Participants & Procedures
The current study was a randomized pilot feasibility and acceptability trial of older adults (N=40) with a diagnosis of OA in the knee or hip comparing Engage-PA to treatment-as-usual plus fitness tracker control. In line with CONSORT guidelines,(37) as a pilot feasibility study, a formal power calculation for sample size was not required. Sample size was selected based on previous successful pilot work conducted by our team with similar feasibility pilot goals and planned secondary outcome analyses.(35,36) Participants were recruited from September 2020 through August 2021. Duke University Institutional Review Board provided ethical approval (Pro00105573) and the trial was registered on www.clinicaltrials.gov (NCT04490395).
Potential participants were identified by electronic medical record review based on inclusion and exclusion criteria. Potential participants were contacted by trained study staff via telephone and/or email summarizing study information and inviting participants to answer screening questions either online or over the phone. Inclusion criteria included (1) adults aged 65 or older, (2) diagnosis of OA in the knee and/or hip, (3) English speaking, (4) ability to participate in telephone sessions, (5) ability to ambulate even if assisted by a cane or walker, (6) rating worst pain and pain interference within the last week as a 3 or greater out of 10. Exclusion criteria included (1) planned surgery during the study duration that would limit mobility for more than 3 weeks, (2) major surgery requiring limited mobility within the past 3 months, (3) myocardial infarction within the past 3 months, (4) fall(s) within the past 3 months that led to immediate medical treatment, (5) current enrollment in cardiac rehabilitation, (6) presence of a serious psychiatric condition, (7) reported or suspected moderate cognitive impairment, (8) indication by a medical provider that exercise should only be medically supervised, and (9) presence of other unmanaged medical condition (e.g., hypertension, diabetes, asthma, neurodegenerative condition) that might lead to unsafe participation as outlined in the Physical Activity Readiness Questionnaire Plus (PARQ-2020;(38) an evidence-based measure for patient-determined safety for engaging in physical activity) subsequently verified by electronic medical record review and/or via communication with patients’ treating medical team.
Interested and eligible participants were invited to engage in consent procedures. Consent was conducted online (via Research Electronic Data Capture [REDCap], a HIPAA-compliant secure online data collection and management tool), or via telephone or mailed paper packets. After consent, participants completed baseline and post-treatment assessments online via REDCap or via telephone or mailed paper packets. All participants received a personal fitness tracker device which provided participants an objective measure of their daily steps. Study staff managing recruitment, assessments, intervention sessions, and fidelity checks were not blinded to condition. Assessment data was entered into REDcap directly by participants wherever possible. Following completion of baseline assessment, participants were randomized to either active treatment (Engage-PA) condition or treatment-as-usual with fitness tracker (TAU+) control group. Randomization was conducted using Randomization Windows Version 5.0 with equal allocation to groups. Study staff who had no contact with study participants conducted randomization; condition assignment was populated in REDcap automatically.
Participants in the intervention group received a study workbook to guide their two 45-minute telephone-delivered treatment sessions, spaced approximately 2 weeks apart. To set up later implementation efforts, the intervention sessions were delivered by a master’s level study therapist, guided by a written treatment manual, audio recorded, and checked for fidelity by a senior researcher. Participants in both groups reported on their daily steps as collected by the personal fitness tracker. All participants were compensated $30 for each completed assessment. Participation in this study lasted approximately 6 weeks. Figure 1 CONSORT diagram provides additional details about participant flow through the study.
Intervention
In Session 1, participants first discussed the fitness tracker and their baseline level of daily steps. The study therapist then explored the participant’s experience with OA pain and physical activity, and introduced personal values and activity-rest cycle skills. Personal values were defined as guiding principles that are likely to bring meaning and purpose. Participants were asked to identify two personal values: one related to physical activity and one related to another life domain (e.g., relationships). Personal values were linked to goal setting to allow for more purposeful action, which may increase meaning and satisfaction. Participants were then asked to reflect on their experience with over-doing activity (often resulting in increased pain) or under-doing activity (often resulting in deconditioning), and were subsequently introduced to the activity-rest cycle. The study therapist worked with each participant to identify timed activity bouts and timed strategic rest breaks so as not to exacerbate pain.Participants used the activity-rest cycle to generate specific, timely, and measurable goals related to physical activity (e.g., walk for 10 minutes) and another value-guided action (e.g., observe a grandchild’s sports event).
In Session 2, conducted two weeks after Session 1, participants reflected on their successes and barriers in doing their specific goals for physical activity and other value-guided activity. The activity rest cycle was troubleshooted in this session, with challenges, successes, and alternative strategies presented. Participants’ specific values were again discussed, and time was spent considering ways to refine values and actions to promote success (or increase physical activity). Continuing to engage in valued activity even if pain or psychological distress increased was highlighted, and the activity-rest cycle was presented as one way to do this.
Primary Outcome Measures
Feasibility and Treatment Acceptability. Feasibility was measured by participant enrollment (accrual), attrition, and session attendance. Acceptability was measured by The Client Satisfaction Questionnaire (CSQ)(39,40)post-treatment. The CSQ is commonly used in treatment acceptability studies and includes items such as, “How satisfied are you with the amount of help you received?” and “To what degree did the program teach you skills that are helping you to better manage your symptoms?” rated on a Likert scale.
Secondary Outcome Measures
Arthritis-Related Pain and Functioning. The Arthritis Impact Measurement Scale (AIMS)(41) was used to assess arthritis pain and functioning. Respondents answer questions on a 5-point Likert scale where higher scores indicate greater arthritis-related impact on functioning and health. The Symptom subscale includes items of pain frequency, pain severity, sleep interference from pain, and morning stiffness. The Physical Functioning subscale includes items of arthritis-related mobility, walking and bending, self-care activities of daily living, and household tasks. The AIMS is commonly used to assess functioning related to OA(41).
Physical Activity (objective and self-report). Physical activity was measured via both self-report and objective daily step measures.
Rapid assessment of physical activity (RAPA).(42) The RAPA is a 9-item self-report measure where respondents answer a series of Yes/No questions about their engagement in a range of aerobic (e.g., light, moderate, vigorous) and non-aerobic (e.g., yoga/flexibility, strength-training) activities, as well as frequency of their engagement in these activities. Higher scores indicate more activity. The RAPA was designed and validated for use with older adults.(42)
Daily steps.Participants used a wearable watch-style consumer-grade fitness tracker, a Garmin VivoFit 4.0 ©, upon waking and until bedtime daily throughout the study. Daily step counts were compiled for seven consecutive days at baseline and seven consecutive days post-treatment. Participants were mailed a daily step log, which they used to write down the daily step reading at bedtime from the Garmin device. Participants communicated with study staff via phone or email to report daily step counts.
Psychological Distress. The AIMS(41) Affect subscale was used to measure overall psychological distress related to anxiety, tension, and depressed mood. As in the other AIMS subscales, respondents answer questions on a 5-point Likert scale where higher scores indicate greater arthritis-related impact on functioning and health.
Psychological Flexibility. The Acceptance and Action Questionnaire-II (AAQ-II)(43) was used to assess overall psychological flexibility, or the ability to accept difficult psychological experiences as one lives in line with important life goals. In the AAQ-II, participants report on their emotional experience (e.g., “I worry about not being able to control my worries and feelings”) on a scale from 1 = “Never true” to 7 = “Always true”. The AAQ-II has commonly been used as a process of change measure in ACT-based studies and has been related to changes in symptom severity.(44,45)
Valued-Living. The Bulls-Eye Values Survey(46,47) was used to assess how consistently patients had been living in line with their chosen values. First, patients write about their personal values for living life with meaning and purpose across domains of relationships, education/work/community, leisure, and health/well-being, and then rate how successful they have been living in line with these values in the last month. A dartboard image is provided and responses ranged from 1 = “a perfect bulls-eye, and great success living in line with a value” to 8 = “very far away from living in line with a value”. This measure has demonstrated utility for pain populations(36) and other behavioral medicine populations.(46,47)
COVID-19 Impact Measure
The impact of the COVID-19 pandemic was assessed at baseline for all participants using the COVID-19 Impact Scale (CIS)(48) which was developed and distributed as part of the National Institutes of Health Office of Behavioral and Social Sciences Research toolkit. The CIS was used to assess overall COVID-19 impact on routines, health, mental health, personal or family member COVID-19 illness/death, and areas of life insecurity (e.g., finances). Additional items were developed for this study, which assessed changes since the pandemic in amount of exercise, amount of time spent in sedentary activity, and other ways in which the pandemic affected exercising (e.g., access to fitness centers, social support for exercising).
Changes to Assessments and Measurement
Two changes to study assessments and measures occurred after the pilot trial began. First, ActiGraph© accelerometers were initially used as an additional objective measure of physical activity (daily steps, time spent in moderate, vigorous or light/sendentary activity), but due to pandemic related challenges with remote procedures, two-thirds of the participants did not wear accelerometers. Additionally, a third assessment (1 month follow up) with both self-reported and objectively-measured (i.e., daily step count) secondary outcome measures was discontinued after the pilot trial began.
Analytic Strategy
Primary feasibility and acceptability outcomes were examined using descriptive statistics. For secondary outcomes, paired sample t-tests were calculated to examine between group changes from pre- to post-intervention in arthritis-related pain and functioning, psychological wellbeing, valued living, and daily steps. Secondary outcome analyses were conducted to examine possible indicators of improvement, and given the small sample size should be interpreted with caution.(49)