Background
Musculoskeletal symptoms are prevalent in the adolescent population and often have significant impact on their future musculoskeletal health. One study identified musculoskeletal pain as the second most reported physical symptom after headaches, and up to 7% of adolescents report this type of symptom often or on a daily basis [
1]. In Ontario (Canada) alone, 380 000 adolescents and pre-adolescents consulted a health provider for musculoskeletal disorders over the course of a year, which represents a consultation rate of 122 visits per 1000 youths [
2]. Low back pain, more precisely, is one of the most prevalent complaints in the adolescent population. Studies found annual prevalence of low back pain varying between 20.5 and 50% during adolescence [
3‐
6]. Adolescents who have low back pain are also more likely to develop chronic low back pain as adults [
7‐
9]. The development of musculoskeletal disorders affecting the spine at a younger age is to be taken seriously considering the potential risk of developing chronic low back pain.
In the active adolescent population, sport and recreational injuries are also common. According to a United-States National medical care survey conducted over a year, the number of emergency department visits for the treatment of injuries due to physical activity and sports was estimated at 2.6 million for people aged between 5 and 24 years [
10]. In Canada, 27% of adolescents aged between 12 and 19 years old suffered from at least one injury in 2009, and 66% of these injuries occurred during physical or sporting activities [
11]. Sports injuries in the adolescent population are on the rise with the proportion of injuries due to sport and physical activity having increased by 5% between 2001 and 2009 [
11].
The survey of musculoskeletal symptoms in the adolescent population, be they related to a sport injury or not, is an important component in the detection and the prevention of musculoskeletal injury or pain and their related consequences. Furthermore, this type of screening tool is often used in the development and assessment of prevention strategies for work related pain and symptoms. [
12,
13]. Surveillance programs similar to those seen in work safety and ergonomics could be useful for school physical activity programs or for individual and team sports.
Epidemiological research on symptoms or injuries can be conducted using clinical records, nationwide surveys or questionnaires. Some of these methods, however, can underestimate the prevalence and incidence of symptoms. Surveillance through hospital records rarely accounts for symptoms or injuries that are treated by other health practitioners (chiropractors, physiotherapists, etc.), nor does it account for the minor disorders that remain untreated [
13‐
15]. Retrospective questionnaires are another method of estimating the actual prevalence of musculoskeletal symptoms, which offer a smaller risk of underestimating or even leaving out minor symptoms. Questionnaires are also a good method to obtain information from a large, and therefore more representative, population sample. Various questionnaires and assessment tools have been shown to be valid [
16,
17], reliable [
17,
18] and cost-efficient [
19] when collecting injury data in the youth population. However, to our knowledge, questionnaires measuring musculoskeletal symptoms that are also adapted to the adolescent population are not currently available.
In order for retrospective questionnaires to be valid and reliable, the recall period must be reasonably short, as recall bias limit data validity depending on the level of detail requested and the severity of the injuries [
20,
21].
Harel et al. [
21] assessed the recall capacities of parents when reporting their children or adolescents injuries over a period varying between 2 weeks to 12 months. The authors of this study concluded that severe injuries resulting in either hospitalisation or one full school day loss are less likely to be affected by recall bias due to memory decay [
21]. On the other hand, minor injuries are affected by memory decay, especially if the recall period exceeds 5 months [
21]. These studies, however, refer to injuries rather than symptoms, and were conducted in an adult population. It is to be noted that recall bias may be slightly different in adolescent populations and when musculoskeletal symptoms are assessed, rather than injuries.
Despite the numerous studies using questionnaires to collect epidemiological information, few validated musculoskeletal symptoms survey instruments exist. One of the commonly used tools is the Nordic Musculoskeletal Questionnaire (NMQ), a validated instrument that was originally developed to study the prevalence and impact of work related musculoskeletal symptoms [
22]. The NMQ, in its extended version (NMQ-E), measures the point, 12-month and lifetime prevalence of musculoskeletal symptoms [
23]. The NMQ-E also measures the severity of the symptoms by assessing the impact of the disorder on work and leisure activities. Questions on the treatment of the disorder such as hospitalisation are also used to estimate symptom severity [
13]. Finally, the NMQ-E is an easy to use, one-page questionnaire designed to obtain wide-ranging information on musculoskeletal symptoms over nine body regions in a short time frame. This questionnaire has, however, never been adapted to younger populations. An adapted version of this questionnaire would be an easy-to-use tool to survey symptom prevalence and severity, thus making it easier to identify and prevent musculoskeletal problems in the adolescent population.
Given the lack of validated musculoskeletal symptom survey instruments, the first objective of this study was to develop a musculoskeletal symptom screening tool for younger populations derived from the NMQ-E and NMQ French versions. The second objective of this study was to determine both the reliability and validity of this adapted version of the NMQ-E.
Discussion
The purpose of this study was to develop a musculoskeletal symptom screening tool for younger populations derived from the NMQ-E [
23] and other NMQ French versions [
13,
24], and to assess the reliability and the validity of this instrument. The final version of the study’s questionnaire includes three questions for each of the 9 body regions: the 6-month prevalence of musculoskeletal symptoms, the impact of these symptoms on school/work attendance as well as their impact on sport/leisure activities. The recall period chosen for the final version was 6 months because, as discussed earlier, a 12-month recall period is often affected by memory decay, and minor injuries are more likely to be forgotten [
20,
21].
The test-retest results with the 6-month recall period were encouraging and showed only slight fluctuations in responses. These fluctuations were presumably not due to changes in the participant’s health status, since the time interval between tests was short (28.1 ± 8.0 hours). It is however possible that the participants’ answers were slightly influenced by the clinician they consulted, since most participants completed one questionnaire before their consultation and the second one after their appointment. However, reliability results suggest that the questionnaire demonstrated a good overall stability of responses between tests with kappa values at moderate to perfect agreement beyond chance.
Dawson et al. assessed the reliability of the NMQ-E in an occupational cohort of 59 nursing students at a 24 hour interval [
23]. This study had similar values for Po and slightly lower kappa reliability results for their 12-month prevalence question with a Po = 0.83-1.00 and k ≥ 0.55. These differences could be attributed to the recall period being longer (12 months) than the one used in the present study (6 months).
The TNMQ-S obtained reliable results regarding musculoskeletal symptoms described as ache, pain or discomfort. These results are not surprising, since adolescent self-reported past pain or injury has been shown to be reliable.
Grimmer et al. 2000, found a very strong positive relationship between adolescents and their parents when reporting injuries occurring one week earlier [
17]. Likewise,
Sundblad et al. 2006 reported high child–parent agreement results over a recall period of 7 to 11 weeks, either when adolescents were in absence of pain or injuries, when pain or injuries were severe or when the adolescents’ complaints were frequent. However, according to the same authors, minor injuries or pain and less frequent complaints were under-reported by the adolescents’ parents [
18]. Another study found that adolescents rarely seek medical attention for their pain and injuries [
31], which means that medical records are likely to underestimate the prevalence of minor pain or injury in the adolescent population. These findings suggest that questioning adolescents to obtain past pain and injury data would be more accurate for detecting minor pain or injuries or less frequent complaints than parental reports or medical records.
In the present study, concomitant validity as a measure of the criterion validity was assessed by comparing the 6-month symptoms prevalence question to the participants’ clinical record. Most diagnosed problems were detected by the questionnaire, with observed agreement of Po = 0.71 for the complete concordance (i.e. including all diagnosed symptoms) and Po = 0.86 for the partial concordance. Kappa values obtained for the criterion validity indicated substantial agreement beyond chance, indicating that there was good agreement between questionnaire items and clinical records [
30]. However, some symptoms found in the questionnaire could not be linked to the clinical records. It is possible that the symptoms found in the questionnaires and not linked to the clinical records were minor problems undeclared to the clinician. In fact, when comparing diagnosed symptoms to symptoms not found in the clinical record, significantly more (P < 0.05) adolescents had reduced their physical or leisure activities due to symptoms associated with a diagnosis. Thus, it seems that adolescents only seek medical attention for pain and symptoms severe enough to have an impact on their physical activity. Indeed,
Watson et al. [
31] assessed the prevalence of back pain in schoolchildren of the United Kingdom, and found that adolescents were more likely to report pain of a greater intensity. The Teen Nordic Musculoskeletal Screening Questionnaire (TNMQ-S) was therefore useful to detect minor symptoms that adolescents did not necessarily report to their clinicians.
Concurrent validity for the impact of symptoms on school and physical activity questions could not be measured due to the lack of concordant information in the clinical records. Only a few clinical records had clear recommendations for sports and activity restrictions. Similarly, information regarding the number of school days lost could not be found in the clinical records.
Study limitations
As mentioned earlier, criterion validation is measured by comparing the responses to questionnaires to a recognised gold standard. Since no recognised gold standard questionnaire measuring musculoskeletal symptoms in the adolescent population was found, the University’s outpatient clinic’s clinical records were used as the comparison for the criterion validation. Even though these records are standardised, regularly audited and provide an adequate level of details regarding patient musculoskeletal symptoms, they cannot be considered as a gold standard. Future studies should assess the construct validity rather than the criteria validity when no gold standard questionnaire is available.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
EPL participated in the study design, data analysis, experimentation and manuscript writing. MD participated in study design, manuscript writing and revision. VC participated in study design, manuscript writing and revision. All authors read and approved the final manuscript.