND group, and the number of unique individuals choosing at least one goal in that category were tallied. Goals focused exclusively on increasing dedicated bouts of exercise or intense physical activity were excluded to maintain a focus on SB-specific goals.

Participant Satisfaction Analysis.

Summary and analysis of patient satisfaction items followed the same procedures described for Fidelity measures, comparing differences in the distributions between I-STAND and control groups using Chi-square (categorical) or two-sample t-tests (continuous).

Results

Participant Characteristics and Trial Outcomes

From a starting sample of N=283 (140=I-STAND; 143=control), n=273 participants were still enrolled in the study at the 6-month primary outcomes timepoint. Baseline participant characteristics are summarized in Table 1. The sample was 66% female, 63% Non-Hispanic White, and had a mean age of 68.8 years (range: 60–88) and BMI of 34.9 kg/m². Approximately 28% had diabetes and 52% had hypertension. The sample was highly sedentary with a mean baseline sitting time of 10.9 hours/day. Previously reported trial outcomes documented statistically significant decreases in both sitting time (−32 min/day adjusted mean difference) and systolic blood pressure (−3.5 mmHg adjusted mean difference) for I-STAND vs. control.¹⁸

Table 1.

Baseline participant characteristics, overall and by study arm

	Overall (N=283)	Attention Control (N=143)	I-STAND (N=140)
	Mean (SD) N (%)	Mean (SD) N (%)	Mean (SD) N (%)
Age	68.8 (6.3)	69.0 (6.5)	68.6 (6.0)
Gender
Female	186 (65.7%)	99 (69.2%)	87 (62.1%)
Male	96 (33.9%)	44 (30.8%)	52 (37.1%)
Non-Binary	1 (0.4%)	0 (0%)	1 (0.7%)
Race and Ethnicity
Asian, non-Hispanic	9 (3.2%)	5 (3.5%)	4 (2.9%)
Black, non-Hispanic	41 (14.5%)	16 (11.2%)	25 (17.9%)
Hawaiian or Pacific Islander, non-Hispanic	6 (2.1%)	2 (1.4%)	4 (2.9%)
Hispanic	16 (5.7%)	10 (7.0%)	6 (4.3%)
Multiracial, non-Hispanic	12 (4.2%)	6 (4.2%)	6 (4.3%)
Native American/American Indian, non-Hispanic	3 (1.1%)	2 (1.4%)	1 (0.7%)
Other, non-Hispanic	1 (0.4%)	0 (0%)	1 (0.7%)
White, non-Hispanic	195 (68.9%)	102 (71.3%)	93 (66.4%)
Retired *	157 (55.5%)	79 (55.2%)	78 (55.7%)
Marital Status *
Married/living as married	175 (61.8%)	90 (62.9%)	85 (60.7%)
Single never married	23 (8.1%)	14 (9.8%)	9 (6.4%)
Widowed/ separated/ divorced	76 (26.9%)	34 (23.8%)	42 (30.0%)
Education *
College or higher	149 (52.7%)	79 (55.2%)	70 (50.0%)
Some college/ vocational school	113 (39.9%)	51 (35.7%)	62 (44.3%)
High School or GED	10 (3.5%)	5 (3.5%)	5 (3.6%)
Some high school or less	2 (0.7%)	1 (0.7%)	1 (0.7%)
Elixhauser Comorbidity Index *	1.73 (1.69)	1.78 (1.92)	1.67 (1.43)
Self-Reported Arthritis	122 (43.1%)	52 (36.4%)	70 (50.0%)
Diabetes	80 (28.3%)	41 (28.7%)	39 (27.9%)
Baseline body mass index (BMI)	34.9 (4.7)	35.2 (4.8)	34.7 (4.6)
Hypertension *	147 (51.9%)	73 (51.0%)	74 (52.9%)
Daily average sitting time, mins/day (activPAL)	652 (117)	654 (116)	650 (118)

^*Summaries reflect missingness due to participant non-response: 8 Retirement; 9 Marital Status; 9 Education; 9 Self-rated health; 7 Fall in last 12 months; 5 Able to walk at a normal pace; 19 Depression Symptoms; 11 Current Smoker; 3 Elixhauser Comorbidity Index; 1 Hypertension; 6 activPAL daily average sitting

Coaching Session Fidelity

A total of 221 (136 I-STAND and 85 control) coaching sessions were coded for fidelity, representing 8.4% of conducted sessions (9.5% of I-STAND sessions, 6.8% of control sessions). Detailed summaries of fidelity results are available in Table 2.

Table 2.

Summary of fidelity measures for sessions reviewed overall and by study arm

	Overall %a	I-STAND %a	Attention Control % a	p-value+
Sessions Reviewed, n	221	136	85	-
Unique Participants, n	96	58	38	-
Overall Session Rating b, mean (SD)	2.9 (0.2)	2.9 (0.3)	3.0 (0.2)	0.242
Motivation Interviewing Techniques
Open-Ended Questions	100	100	100	-
Affirmations	99.6	99.3	100	0.428
Reflective Listening	91.4	90.4	92.9	0.519
Summaries	89.1	90.4	88.2	0.614
Positive Framing	99.6	100	98.8	0.205
Discussed Progress c	100	100	100	-
Problem Solved Encountered Barriersc	92.7	93.3	91.9	0.85
Tied Content to Values	83.3	79.4	89.6	0.052
Set SMART Goalsd	96.6	97.6	95.2	0.343
Proactive Problem Solving for new goalsd	81.6	89.2	70.4	0.003
Sedentary Behavior (SB) Content
Discussed any SB Content	62.9	100	3.5	<0.001
Reviewed Feedback Chart e	32.6	52.6	1.2	<0.001
Discussed Sitting Habits/ Routines	60.6	98.5	0	<0.001
Sitting Break Reminder Strategies	60.6	94.9	5.9	<0.001
Wrist Prompting Band	50.2	81.6	0	<0.001
Standing Desk	43.0	69.9	0	<0.001
Home Environment	28.5	45.6	1.2	<0.001
Light PA/Chores for breaks	43.9	70.6	1.2	<0.001
Break from sitting on call	25.8	42.2	0	<0.001
Physical Activity (PA) & Exercise
Discussed PA	75.1	89.7	51.8	<0.001
Who Initiated PA Discussion? f
Participant	68.7	59.8	93.2	<0.001
Coach	31.3	40.2	6.8
Was PA discussion sustained? f
Coach Sustained	86.1	100	47.8	<0.001
Coach Redirected	13.1	0	52.3
PA Goal Set	74.2	97.8	37.6	<0.001
Other Content
Discussed other health topics	40.3	6.6	94.1	<0.001
Discussed relevant safety Information	100	100	100	-
Discussed study logistics	100	100	100	-

^aPercentage of sessions reviewed meeting the indicated criteria.

^bOverall session rating scale: 0 = Did not follow content; 1=addressed some content but skipped key concepts; 2= mainly followed content with minor changes; 3=followed content as laid out

^cCriteria only applicable to non-first sessions, n=194 overall (120 I-STAND, 74, control)

^dCriteria only applicable to non-lasts sessions, n=208 overall (125 I- STAND, 83, control)

^eFeedback charts from activPAL wear were only available at sessions following a measurement activPAL wear, typically sessions 1, 2 (pre-pandemic), session 6, and session 10.

^fOnly applicable to sessions that discussed PA, n=166 overall (122 I- STAND, 44 control)

⁺p-values derived from a Chi Square test (categorical) or t-test (continuous) of differences in the distributions between I- STAND and attention control sessions. P<0.05 considered statistically significantly different and indicated in bold.

Among coded sessions, use of motivational interviewing principles and techniques was high (≥80% for all coded techniques) and did not differ significantly (p>0.05) between arms. Participants set specific, measurable, achievable, relevant, and time-bound (SMART) goals²⁴ in 96.6% of coded sessions, and coaches used both retroactive (92.7%) and proactive (81.6%) problem-solving of barriers across groups to help participants address possible barriers to meeting their chosen goals.

All coded I-STAND sessions discussed SB content and behavior change strategies; while SB topics were discussed in 3.5% of coded control sessions. Among the I-STAND group, the participant’s daily sitting habits and routines (98.5%), reminder strategies to take breaks from sitting (94.9%), and the study-provided wrist prompting band (81.6%) were the most common SB topics discussed. Use of the study-provided (or participant owned) standing desks (69.9%) and the use of light physical activity or household chores to break up sitting time (70.6%) were also common topics.

The topic of physical activity and exercise was prevalent in both study arms, arising in 89.7% of coded I-STAND sessions and 51.8% of control sessions (p<0.001). When discussed in either condition, it was most-often initiated by the participant (59.8% I-STAND, 93.2% control). When raised, health coaches sustained the topic in all intervention sessions, but redirected participants in approximately 52.3% of control sessions. Physical activity goals were set in 97.8% of coded I-STAND session compared to 37.6% of control sessions. Among control sessions reviewed, healthy eating, keeping the brain active, and coping with stress and anxiety (particularly from the COVID-19 pandemic) were the most discussed topics.

Goals

As expected, goals in the I-STAND group were focused exclusively on strategies to sit less, take more standing or moving breaks, or introduce more exercise and physical activity. Participants chose goals that leveraged all three reminder strategy types (i.e., outward, inner, and habit). Detailed goal examples in all domains are available in Appendix 2.

I-STAND Goals – Outward Reminders.

Outward reminders were the most common goal type set by participants. Table 3 summarizes common outward reminder goal categories observed among I-STAND sessions and the number of participants that chose a goal in that category during at least one session. Outward reminders were often tailored for different contexts (e.g., home vs. work), depending on individual participant needs and preferences. Outward reminder goals to use a prompting device (either the study-provided or a personal device), use a standing desk for work or personal activities, or break up or reduce TV and other media use time were the most frequently chosen goals among I-STAND participants with 110, 104, and 99 unique individuals, respectively, choosing goals from these categories. Of note, goals targeting TV and other media viewing time were popular among participants, but most chose goals focused on introducing more standing, moving, or exercise to their TV time, rather than reducing total TV time or replacing TV with other activities.

Table 3.

Frequency of outward reminder goals selected from distinct I-STAND goal categories

Context	Goal Category	Number of individuals (N=140)
Context independent (used for both work and home)	Use study-provided prompting band (or personal fitness tracker) for standing reminder	110
	Use standing desk (for various work or personal activities)	104
	Use other electronic devices/phone to remind to stand	73
	Use cue card for standing reminder	20
	Use kitchen timer as reminder	14
Home	Standing/walking related to dog/cat/pet care and exercise	23
	Media-related goals (reducing or breaking up TV, streaming, YouTube, movies, sports, or computer game time)	99
Work	Work-related reminder strategies (calendar reminders, zoom breaks, etc.), other than using a standing desk	44

Note. Counts presented in table are not mutually exclusive. Individuals could be counted in multiple goal categories if they made goals in both domains. Additionally, some goals may be represented in multiple categories if they crossed domains (e.g., using a cue card or timer to break up TV time).

I-STAND Goals – Habit Reminders.

Habit reminders were also commonly used by participants in both home and work settings. Most commonly, participants set habit reminders that attached small bouts of standing or moving to daily self-care or household tasks like brushing their teeth, paying bills, or engaging in regular hobbies. Attaching more standing and moving to mealtimes and regular TV habits, like watching the morning news or changing the channel, were also common strategies. Some examples of habit reminder goals in the home setting were: “stand while pet eats”, “stand after first morning coffee”, “stand until tea water boils”, and “stand each time the TV remote is touched.” Habit reminders were also used in workplace contexts by attaching small standing or moving bursts to regular breaktimes or workplace tasks like greeting clients or talking with co-workers.

I-STAND Goals – Inner Reminders.

Inner reminders were the least common strategy used by participants but were employed by some. Unlike outward and habit reminders, inner reminders were not specific to home vs. work contexts but were used throughout various settings and situations. Examples of inner reminder goals included: “Note pain in back, bum; stand when notice this”, “Change thinking about efficiency; instead make several trips or go the long route.”

Goal-Related Barriers.

While barriers to achieving SB-focused goals were not systematically captured, some common barriers were noted in goal text. A summary of these spontaneously reported barriers is available in Appendix 3. Weather, such as cold, heat, and rain, were common barriers and goals were often constructed with a weather contingency such as “Walk when weather allows.” Barriers related to work, health, technology, the home or neighborhood built environment, and other life circumstances were also noted spontaneously in analyzed goal text.

Attention Control Goals.

Goals in the control group covered a broader range of topics. However, as noted in the fidelity coding, despite coach redirection, some control participants chose to set goals related to physical activity or reducing sitting time. The most common activity-related goal was to begin or increase walking (n=23), though several participants also set goals focused on completing physical therapy exercises recommended by their health care team (n=13) and on doing yoga or Tai Chi (n=15). Stretching and balance exercises, exploring new exercise programs or gyms, and engaging in physically active household chores and projects were also documented.

Participant Satisfaction

Participant satisfaction items at 6-months are displayed in Table 4. Overall participant satisfaction was high with more than 90% of respondents feeling “satisfied” (36%) or “very satisfied” (57%) with the study overall. Notably, approximately 20% of participants did not complete satisfaction items due to drop-out or personal choice. More than 90% of responding participants reported coaching sessions and goal setting to be helpful, with no significant differences between study arms. Participants indicated that written material like the study workbook and goal tracking log were less helpful than coaching and goal setting, though most participants still found these parts of the program helpful (90% and 75% for the workbook and goal log, respectively). Satisfaction with goal tracking logs differed by study arm (81% I-STAND vs. 69% control, p=0.029). Around 90% of participants also reported satisfaction with the frequency of study contacts and the topics of sessions covered, though fewer control participants were satisfied study contact frequency vs. I-STAND participants (p<0.001). Overall, 69% of participants reported noticing improvements in their health over the course of the study, though this differed by study arm (84% I-STAND vs. 54% control, p<0.001). Few participants (12%) noted worsening health over the course of the study, with no difference by study arm.

Table 4.

Participant satisfaction, overall and by study arm

	Overall	I- STAND	HL
	N=2731	N=1371	N=1361
	N (%)	N (%)	N (%)	p-value2
Study Components
Workbook				0.261
Very or Somewhat Helpful	200 (89.7)	103 (92.0)	97 (87.4)
Not Helpful or Did Not Use	23 (10.3)	9 (8.0)	14 (12.6)
Coaching Sessions				0.094
Very or Somewhat Helpful	209 (93.7)	108 (96.4)	101 (91.0)
Not Helpful or Did Not Use	14 (6.3)	4 (3.6)	10 (9.0)
Setting Goals				0.12
Very or Somewhat Helpful	208 (92.9)	107 (95.5)	101 (90.2)
Not Helpful or Did Not Use	16 (7.1)	5 (4.5)	11 (9.8)
Goal-Tracking Logs				0.029
Very or Somewhat Helpful	165 (75.0)	91 (81.3)	74 (68.5)
Not Helpful or Did Not Use	55 (25.0)	21 (18.7)	34 (31.5)
Frequency of Study Contact				<0.001
Very Satisfied or Satisfied	199 (90.5)	110 (99.1)	89 (81.6)
A little or Not Satisfied	21 (9.5)	1 (0.9)	20 (18.4)
Topics of Sessions				0.053
Very Satisfied or Satisfied	194 (88.6)	102 (92.7)	92 (84.4)
A little or Not Satisfied	25 (11.4)	8 (7.3)	17 (15.6)
Wearing activPAL				0.772
Very Satisfied or Satisfied	188 (90.4)	97 (89.8)	91 (91.0)
A little or Not Satisfied	20 (9.6)	11 (10.2)	9 (9.0)
Study Overall
Overall Satisfaction				0.194
Very Satisfied	121 (56.5)	63 (58.3)	58 (54.7)
Satisfied	76 (35.5)	40 (37.0)	36 (34.0)
A Little Satisfied	17 (7.9)	5 (4.6)	12 (11.3)
Not Satisfied	0 (0)	0 (0)	0 (0)
Noticed Improvements in health/ADLs				<0.001
Yes	152 (68.5)	92 (83.6)	60 (53.6)
No	70 (31.5)	18 (16.4)	52 (46.4)
Noticed Worsening in health/ADLs				0.686
Yes	28 (12.5)	15 (13.4)	13 (11.6)
No	196 (87.5)	97 (86.6)	99 (88.4)

¹Total participants enrolled at 6 months was N=273 from the baseline N=283 due to participant drop-out.

²p-value derived from a chi-square test for differences in the distribution of each categorical satisfaction variable between the I- STAND and Health Living groups.

Note: Reported Ns for some items do not sum to N=273 due to participant non-response to the 6-month survey in its entirety or on particular items. Number of missing or “don’t know” values for each item as follows: workbook n=50 (n=25 I-STAND, n=25 HL); coaching sessions n=50 (n=25, n=25); setting goals n=49 (n=25, n=24); goal tracking logs n=53 (n=25, n=28); frequency of study contact n=53 (n=36, n=27); topics of sessions n=54 (n=27, n=27); wearing activPAL n=65 (n=29, n=36); overall satisfaction n=57 (n=28, n=29); health improvements (n=51 (n=24, n=27); worsening health n=49 (n=24, n=25).

Discussion

These data demonstrate that a highly tailored, goal-driven SB reduction intervention for older adults can be administered with high fidelity and high participant satisfaction among a population of older adults. The intervention fidelity data underscore that coaching sessions leveraged planned motivational interviewing techniques and were highly focused on participant individualization, with heavy emphasis placed on reviewing progress, troubleshooting individual barriers, and setting new goals to maintain and progress behavior change. Importantly, while most participant goals leveraged similar reminder strategies and shared common high-level behavioral targets (e.g., breaking up TV time, using a standing desk for computer tasks), goals were highly individualized to fit the unique needs and context of each participant. There are infinite ways to break up sitting behaviors so tailoring to individual circumstances, preferences, and abilities appears to be a key reason older adults are able to reduce their sitting time. Outward reminders, which leveraged modifications to or cues from the participant’s environment, and habit reminders, which tied new sitting reduction behaviors to existing habits and routines, appeared to be the most approachable for participants and were used most frequently in their goal setting. Yet, future studies could do more work to leverage inner reminder strategies, since internal cues are available anytime when they are attended to. Furthermore, participants in both groups, but particularly in the I-STAND group, expressed very high satisfaction with the individual health coaching and emphasis on goal setting.

These data also highlight some important challenges in the use of attention control conditions in behavioral interventions targeting SB. Attention control groups are encouraged in many behavior change studies, to ensure that any observed intervention effects are truly due to the active ingredients of the intervention rather than attention from the study team.^25,26 Here control participants frequently expressed interest in discussing and setting goals focused on increasing daily movement, exercise and physical activity, often with a goal of weight loss. Despite coach training and efforts to redirect focus to other healthy living topics, many attention control participants still set goals related to exercise in some way. These participant desires may be exacerbated in a population with high BMI and chronic conditions, who are frequently encouraged by health care professionals and broader health messaging that exercise and weight loss are key for improved health.^27–29 Overall trial results, published elsewhere, suggested small changes in daily sitting time and daily steps among the attention control group (−8.5 min/day and +352 steps), though not statistically significant, this could have biased overall trial results towards the null.¹⁸ Prior literature highlights potential drawbacks to the use of attention control conditions in behavioral intervention RCTs²⁶; we believe the data here underscores some of those drawbacks. However, attention control groups also provide important benefits borne out here, including high retention and satisfaction and assurance that observed effects are derived from the intervention content rather than attention. Careful consideration of more naturalistic or inactive control conditions, either as sole comparator or in a three-armed trial design, could be considered in future trials intervening on SB.

This work has important limitations. The HART trial was specifically targeting older adults with a high BMI, and the HART sample was not representative of the general US older adult population; these learnings may not be applicable to SB change in other populations. The HART trial was also conducted both prior to and during the COVID-19 pandemic. Interruptions to participants’ daily life and activities may have impacted the goals participants chose, and, importantly, may have supported an increased awareness and focus on finding ways to move more in both intervention and control groups. The qualitative analysis of goals data primarily relied on a single analyst’s review and grouping of all goals data, which could introduce concerns about reliability. However, to minimize this risk, a small analytic team collaboratively refined thematic goal groupings after preliminary coding.

Conclusion

These data from the HART trial, which reduced both daily sitting time and systolic blood pressure in the I-STAND group at 6 months (difference in mean change = −32 min/day and −3.48 mmHg, respectively),¹⁸ afford a unique look behind the scenes of a successful sedentary behavior reduction intervention for older adults in a real-world context.^30,31 These findings underscore that supporting high levels of individualization to participant needs and preferences in SB intervention design may be critically important to success. Future interventions targeting SB should include an emphasis on individualization rather than one-size-fits-all strategies, and researchers should carefully consider the benefits and drawbacks of attention control comparison groups when selecting a trial design.