Limitations
Our results are based on a small clinical population from one Swedish county. Nevertheless, the prevalence of CP was similar to figures reported elsewhere, indicating that it is a representative sample. The youngest children were only 2 years of age when we conducted the initial search, which entails the risk that not all children with CP were identified. However, no child was identified after two years, and only two children meeting the inclusion criteria were identified in the complementary search. The retrospective design limited our access to the entries made at the children’s hospital, possibly resulting in missed information from other sources, such as the CHS. The review was mainly performed by the first author, and to reduce the risk of bias, a template was used when collecting the data.