Authors :
Presenting Author: Tim Benke, MD PhD – Children’s Hospital Colorado, Pediatric Neurology, University of Colorado School of Medicine
Jacinta Saldaris, PhD – Telethon Kids Institute, Centre for Child Health Research, The University of Western Australia; Peter Jacoby, PhD – Telethon Kids Institute, Centre for Child Health Research, The University of Western Australia; Helen Leonard, MBBS – Telethon Kids Institute, Centre for Child Health Research, The University of Western Australia; Eric Marsh, MD PhD – Division of Neurology, Children’s Hospital of Philadelphia, School of Medicine, University of Pennsylvania; Gina VanderVeen, MPH – Children’s Hospital Colorado, Paediatric Neurology, University of Colorado School of Medicine; Scott Demarest, MD – Children’s Hospital Colorado, Paediatric Neurology, University of Colorado School of Medicine; Jenny Downs, PhD – Telethon Kids Institute, Centre for Child Health Research, The University of Western Australia
Rationale:
CDKL5 Deficiency Disorder (CDD) is a rare genetically caused Developmental Epileptic Encephalopathy (DEE). Refractory seizures and developmental delay are core features of CDD. Clinical trials for new medicines that could substantially change disease status (e.g., gene therapies) are imminent but few validated outcome measures for CDD exist. This study modified an existing gross motor measure and evaluated its psychometric properties in CDD.
Methods:
We modified the Rett Syndrome Gross Motor Scale with new items to measure head and trunk control and developed a protocol for parents to upload video clips of the tasks from their smart phone to a protected server. After pilot testing with six parents with a child with CDD, the CDD gross motor scale comprised 17 items that described head and trunk control, sitting, transfers, standing and walking activities, each rated on a four point scale for level of assistance. We collected 137 sets of videos from families participating in the International CDKL5 Disorder Database (ICDD) and the CDD Centers of Excellence network in the USA (COMIRB 19-2756, UWA 2019/RA/4/20/6198). Assuming the items are ordered on the same unidimensional latent trait, a Rasch analysis was conducted to evaluate targeting of items, item fit, internal consistency, unidimensionality, and category thresholds. Intra- and inter-rater reliability was evaluated on a subset of 50 sets of video clips using intra-class correlations (ICCs) and kappa statistics.
Results:
The median (range) age of the individuals with CDD was 6.9 years (1.5-34.1). The majority (85%) were female. The mean (SD) score was 22 (15) from a total possible score of 51. There was a slight right skew (0.3). Rasch analysis confirmed that the items covered the full range of gross motor abilities; Person Separation Reliability was 0.947 indicating good internal consistency; and there was evidence for a single underlying gross motor dimension. One item (prone neck extension) showed misfit to the Rasch model while fit statistics for three items were lower than recommended indicating some redundancy. Category thresholds were disordered for many items because few scores for these items showed any level of assistance. There was excellent agreement (Kappa >0.80) between two raters for most of the 17 items and all had a Kappa >0.60. Intra-rater reliability showed excellent to perfect agreement (Kappa >0.80) for all items. The ICC of the total score was excellent (0.98).
Conclusions:
Findings support that the CDD gross motor scale measured the range of gross motor skills in CDD with sensitivity to small increments in skill. One item should be dropped due to misfit giving a 16-item scale. Reliability was good. Although there were disordered category thresholds for many items, this is unlikely to affect the integrity of this scale because collapsing categories would lose valuable trait information and responsiveness to change. The CDD gross motor scale appears to be a suitable remotely administered measure that avoids a floor effect for individuals with severe gross motor impairments.
Funding:
NIH/NINDS U01NS114312 (Benke/PD)