Design and sample
Cross-sectional data from two related studies using identical outcomes were combined. Firstly, the cross sectional Health in Adapted Youth Sports (HAYS) study that includes children and adolescents aged 10–19 years old with a chronic disease or physical disability (CDPD) [31]. Secondly, the baseline data from the Sports-2-Stay Fit (S2SF) study which is a clinically controlled trial among children and adolescents aged 6–19 years with a CDPD [32].
For both studies, children and adolescents were recruited through patient organisations, paediatric physical therapy practices, Wilhelmina Children’s Hospital in Utrecht, De Hoogstraat Rehabilitation Center, Fitkids and schools (for special education) in the Netherlands.
Inclusion criteria were having a physical disability or a chronic disease (cardiovascular, pulmonary, musculoskeletal or neuromuscular disorder), aged between 10 and 19 years (HAYS) or between 6 and 19 years (S2SF), ability to understand simple instructions, able to perform physical fitness tests. Children and adolescents were not eligible for participation in these studies if they had a progressive disease, used an electric wheelchair, participated in other (research) projects that may influence the results of the current studies, or did not sign the informed consent form.
For the current study, children and adolescents were included in the analyses if they had valid data on sports participation and valid data on at least one of the outcome variables (quality of life, self-efficacy, self-perceptions, general self-worth). Eight children and adolescents from the S2SF-study did not complete the online questionnaires during their first assessment, but did during the second assessment (8 weeks later). For that reason, we included data from these eight children or adolescents from their second assessment. We assumed that even though their scores may have been improved in that 8-week period, this will not affect the association between sports participation and these variables.
In total, 197 participants had complete data on sports participation, of whom two had no valid data on the outcome variables. Of the remaining 195 participants, 177 participants had complete data on all outcome variables, another 13 participants had valid data on two of the three outcome variables and 5 participants had valid data on only one outcome variables. The distribution of valid data on the outcome variables did not differ by sport participation status (p = 0.428) or by diagnosis (ACSM category) (p = 0.346) or by gender (p = 0.393). We included 195 children and adolescents who had complete data on sports participation and at least one outcome variable. Of those included children and adolescents, 145 participated in the HAYS study, while 50 participated in the S2SF study.
Procedures
The procedures and protocols for the HAYS and the S2SF study have been described elsewhere in more detail [31, 32]. Briefly, participants who agreed to participate and met the inclusion criteria were scheduled for an assessment at the lab (University of Applied Science, Utrecht). 1 week before this visit to the laboratory, the participants or their parents received a secured link to online questionnaires assessing exercise self-efficacy and quality of life. These questionnaires took about 10–15 min to complete. When the children visited the laboratory, they were first asked to complete the questionnaires on self-perceptions and global self-worth in the presence of one of the researchers or research assistants, which took about 10 min. When they finished the questionnaires their physical fitness, cognition and cardiovascular health was assessed, which took on average about 2 h [31, 32].
The studies were approved by the Medical Ethics Committee of the University Medical Center Utrecht, the Netherlands. (METC number: 14-332/c and 14-118/m). All participants and the parents of participants under 18 years of age provided their informed written consent. Studies were conducted in accordance with the Helsinki Declaration.
Measurements
Independent variable
Sports participation
Sports participation was assessed by means of a questionnaire that was completed before the start of the physical fitness tests. Sports participation was identified using three standardised questions used by the National Institute for Public Health and the Environment (RIVM) [33] (1) “do you participate in organized sports?” (2) “what is/are the type of organized sport(s)?” and (3) “what is the frequency of participation in organized sports per week?”. When participants indicated that they participate in organised sports at least two times per week, they were classified as ‘participating in sports’, all others were classified as ‘non-sporting’. This cut-off was based on the Consensus statement for the Dutch Guidelines for Physical Activity for youth (< 18 years old) [34]. This guideline was developed for typically developing children and was in place at the time of the start of the current research project. It advises that children and adolescents should be physically active for at least 1 h per day at at least a moderate intensity. In addition, children and adolescents should engage in activities that specifically address physical fitness at least two times a week. Sport activities are typically activities in which physical fitness, including strength, flexibility and coordination, are addressed. However, there are no universally accepted guidelines for children and adolescents with a chronic disease of physical disability and Van Brussel and colleagues [35] advise to use a training frequency of minimally two times per week. This is in line with recommendations for people with cerebral palsy that were based on a comprehensive literature review, expert opinions and extensive clinical experience [36].
Dependent variables
Health-related quality of life
To evaluate the quality of life, the Dutch version of the Disabkids (DCGM-37) was used, which was completed online by either the participant alone (n = 45), together with one of the parents/caregivers (n = 54) or the parent alone (n = 9) (unknown for n = 89). This questionnaire measures the quality of life and the independence of children and adolescents with chronic health conditions. The questionnaire has been used in other paediatric populations and has been tested for internal consistency and validity [37,38,39] The questionnaire includes 37 items that cover six subscales, i.e. mental independence, mental emotion (inner strength), social inclusion, social exclusion (social equality), physical limitations (physical ability) and physical treatment (i.e. the impact of taking medication, receiving injections, etc.) [38]. All items were scored on a 5-point Likert scale ranging from 1 = never to 5 = always. For each scale, a sum score was calculated by following the instructions of the developers of the DCGM-37 and the provided syntax file [38]. If one item was missing, this missing value was substituted by the mean of the non-missing values. If more than 1 item of a domain was missing, no sum score was calculated. These scores were transformed so that they had a range from 0 to 100 with higher scores indicating a higher perceived quality of life.
Self-perception and global self-worth
To evaluate self-perception, the Dutch translations of the self-perception profile for children (SPPC) and for adolescents (SPPA) were used [40]. We used the children’s version for children of 12 years old and younger, or with cognitive capabilities of this age group, and the adolescent version for those older than 12 years. Based on the researcher’s expertise, an older child was provided the children’s version or when the child clearly was not able to complete the adolescent version or did not understand the questions. The children’s version uses 36 items to measure five domains, i.e. ‘scholastic competence’, ‘social acceptance’, ‘athletic competence’, ‘physical appearance’, ‘Behavioural conduct’, and the general concept ‘global self-worth’ [41, 42]. All six scales consist of 6 items that include bipolar statements (e.g. ‘some kids feel they are very good at school, but other kids worry about whether they do the schoolwork assigned to them’). The children have to choose which of the two statements resembles them, and how much, i.e. ‘sort of true for me’ or ‘really true for me’. All items were scored on a 4-point scale and sum scores were calculated. The questionnaire has been tested in a Dutch norm group and showed reasonable to good internal consistency and test-retest reliability [42].
The adolescent version includes 35 items which make up 7 scales, i.e. ‘scholastic competence’, ‘social acceptance’, ‘athletic competence’, ‘physical appearance’, ‘behavioural conduct’, ‘close friendship’ and the general concept ‘global self-worth’. Similar to the child version of the questionnaire, all items included bipolar statements that were scored on a 4-point scale by which the adolescents indicated which statement resembled them ‘sort of true’ or ‘really true’ [43].
If one item of the scale was missing, the scale score was not calculated (n = 3). All scores on each scale were compared to Dutch norm scores to indicate whether the participants scored ‘below average’, i.e. below the 15th percentile, or above average, i.e. at or higher than the 85th percentile. For the child version, different norms have been formulated for boys and girls [41]. For the adolescent version, different norms were formulated by school level for the scales ‘scholastic competence’ and ‘behavioural conduct’ and a distinction by sex was made for all scales except ‘social acceptance’ [43]. In the current study, results regarding these norm scores were only used for descriptive purposes.
Exercise self-efficacy
To assess whether sport participation was associated with exercise self-efficacy, the Dutch version of the Exercise Self-Efficacy Scale [44, 45] was filled out digitally at home by the child. Self-efficacy is a well-known behavioural determinant and described in several behavioural theories such as the Social Cognitive Theory by Bandura [46] in which self-efficacy is seen as important influencer of behaviour. In the current study, we hypothesise that behaviour, in this study engaging in sport, can have a positive impact on self-efficacy in relation to physical activity and exercise. This hypothesis is based on the fact that mastery experiences and vicarious experiences are important sources of self-efficacy [47] and when participating in sport, children may learn new skills and may see others like themselves performing specific tasks or behaviours. The questionnaire takes approximately 2 min to complete. The questionnaire consists of 10 items about the level of self-confidence with regard to performing regular physical activities and exercise that could be rated on a 4-point Likert scale ranging from ‘not true at all’ to ‘always’. A sample item is “I am confident that I can overcome barriers and challenges with regard to physical activity and exercise if I try hard enough”. A sum score was calculated when all 10 items were answered. Internal consistency of the scale was high, Cronbach’s alpha = 0.99. A higher score indicates a higher exercise self-efficacy. Validity and reliability of the scale have been tested in a sample with spinal cord injury and showed good validity and adequate reliability [44].
Potential confounders and co-variates
Demographic variables
The demographic questionnaire that included questions about date of birth, sex and school level.
Medical diagnosis
The general questionnaire assessed the medical diagnosis. The medical diagnoses were further classified in categories according to The American College of Sports Medicine (ACSM) [48]: 1. Cardiovascular (e.g. ventricular septal defect, tetralogy of Fallot, cardiomyopathy), 2. Pulmonary (asthma), 3. Metabolic (diabetes), 4. Musculoskeletal or orthopaedic (amputation, clubfoot, hereditary multiple exostoses-multiple osteochondromas, congenital anomalies), 5. Neuromuscular (cerebral palsy, spina bifida, neurofibromatosis, Kabubi syndrome, centronuclear myopathy, psychomotor retardation, Martsolf syndrome, acquired brain injury), 6. Immunological or haematological (rheumatism, Fanconi anaemia), 7. Cancer (tumour in hypophysis), 8. Epilepsy. In the regression analyses, categories were merged distinguishing disabilities or disease related to the metabolic system or oxygen transport (1. Cardiovascular, 2 Pulmonary, and 3. Metabolic) from the other disabilities or diseases, mostly related to the musculoskeletal or neurological system.
School type
Children participating in the current study attended either regular education or special education for children with physical disabilities (so called mytyl schools). These schools, dedicated to children with CDPD, have similar learning objectives as regular schools, but the children receive additional attention and support. Children attending special education may have different references for social comparison which may influence how they perceive themselves. Therefore, school type was added as a potential confounder in the analyses.
Statistical analyses
Continuous data was described by means or medians and standard deviations or interquartile ranges and categorical data by frequencies and proportions. Crude comparisons on the outcome variables between the sporting and non-sporting participants were made using ANOVA. Adjusted associations between sports participation and the outcome variables were estimated by linear (or logistic) regression analyses. Assumptions were checked, and residuals showed an acceptable normal distribution, except for the Exercise Self-Efficacy scale. Therefore, the scale was dichotomized at the third tertile (score ≥ 37) and associations with sports participation were assessed by means of logistic regression analyses. Sensitivity analyses were run with and without outliers, i.e. those with standardised residual scores below − 3 or higher than 3.
Different models will be presented, unadjusted as well as adjusted for potential confounders’ sex, age, school type, diagnosis or ACSM category. All analyses were checked for potential interaction by age. In case of significant interactions, results will be presented for two age groups. For the analyses, the ACSM categories were merged into two categories, i.e. one including those with cardiovascular, pulmonary and metabolic conditions, and another including those with musculoskeletal, neuromuscular, immunological, cancer or epileptic conditions. Results are reported as regression coefficients or odds rations and their 95% confidence intervals.
