The search for risk factors for cerebral palsy (CP) is a never-ending quest. Low gestational age and adverse events during gestation, birth, and the neonatal period are most often studied, but pre-conceptional and post-neonatal factors must also be considered. Causal pathways leading to CP have been described,¹ although it has been suggested that these factors may interact in an intricate web rather than following separate avenues. The panorama of risk factors may also vary by socioeconomic factors and access to health care in a country or region, and that panorama may change over time. In the search for risk factors for CP and other neurodevelopmental conditions, different approaches complement each other. In current studies techniques such as directed acyclic graphs and machine learning are employed.^{2, 3} Large register studies are also warranted to reveal the risk factors for CP, to establish their occurrence and importance, and in which subgroup to look for them. Besides issues with completeness that may occur in any register, the nature of the registers providing data for such studies may vary. Some are population-based registers dedicated to surveillance and/or follow-up of CP, while some register data may be derived from administrative health care registers or health insurance databases – data collected for a different purpose. Studies including the combination and linkage of different types of registers may be particularly challenging. Register studies may, however, be very useful.

Chen et al. present results from the National Health Insurance Research Database, covering nearly all of Taiwan's population.⁴ Children were born in 2010 to 2016 and diagnosed with CP from birth up to 3 years of age, an age when most children with CP will have received the diagnosis. A wide range of risk factors for CP were explored. Besides more established risk factors in the child, such as periventricular leukomalacia or intraventricular haemorrhage in the child, maternal, but also paternal factors such as diabetes mellitus, drug abuse, and seizure disorders presented as significantly associated to the outcome of CP in the offspring. Importantly, the authors point out that they could not establish causality, as this was a retrospective cohort study.

The trend of later childbearing in high-income countries has previously elicited research mainly focussed on the older age of mothers and less on paternal age. Recently, older age was highlighted among paternal risk factors in a case–control study.⁵ In the present paper by Chen et al.,⁴ older paternal age had no significant association to CP in the child. Parental factors that are suggested to affect both fertility and the outcome of CP in the offspring can be either maternal, paternal, or combined. Epilepsy in the father may suggest a genetic background of CP as speculated by Chen et al.,⁴ and de novo mutations in the older father or mother may affect fetal development,⁵ but other examples are lifestyle issues such as drug abuse and infections.

Large register studies broaden our views, point us in new directions, and provide guidance for new studies to assess the suggested risk factors. The identification of risk factors leading to CP may initiate preventive actions, but also guide targeted assessments and facilitate early interventions in the child. This underlines the importance of comprehensive and careful research in this area, adding one piece after another to the puzzle, waiting for the image to emerge.