Intervention with an educational video after a whiplash trauma – a randomised controlled clinical trial

2.1 Study design and settings

This was a controlled, randomised clinical trial. The study population consisted of acute patients seeking care for neck pain after a motor vehicle collision (MVC). The patients all presented at the ED at Södersjukhuset, a teaching hospital in Stockholm, Sweden. The hospital has well over 110,000 emergency visits yearly and a catchment area of over 700,000 inhabitants.

The primary outcome measure was self-reported non-recovery rate, defined as a “no” answer to the question “Do you feel recovered from the neck injury after the MVC?”. Secondary outcomes were neck pain evaluated on a numerical rating scale (NRS) and disability according to a modified version (12 questions) of the whiplash disability questionnaire (WDQ) [23]. Follow-up time was set at 6 months after the ED visit following the MVC. The two arms of the study were an educational video (intervention group) and a standard information sheet (control group). Inclusion criteria were patients with neck pain after a motor vehicle accident, age of 16–65 years, WAD grade I–III and reported continued neck pain when contacted 10 days after their ED visit. The exclusion criteria were fractures or hospitalisation due to the accident, lack of fluency in the Swedish language or other factors limiting comprehension, patients seeking care more than 2 weeks after the trauma, participants in another scientific study or non-Swedish residents (Fig. 1).

Fig. 1:

Flow chart of the enrollment. *=followed up with questionnaire (n=62). Followed up with telephone interview (n=31). **=followed up with questionnaire (n=63). Followed up with telephone interview (n=40). ITT=intention to treat, PP=per protocol, NRS=numeric rating scale, WDQ=whiplash disabilty scale.

2.2 Enrolment and randomisation

Research nurses reviewed the daily ED logs for the period December 2015–December 2017 with disruptions from June to August 2016 and 2017. Patients who were admitted to the ED for neck pain after a MVC were considered eligible for the study and were contacted by telephone within 10 days after their ED visit. During this first contact, inclusion and exclusion criteria were checked and the patients were informed orally about the study. Patients agreeing to participate were enrolled, and a first interview was conducted. The interview included questions about earlier MVCs, the presence of neck pain before the current accident, and coexisting fractures or injuries requiring in-house hospital care. The patients were randomised to either the intervention group or the control group with a 1:1 allocation. The sequence generation was performed by two study nurses using manually generated blocks of twenty identical envelopes, 10 including a double folded note indicating educational film and 10 with one note indicating written information. The same two study nurses performed allocation concealment by sealing the opaque envelopes and then mixing them thoroughly. The envelopes were then sequentially numbered. Implementation was assigned after successively opening a sealed and numbered envelope and linking the intervention to the concerned participant’s number. The participants then received the intervention or control material on a link via e-mail, along with written information about the study and the baseline questionnaire. Written consent for participation obtained from all included patients.

2.3 Study intervention

The educational video was inspired by the analogue video used by Brison et al. [18], based on pain mechanisms and reassurance. The 15-min video is a professional production and presented in Swedish by a well-known Swedish TV-host, who interviews well-regarded and experienced medical personnel: an orthopaedic surgeon, a physiotherapist and a psychologist. The video explains the underlying pathophysiology of whiplash injury and presents recommendations regarding the benefits of staying active and of participating in physical exercise, with a focus on deep cervical flexor training [24]. It also includes information about possible psychological reactions and it explains the body’s response to pain through animated graphics (Supplementary Appendix 1).

2.4 Control group

The information sheet displays six simple exercises for staying active and encouraging motion of the cervical spine and musculoskeletal system. It is the same sheet that the local ED staff normally hands out to patients with neck pain after a WAD. Similar pamphlets have previously reported not to affect the outcome after whiplash injury when compared to usual care (i.e. care without standardised patient information) [25] (Supplementary Appendix 2).

2.5 Data collection

Along with the educational video or information sheet, all patients received a baseline questionnaire with questions about socioeconomic background, previous health, sick leave, the modified WDQ and WAD-associated symptoms, such as pain level and mental distress (according to a NRS). Levels of pain and mental distress during the last 3 days were measured on a 10-grade NRS, where zero represented no notable pain/distress and 10 represented worst imaginable pain/distress. The questionnaire was estimated to require 20 min to fill in. Patients not responding to the first questionnaire received two reminding e-mails and, if needed, a telephone reminder from the study nurses.

Six months after the first questionnaire, a second similar questionnaire was sent to all patients. This follow-up also included questions regarding their participation in the treatment, as well as the binary question, “In total, do you feel recovered from the neck-related problems that you experienced after the accident?” to be answered with a “yes” or “no” answer. The same procedure used for reminders for the baseline questionnaire was repeated for the second questionnaire. Non-repliers to this follow-up questionnaire were contacted for a brief telephone interview and asked questions about the intervention, non-recovery and current pain level.

2.6 Statistical analysis

The level of pain/distress was categorised into three groups, analogously with previous research [26]. Fischer’s exact test was used to evaluate the difference in the primary outcome of non-recovery between the two intervention groups. For comparison of baseline characteristics and analyses of secondary outcomes, the independent samples t-test and Fischer’s exact test were performed to study differences between the groups. A p-value of <0.05 was considered statistically significant. We also used logistic binary regression to study the association between the primary outcome (non-recovery) and the intervention. First, a crude analysis was conducted to compare non-recovery and the intervention, and other factors one at a time. Second, to adjust the association between non-recovery and the intervention for differences between the intervention groups, we estimated a model with the eight factors shown in Supplementary Table 1. These factors were the intervention and a further seven factors that had been reported earlier to be associated with poor outcome after whiplash trauma [8], [11]. The odds ratios (OR) and corresponding 95% confidence intervals (95% CI) are also shown.

All statistical analyses were conducted using IBM SPSS Statistics, version 23 (SPSS Inc., Chicago, IL, USA). The analyses were performed according to both the Per Protocol (PP) and the Intention To Treat (ITT) principles. Since the results did not differ, we only present the results based on the ITT principle.

2.7 Power analysis

We performed a priori power calculations using Sample Power version 3. A total of 206 patients (103 in each arm) was needed to detect an absolute clinically relevant difference of 20% (60% vs. 40%) in the non-recovery rate between the two arms of the study, with a power of 80% and a two-sided significance level of 5% after 6 months. In addition, we assumed that 25% of the patients would be lost to follow-up, so the recruitment goal was set to 280 (140 in each arm).

2.8 Ethics

The study was approved by the Swedish Ethical Review Authority’s review board in Stockholm, Sweden (decision 2010/1269-32) and conducted in accordance with the Helsinki II declaration.

4.1 Strengths and limitations

This study has several important strengths. The included patients all attended a single-centre ED, so they represented a homogenous group. We randomised the subjects and their assignment to the intervention and the primary treating physicians were blinded to the study intervention. The primary outcome measure was self-reported non-recovery, since we preferred a person-centred outcome as opposed to objective scales and found this useful as a simple and confirming way to measure recovery from musculoskeletal injuries. Self-reported recovery has previously been suggested to be a good reflector of other validated indices of recovery outcome in WAD [31], [32].

One limitation of using an educational video as the intervention is the difficulty in verifying compliance. The patients in this study reported at the follow-up whether they had faithfully participated in the study by viewing the information as required. However, we do not know how many subjects had actually watched the video (or even if they had watched it). We also could not know to what extent the patients followed the recommendations provided in the video. Even if the frequency of self-reported participation in the intervention (74%) is true, this compliance bias undermines the results.

Another important limitation is the higher than expected dropout ratio, resulting in a final analysis of 68% of the patients out of all that had been randomised (196/289). A secondary catchment mechanism by the study nurses had been planned prior to the initiation of the study, but this could not ensure an adequate number of patients based on our preceding power calculation. Concurrently, we were only able to analyse 93 out of the required 103 patients allocated to the written information.

We had a 70% response rate to the baseline questionnaire (206/289), but only a 43% response rate to the complete follow-up questionnaire (125/289). We therefore chose to contact the non-responders to the second questionnaire and to perform a shorter telephone-based second questionnaire to reach a 68% response rate in total (196/289). The high rate of early dropout (i.e. dropout before the baseline questionnaire) could possibly have led to a selection bias. Another possibility is that the subjects who were already recovered did not respond to the first questionnaire, although this would probably be independent of randomisation.

We have no reports of potential crossovers between the groups. Since the written information pamphlet used as a control in this study previously has been a part of the standard care at the ED, presumably some of the patients allocated to the interventional group received the same material. The risk is probably lower for contrary crossover, since the video had not previously been distributed. Additionally, a considerable risk existed for patients receiving patient information from third-party sources. The numerous internet sites with information about whiplash injuries are steadily increasing.

The decision to use the information pamphlet as the control was based on the “treatment as usual”-principle. It could be debated if a third arm of the trial with no information would have been beneficial.

A presumable selection bias exists within this material. Only patients with persisting neck pain at the inclusion interview, which was conducted within 10 days of the MVC, were included. Those patients with the most rapid recovery were, accordingly, not included in the study.

The modified version of WDQ is not a validated outcome measure. The reason for leaving out one question (“Do your whiplash symptoms interfere with your non-sporting leisure activity?”) was that it was considered vague in its constitution and was likely to be misinterpreted by the patients. The original WDQ score has a maximum sum of 130, while the modified score has a maximum sum of 120.

The follow-up time of 6 months is rather short. However, the highest frequency of recovery has been reported to occur during the first 3 months; thereafter, the non-recovery rates stay unchanged to a large extent [33]. Furthermore, previous similar studies have used the same follow-up time [18], [19].

The lack of a standardised definition of recovery means that the interpretation of results in whiplash studies is difficult [31]. Additionally, there is no consensus regarding the most accurate outcome measure for WAD [34]. We used the question “In total, do you feel recovered from the neck-related problems that you experienced after the accident?”. This subjective measure has the disadvantage of not being robust for observer bias. However, we believe that our question represents a reliable clinical assessment of recovery.

In summary, we consider that scientific support is low for recommending educational video intervention over regular information for patients with acute WAD.