Intervention
The research team partnered with Newcastle and Lake Macquarie Local Government municipal districts that had previously installed outdoor gym equipment at 12 locations (
Appendix Table 1, available online, for the characteristics of included facilities). This effectiveness trial aimed to increase physical activity (particularly RT) by prescribing workouts through the ecofit smartphone app using the installed outdoor gym equipment. The ecofit app included self-monitoring, goal setting, home-challenge functions. On the basis of the physical activity guidelines, each ecofit workout was built on a standardized structure of 8 RT exercises targeting major muscle groups. Participants could choose between different workout types, training categories, and difficulty levels (see
Appendix Table 2, available online, for more details about the workouts). Participants were encouraged to complete 2 ecofit workouts per week (to meet the minimum RT guidelines).
All intervention participants were asked to attend a 90-minute introductory session before receiving access to the ecofit app. The introductory session (
n=53 sessions in total; delivered to
n=119 participants) was initially conducted face to face; however, owing to the coronavirus disease 2019 (COVID-19) pandemic over half of the sessions (54%) were completed through Zoom video conference to almost half of the participants (46%). The content between the face-to-face and Zoom sessions was very similar, except for some minor deviations (
Appendix Table 3, available online).
All intervention group participants were invited to join the ecofit Facebook group, which was set up for participants to find workout partners and socialize. The Facebook group was facilitated by participants (i.e., the researchers did not facilitate or encourage conversation); however, the researchers monitored the site to ensure appropriate use.
Appendix Figures 1 and 2 (available online) depict the home page of the app and examples of a workout location, work out, and exercise. See
Appendix Table 2 (available online) for more details of the ecofit intervention components.
Measures
The primary and secondary outcome measures are outlined below. Data collection occurred at baseline (September 2019 to March 2021), 3 months (primary time point; December 2019–August 2021), and 9 months (follow-up; May 2021–March 2022). Owing to the COVID-19 pandemic, some of the 3-month (2%) and 9-month (23%) physical assessments were conducted through Zoom as opposed to face to face (as per the original protocol). Body composition was only assessed at baseline and at the 3-month time point to minimize participant burden. Assessors conducting the face-to-face physical assessments were blinded to participant group assignment; however, body composition scans and the physical assessments undertaken through zoom were not blinded. For an outline of deviations from the original protocol because of the COVID-19 pandemic, see
Appendix Table 3 (available online).
The coprimary outcomes were upper and lower body muscular fitness. Upper body muscular fitness was evaluated using the validated 90-degree push-up test.
27
Fitnessgram/Activitygram Reference Guide.
,
28
- Hashim A
- Ariffin A
- Hashim AT
- Yusof AB.
Reliability and validity of the 90° push-ups test protocol.
The push-up test assesses the maximum number of repetitions that can be performed correctly in rhythm (i.e., 40 beats per minute) without breaking form for >2 consecutive or nonconsecutive push-ups. Lower body muscular fitness was measured using the validated 60-second sit-to-stand test.
29
- Strassmann A
- Steurer-Stey C
- Lana KD
- et al.
Population-based reference values for the 1-min sit-to-stand test.
,
30
1-minute Sit-to-Stand Test: systematic review of procedures, performance, and clinimetric properties.
The sit-to-stand test measures lower body muscular strength and endurance by the number of times the person can stand up and sit down on a regular chair in 1 minute.
31
Sit-to-stand test for measuring performance of lower extremity muscles.
Secondary outcomes included device-based and self-reported physical activity, body composition (i.e., fat and lean tissue, visceral adipose tissue), aerobic fitness, body mass index (BMI), resistance-based physical activity, happiness, and mental health outcomes. Validated social-cognitive outcomes were also assessed, including physical activity self-efficacy, RT self-efficacy, implementation intentions for RT, and social support.
Device-measured physical activity was measured using the Actigraph GT9X Link wrist-worn accelerometers, which have shown acceptable validity and reliability compared with other monitors.
32
- Sasaki JE
- John D
- Freedson PS.
Validation and comparison of ActiGraph activity monitors.
,
33
- Santos-Lozano A
- Santín-Medeiros F
- Cardon G
- et al.
Actigraph GT3X: validation and determination of physical activity intensity cut points.
,
34
- Santos-Lozano A
- Marín PJ
- Torres-Luque G
- Ruiz JR
- Lucía A
- Garatachea N.
Technical variability of the GT3X accelerometer.
Daily wear time and periods of sleep were identified using validated methods,
35
- Cole RJ
- Kripke DF
- Gruen W
- Mullaney DJ
- Gillin JC.
Automatic sleep/wake identification from wrist activity.
,
36
- Suorsa K
- Pulakka A
- Leskinen T
- et al.
Objectively measured sedentary time before and after transition to retirement: the Finnish retirement and aging study.
,
37
- Choi L
- Ward SC
- Schnelle JF
- Buchowski MS.
Assessment of wear/nonwear time classification algorithms for triaxial accelerometer.
with the duration of physical activity during periods of wear and nonsleep classified on the basis of the following Vector Magnitude criteria: sedentary (0–2,859 per minute), light sedentary (2,860–3,940 per minute), and moderate to vigorous (>3,941 per minute).
38
- Montoye AHK
- Clevenger KA
- Pfeiffer KA
- et al.
Development of cut-points for determining activity intensity from a wrist-worn ActiGraph accelerometer in free-living adults.
A minimum of 10-hour wear during nonsleep periods on at least 5 days was classified as valid wear time.
Body composition was measured using the GE Lunar Prodigy Scanner (Model part: Spellman, Lunar 8,743), which is a validated dual energy x-ray absorptiometry.
39
- Haarbo J
- Gotfredsen A
- Hassager C
- Christiansen C.
Validation of body composition by dual energy X-ray absorptiometry (DEXA).
Participants lay on a scanner bed while scanning arms passed over their bodies measuring muscle and fat composition. Each scan took approximately 6–11 minutes to complete.
Aerobic fitness was measured using the validated 3-minute YMCA step test.
40
YMCA Fitness Testing and Assessment Manual.
,
41
- Beutner F
- Ubrich R
- Zachariae S
- et al.
Validation of a brief step-test protocol for estimation of peak oxygen uptake.
The step test measured cardiorespiratory fitness by stepping up and down a platform at the rate of 24 step-ups per minute for 3 consecutive minutes. On completion of the test, participants sat quietly for 1 minute. Participants’ performance level was determined by the recovery heart rate at the 1-minute mark after completion.
BMI was calculated using the standard equation (weight [kg]/height[m]²). Weight and height were estimated to the nearest 0.1 kg and 0.1 cm and were measured using a portable digital scale and stadiometer, respectively.
Self-reported aerobic and resistance-based physical activity were measured using a modification of the validated Godin Leisure-Time questionnaire.
42
The Godin-Shephard leisure-time physical activity questionnaire: validity evidence supporting its use for classifying healthy adults into active and insufficiently active categories.
, The modification included the average number of minutes per session
44
- Courneya KS
- Jones LW
- Rhodes RE
- Blanchard CM.
Effects of different combinations of intensity categories on self-reported exercise.
,
45
- Plotnikoff RC
- Taylor LM
- Wilson PM
- et al.
Factors associated with physical activity in Canadian adults with diabetes.
,
46
- Plotnikoff RC
- Lippke S
- Courneya KS
- Birkett N
- Sigal RJ.
Physical activity and social cognitive theory: a test in a population sample of adults with type 1 or type 2 diabetes.
as well as adding an additional question regarding RT, that is, average times per week participating in RT.
21
- Plotnikoff RC
- Wilczynska M
- Cohen KE
- Smith JJ
- Lubans DR.
Integrating smartphone technology, social support and the outdoor physical environment to improve fitness among adults at risk of, or diagnosed with, type 2 diabetes: findings from the ‘eCoFit’ randomized controlled trial.
Active travel was measured using 3 active travel items from the Global Physical Activity Questionnaire.
47
- Bull FC
- Maslin TS
- Armstrong T.
Global physical activity questionnaire (GPAQ): nine country reliability and validity study.
Items included, Do you walk or use a bicycle at least 10-minutes continuously to get to and from places, other than work, In a typical week, on how many days do you walk or bicycle for at least 10-minutes continuously to get to and from places, other than work?, and How much time in minutes and hours do you spend walking or cycling for travel on a typical day?.
Self-reported happiness was measured using the item
In the past month, have you felt happy.
48
- Richards J
- Jiang X
- Kelly P
- Chau J
- Bauman A
- Ding D.
Don’t worry, be happy: cross-sectional associations between physical activity and happiness in 15 European countries.
Participants chose from 6 response options ranging from never (1) to all the time (6).
Mental ill health was measured using the 21-item Depression, Anxiety, and Stress scale.
49
- Antony MM
- Bieling PJ
- Cox BJ
- Enns MW
- Swinson RP.
Psychometric properties of the 42-item and 21-item versions of the Depression Anxiety Stress Scales in clinical groups and a community sample.
Each item was scored using a 4-point Likert-type scale, ranging from 0 (item did not apply to them at all) to 3 (item applied to them very much or most of the time), to rate the extent to which participants experienced each state over the past week. For example, over the last week, “I have found it hard to wind down” (0=not at all, 1=some of the time, 2=a good part of the time, and 3=most of the time). Scores were determined by calculating the sum of scores of the relevant items within each of the 3 scales.
Physical activity self-efficacy refers to an individual’s confidence to engage in physical activity and was measured using 10 items from a 13-item scale for physical activity.
46
- Plotnikoff RC
- Lippke S
- Courneya KS
- Birkett N
- Sigal RJ.
Physical activity and social cognitive theory: a test in a population sample of adults with type 1 or type 2 diabetes.
For instance, “I am confident that I can participate in regular physical activity when I am a little tired.” Participants responded on a 5-point Likert-type scale (α=0.91) ranging from (1) not at all confident to (5) extremely confident.
RT self-efficacy refers to an individual’s belief in their capacity to perform RT and was measured using a 4-item scale.
50
- Lubans DR
- Morgan P
- Callister R
- et al.
Test-retest reliability of a battery of field-based health-related fitness measures for adolescents.
Questions queried participants about their confidence in engaging in muscle-strengthening activities, for example, “If I don’t have access to a gym I can still do RT activities.” Each item was scored on a 5-point Likert-type scale (α=0.77), ranging from (1) strongly disagree to (5) strongly agree.
Implementation intention for RT was measured using a 7-item scale, which adopted Gollwitzer’s principle of implementation intentions for physical activity behavior.
51
Implementation intentions: strong effects of simple plans.
The questions had been altered to assess RT as opposed to physical activity. Four items of the 7-item scale were scored on a 5-point Likert-type scale (α=0.87), ranging from 1 (not at all) to 5 (completely).
Perceived environment in relation to physical activity was measured using a 7-item scale.
52
- Sallis JF
- Kerr J
- Carlson JA
- et al.
Evaluating a brief self-report measure of neighborhood environments for physical activity research and surveillance: Physical Activity Neighborhood Environment Scale (PANES).
Participants were asked to read statements about physical activity and indicate how much they agree or disagree with each statement, for instance, “there are walking paths on most of the streets in my local area.” Each item was scored on a 4-point Likert-type scale (α=0-.68), ranging from (1) strongly disagree to (4) strongly agree.
Social support in relation to physical activity was measured using a 2-item scale
46
- Plotnikoff RC
- Lippke S
- Courneya KS
- Birkett N
- Sigal RJ.
Physical activity and social cognitive theory: a test in a population sample of adults with type 1 or type 2 diabetes.
,
53
- Courneya KS
- Plotnikoff RC
- Hotz SB
- Birkett NJ.
Social support and the Theory of Planned Behavior in the exercise domain.
: “People in my social network are likely to help me participate in regular physical activity” and “I feel that someone in my social network will provide the support I need to get regular physical activity.” The items were scored on a 7-point Likert-type scale (α=0.84), ranging from 1 (strongly agree) to 7 (strongly disagree). For all social-cognitive outcomes, mean scale scores were determined by calculating the sum of scores divided by the number of items.
Statistical Analysis
A priori power calculations (2 tailed) were conducted for the 2 primary outcomes (i.e., lower and upper body muscular fitness) at the primary time point of 3 months. The effect sizes for upper body muscular fitness are much larger than those for the lower body,
21
- Plotnikoff RC
- Wilczynska M
- Cohen KE
- Smith JJ
- Lubans DR.
Integrating smartphone technology, social support and the outdoor physical environment to improve fitness among adults at risk of, or diagnosed with, type 2 diabetes: findings from the ‘eCoFit’ randomized controlled trial.
so the sample size was conservatively based on the effect size for lower body muscular fitness. Power calculations were based on the ecofit efficacy study,
21
- Plotnikoff RC
- Wilczynska M
- Cohen KE
- Smith JJ
- Lubans DR.
Integrating smartphone technology, social support and the outdoor physical environment to improve fitness among adults at risk of, or diagnosed with, type 2 diabetes: findings from the ‘eCoFit’ randomized controlled trial.
with the alpha set at 0.025 (adjusting for 2 coprimary outcomes) and power=80%. To detect an effect of half a SD, 77 participants were required in each arm. Allowing for approximately 25% attrition (on the basis of the ecofit efficacy trial
22
- Wilczynska M
- Lubans DR
- Cohen KE
- Smith JJ
- Robards SL
- Plotnikoff RC.
Rationale and study protocol for the ‘eCoFit’ randomized controlled trial: integrating smartphone technology, social support and the outdoor physical environment to improve health-related fitness among adults at risk of, or diagnosed with, Type 2 Diabetes.
), 200 individuals (100 per arm) were required. Given that participants were able to enroll in clusters of up to 4 people, an average cluster size of 2–3 was assumed, with an interclass correlation=0.1, giving us an estimated design effect of up to 1.2. To detect the effect size for lower body muscular fitness mentioned earlier, a total sample size of 240 (120 per study group) was calculated.
Statistical analyses were conducted in April 2022 using IBM SPSS Statistics, Version 27. Data are presented as means (SD) or mean and 95% CIs for continuous variables and as counts (percentages) for categorical variables. Linear mixed models were used to assess intervention effectiveness for both primary and secondary outcomes. The models included a random effect for participant cluster and participant (to account for repeated measures) and fixed effects for group (intervention versus control), time (treated as a categorical variable; baseline, 3 months, and 9 months), and the group-by-time interaction. Analyses followed the intention-to-treat principle, and moderators were defined a priori. Moderators of intervention effects (i.e., sex, age [<65 years versus ≥65 years], BMI [≤30 versus ≥30]) were also explored using linear mixed models with interaction terms (i.e., group-by-time moderator). Sensitivity analyses (i.e., completed cases only and last observation carried forward at the primary time point [i.e., baseline and 3 months] and follow-up [i.e., baseline, 3 months, and 9 months]) were conducted for the 2 primary outcomes. A Bonferroni correction was applied to the p-value for the coprimary outcomes. Cronbach’s alpha (α) was calculated for each of the social-cognitive outcomes to determine the internal consistency of these measures.
