Functional severity

The GMFCS has internationally established validity, reliability and stability for the classification and prediction of motor function of children with CP aged 2–12 years.50–52 It has an acceptable inter-rater and intrarater (test–retest) reliability (generalisability coefficients 0.93 and 0.68, respectively).51 Two physiotherapists, trained in the use of the GMFCS, independently observed and classified children in one of five functional categories.50

Classifications of gross motor abilities change with age and therefore separate descriptions were used for different age bands. In the present study, the 6–12 years descriptions from the extended and revised GMFCS (GMFCS-ER) will be  used.53 The GMFCS has been correlated with a number of motor scales, as well as CP motor type and distribution.54

Motor type and distribution

Motor type was classified as spastic, dystonic, ataxic, hypotonic, choreoathetosis, mixed CP or unclassifiable according to Surveillance of Cerebral Palsy in Europe guidelines.55 Distribution was classified by number of limbs impaired, unilateral and bilateral distribution (hemiplegia, diplegia, triplegia, quadriplegia) by at least two independent raters. The Dyskinesia Impairment Scale56 will be undertaken for those participants with a motor-type (primary or secondary) diagnosis of dystonia and/or choreoathetosis. This is an important assessment to measure the motor capacity and function of children with these particular motor types.57

Functional performance

The Functional Mobility Scale58 at 5 m (home), 50 m (school) and 500 m (community)58 will be used to evaluate functional performance. This is a valid and reliable measure of a child’s usual walking ability at three distances (5 m, 50 m and 500 m), representing their home, school and wider community, respectively.59

Gait pattern

Gait patterns will be classified according to the classification by Rodda and Graham,60 which demonstrated validity and reliability.61 Gait patterns for bilateral ambulant CP will be classified as either: i) true equinus, ii) jump knee, iii) apparent equinus or iv) crouch gait. For children with unilateral CP, gait patterns were classified according to the classification by Winters et al.48 This classification considered the sagittal plane joint movements: i) type I—foot drop during swing phase (apparent equinus); ii) type II—persistent ankle plantarflexion (true equinus); iii) type III—maintained plantar flexion through gait cycle plus limited knee flexion-extension and iv) type IV—similar to III, plus reduced hip flexion-extension. The classification by Winters et al had good inter-rater reliability using written reports (weighted kappa, wκ=0.76) and videos (wκ=0.63).61–63

Upper limb function

Upper limb function is classified using the MACS.64 The MACS is an international system to classify hand function based on the child’s typical performance when handling objects in daily activities. The MACS is a five-level classification of how well children with CP use their hands to handle objects in day-to-day activities.64 This classification system was developed for children aged 4–18 years, and has good reliability for use in children as young as 2 years.64 The MACS has reported construct validity, and excellent inter-rater reliability (intraclass correlation coefficient=0.97 between therapists and 0.96 between therapists and parents) for children with CP.65 Children will be classified using the MACS by an occupational therapist in discussion with the child’s carer.

Communication function

Communication function will be classified on three distinct but overlapping systems:

1. Communication Function Classification System classifies children’s performance in sending and receiving communicative messages using their typical communication means (considering all communication methods including Augmentative and Alternative Communication). It has been validated in children with CP aged 2–18 years. Reliability between professionals was moderate (κ=0.66), professional-parent fair (κ=0.49) and test–retest strong (κ=0.82).66

2. Functional Communication Classification System classifies children’s performance only in sending communicative messages, and also considers their typical communication (including all communication methods including Augmentative and Alternative Communication). It has excellent inter-rater reliability between professionals (κ=0.94) and parent-professional (κ=0.59).67

3. The Viking Speech Scale (VSS) was used to classify children’s speech production.68 The VSS is a four-level classification system, which can be used to classify speech intelligibility for strangers and unfamiliar conversation partners of children with CP aged 4 years and above. It has strong content validity, and moderate-to-substantial inter-rater reliability between pairings of speech pathologists, healthcare professionals and parents (κ=0.58–0.81).68

Eating and drinking function

The Eating and Drinking Ability Classification System (EDACS) classifies the eating and drinking abilities of children with CP aged 3 years and above. Classification is I–V and describes children’s safety and efficiency predominately focusing on food and fluid textures.69 The EDACS has strong inter-rater reliability between professionals (ICC=0.93), but fair reliability between parings of professionals and parents (ICC=0.45).69

Brain structure on MRI

The American Academy of Neurology practice parameter has recommended  that brain MRI should be part of the diagnosis of CP.70 Early MRI at 0–3 years was classified according to the nature and presumed timing of the lesion11 and analysed for brain lesion severity on the semi-quantitative scale of Fiori.8 Aetiology of CP was evaluated using MRI (location, nature and structure of the brain lesion).11 The brain lesion was classified by three main criteria:

1. The anatomical features of the lesion:

2. localisation by tissue (eg, cortical, white matter, deep grey matter, etc)

3. localisation by region (eg, lobes involved, laterality, etc)

4. extent of lesion (eg, generalised, hemispheric, lobar, etc)

5. The presumed aetiology of the lesion: i) genetic; ii) ischaemic; iii) infective and iv) other.

6. The presumed timing of the insult that caused the lesion:

7. Prenatal by trimester or by stage of brain development;

8. Perinatal;

9. Postnatal.

All MRIs were classified by a neurologist (SF) together with a neuroradiologist (AC) using a standardised method of image evaluation and classification. Following these evaluations, consensus was reached regarding the above three criteria. Based on preliminary data, it was estimated that >60% of children currently receiving a diagnosis of CP had early brain MRI as part of their clinical workup. All children (n=245) will be offered a repeat brain MRI at 8–12 years at 3 T. The majority will have their imaging performed and reported through the Herston Imaging Research Facility, on a Siemens 3.0 T MR scanner. The current minimum imaging protocol for patients with suspected CP consists of axial fast spin echo and coronal fast spin echo sequences and three-dimensional (3D) inversion prepared fast spoiled GRASS sequence. The 3D acquisitions will be reformatted in axial, coronal and sagittal planes, with additional oblique and curved reformatting. Age-specific protocols will be used to maximise the ability to detect cortical and white matter abnormalities at different stages of myelination. All neuroimagings will be reviewed by a neurologist (SF, AG) and a neuroradiologist (AC) familiar with the features of lesions that resulted in CP. This approach is consistent with a clinical practice guideline suggesting that all patients with the label of CP had high-quality MRI on at least one occasion.70 MRI scans will be performed predominantly awake, without anaesthesia and after informed consent. Preparation for the MRI will be offered to families in the form of a training DVD explaining the scanner experience and practice in a ‘mock scanner’ (0.0 T) where required.

Brain lesion severity will be assessed using a structured scoring proforma8 based on the CH2 template,71 a highly detailed single-subject T1 template in Montreal Neurological Institute MNI space, which is the international standard for brain mapping (International Consortium of Brain Mapping). Lesions will be transcribed onto the proforma and the following measures  obtained: i) number of anatomical lobes involved, ii) number of slices on the template that were affected and iii) size and distribution of the lesion measured by a global lesion score and lesion subscores. The score (maximum of 40) is based on: i) anatomical lobes involved; ii) number of affected slices and iii) size and distribution of the lesion. The number of lobes and slices affected will be the average of summed right and left hemispheres. To calculate total lesion score, each frontal, parietal, temporal and occipital lobes will be first considered in three sections: periventricular, middle and subcortical matter. Each section scored 0.5 if <50% of area is involved; or 1, for >50% involvement, with a maximum lobar score of 3. Lobar scores for each hemisphere are summed, with a maximum hemispherical score of 12. The total lesion score is the sum of right and left hemispherical scores (maximum score of 24). A 1-point score (involved/not involved) wis also attributed to 16 anatomical structures including the corpus callosum, the cerebellum and the main subcortical structures. The final maximum score of the scale is, thus, a maximum of 40 (24+16).72 The Fiori scale method has strong inter-rater and intrarater reliability72 and strong construct validity based on dMRI and functional severity in children with unilateral CP.73

At 8–12 years, structural MRI (sMRI)-guided and functional MRI-guided dMRI scans suitable for connectivity analyses will be undertaken on the 3T scanner at Herston Imaging Research Facility (or The Lady Cilento Children’s Hospital at 3T for children requiring general anaesthesia ≈ 5%). dMRI for white matter fibre tracking and whole brain connectomes will be acquired using our published protocol.74 The sMRI images will be acquired using an Magnetisation Prepared Rapid Acquistion Gradient Echo(MPRAGE) sequence at an isotropic resolution of 1 mm. dMRI data will be preprocessed to reduce image artefacts,75 and the fibre orientation distribution was estimated using constrained spherical deconvolution.76 Probabilistic tractography will be conducted using MRtrix software and connectivity matrices generated using previously described methods.74 Quantitative diffusivity indices Fractional Anisotrophy (FA) and Mean Diffusivity (MD) will be encoded within the connectome to assess reorganisation.74 Network-based statistics77 will be performed between FA and MD connectomes to identify significant cortical networks associated with neural reorganisation. A second analysis will investigate brain maturation by comparing serial sMRI data acquired around 2 years with scans in the same children at 8–12 years to develop a predictive model of brain structure and functional outcome using spatiotemporal analysis of the longitudinal imaging data.77

Gross motor function

Gross motor function will be evaluated using the GMFM-66 and GMFM-8886 by experienced research physiotherapists. The GMFM-88 assessed a child’s motor abilities in lying to rolling, sitting, crawling to kneeling, standing, walking, running and jumping. The GMFM-66 comprised a subset of the 88 items identified (through Rasch analysis) as contributing to the measure of gross motor function in children with CP. The GMFM-66 will provide an overall measure of gross motor function and the GMFM-88 provides domain scores to explore specific motor skills.86

Upper limb performance

Children with unilateral CP whose manual ability is MACS I–III will be assessed on the school kids AHA, a Rasch measure of effectiveness of impaired hand in bimanual activities. Test–retest reliability was high (ICC 0.98) and there was predictive validity of future assisting hand use.87 BoHA will be used for children with bilateral CP who were manual ability MACS I–IV. The BoHA test content was developed by researchers in Norway and Sweden through modification of the AHA test items and by generation of new items.88 Associations between BoHA measures and MACS levels showed strong correlation (Spearman’s rho: 0.74). The person separation ratios (4.36 and 5.19) and the person reliability (0.95 and 0.96) for the subscales indicated that the children’s hand function could be separated into six and seven ability levels.88 BoHA is the first observation-based assessment of effective use of the hands in bimanual activities for children with bilateral CP.

Hand dominance will be assessed using the Edinburgh Handedness Inventory laterality quotient.89 The Edinburgh Handedness Inventory questionnaire consists of 10 items regarding hand preference (right or left) in performing a number of everyday tasks requiring one (eg, writing, drawing, throwing and using scissors) or two hands (eg, using a broom or opening a box). The laterality quotient is calculated using the following formula: laterality quotient = (right hand–left hand/ (right hand+left hand)×100). The Edinburgh Handedness Inventory is included to objectively determine upper limb dominance. This classification system consists of a table that requires the participants to indicate which hand they use to perform a selection of everyday tasks.

Stereognosis relates to a participant’s ability to perceive and recognise objects by using only tactile information90 91 and will be assessed on the impaired and unimpaired limbs, using the approach described by Sakzewski et al.91 Participants will be required to identify objects placed in their hand, without any visual cues. A total of nine objects will be placed in the hand one at a time. Three familiar objects (teaspoon, key, peg) and six similar matched objects (safety pin and paperclip; pen and pencil; coin and button) will be used. With vision occluded, participants are presented with each item. If a participant is unable to grasp, manipulate or release an object, the occupational therapist assisted the participant and will move the object for them within their hand. A corresponding set of items will be used to allow participants to identify the object in order to minimise any errors due to incorrect naming of the object. Scores ranged on a scale from 0 to 9, where participants scoring below 9 are considered to have impaired stereognosis.90 91

Radiological measures of hip displacement and spine

Hip surveillance, including AP pelvis X-ray, is recommended for all Australian children with CP to facilitate early detection and treatment of severe or progressive hip displacement.19 92 93 The migration percentage is widely accepted as the gold standard measure in hip surveillance,78 94 measuring femoral head displacement.95 Other measures included the Acetabular Index, assessing acetabular dysplasia,96 the Hilgenreiner's epiphyseal angle (HEA)96 and the femoral neck-shaft angle.95 97 The HEA96 is a radiographic measure describing the proximal femoral epiphysis and has been previously applied to assessment of coxa valga,98 99 but may offer prognostic information for hips at risk in CP. The HEA represents the acute angle between a line drawn parallel to and through the proximal femoral epiphysis and Hilgenreiner’s line.81 82 Physiotherapists will perform a clinical examination of spinal alignment and mobility to screen for evidence of a potential scoliosis or kyphosis. Where indicated, an AP spine radiograph (for scoliosis) or lateral (for kyphosis) will be performed. Spines where scoliosis was evident will be measured according to the Cobb angle.100

Body composition and bone health

Body composition measures and bone parameters will be acquired using a Lunar Prodigy DXA (GE Medical Systems, LUNAR, Madison, Wisconsin, USA). Body composition measures included FM (g) and FFM (g). Bone parameters include areal bone mineral density (g/cm²) and bone mineral content (g) for all total body, bilateral proximal and lateral distal femur sites. The lateral distal femur is a common site of fracture,101 102 with the technique previously described,102 and measurements are reproducible in children with CP.103 104 The analysis involves creating three regions of interest, each containing different proportions of trabecular and cortical bone with results for each region of interest (ROI), therefore, treated independently.101 102 Additionally, the proximal and distal femoral sites are used to calculate bone mineral apparent density (g/cm³), derived from the projected bone area (cm2) to provide an approximation of volumetric BMD.105 All scans in this research are a ‘one-off’ occurrence, with the total radiation dose for these five DXA scans being <15 μSv.106 This is equivalent to approximately 1–2 days natural background radiation exposure, and only equivalent to 3% of the dose constraint limit for children as research volunteers, up to the age of 18 years.¹⁰⁶ The total estimated time for all DXA scans is 30 min, performed at the University of Queensland Children’s Nutrition Research Centre.

Fracture rate

Fractures will be diagnosed radiologically. Parents will report by telephone within 24 hours of fracture occurrence and will bring X-ray films and details of management to their study visit. Vertebral fracture will be diagnosed on lateral X-rays of the thoracic and lumbar spine when indicated. Children who were GMFCS III–V will undergo thoracic and lumbar spine (AP/lateral) at 8–12 years if there were clinical signs of fracture and/or scoliosis/kyphosis. Radiographs are minimised to reduce the radiation exposure.

Sexual maturation

Legal guardians of participants will be provided with standardised Tanner stage puberty diagrams, and parents will be asked to evaluate the child’s current pubertal stage.107 Parental pubertal assessment will be reviewed by a physician for precocious puberty. In cases of precocious puberty, a left hand/ wrist X-ray will be conducted to determine the bone age and relative skeletal maturity of children. The bone age will be used to determine if the CP condition is interfering with the proper growth and bone development of the child.

Pain

Children will complete the Paediatric Pain Questionnaire (PPQ) with adult help if required.108 The PPQ asks children to report their pain now (severity, type and location), as well as the severity of the worst pain they had in the previous week. The PPQ’s visual analogue scale for pain rating provides a valid and stable measure of pain intensity in children and adolescents with chronic musculoskeletal pain.109

Three-dimensional gait analysis and in vivo muscle mechanics

A full 3D gait analysis, including synchronised measurement of muscle activation using electromyography (EMG) and calf muscle mechanics using two-dimensional (2D) ultrasound, will be performed for all children functioning at GMFCS I–III. Participants will walk unaided and barefoot at a self-selected speed over a level walkway (10 m in length) with four force platforms embedded in the laboratory floor in the centre of the walkway. Reflective markers will be attached to the trunk, pelvis and lower limbs according to the modified ‘Plug in Gait’ marker set, with additional clusters of three markers on each thigh and shank segment, and a marker on the fifth metatarsal head.110 Marker trajectories will be recorded at 100 Hz using an 10-camera, 3D motion capture system (Vicon Motion Systems, Oxford, UK) and ground reaction force data will be acquired at 1 kHz using four 510 mm×465 mm force platforms (AMTI, Watertown, Massachusetts, USA) arranged in series. Lower limb muscle activations of the rectus femoris (RF), medial hamstrings (MH), medial gastrocnemius (MG), lateral gastrocnemius (LG), soleus (SOL) and tibialis anterior (TA) is  recorded for both legs at 1 kHz using a wireless surface EMG system (Aurion ZeroWire, Milan, Italy). Raw EMG signals will be high-pass filtered (Butterworth, zero-lag, fourth order, 30 Hz) to remove movement artefact, full wave rectified and low passed filtered (Butterworth, zero-lag, fourth order, 6 Hz), and interpolated to 101 points per cycle. Non-negative matrix factorisation will be applied to extract muscle synergies,111 which represented neuromuscular control during gait. Whole body 3D gait kinematics, joint moments at ankle, knee and hip joints and musculotendinous lengths for MG, LG, SOL, RF and MH will be computed across at least five trials using OpenSim112 and normalised to length in standing. The Gait Profile Score113 will be calculated as an index of overall gait pathology. A digital output signal from the ultrasound system will be used to synchronise acquisition of all 3D marker, force plate and EMG data.

Two-dimensional B-mode ultrasound will be used to examine MG and SOL muscle function during walking by attaching a flat ultrasound transducer (LV7.5/65/64D, Telemed Echo Blaster 64 EXT-1T, Vilnius, Lithuania) to the surface of the skin above the MG muscle and recording muscle fascicle length and pennation angle changes, as described previously.114 Muscle fascicle behaviour during walking will be analysed using a semi-automatic process, which has been shown to be highly repeatable (coefficient of multiple correlation 0.88).115 The average of five complete strides will be used in the analysis for each participant to ensure the overall reliability of muscle fascicle length data.116 Freehand 3D ultrasound will be used to measure muscle size and structure of the lower leg muscles: MG, LG, SOL and TA.117 This method of 3D ultrasound is valid (within 1.3%) and reliable (ICC >0.99) for measuring gastrocnemius muscle volume and length in vivo.117 Calf muscle physiological cross-sectional area will be measured as the ratio of muscle volume muscle fascicle length, corrected for fascicle pennation angle.

6 min walk test

This simple, submaximal test measures the distance walked over 6 min, providing information about endurance during functional activities.118 The 6MWT has excellent test–retest reliability (ICC=0.98) in CP.119 Percentile curves have been created on 1445 children with TD aged 7–16 years.120 The test will be performed according to guidelines of the American Thoracic Society on a 10 m course.121

Muscle power sprint test

The muscle power sprint test (MPST) provides an estimate of anaerobic power.122 The MPST requires participants to complete six 15 m runs as fast as possible with 10 s rest between each lap. Power output is calculated as the product of body mass and distance, divided by time.122 The MPST has been validated against the Wingate Anaerobic cycling test,123 and has excellent test–retest reliability (ICC 0.98) in children with CP.122

10 m fast walk test

The 10 m fast walk test (10 mFWT) is a test of maximal walking speed over a distance considered the minimum for functional ambulation. The 10 mFWT has moderate test–retest reliability in children with CP (ICC 0.81).124

Lower limb functional strength

Thirty second repetition maximum (rep_max) of functional strength exercises (including sit-to-stand, lateral step-ups and half-kneel to stand) will be tested according to published recommendations.125 Functional strength tests demonstrate acceptable inter-tester reliability (ICC≥0.91; coefficient of variation (CV) 12.1%–22.7%) in children with CP.125 For each lower limb functional strength exercise, participants will be given verbal and visual instructions as well as two practice repetitions prior to testing. The exercises were assessed in the following order: sit-to-stand, lateral step-up dominant leg, lateral step-up non-dominant leg, half-kneel to stand dominant and half-kneel to stand non-dominant. Participants will be given verbal encouragement throughout. Participants will be given 180 s rests between exercises. If a participant does not complete an exercise while performing the practice attempts, they will be assigned a score of 0 and will not proceed to testing.

Habitual physical activity

Triaxial accelerometers (ActiGraph GT3X+, Pensacola, Florida, USA) will be used to evaluate the frequency, intensity and duration of physical activity.126 ActiGraph accelerometers have evidence of validity and interinstrument reliability in children with TD compared with heart rate monitoring, direct observation, indirect calorimetry, whole-room calorimetry and doubly labelled water.126 The ActiGraph has been validated for measurement of physical activity intensity in adolescents with CP using oxygen uptake as the criterion measure.127 128 ActiGraphs will be fitted during assessment and worn during waking hours for 7 days.¹²⁶ Participants’ caregivers completed a 7-day physical activity monitor log to record wear and non-wear times (see online supplementary Appendix 2). Stored data will be uploaded to an excel macro to determine daily wear time, average counts per min, daily time spent in sedentary, light, moderate and vigorous activity. Counts will be classified using established cut-points for children with CP.128

Blood samples for growth hormone and vitamin D

Blood will be collected and tested for hormones and other markers required for optimal growth, bone and metabolic health. Specifically, these tests are liver function, kidney function, full blood count, insulin-like growth factor-1, thyroid hormone, parathyroid hormone, vitamin D3, calcium, phosphate and iron studies. As described above, these parameters of growth, bone health and body composition are often altered in children with CP and are related to gross motor function classification, body composition, growth velocity and nutritional status. Blood tests are optional and consent will be obtained from the parent and assent from the child where possible. Blood samples will be collected by qualified phlebotomists, who are familiar with collecting blood from paediatric subjects using their standard procedures. Where preferred, samples will be collected under general anaesthesia, if a patient undergoes an unrelated and non-emergency surgical procedure (eg, orthopaedic surgery, Botulinum toxin A injections, MRI under anaesthesia). Parents of participants will be advised if these results fall outside the relevant reference ranges in relation to age, gender and pubertal status. Parents will provide informed consent for information to be provided to their treating clinician who will take responsibility for ongoing care and follow-up.

Dietary intake

Dietary energy intake will be recorded using a 3-day weighed food record as validated29 using our published methods.26 Food records will be analysed using FoodWorks. Mean energy intake will be expressed as megajoules per day and as a percentage of age-specific and gender-specific recommendations.129

Vitamin D intake

A food frequency questionnaire will be completed by parents/caregivers to determine the habitual intake of vitamin D-containing foods by the participants. The questionnaire consists of a table that requires parents to tick a frequency box and record the brand of a simple list of foods.130

Sun exposure

A 7 day sun exposure diary will measure daily sun exposure in the participants to measure ultraviolet radiation (UV) exposure for vitamin D adequacy. Each day, participants will record the amount of time spent in the sun during each 1 hour interval (0, <15, 15 to <30, 30 to <45 or 45–60 min) between 05:00 and 19:00 hours. Clothing cover (based on a clothing cover guide provided with the dairy) and use of sunscreen (frequency and application site) using established methodology will also recorded131 (see online supplementary Appendix 3). It is proposed that the sun exposure diary is done at the same time as the physical activity monitor record to lessen the burden on the participants. Sun exposure diaries will be done within 2 weeks of serum vitamin D levels being collected, to allow for meaningful interpretation of sun exposure and vitamin D levels.

Oropharyngeal dysphagia

OPD (feeding and swallowing difficulties) will be evaluated during a digital video-recorded snack of 20 min. Children will be presented with three standardised boluses of five textures; puree, semi-solid, chewable, tough chewable and fluid. The following measures will be used to rate the mealtime:

1. The Dysphagia Disorders Survey (DDS)—part 2 consists of a series of binary judgements on eight ingestion functions across the oral preparatory, oral, pharyngeal and gastro-oesophageal phases (maximum raw score of 22). The DDS has good reliability132 133 and convergent validity.132–136

2. The Schedule for Oral Motor Assessment (SOMA) consists of seven oral motor challenge categories corresponding to four food textures and three fluid utensils. The SOMA has been validated on 127 young infants; 58 comparison children with typical oral skills, 56 with non-organic failure to thrive (aged 8–24 months) and 13 children with CP and overt feeding difficulties (aged up to 42 months).137 It has strong inter-rater reliability (κ=1.0 in 68% of fluid category items and 58% of food category items) and test–retest reliability between boluses (κ=1.0 in 84% of items).137

3. Observations of 16 clinical signs suggestive of pharyngeal phase impairment (eg, cough, gurgly phonation, wet respiration) will be rated premealtime and postmealtime by a trained researcher, and rated according to each food/fluid texture from video by a speech pathologist.138–140

4. The Thomas-Stonell and Greenberg scale will be used premealtime and postmealtime to rate saliva loss.141 This consisted of two observational ordinal scales (1–5), based on severity and frequency of loss.

5. The Cerebral Palsy Child Feeding Questionnaire, used in the CP-Child Study142 gathered information on the child’s typical mealtime performance based on parent report, which supplemented the data obtained from the participant’s clinical feeding assessment.

6. The Feeding/Swallowing Impact Scale (FS-IS) addresses questions of carer QoL and how to incorporated into the economic evaluation. The FS-IS is a validated tool to measure the impact of caring for a child with dysphagia and concerns on caregiver QoL.143 It is an 18-item, parent questionnaire divided into three subsections: i) daily activities; ii) worry and iii) feeding difficulties. The tool was validated on the caregivers of 164 children (median age 14 months, mean: 32±44 months) with varying comorbidities including prematurity (born <37 weeks) in 66 (40%) children, 144 (88%) were medically complex with conditions in more than one diagnostic-based category and 77 (47%) of children had feeding tubes.143

Communication

All children will have language assessed using the core language subtests of the Clinical Evaluation of Language Fundamentals screener (CELF-4),144 in addition to the Peabody Picture Vocabulary Test (PPVT). The CELF-4 is a criterion referenced assessment of language skills in children aged 5–21 years, with Australian norms available. Children who were non-verbal only will complete the receptive subtests of the CELF-4 and the PPVT. Children unable to participate in standardised assessment (eg, due to significant cognitive, visual or motor limitations) will have language classified using the Triple C, a parent-reported observational assessment. Communication performance will be indicated by parents on a communication questionnaire developed for this study (see online supplementary Appendix 4), which includes information on augmentative and alternative communication system type, use, frequency and access.144

Speech production will be assessed using the verbal motor production assessment for children (VMPAC). The VMPAC145 is a diagnostic tool for the systematic assessment of neuromotor integrity of the motor speech system validated in 1434 children aged 3–12 years146. The following subtests of the VMPAC will be administered:

1. Oromotor production in word sequences and sentences (six items): items consisted of three-word and four-word sequences and five-word sentences. These items are designed to evaluate the child’s ability to sequence oromotor movements across different plane within a linguistic context.

2. Oromotor production in connected speech and language (five items): this subtest assesses the child’s motor control (eg, motor precision) as it varied in the context of higher level language formulation.

3. Oromotor production in automatic verbal sequences (one item): this subtest allows evaluation of speech characteristics including pitch, resonance, vocal quality, loudness, prosody, intonation and rate during production of an automatic speech task (eg, counting to 10, saying the alphabet).

4. Oromotor production in word sequences and sentences (six items): items consisted of three-word and four-word sequences and five-word sentences. These items are designed to evaluate the child’s ability to sequence oromotor movements across different plane within a linguistic context. Oromotor production in connected speech and language (five items): this subtest assesses the child’s motor control (eg, motor precision) as it varied in the context of higher level language formulation.

5. Oromotor production in automatic verbal sequences (one item): this subtest allows evaluation of speech characteristics including pitch, resonance, vocal quality, loudness, prosody, intonation and rate during production of an automatic speech task (eg, counting to 10, saying the alphabet).

Paediatric evaluation of disability inventory computer adapted test

Performance of self-care will be evaluated using the parent-reported PEDI-CAT for the domains of self-care, mobility and social functioning using scaled scores (Rasch), which have good validity and reliability.147 The PEDI-CAT was developed on the basis of the original PEDI.148 Scaled scores (possible range 20–80) for each domain provide an indication of the child’s performance along a continuum of item difficulty and are most suitable for research.147 Scaled scores give more precise results in the extreme ranges than normative standard scores.147 Scaled scores are recommended to track functional progress in children who are substantially delayed.147

Participation and environment measure for children and youth

The participation and environment measure for children and youth (PEM-CY) is a parent-reported instrument that examines participation and environment across three settings: i) home; ii) school and iii) community.149 There are 10 items in the home section, 5 in the school section and 10 in the community setting. For each item, the parent is asked to identify how frequently (over the past 4 months) the child has participated (eight options: daily to never); how involved the child typically is while participating (five-point scale: very involved to minimally involved) and whether the parent would like to see the child’s participation in this type of activity change (no or yes, with five options for the type of change desired). For each setting, the parent is then asked to report on whether certain features of the environment make it easier or harder for the child to participate. The PEM-CY has reported moderate-to-good internal consistency (0.59 and above) and test–retest reliability (0.58 and above) in a population of children (aged 5–17 years) with and without disabilities residing in the USA and Canada (n=576).149 The PEM-CY will be collected using either a paper or online questionnaire format to gain an understanding of the participation of children and adolescents and the impact of environmental barriers and facilitators.

Strengths and Difficulties Questionnaire

The Strengths and Difficulties Questionnaire150 151 is a 33-item parent-rated questionnaire that is used to assess parents’ perceptions of prosocial and difficult behaviours in their child or child adjustment. Parents responded to 25 questions about their child’s behaviour in the last 6 months using a three-point Likert scale (ie, ‘0’=not true to ‘2’=certainly true). These 25 questions are combined to create five subscales of: frequency of emotional symptoms; conduct problems; inattention/hyperactivity; peer problems and prosocial behaviour (eg, ‘considerate of other people’s feelings’). A total score for each scale (0–10) and overall total difficulties score (0–40) is calculated, with higher scores indicating more distress on all scales except prosocial behaviour. Scores of 17 or above for the total difficulties scale are used as a clinical cut-off point. Scores from the five subscales and the overall difficulties scale will be used as a measure of the child’s psychological functioning. The overall total difficulties score has been demonstrated to have moderate-to-high internal consistency (Cronbach's α=0.73–0.82) and test–retest reliability (r=0.77–0.85).152

School attainment

The Australian Early Development Index is a population measure of development at school entry, assessing school readiness.153 The National Assessment Program—Literacy and Numeracy (NAPLAN) measures literacy/numeracy achievement and individual reports will be obtained from families as a measure of early school achievement.154 Where applicable, consent will be obtained from parents to the release of NAPLAN result by the Queensland Curriculum and Assessment Authority.

Raven’s coloured progressive matrices

The Raven’s coloured progressive matrices (RCPM) is an assessment of non-verbal intelligence for children aged 5–11 years with intellectual delay or physical disability. It consists of 36 items (15–30 min) of increasing difficulty in which the child has to complete a pattern. RCPM has validity with CP and Australian norms.155 156 In 618 Australian children (aged 6–11 years), the RCPM had good internal consistency (0.76–0.88) and split-half reliability (0.81–0.90).156

Behaviour rating inventory of executive function

The behaviour rating inventory of executive function is a parent completed 86-item measure of executive function in their child’s everyday life, yielding two scores: i) the behavioural regulation index and ii) the metacognition index to form a global executive composite score. Both scores can be converted into T scores with ≥65 indicative of dysfunction.157 It has good convergent and divergent validity with Child Behaviour Checklist and the Behaviour Assessment System for Children,158 high internal consistency, (Cronbach’s α=0.80–0.98) for parent form.158 159

Conners three parent short form (Conners 3)

The Conners three parent short form (Conners 3)¹⁶⁰ is a thorough assessment of attention deficit hyperactivity disorder (ADHD) and its most common comorbid problems and disorders in children and adolescents aged 6–18 years. The Conners 3 will be completed by the participant’s parents or guardian and consisted of 110 statements and takes approximately 20 mins to complete. Parents or guardians rate each statement using a four-point scale ranging from ‘0, not true at all (never, seldom)’ to ‘3, very much true (very often, very frequently)’. The Conners 3 measures the seven key areas of inattention, learning problems, aggression, family relations, hyperactivity/impulsivity, executive functioning and peer relations. Raw scores are converted into T scores based on a large representative normative sample based on the US consensus data. In addition, the Conners 3 calculates T scores for symptom scales including ADHD hyperactive/Impulsive, ADHD combined, oppositional defiant disorder, ADHD inattentive and conduct disorder. Both internal consistency coefficients (α=0.83–0.94) and test–retest reliability (r=0.52–0.94) were good for the Conners 3 parent version total sample age range.160

The autism spectrum quotient-child

The autism spectrum quotient-child (AQ10-child) is a 10-item parent-reported autism screening measure for children aged 4–11 years.161 It was developed from the Quantitative Checklist for Autism in Toddlers. Parents respond to the 10 items on a four-point Likert scale from definitely agree to definitely disagree. Items will be  summed and the scale had a cut-off of 6. It had high internal consistency (Cronbach’s α=0.90), sensitivity (0.95) and specificity (0.89).161

Sleep Disturbance Scale for Children

Sleep disorders are up to four times more common in children with CP compared with the general population and are linked to both physical (total body involvement, severe visual impairment) and environmental factors (single-parent household, bed-sharing). The most commonly reported disorders include difficulty initiating and maintaining sleep, sleep-wake transitions and sleep breathing disorders.162 The Sleep Disturbance Scale for Children is a 27-item parent-reported questionnaire that assesses sleep disturbance in children within the past 6 months.163 Each item is responded to on a five-item Likert scale with higher values representing greater clinical severity. Items are summed to produce a total score and six subscores representing different facets of sleep disturbance: disorders of initiating and maintaining sleep, sleep breathing disorders, disorders of arousal, sleep-wake transition disorders, disorders of excessive somnolence and sleep hyperhidrosis. It has good internal consistency (Cronbach’s α=0.71–0.79), test–retest reliability (r=0.71) and discriminative validity in distinguishing between clinical and community samples.163

Condition-specific QoL measures

The cerebral palsy quality of life-child assesses well-being across multiple domains using parent-report (aged 4–12 years) and child self-report from 9 years.164–166 Psychometrics are excellent with Cronbach’s α 0.74–0.92 parent-proxy report and 0.80–0.90 child self-report. Test–retest was adequate (ICC 0.76–0.89) and moderately correlated with health (r=0.30–0.51).164–166

Generic QOL

The Child Health Utility-9D is a generic instrument for children aged 7–11 years for which there is an algorithm to give a single preference-based utility index for health states (giving a single generic preference-based indicator of each individual’s health state), making the data amenable for economic evaluations for interventions.167–170 The EuroQoL five-dimensional descriptive system (EQ-5D-5L) is a generic instrument designed to describe and value health based on five dimensions: mobility, self-care, usual activities, pain/discomfort and anxiety/depression.171 The EQ-5D-5L will be completed by the parent/caregiver, about themselves and scored using the Australian algorithm.172 173 This is relevant (alongside the CES) to address questions such as carers QoL and how to incorporate it into economic evaluation.

Demographic questionnaire

Parents will be required to complete questionnaires related to various aspects of their child's development/progress including participation, feeding, HRQoL, sleep, fracture history, pain, executive functioning in everyday life, clinical comorbidities including epilepsy, psychological functioning, treatments and interventions the child has received since the completion of the CP-child projects (see online supplementary Appendix 5). The socioeconomic status of Australian families will be classified into tertiles using scores on the Socioeconomic Indexes for Areas Index of Relative Disadvantage.174

Carer Experience Scale

The Carer Experience Scale (CES) will be completed by primary caregivers at their study visit. This validated measure of care-related QoL has six domains (activities, support, assistance, fulfilment, control and relationship with the care recipient) and takes approximately 3 min to complete.175 The CES is scored from an algorithm derived from preferences of the general population and can be used to value carer outcomes in economic evaluation using index values.176

The McMaster family assessment device general functioning scale

Caring for a child with a disability can have an impact on the health and functioning of the caregiver and family unit.177 The family assessment device is a 12-item measure of general family functioning.178 Each item is rated on a four-point Likert scale from strongly agree to strongly disagree and the items are summed to create the total score. It has good internal consistency (Cronbach’s α=0.92), construct validity and reliability.178–180

Monitoring of resource use and direct costs of treatments

To determine the relationship between motor prognosis and healthcare costs, resource use and direct costs of treatment will be monitored using the Health Resource Use questionnaire14 (see online supplementary Appendix 6). Associations between costs (dependent variable) and all other outcome variables (independent variables) including those related to growth, body composition and HPA will be assessed, with adjustment for confounders such as brain lesion severity. Health-related resource use data will be collected including therapy frequency and duration (traditional/alternate), hospital admissions, GP and medical specialist visits, medications (eg, Botulinum toxin A) and equipment (eg, orthoses). Data will be collected via questionnaire,14 supplemented by consented access to individual hospital, Medical Benefits Scheme (MBS) and Pharmaceutical Benefits Scheme (PBS) records. Standard cost sources (MBS, PBS) will be used to apply unit costs to resources. Statistical approaches, which consider the likely skewed distribution of cost data (such as generalised linear models, with extensions to allow for correlations in the data across observations from multiple time points from each individual as described below), and diagnostic-related group costs for admissions to hospital) will be employed. As this cohort is embedded in a state-wide clinical service, we have consistency of interventions based on best practice guidelines for Botulinum toxin A, bone health and hip surveillance generalisable across Australia. All children are offered best practice treatment across the state-wide service.