Common data elements and minimum data sets in cerebral palsy: Start small to grow big

Short abstract

This commentary is on the original article by Wilson et al. on pages 1470–1476 of this issue.

Common data elements (CDE) are precisely defined key words or items in clinical research to allow for a standardized collection of data across multiple sites. Mandatory items of the CDEs are considered as a minimum data set (MDS). The article by Wilson et al. ¹ presents the results of an online Delphi survey carried out to establish CDEs and a MDS with the aim to standardize and harmonize phenotypic data in view of further genomic studies in cerebral palsy (CP). Therefore, this is a preliminary step to establish phenotype–genotype and maybe even phenotype–epigenotype correlations based on multiple federated cohorts of individuals with CP. Indeed, genetic, genomic, and epigenomic studies have recently made very good progress in this field.

At the end of this substantial multidisciplinary and international survey, the participants selected 10 items out of 107 as ‘mandatory’. At first sight, this seems insufficient to ensure that this MDS could really help enhance deeper phenotyping and further our understanding of the causes and risk factors of a condition as complex as CP. All the more so as half of these items are only administrative data rather than clinical data.

The definition of CP has evolved over time. By relying more on function than anatomical or aetiological impairment, it now includes medical and social parameters that occur in the context of a prenatal or perinatal non‐progressive brain injury. ² As such, and despite a rather small proportion of individuals with CP having an underlying monogenic cause, researchers and practitioners are calling for CP to be viewed as a clinical entity independently of the underlying cause. ³ This reasonable approach focuses on clinical management that currently does not differ significantly between individuals. The same approach is used in other complex disorders, such as autism spectrum disorders, arthrogryposis multiplex congenita, or congenital myopathies. Fortunately, this approach does not exclude the possibility to apply other types of classification according to hypothesis‐driven questions in the field. ⁴ Furthermore, we learned from neuromuscular diseases that going deeper into genetics does not necessarily lead to better understanding but rather higher complexity. ⁵

Therefore, at second sight, the relatively small number of items selected for the MDS for CP is highly relevant to (1) ensure that many contributors can collect these items without missing data and (2) help better stratify and classify CP on a higher hierarchical level than aetiology or genetics alone. This approach most probably will open access to CP as a concept to a broader medical community including clinical geneticists. Since some of these patients have overlapping clinical features but distinct classifications, the MDS may represent a unique opportunity to create overarching interactions between caregivers involved in the management and diagnosis of these patients but working in different networks. From a clinical perspective, this is also a first step towards personalized medicine in the treatment, prognosis, and genetic counselling of these individuals and families. Above all, if the MDS and by extension the CDE for CP could help patients being guided towards the right diagnostic and treatment networks with the relevant keywords for diagnosis between the medical specialties they depend on, this work will likely be a success.