The impact of diagnostic labelling in population-based research into cerebral palsy.

This study explores and quantifies the impact on the estimation of prevalence rates and aetiological hypotheses of inclusion and exclusion of different diagnostic labels and types of cerebral palsy (CP). The study was based on data from a CP register which had been established in the English North East Thames Regional Health Authority (NETRHA). As a deliberate policy, no definition of CP was given to notifiers and no inclusion or exclusion criteria were specified. Clinical information, including known malformations, genetic disorders, and features that made the diagnosis doubtful, was requested. Rates and relative risks for different inclusion and exclusion criteria were calculated. The crude rate of CP as defined above was 1.6/1000 (95%CI 1.5, 1.7). Exclusion of all cases with a known or potentially causal association reduced this rate to 1.2/1000 (95%CI 1.0, 1.3). Comparison with an intensively investigated sample from Germany and Sweden showed that more of the same causal associations (or diagnostic labels) were found, particularly where MRI studies had been carried out. Future comparative studies in CP will need to be very precise in specifying inclusions and exclusions and in estimating the effects they will have on monitoring trends over time and aetiological hypotheses.