2.1

Participants and establishments

Children were recruited from 2 paediatric rehabilitation centres with a total of 37 inpatient beds and 18-day hospital places. The centres admitted and followed children and teenagers with varied neurological and orthopaedic pathologies, including both transitory and permanent motor disability. The teams of both rehabilitation centres were aware of pain assessment and management procedures before the beginning of the study (e.g., through undergraduate training, transdisciplinary post-graduate training, visual reminders posted in the nurse’s station and rehabilitation rooms, pain committees integrated in the centres, auditing of professional practices and identification of pain).

Participants were randomly drawn from children admitted to either centre for at least 1 week (as in- or outpatients) with transitory or permanent motor handicap. To facilitate and ensure the feasibility of the study, a maximum of 3 children was simultaneously included per centre. Since the frequency of care-induced pain was not known, before beginning the study, we estimated including an arbitrary number of at least 30 children.

2.2

Procedure

Data were collected over 5 consecutive days for both in- and outpatients and during 1 night for inpatients by using a case report form that the child carried everywhere in the department. All care staffs were trained in data collection and pain evaluation and were told which children had been included. The aim was to evaluate all care activities that required physical contact with the child.

2.3

Data collection and measures

The case report form was used to collect relevant details regarding the child, the nature of the care activity and the pain assessment. Demographic data included age, gender and the type of admission (in- or outpatient). Clinical and medical data included the pathology, impairments of the upper or lower limbs, the type of disability (transitory or permanent) and the level of dependence. Level of dependence was measured by using the Programme de médicalisation des systèmes d’information (PMSI) dependence scale (French hospital activity database), which is based on Diagnosis Related Groups . This scale provides a broad estimation of autonomy for 6 different activities: washing, dressing, feeding, continence, behaviour and communication. Each activity is rated as 1, independent, 2, needs supervision, 3, needs partial assistance or 4, needs total assistance . This score must be considered purely indicative because the scale has not been validated for use in this population.

Characteristics of care activities ( Appendix A ) included the day and time, the type of activity, the duration in minutes, the person(s) involved (with symbols; e.g., PT for physiotherapist) and any specific measures used to prevent pain during the activity from a predefined list (none, paracetamol, morphine, etc.). The type of care activity was selected from a closed list of potentially painful activities developed during 3 formal meetings and intermediary exchanges involving professionals from all domains (e.g., doctors, nurses, auxiliaries, child-care assistants, physiotherapists, occupational therapists). The list was representative and exhaustive for all care activities carried out in the centres. It contained 30 activities involving nursing care (e.g., samples, dressings, mouth care), daily living care (washing and dressing), physiotherapy (e.g., mobilisation, walking), occupational therapy (e.g., orthoses), speech and language therapy and adapted physical activity ( Appendix A ). An activity not on the list could be reported but had to be detailed.

A proxy measure of pain was recorded by the professional involved, who used a behavioural pain scale, the French version of the Face, Legs, Activity, Cry, Consolability-revised scale (FLACC-r) and a visual analog scale (VAS caregiver). The FLACC-r has been validated in children who cannot communicate pain orally . It is based on 5 items: facial expression, leg movements, activities, vocal expression and the consolable nature of the person. Each item is rated from 0 to 2. A total score ≥ 4 was considered the experience of pain and 1–3 discomfort. The scale is well correlated with the VAS, the gold standard for evaluating pain in children who can communicate . It is reliable and easy to use for all caregivers . The VAS involves evaluating pain intensity on a 100-mm scale by the participant placing a cursor along its length (0 = no pain, 100 = maximal pain). Many studies of children have shown that this scale is valid as a proxy measure recorded by an experienced person such as a caregiver . The FLACC-r and the VAS are complementary and provide an indication of the intensity of pain experienced by the child in the most exhaustive and objective manner possible.

Self-reported pain was assessed in children > 6 years old if they had sufficient cognitive and verbal capacity to use a VAS (VAS patient) as previously described. This scale is considered the gold standard in participants > 5 or 6 years old. Younger children tend to use only the extremities of the scale .

2.4

Statistical analysis

Data are described with mean ± SD or median (interquartile range). We analyzed the number of times an activity caused pain or discomfort among the total number of evaluations, the number of times an activity caused pain or discomfort over the number of times this activity was performed (the “rate of pain”) and the pain intensity. Validity of the assessment procedure was evaluated by using Statistica v6.0 (StatSoft, Tulsa, OK, USA). Pearson’s correlation coefficient was used to evaluate the concurrent validity between the FLACC-r and the VAS caregiver and VAS patient. The level of significance was fixed at P < 0.05.