Rethinking controlled trials for evaluating wheelchair skills training in children

Abstract

Wheelchair skills training is one of four wheelchair service steps recommended by the World Health Organization (https://www.who.int/publications/i/item/9789240074521). The Wheelchair Skills Program is the most cited approach and was named an international gold-standard (https://wheelchairskillsprogram.ca/en/skills-manual-forms/). Accessed by 197 488 users in 200 countries, there is high-level evidence from two meta-analyses supporting effectiveness of the Wheelchair Skills Program for improving wheelchair skills.^{1, 2} Although most studies have focussed on manual wheelchair use in adults, there is growing evidence for power wheelchair skills training and applications in children.

The small, well-conducted pre-post study by Naaris et al. supports wheelchair skills training for improving power wheelchair skills and satisfaction with participation in children with cerebral palsy.³ Historically, the results of small pre-post studies have been undervalued compared to larger randomized controlled trials (RCTs). However, findings from the current study (and similar studies) provide important contributions that should not be overlooked.

Although authors should caution readers when interpretating the results and should not use the term ‘effectiveness’ without a controlled trial, selecting appropriate terminology can highlight within-subject improvements that are arguably more important than statistically significant between-group differences. Attainment of individual goals and successful social participation are at the heart of rehabilitation. In contradistinction to RCTs, ensuring goal attainment and participation in meaningful activities requires personalized approaches to intervention. For example, Naaris et al. mentioned that the delivery of the wheelchair skills training program was standardized, yet each session was personalized according to individual needs and goals as recommended by the Wheelchair Skills Program. If meaningful goals are not targeted, there is a risk of low motivation to learn and unsustained use.⁴

Statistically significant between-group differences do not outweigh the clinical significance of wheelchair skills training. Improving a child's score by a single level (e.g. from 1 to 2 on the Wheelchair Skills Test [WST]) by becoming independent on just one wheelchair skill (i.e. an improvement of only 1.1% on the total percentage WST score) may facilitate getting out of the house to play with friends.

In supporting the need for study designs other than RCTs, authors should provide explicit details when reporting small pre-post trials. Specific to wheelchair skills training, this may include training location, skills practiced, frequency and duration of sessions, individual adaptions or ‘tips and tricks’ used, and involvement of families or caregivers. Moreover, authors should select outcomes that cover various constructs that are important to wheelchair use (e.g. satisfaction, as documented by Naaris et al.).

It is also useful to provide both absolute and relative scores to assist with the interpretation of findings. For instance, a child with a total percentage WST score of 40% pre-training and 50% post-training will have had an absolute improvement of 10% (that is impressive on its own, as noted above), but the relative increase (post-training/pre-training x 100%) of 25% is of a magnitude to capture the attention of policymakers.

Personalized interventions that consider personal, environmental, and social factors are important for achieving social participation. Wheelchair skills are fundamental to wheelchair mobility and social participation. As Naaris et al. concluded, customized power mobility skills training completed in natural environments could be recommended and, we would add, should be recommended. Respecting internal and external validity is difficult in rehabilitation sciences and is nearly impossible with children. With limited resources and budgets, we need to encourage the evaluation of promising interventions using alternatives to RCTs, such as case studies, single-subject designs, and qualitative and participatory research approaches. Such approaches will facilitate deeper exploration of individual nuances.