Physical Activity Intervention for Loneliness

Attendance will be calculated as the total number of attended sessions divided by the total number of sessions of the intervention and recorded as a percentage.

Percent of subjects responding to ads and friends' referrals out of total number of invitations given out as a percentage

Number of participants completing the study as a proportion of those randomised

Is intervention effective for reducing loneliness, is intervention practical in the proposed design settings as assessed during the focus group interviews and analysed qualitatively

Focus group interviews using phenomenological inductive approach will be aimed to help the research team to improve the quality of the delivered intervention by making changes in the methodology and design of the intervention for the future implementation in a consequent study

The assessment rate of questionnaires will be evaluated as the total number of completed questionnaires divided by the total number of questionnaires and recorded as a percentage

The appropriateness of statistical methods of data analysis will be analysed by research team

A power calculation and sample size estimation will be calculated for meaningful outcomes (e.g. loneliness or social support) using the method based on the differences in means between the intervention and control group using the G-power software

Means (M) and standard deviations (SD) will be used to investigate the effect size for change in loneliness using mixed between (intervention group) and within (time) repeated-measures analysis of variance (ANOVAs) with post-hoc comparisons

Means (M) and standard deviations (SD) will be used to investigate the effect size for change in social support using mixed between (intervention group) and within (time) repeated-measures analysis of variance (ANOVAs) with post-hoc comparisons

Means (M) and standard deviations (SD) will be used to investigate the effect size for change in social networks using mixed between (intervention group) and within (time) repeated-measures analysis of variance (ANOVAs) with post-hoc comparisons

Means (M) and standard deviations (SD) will be used to investigate the effect size for change in depression using mixed between (intervention group) and within (time) repeated-measures analysis of variance (ANOVAs) with post-hoc comparisons

Means (M) and standard deviations (SD) will be used to investigate the effect size for change in anxiety using mixed between (intervention group) and within (time) repeated-measures analysis of variance (ANOVAs) with post-hoc comparisons

Means (M) and standard deviations (SD) will be used to investigate the effect size for change in self-efficacy for exercise using mixed between (intervention group) and within (time) repeated-measures analysis of variance (ANOVAs) with post-hoc comparisons

Means (M) and standard deviations (SD) will be used to investigate the effect size for change in satisfaction with level of social contacts using mixed between (intervention group) and within (time) repeated-measures analysis of variance (ANOVAs) with post-hoc comparisons

Means (M) and standard deviations (SD) will be used to investigate the effect size for change in expected outcomes and barriers for exercise using mixed between (intervention group) and within (time) repeated-measures analysis of variance (ANOVAs) with post-hoc comparisons

Loneliness will be assessed using the 8-item UCLA (University of California, Los Angeles) Loneliness Scale (UCLA-8) (Hays and DiMatteo, 1987). Items will be rated on the 4-grade Likert scale ranging from 1 (never) to 4 (often). The total scoring is ranged from a minimum of 8 to a maximum of 32 with the higher score indicating more loneliness considered to be a worse outcome. Questions "I am an outgoing person" and "I can find companionship when I want it" will be reverse scored.

Social support will be assessed using the 20-item Medical Outcomes Study Social Support Survey (MOSSSS) (Sherbourne and Stewart, 1991). The questionnaire will consist of 20 items including the assessment of a structural social support (number of close friends - item 1) and assessment of functional social support (items 2-20). Participants will be asked to complete the questionnaire using the five-point Likert scale ranging from low social support "None of the time" (1 point) to a high social support "All of the Time" (5 points). The total score will be calculated as the sum of the scores. Higher scores of support will show better social support.

Social networks will be assessed using the 6-item Lubben's Social Network Scale (LSNS-6) (Lubben et al., 2006).LSNS-6 has total 6 items divided into two subscales: family and friendship (three items for each). The total score will be calculated by sum of the six items with scores ranging from 0 to 30. A sum of scores for six items less than 12 will indicate that the respondent had fewer than two people for social interactions. Similarly, a cut-point score of 6 on the three-item will indicate as being at risk for social isolation (Lubben et al., 2006). The family subscale will indicate that the participants had less than two family members to perform social interactions, similarly for the friendship subscale.

Depression and anxiety will be assessed using the 14-item Hospital Anxiety and Depression Scale (HADS) (Zigmond and Snaith, 1983).The questionnaire consists of 14 items with the assessment on four-point scale (0-3). The maximum total score for each subscale will be 21 points. The cut-offs for the total score for each sub-scale will be: 0-7 - normal, 8-11 - mild, 12-14 - moderate, 15 and over - severe.

Self-efficacy for exercise will be measured using the revised 9-item Self-Efficacy for Walking/Exercise Scale (SEE) in a paper-and-pencil format (Resnick and Jenkins, 2000). Using numbers from 0 (not confident) to 10 (very confident) participants will be asked to rate their confidence in their ability to walk/exercise for 20 minutes three times per week (Resnick and Jenkins, 2000). The scoring will be calculated by summing the numerical ratings for each response and dividing by the total number of items.

Satisfaction with level of social contacts will be measured with the question "How satisfied are you with your social contacts?" (Fokkema and Knipscheer, 2007). Participants will rate their answer on a scale from 0 (entirely dissatisfied) to 10 (entirely satisfied).

Expected outcomes and barriers for exercise will be administered using the Expected Outcomes and Barriers for Habitual Exercise scale (Steinhardt and Dishman, 1989). Participants will be replying using the five-point Likert scale ranging from "Strongly disagree" (1 point) to "Strongly agree" (5 points). Scoring for each subscale will be provided by summing all numerical ratings for each item and divided by the total number of items. The total score of expected outcomes and barriers for exercise will be ranged from 1 point (low) to 5 points (high).

Physical activity will be measured using ActivPAL accelerometers (PAL Technologies Ltd. Glasgow, UK). Data will be exctracted for average step counts (steps/day), time stepping (hours), time sitting/lying (hours), and energy equivalent (METs) and analysed as continuos variable for pre-to post-intervention differences.