Effectiveness of an injury prevention video on risky behaviours in youth snow sports: A randomized controlled trial

Abstract

Objectives

Ski and snowboard-related injuries are common among Canadian youth. Analyzing the role of risky behaviours that contribute to injury risk is essential for gaining an understanding of injury prevention opportunities. The objective was to determine if rates of risky behaviour seen at the ski hill were lower for children and adolescents exposed to an educational injury prevention video.

Methods

This single-blinded cluster randomized controlled trial included students (ages 7–16) from 18 Calgary schools who were enrolled in novice levelled school-sanctioned ski and snowboard programs. Consenting schools were randomly assigned to the intervention or control. The control group followed standard preparation including watching a general ski hill orientation video that was created by the ski hill. The intervention group viewed the intervention video focussed on injury prevention. The Risky Behaviour and Actions Assessment Tool was used by blinded research assistants to observe and record students’ risky behaviours at an Alberta ski hill.

Results

In total, 407 observations estimated the rate of risky behaviour. The overall rate of risky behaviour was 23.31/100 person runs in the control group and 22.95/100 person runs in the intervention group. The most commonly observed risky behaviours in both groups were skiing too close to other skiers/snowboarders and near collision with an object/person.

Conclusions

Both groups showed similar rates of risky behaviour and demonstrated the same most common type of behaviour. Practical applications: future work should focus on mitigating common risky behaviours.

INTRODUCTION

Skiing and snowboarding are common winter activities, especially in locations such as Alberta, Canada. Recent data estimates 2.5 million ski area visits in Alberta annually and 78 million annual visits at ski hills across North America (1). Participants face potential injury risk due to unpredictable terrain and high speeds, which may lead to multiple risky behaviours and, potentially, injury (2). Injuries can range from minor scrapes to fatal and catastrophic trauma. Ski and snowboard helmets have been the primary injury prevention strategy to date; however, few studies have considered the actual risky behaviour in these sports (3,4). Poor decisions, actions, and human error can be potential causes of snow sport injury (5). Greater insight about risky behaviours is essential for gaining a comprehensive understanding opportunities for sport injury prevention (4).

Risky behaviour is any action that has the potential to result in harm to oneself or others. Risky behaviour can contribute to the injury mechanism (4). Compared with adults, children and adolescents are more likely to engage in risk taking behaviours. Adolescents have the highest risk of injury (6). It is important to investigate ways to prevent risky behaviours and decrease injury risk in popular sport and recreational activities such as skiing and snowboarding.

The most common risky behaviours and the rate of these behaviours while skiing and snowboarding are largely unknown. This limits our understanding on how to promote safe behaviours and prevent injuries. Therefore, the aim of this study was to use the Risky Behaviour and Actions Assessment Tool (RBAAT) to determine the incidence of risky behaviours among children and adolescents at a ski hill and to determine if a safety video intervention reduced the incidence of risky behaviours while skiing or snowboarding compared to a general safety and ski hill orientation video.

METHODS

Ethics approval was attained from the University of Calgary Conjoint Health Research Ethics Board (ID: REB15-0749) and the trial details were registered at ClinicalTrials.gov (identifier: NCT03184779). All schools consented to the participation of their students within the study. The study was a single-blinded cluster randomized controlled trial to examine the effect of a ski and snowboard safety video on outcomes including risky behaviours. The randomization procedure was conducted by the research coordinator who knew the intervention status of all schools and coordinated administration of videos but did not participate in data collection. Schools were stratified by Junior High school (typically grades 7–9) and elementary school (typically Kindergarten to grade 6) with block sizes of 2, 4, and 6. Students at the ski hill on an observation day were either all assigned to the intervention, or all assigned to the control group. Allocation concealment was ensured as the organizer did not inform the research assistants (RAs) observing the risky behaviours which schools were allocated on what days and nor were they present when the video was shown.

The intervention video was built on the Health Action Process Approach model, the Alpine Responsibility Code, previous literature, feedback from co-investigators, focus group research, and existing evidence from ongoing local research in ski and snowboard injury prevention (7–15). A professional multimedia company developed the video. The intervention was aimed at educating children and youth about the risk of injuries and effective strategies to prevent them, while still promoting participation in snow sports. This 12-minute video was viewed by students at school prior to attending the ski hill.

The control video was the general ski hill orientation video created by the ski site for a wide audience, with minimal injury prevention focus. Students viewed this 9-minute video at their school prior to attending the ski hill. This video outlined what the students should expect during their visit to the ski hill.

During the 2016–2017 winter season, all schools registered in the ski area school program were invited to participate. The participants were students in grades 2–9 (ages: 7–16 years) and were enrolled in beginner ski or snowboard lessons at their local ski hill.

The RBAAT was created to serve as an accurate and reliable way of assessing risky behaviours on the ski hill (16). The tool consisted of a checklist of known risky behaviours that were identified in the literature, through personal communication with experts within the ski and snowboard industry, and through targeted focus groups. Literature identified risky behaviours such as obstructing the hill in any way, cutting off other participants, skiing too close to others, not following hill etiquette, any action that resulted in a collision, excessive speed, jumping unsafely, lack of awareness of surroundings, skiing backwards, not following rules, incorrect use of equipment, lack of control, and other actions that could result in severe injuries (5,17–20). Information from the ski hill’s guidelines and the Alpine Responsibility Code were incorporated into the final draft of the tool. The reliability of this tool was assessed by examining the inter-rater reliability between an experienced ski instructor and ski patroller, comparison of their consensus with a non-expert RA and finally between two RAs (16).

To identify participating students, students wore a yellow armband that was visible to the RAs. Observations were made in beginner areas for 90-minute periods. At times participants were in more than one area for the day. If the students with armbands were present in multiple areas, the RAs split their time between the areas. RAs evaluated the behaviour and characteristics of school participants using the RBAAT. The first RA (RA1) made observations on 15 days and a second RA (RA2) was present for 10 of those days.

When both RAs were ready to begin the observation period, they gave each other an auditory cue (e.g., “ready”). The first student observed was the next child exiting the magic carpet lift with an armband. The RAs discussed visual cues to ensure they were observing the same student (e.g., “green jacket”) before the student went down the hill. After the student was identified, the second RA confirmed the assigned ID number and said “GO” to begin data collection and “END” once the student was at the bottom of the run. Both RAs independently observed any of the listed risky behaviours by the student. On the 5 additional days, only one RA was able to collect data and they followed the same protocol.

We assumed a risky behaviour rate of 10 behaviours per 100 student runs among the control group and the video intervention would reduce risky behaviours by 50%. Therefore, 21 clusters per arm were needed for 80% power to detect a rate ratio of 0.5 (21). We estimated that 42 observation days were needed to obtain the required 2436 students in each arm (total of 4872).

The proportions of risky behaviours were calculated. Rates of risky behaviours performed per 100 student runs with 95% confidence intervals (CI) calculated for the intervention and control groups. Rate ratios were also calculated adjusting for clustering by school. The rates of risky behaviour were reported for two different definitions of risky behaviour: 1) one or both observers reported any risky actions on the hill or 2) both observers reported any risky actions on the hill.

RESULTS

A total of 100 schools (12,080 students) registered in the ski and snowboard school programs at the time of recruitment for the 2016–2017 season. Eighty schools were excluded: schools and school boards did not agree to participate, could not schedule the initial school visit before the outing, or schools enrolled students outside the age criteria. Twenty schools (2728 students) agreed to participate but video watching could not be scheduled prior to two schools’ first outing to the ski hill. Eighteen schools in the Calgary area (2348 students) were randomized to the intervention (8 schools, 1040 students) or control (10 schools, 1308 students) video (Figure 1). There were 31 potential observation days where included schools attended the ski hill. Based on inclement weather, school cancellations, and unavailability of the RAs, there were a total of 15 days of school observations at the hill by RA1. RA2 accompanied RA1 on 10 of those days. As a result of the scheduling, observations were made exclusively on elementary school students up to grade 6. Overall, 43% of ski and snowboard students were male and 54% were skiing (Table 1). RA1 observed 407 participant runs on 15 days with RA2 present to observe 227 of these runs on 10 days. Based on the data from RA1, most students (77%) were not observed engaging in any risky behaviours. When both RAs recorded at least one risky behaviour, the overall incidence of 23.10 (95% CI: 15.21–27.53) risky behaviours/100 runs and 24.57 (95% CI: 22.49, 36.99) risky behaviours/100 runs where either one of the RAs recorded at least one risky behaviour.

Figure 1.

Flowchart of schools and participants involved within the study.

Table 1.

Basic descriptors: sex, activity, proportions of risky behaviours, and overall risky behaviour

	Number (proportions)
Sex
Male	176 (0.43)
Female	125 (0.31)
Unknown sex	106 (0.26)
Activity
Ski	218 (0.54)
Snowboard	189 (0.46)
Number of risky behaviours in one ski run hill
0	312 (0.77)
1	36 (0.09)
2	28 (0.07)
3	16 (0.04)
4	12 (0.03)
5	2 (<0.01)
6	1 (<0.01)
Whether there was at least one risky behaviour or not
No	312 (0.77)
Yes	95 (0.23)

Based on the data from RA1, for the intervention video, there was an incidence of 22.95 risky behaviours/100 runs (95% CI: 17.63–28.26) and the incidence rate for the control video was 23.31 risky behaviours/100 runs (95% CI: 16.75–29.87). When defined as both RAs observing a risk behaviour, the incidence was 19.28 (95% CI: 13.19–27.21) risky behaviours/100 person runs for the intervention video and 24.59 (95% CI: 13.76–40.56) risky behaviours/100 person runs for the control video. For risky behaviour observed by at least one RA, the intervention group incidence rate was 29.52 (95% CI: 21.84–39.02) risky behaviours/100 person runs and 27.87 (95% CI: 16.23–44.62) risky behaviours/100 person runs for the control group.

The most common observed risky behaviours for both RAs were skiing too close to others (RA1 = 41; RA2 = 27) and near collision with an object or person (RA1 = 33; RA2 = 21). The next most common risky behaviours seen by RA1 were visibly unable to stop/stay in control (n = 20), excessive speed (n = 17), and stopping in an unsafe area (n = 17). For RA2, the next most common risky behaviours were visibly unable to stop/stay in control (n = 14), stopping in an unsafe area (n = 9), and skiing backwards (n = 9).

Focussing on the observations of RA1, there were no significant differences in the incidence rate of risky behaviours by intervention status, sex, time of day students were observed, or days since the students watched the video (Table 2). Regardless of the intervention status, snowboarders engaged in a significantly higher rate of risky behaviours (IRR: 1.94; 95% CI: 1.28–2.95) than skiers. However, among students who watched the video less than 8 days before their outing, those in intervention group engaged in significantly more risky behaviours (IRR: 7.00; 95% CI: 1.19–282.98) than those who had watched the control video (Table 3).

Table 2.

Number of outcomes, exposure, rate, and rate ratios by covariate category (i.e., sex, activity, time period, when participants watched the video)

95% CI	Number of risky behaviours seen	Total runs seen	Rate (per 100 runs	Rate ratio (95% CI)
Intervention status
Control	38	163	23.31)	1	--
Intervention	56	244	22.95	0.99 (0.64–1.52)
Sex
Female	27	125	21.60	1
Male	50	176	28.41	1.32 (0.82–2.10)
Unknown	17	106	16.04	0.74 (0.40–1.36)
Activity
Ski	35	218	16.06	1
Snowboard	59	189	31.22	1.94 (1.28–2.95)
Time period
Morning	81	324	25.00	1
Afternoon	13	83	15.66	0.63 (0.35–1.13)
How many days after watching the video was the group on a ski hill
Within 7 days	43	203	21.18	1
More than 7 days	51	204	25.00	1.18 (0.79–1.77)

Table 3.

Number of outcomes, exposure, rate, and rate ratios for the intervention group compared the control group by sex, activity, days since watching the video

	Number of risky behaviours (Numerator)	Number of exposures (denominator)	Rate	Rate ratio (95% CI)
Control: male	17	72	23.61	1
Intervention: male	33	104	31.73	1.34 (0.73–2.57)
Control: female	11	58	18.97	1
Intervention: female	16	67	23.88	1.26 (0.55–3.00)
Control: unknown	10	33	30.30	1
Intervention: unknown	7	73	9.59	0.32 (0.10–0.92)
Control: ski	24	111	21.62	1
Intervention: ski	11	107	10.28	0.47 (0.21–1.01)
Control: snowboard	14	52	26.92	1
Intervention: snowboard	45	137	32.85	1.22 (0.66–2.41)
Control: <8 days	1	29	3.45	1
Intervention: <8 days	42	174	24.14	7.00 (1.19–282.98)
Control: ≥8 days	37	134	27.61	1
Intervention: ≥8 days	14	70	20.00	0.72 (0.36–1.37)

DISCUSSION

This study determined the incidence of risky behaviours using a standardized tool among novice skier and snowboarder youth and examined whether these behaviours were mitigated by a video intervention. The rates were similar regardless of intervention status. There was a significant difference in rates of risky behaviours between the intervention and control group based on when the students watched the video. Among those who watched the video less than 8 days prior to their ski hill outing, the intervention group exhibited 7.00 times more risky behaviours than those in the control group. However, only one student was observed performing a risky behaviour in the control group, which resulted in an imprecise estimate and the inability to control for potential confounders or clustering by school. This finding should be interpreted cautiously.

Previous research identified younger age, first time participants, higher skiing ability, males, and lower body mass index as predictors of risky behaviour while skiing and snowboarding (3). First time participants, beginners, children and adolescents, males, those with improper ski binding adjustment, using rented equipment, and snowboarding have been associated with injury (12,15,22–26). We found that snowboarders engaged in a risky behaviour at almost twice the rate as skiers. Sulheim et al. found a twofold greater rate of injury in snowboarders than skiers (15). Additional work is needed to determine if performing risky behaviours results in an increased risk of injury.

The majority of snow sports injuries are perceived to be related to poor decisions, actions, and human error (5). Understanding risky behaviours is needed to optimize promotion of safe behaviours, prevent injuries, and properly incorporate injury prevention strategies to reduce injury risk (4). This is especially true for adolescents and young adults, as they typically take the most risks (27, 28). Previous studies that have assessed risky behaviour in sport have mainly used self-reported information, which may be unreliable (3, 29). A strength of this study was using a standardized assessment tool to evaluate the incidence of risky behaviours.

Although a previous video intervention reduced ski and snowboard injuries (30), this intervention video did not appear to reduce the risk of performing risky behaviours that could lead to injury. Children and adolescents have a variety of learning strategies (31). The video did improve immediate knowledge (32), arguably a minimum prerequisite for behaviour change and injury prevention. Perhaps not all students were able to retain or implement knowledge gained from the video. Peer-pressure and social deviance may also play a role. Students may have acted in a riskier manner in an attempt to impress their classmates, even though the intervention video advised against this (29). Finally, students were skiing and snowboarding within a controlled, supervised environment; this may have resulted in students engaging in less risky behaviours regardless of intervention status.

There were several limitations. First, students may have been absent from school on the day the assigned video was shown, or they may not have paid attention. It was not possible to directly approach students on the hill and ask if they had watched their assigned video. In the future, feedback should be sought to ensure that students watched their video and to also find ways to improve the video content to reduce risky behaviours. Second, RAs were not always able to capture all unsafe behaviours during the designated observation period. Only one child was observed at a time. While the risky behaviours and risk factors listed on the RBAAT were identified using multiple sources (17–20), some of the observed behaviours that were not listed might have been missed. However, the form has an “other” section where additional risky behaviours could be recorded. Third, the yellow armbands on the children were more difficult to view than anticipated. Better ways to identify students would have potentially increased the number of observations. Also, the verification processes were unexpectedly time consuming, leading to a small sample size.

This study focussed on elementary students within a controlled, beginner lesson environment. The incidence rate of risky behaviour may be different for unsupervised skiers and snowboarders outside the school program lessons, older children, or more experienced skiers and snowboarders. Future studies should explore the generalizability of this tool to other populations.

In conclusion, similar rates of risky behaviours were reported in the intervention and control groups. The most common types of risky behaviours engaged in by students in the school programs were skiing too close to other skiers/snowboarders and near collision with an object or person. Overall, higher risky behaviour rates were observed in snowboarders compared with skiers. Future research should focus on determining whether the intervention video can reduce injuries and the rates of risky behaviour within other populations.

FUNDING

This project was funded by Alberta Innovates Health Solutions Collaborative Research and Innovation Opportunities program grant for the Alberta Program in Youth Sport & Recreational Injury Prevention.

POTENTIAL CONFLICTS OF INTEREST

Unrelated to the manuscript, BH’s institution received funding from CIHR, the NFL Scientific Advisory Board (SHRED Concussions Funding), the city of Calgary, Maternal Newborn Child & Youth Strategic Clinical Network Health Outcomes Improvement Fund, and the International Olympic Committee: Research Centres for Prevention of Injury and Protection of Athlete Health. BH’s institution also received the BMO Endowed Research Award in Healthy Living in the Branch: Child Health and Wellness Grand Challenge Catalyst competition, Bone and Joint Health Strategic Clinical Network Targeted Seed Grant Competition, Department of Paediatrics Innovation Award, Alberta Children’s Hospital Research Institute Healthy Outcomes Collaborative Research and Innovation Grant, Calgary Centre for Clinical Research, Clinical Research Fund Seed Grant. Unrelated to the manuscript, CE’s institution received funding from Canadian Institutes of Health Research, National Football League Scientific Advisory Board, International Olympic Committee Medical and Scientific Research Fund, World Rugby, University of Calgary. CE also received a travel stipend from Publi Creations to attend a meeting and is an Associate Editor (unpaid) for the journal British Journal of Sports Medicine (BJSM). There are no other author disclosures.

CONTRIBUTORS

BH, CE, ANA, KR, JC, MD, and TM conceptualized and designed the study. MD coordinated the project and played a vital role in the students viewing the videos. TM and SM collected observational data at the ski hill. TM and ANA conducted data analysis. TM drafted the manuscript. All authors contributed to the interpretation of the results, critically revised the manuscript, and gave their approval of the final version.