This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in the Journal of Medical Internet Research, is properly cited. The complete bibliographic information, a link to the original publication on https://www.jmir.org/, as well as this copyright and license information must be included.
Outdoor play is critical to children’s healthy development and well-being. Early learning and childcare centers (ELCCs) are important venues for increasing children’s outdoor play opportunities, and early childhood educators’ (ECE) perception of outdoor play can be a major barrier to outdoor play. The OutsidePlay-ECE risk-reframing intervention is a fully automated and open access web-based intervention to reframe ECEs’ perceptions of the importance of outdoor play and risk in play and to promote a change in their practice in supporting it in ELCC settings. We grounded the intervention in social cognitive theory and behavior change techniques.
The aim of this study is to evaluate the effectiveness of the OutsidePlay-ECE web-based risk-reframing intervention.
We conducted a single-blind randomized controlled trial in Canada between December 2020 and June 2021 to test the OutsidePlay-ECE risk-reframing intervention for ECEs. We recruited participants using social media and mass emails through our partner and professional networks. We invited ECEs and administrators working in an ELCC, who can speak, read, and understand English. We randomized consented participants to the intervention or control condition. The participants allocated to the intervention condition received a link to the OutsidePlay-ECE intervention. Participants allocated to the control condition read the Position Statement on Active Outdoor Play, a 4-page document on research and recommendations for action in addressing barriers to outdoor play. The primary outcome was a change in tolerance of risk in play. The secondary outcome was goal attainment. We collected data on the web via REDCap (Vanderbilt University) at baseline and 1 week and 3 months after intervention.
A total of 563 participants completed the baseline survey, which assessed their demographics and tolerance of risk in play. They were then randomized: 281 (49.9%) to the intervention and 282 (50.1%) to the control condition. Of these, 136 (48.4%) and 220 (78%) participants completed the baseline requirements for the intervention and control conditions, respectively. At 1 week after intervention, 126 (44.8%) and 209 (74.1%) participants completed follow-up assessments, respectively, and at 3 months after intervention, 119 (42.3%) and 195 (69.1%) participants completed the assessments, respectively. Compared with participants in the control condition, participants in the intervention group had significantly higher tolerance of risk in play at 1 week (
The results of this randomized controlled trial demonstrated that the OutsidePlay-ECE intervention was effective and had a sustained effect in increasing ECEs’ and administrators’ tolerance of risk in play. It was not effective in increasing goal attainment.
ClinicalTrials.gov NCT04624932; https://clinicaltrials.gov/ct2/show/NCT04624932
RR2-10.2196/31041
Children’s opportunities for outdoor play have been decreasing over successive generations in many developed countries [
Over the past few decades, several factors have been proposed to explain the overall decline in children’s outdoor play. Increasingly, risk-averse cultural norms have resulted in ubiquitous adult supervision and playground equipment that offer little challenge [
To address these concerns and influence parents’ perception of the importance of outdoor play, reduce their fears regarding the risks taken in play, and help them develop supportive parenting behaviors toward outdoor play, we built the OutsidePlay web-based risk-reframing intervention [
For many children, most of their waking hours are spent in an ELCC, which can be an invaluable opportunity to provide them with high-quality opportunities for outdoor play, particularly for children who may have limited access to outdoor play in their home environments [
The aim of this study was to report the results of a randomized controlled trial (RCT) evaluating the effectiveness of the OutsidePlay-ECE intervention in increasing ECEs’ and ELCC administrators’ tolerance of risk in play and attain a behavior change goal related to providing outdoor play opportunities for children in their ELCC. Given the positive RCT results that we obtained on the effectiveness of the OutsidePlay-Parent intervention previously developed for parents [
We used a single-blind (researchers and outcome assessors), 2-parallel condition RCT. We conducted this study between December 1, 2020, and June 30, 2021, in Canada and collected measures at baseline and at 1 week and 3 months after intervention. Our primary outcome was the change in tolerance of risk in play at either follow-up time point. The secondary outcome was the participants’ goal attainment at either follow-up time point. The details on the intervention’s theoretical framework, development, content, and the RCT study protocol can be found in the study by Brussoni et al [
The RCT was approved by the University of British Columbia and Children’s and Women’s Health Center of British Columbia Research Ethics Board (H19-03644). We conducted the RCT (including the intervention) entirely on the web; thus, there was no human involvement, except when participants had inquiries and reported technical issues via email. The sole identifiable information collected was the participants’ email addresses, which were required for sending allocated study material, follow-up measures, and reminders. We did not export the participants’ email addresses for data analysis. We assigned each participant a study number that did not include identifiable personal information.
We recruited participants between December 1, 2020, and March 15, 2021, via social media posts on Facebook, Facebook advertisements, Twitter, and Instagram. We also circulated mass recruitment emails through partners and professional networks. Potential participants completed a web-based survey in REDCap (Research Electronic Data Capture; Vanderbilt University), an electronic data capture tool hosted by and stored in the British Columbia Children’s Hospital Research Institute server [
We temporarily halted participant recruitment and participation from December 18, 2020, to January 4, 2021, to accommodate the Christmas and New Year holidays, during which most ELCCs were closed. We made this decision to secure more valid participant responses to the goal attainment question in the 1 week after intervention follow-up time point, asking “Did you accomplish your goal?” which concerned their behavior in promoting children’s outdoor play, specifically in their ELCC. We posted a message on the REDCap enrollment survey informing participants of this interruption and the date that RCT recruitment would resume.
Eligible participants were adult ECEs and ELCC administrators currently working in Canada, who could speak, read, and understand English. Given that ECEs work closely with, and are influenced by, ELCC administrators, we included both ECEs and ELCC administrators in this RCT. We did not have any exclusion criteria. We included participants deemed eligible according to the aforementioned eligibility criteria. As the RCT was conducted entirely on the web, computer and internet literacy was an implicit de facto eligibility criterion. Eligible and interested participants provided consent by downloading the consent form for review and selecting a checkbox to participate. We then invited the enrolled participants to complete the baseline survey, which included sociodemographic questions and a questionnaire that assessed participant tolerance of risk in play, and enter their email address.
We automatically randomized the enrolled participants who completed the baseline survey in REDCap to 1 of the 2 conditions: intervention and control. The participants had an equal (50%) likelihood of being assigned to each condition. We generated the randomization schedule beforehand by the Sealed Envelope service (Sealed Envelope Ltd) using randomized permuted blocks of sizes 4, 6, and 8. We then transferred the list to REDCap. We concealed allocation to the researchers during participant assignment and data analysis. We sent participants a unique link to their materials upon completion of the baseline survey and randomization. The nature of the intervention did not permit blinding of the participants. They may have intuited which condition they were allocated to, based on the details of the 2 conditions provided in the consent form: intervention (eg, web-based intervention) and control (ie, a PDF document). In addition, there has always been a risk that 2 or more participants from the same ELCC participated in the study and have become exposed to a condition different from theirs by communicating with their peers. We believe this scenario would be unlikely given that we recruited participants across Canada, and as such, we did not implement any precaution.
The goal of the OutsidePlay-ECE intervention is to reframe ECEs’ perception of the importance of outdoor play and its inherent risks and promote a change in their practice in supporting children’s outdoor play in ELCC settings. We designed OutsidePlay-ECE for ECE as a fully automated web-based risk-reframing intervention. It consists of 3 chapters, which are guided by the animated character of the first author (MB) and include self-reflection questions. We (the study authors) developed the OutsidePlay-ECE intervention following the intervention mapping process [
Screenshots of the OutsidePlay-ECE intervention landing page.
Our protocol paper [
The OutsidePlay-ECE intervention focused on social cognitive theory constructs: outcome expectation; knowledge, barriers and opportunities; observational learning; self-efficacy; behavioral skills; and intentions [
We soft launched the OutsidePlay-ECE intervention on December 1, 2020, for the RCT and froze the content during the RCT (ie, we did not make any changes) and analysis. This means that we released the intervention only for RCT purposes with no publicity push before its full launch to the public. The participants were allocated to the intervention condition with a link to the OutsidePlay-ECE intervention. It could be completed in up to 100 minutes, depending on the participants’ movements through each chapter. Participants could also return to the intervention at their convenience, picking up from where they left off previously, provided that they did not delete their browser cache and http cookies. REDCap sent out a maximum of 3 automated reminders at 24, 48, and 60 hours after completion of the baseline survey and at the 1 week and 3 months after intervention follow-ups.
We asked participants in the control condition to review a PDF of the Position Statement on Active Outdoor Play, a 4-page document with information on research and recommendations for action in addressing barriers to outdoor play [
Our primary outcome measure was change in the total score on the Teacher Tolerance of Risk in Play Scale (T-TRiPS), a validated, reliable 26-item measure with dichotomous yes or no responses on items that reflect the 6 categories of risky play (great heights, high speed, dangerous tools, dangerous elements, rough-and-tumble, and disappear or get lost) [
Our secondary outcome measure was self-reported behavior change, measured by participants’ self-reported progress in attaining the goal they set for themselves. At each follow-up time point, participants were reminded of their goal and asked, “Did you accomplish your goal?” with dichotomous yes or no responses.
We assessed the primary outcome measure at baseline and at 1 week and 3 months after intervention. We only assessed the secondary outcome measure at 1 week and 3 months after intervention because at baseline, they could not have accomplished a goal they had not yet set. We paid the participants US $25 via electronic transfer upon completing the baseline questionnaire and allocated intervention. We then paid them US $16 at each of the follow-up time points at 1 week and 3 months after intervention. In addition, we issued participants in the intervention condition a professional development certificate for 100 minutes upon completion of the OutsidePlay-ECE intervention.
Study flow diagram.
We conducted all statistical analyses in Stata (StataCorp, version 15) [
For a sample size of 206 ECEs and ELCC administrators in total, a linear mixed model examining the impact of intervention relative to control, including an interaction with time, was calculated to have 80% power at a
To compare the raw outcome differences between conditions at each time point, for continuous outcomes (TRiPS scores), we used 1-way ANOVA or Kruskal-Wallis
We concluded that linear and generalized linear mixed effects models with random intercepts and unstructured covariance were fit to analyze the effects of the intervention on T-TRiPS scores and goal attainment, respectively. In other words, we used the mixed effects regression analysis to examine (1) whether T-TRiPS scores changed between 1 week and 3 months after intervention and (2) whether these changes were greater in the intervention condition (ie, the OutsidePlay-ECE intervention) than in the control condition. We used intent-to-treat analysis of T-TRiPS scores that used the last observation carried forward as the method of imputation, because these participants only completed the baseline survey and did not complete the intervention, it is reasonable to expect their T-TRiPS scores to remain the same throughout the study period. Unstandardized (ie, raw)
Similar to the T-TRiPS analysis, we conducted a generalized mixed effects regression analysis to examine the effect of the intervention on goal attainment, when comparing the control condition at the 1 week after intervention follow-up time point with the intervention condition at the 3 months after intervention follow-up time point. An intention-to-treat analysis of goal attainment was not performed because of the absence of baseline data. To establish a goal, participants had to complete either interventions (eg, the OutsidePlay-ECE intervention or the Position Statement on Active Outdoor Play). Consequently, there was no basis for imputing the values of goal attainment. We calculated odds ratios, which were interpreted as the odds of attaining goals for the intervention condition at the 3 months after intervention follow-up time point, divided by the odds of the control condition at 1 week after intervention (relative effect size). We also calculated the absolute effect size, that is, risk differences and the probability of attaining a goal in the intervention group minus the probability in the control condition.
The University of British Columbia/Children’s and Women’s Health Center of British Columbia Research Ethics Board categorized the intervention as low risk and not associated with any harm. No privacy breaches or technical problems affected the participants. Although we tried to accommodate participants’ varying internet bandwidths by automatically adjusting the media resolution (eg, high or low), this did not resolve issues caused by some users accessing the intervention from old or incompatible devices.
We included baseline sociodemographic data from 563 ECEs and ELCC administrators who were randomized to a condition in our analyses (baseline characteristics between the 2 conditions are presented in
We compared the sociodemographic characteristics among those who completed the baseline survey and between those who were randomized (N=563) and those who were not randomized (N=56), as these participants did not provide email address to proceed with the study and found that those who were randomized were more likely to be female (544/563, 96.6% vs 51/56, 91.1%, respectively;
Baseline characteristics between the 2 intervention conditions.
Characteristics of participants who completed the baseline survey | Control (n=282) | Intervention (n=281) | Total (N=563) | |
|
||||
|
Male | 8 (2.8) | 8 (2.9) | 16 (2.8) |
|
Female | 272 (96.4) | 272 (96.8) | 544 (96.6) |
|
Other | 2 (0.7) | 0 (0) | 2 (0.4) |
|
Prefer not to answer | 0 (0) | 1 (0.4) | 1 (0.2) |
|
||||
|
19 to 24 | 26 (9.3) | 33 (11.8) | 59 (10.5) |
|
25 to 30 | 72 (25.6) | 55 (19.6) | 127 (22.6) |
|
31 to 40 | 71 (25.3) | 86 (30.7) | 157 (28) |
|
41 to 50 | 66 (23.5) | 64 (22.9) | 130 (23.2) |
|
51 to 60 | 32 (11.4) | 36 (12.8) | 61 (12.1) |
|
61 to 70 | 13 (4.6) | 6 (2.1) | 19 (3.4) |
|
≥71 | 1 (0.4) | 0 (0) | 1 (0.2) |
|
Prefer not to answer | 1 (0.4) | 1 (0.4) | 2 (0.4) |
|
||||
|
English | 263 (93.3) | 261 (92.9) | 524 (93.1) |
|
Othera | 19 (6.7) | 20 (7.1) | 39 (6.9) |
|
||||
|
ECEb | 203 (72) | 189 (67.3) | 392 (69.6) |
|
ECE administrator | 75 (26.6) | 90 (32) | 165 (29.3) |
|
Otherc | 4 (1.4) | 2 (0.7) | 6 (1.1) |
Working in the field (years; N=530), mean (SD) | 10.22 (9.51) | 10.05 (9.16) | 10.14 (9.33) | |
Working at the current center (years; N=530), mean (SD) | 6.09 (7.41) | 5.18 (5.98) | 5.63 (6.74) | |
|
||||
|
Alberta | 6 (2.1) | 7 (2.5) | 13 (2.3) |
|
British Columbia | 129 (45.9) | 129 (46.2) | 258 (46.1) |
|
Manitoba | 3 (1.1) | 2 (0.7) | 5 (0.9) |
|
New Brunswick | 52 (18.5) | 44 (15.8) | 96 (17.1) |
|
Newfoundland and Labrador | 8 (2.9) | 12 (4.3) | 20 (3.6) |
|
Nova Scotia | 13 (4.6) | 9 (3.2) | 22 (3.9) |
|
Ontario | 64 (22.8) | 66 (23.7) | 130 (23.2) |
|
Prince Edward Island | 0 (0) | 2 (0.7) | 2 (0.4) |
|
Quebec | 2 (0.7) | 4 (1.4) | 6 (1.1) |
|
Saskatchewan | 4 (1.4) | 4 (1.4) | 8 (1.4) |
|
||||
|
Yes | 266 (97.4) | 264 (97.4) | 530 (97.4) |
|
No | 7 (2.6) | 7 (2.6) | 14 (2.6) |
|
||||
|
Small: 1 to 24 | 91 (32.7) | 95 (34.1) | 186 (33.4) |
|
Medium: 25 to 48 | 76 (27.3) | 78 (28.0) | 154 (27.7) |
|
Large: ≥49 | 111 (39.9) | 106 (38.0) | 217 (39.0) |
|
||||
|
Small: 1 to 5 | 100 (36.2) | 102 (37.1) | 202 (36.7) |
|
Medium: 6 to 12 | 85 (30.8) | 91 (33.1) | 176 (31.9) |
|
Large: ≥13 | 91 (33.0) | 82 (29.8) | 173 (31.4) |
|
||||
|
Yes | 270 (97.1) | 270 (96.8) | 540 (97) |
|
No | 8 (2.9) | 9 (3.2) | 17 (3) |
|
||||
|
Very good | 96 (35.6) | 88 (32.7) | 184 (34.1) |
|
Good | 95 (35.2) | 110 (40.9) | 205 (38.0) |
|
Acceptable | 65 (24.1) | 60 (22.3) | 125 (23.2) |
|
Poor | 12 (4.4) | 11 (4.1) | 23 (4.3) |
|
Very poor | 2 (0.7) | 0 (0) | 3 (0.4) |
Time children spent playing outdoors at the center (hours; N=556), mean (SD) | 2.01 (1.15) | 2.14 (1.13) | 2.08 (1.14) | |
|
||||
|
Yes | 253 (89.7) | 250 (90.3) | 503 (90.0) |
|
No | 11 (3.9) | 14 (5.1) | 25 (4.5) |
|
Feeling partially supported | 8 (2.8) | 4 (1.4) | 12 (2.1) |
|
N/Ad | 10 (3.5) | 9 (3.2) | 19 (3.4) |
aArabic (n=3), Cantonese (n=3), Chinese (n=2), Croatian (n=2), Gujarati (n=2), Hindi (n=1), Hungarian (n=1), Korean (n=6), Mandarin (n=1), Minnan (a Chinese dialect; n=1), Nepali (n=1), Punjabi (n=4), Serbian (n=1), Sinhala (n=1), Sinhalese (n=1), Slovak (n=2), Spanish (n=2), Tagalog (n=2), Tamil (n=2), Turkish (n=1), and Dutch (n=1).
bECE: early childhood educator.
cIncludes childcare provider consultant (n=1), child and youth care (n=1), no ECE (n=2), classroom teacher (n=1), and instructor at college (n=1).
dN/A: not applicable; for example, the participant is the only staff member.
Description of the Teacher Tolerance of Risk in Play Scale scores by intervention conditions and time points.
Evaluation period | Sample size, N | Control, mean (SD) | Intervention, mean (SD) | |
Baseline | 563 | 0.040 (1.207) | −0.040 (1.243) | .44 |
Completed intervention | 356 | 0.017 (1.211) | 0.123 (1.196) | .86 |
1 week after intervention | 337 | −0.156 (1.304) | 0.262 (1.117) | .003 |
3 months after intervention | 314 | −0.118 (1.400) | 0.200 (1.211) | .04 |
Mixed effects regression analysis for the Teacher Tolerance of Risk in Play Scale (T-TRiPS) θ score.
Regression and condition comparisons | Coefficients (95% CI) | |||||
|
||||||
|
Intervention effects: intervention versus control | 0.100 (−0.169 to 0.369) | .47 | .02 | ||
|
|
.99 | ||||
|
|
1 week versus baseline | −0.154 (−0.267 to −0.041) | .007 |
|
|
|
|
3 months versus baseline | −0.124 (−0.240 to −0.008) | .04 |
|
|
|
|
.002 | ||||
|
|
Intervention versus control by 1 week versus baseline | 0.320 (0.135 to 0.505) | .001 |
|
|
|
|
Intervention versus control by 3 months versus baseline | 0.251 (0.062 to 0.440) | .009 |
|
|
|
||||||
|
Intervention effects: intervention versus control | 0.019 (−0.217 to 0.254) | .88 | .054 | ||
|
|
.96 | ||||
|
|
1 week versus baseline | −0.156 (−0.268 to −0.044) | .006 |
|
|
|
|
3 months versus baseline | −0.126 (−0.241 to −0.011) | .03 |
|
|
|
|
<.001 | ||||
|
|
Intervention versus control by 1 week versus baseline | 0.335 (0.156 to 0.514) | <.001 |
|
|
|
|
Intervention versus control by 3 months versus baseline | 0.271 (0.088 to 0.454) | .004 |
|
aItalicization denotes two separate sets of analysis.
We asked participants to think of one tangible and achievable goal and created a feasible plan to accomplish it. Participant goals varied widely from “I could add more building tools” to “Bring this topic and learning opportunity up with my colleagues at our daily check-in.”
Goal attainment by intervention condition and time point.
Evaluation period and goal attainment | Control, n (%) | Intervention, n (%) | Sample size, N | ||||||
|
335 | .17 | |||||||
|
Yes | 141 (67.5) | 94 (74.6) |
|
|
||||
|
No | 68 (32.5) | 32 (25.4) |
|
|
||||
|
314 | .18 | |||||||
|
Yes | 163 (83.6) | 106 (89.1) |
|
|
||||
|
No | 32 (26.4) | 13 (10.9) |
|
|
Results of the mixed effects regression analysis for goal attainment by intervention condition and time.
Regression and condition comparisonsa | Relative effect size | Absolute effect sizes | ||
|
Odds ratios (95% CI) | Risk differences (95% CI) | ||
Intervention effects: intervention versus control | 2.046 (0.740 to 5.655) | .17 | 0.071 (−0.019 to 0.162) | .12 |
Time effects: 3 months versus 1 week | 5.749 (2.664 to 12.407) | <.001 | 0.158 (0.098 to 0.218) | <.001 |
Intervention by time effects: intervention versus control by 3 months versus 1 week | 1.124 (0.335 to 3.774) | .85 | −0.018 (−0.115 to 0.079) | .72 |
aN=335, who were randomized to a condition and set a goal at baseline and completed goal attainment measures at a follow-up time point.
Our RCT tested the effectiveness of the web-based OutsidePlay-ECE intervention in changing ECEs’ and ELCC administrators’ tolerance of risk in play and the attainment of their personalized goals for change to support children’s outdoor play within the ELCC. The RCT results partially support our hypotheses. ECEs and ELCC administrators receiving the intervention reported significantly higher increases in their tolerance of risk in play at 1 week after intervention than participants in the control condition. These differences remained significant at 3 months after intervention. There were no significant differences in goal attainment. These results are consistent with the findings of a previous RCT testing the OutsidePlay-Parent intervention, which also found significantly greater increases in tolerance of risk in play for intervention versus control participants at 1 week and 3 months after intervention [
There are several possibilities for the lack of a significant effect of the intervention on the secondary outcome, goal attainment. Findings reported for the OutsidePlay-Parent intervention [
As ECE attitudes toward outdoor play and risk-taking in play have a major impact on children’s outdoor play in ELCCs [
Our study had several limitations. First, although social cognitive theory and respective behavior change techniques have been mapped [
ELCCs are important settings for influencing early childhood, and these childcare experiences can impact lifelong health, development, and well-being trajectories [
CONSORT-eHEALTH checklist (V 1.6.1).
Complete screenshots of the OutsidePlay-ECE intervention.
Consolidated Standards of Reporting Trials of Electronic and Mobile Health Applications and Online Telehealth
early childhood educator
early learning childcare center
randomized controlled trial
Research Electronic Data Capture
Teacher Tolerance of Risk in Play Scale
This RCT study was funded by the Lawson Foundation (grant GRT 2019-79). The Lawson Foundation was not involved in any aspect of the study design or the writing of the manuscript. MB was supported by salary awards from the British Columbia Children’s Hospital Research Institute. The authors are also grateful to the Government of Canada—Early Learning and Child Care—under the Social Development Partnership program, Project 016554719 for funding the development of the OutsidePlay-ECE intervention. The authors thank the Digital Lab at the University of British Columbia Department of Pediatrics for providing partnership and technical support in the development and programming of the OutsidePlay intervention.
All data, password-protected and stored in the secure network at the British Columbia Children’s Hospital Research Institute, will be available from MB upon reasonable request within 5 years of the completion of the study.
MB conceived of the study. CSH, FM, MZ, MW, TC, and EO assisted MB with the development of the OutsidePlay-ECE intervention content. MB and JJ led the development of the intervention design, with contribution from CSH. YL performed the statistical analysis. CSH and YL wrote the first full draft of this manuscript. All authors have read and approved the final manuscript.
The authors declare that they are both the developers and evaluators of the intervention.