Prognostic models for identifying adults with intellectual disabilities and mealtime support needs who are at greatest risk of respiratory infection and emergency hospitalisation.

These are exploratory prognostic models, not validated risk scores. The authors could not split the data into training and validation samples because the sample was small. Only 142 of 726 invited people agreed to take part, and only 127 were used in the models, so selection bias may limit generalisability. Some information was recalled for the year before the baseline interview, so recall bias may affect variables such as prior respiratory infections. Missing data reduced the respiratory-infection model to 119 people and the hospitalisation model to 124 people. The age term was modelled in a quadratic form, and the authors state it cannot be interpreted strictly as an age effect. Epilepsy severity and other details were not collected, so the study cannot say whether epilepsy control or severity explains the associations.

None declared. The research was commissioned by the National Institute for Health Research under its Research for Patient Benefit programme, grant reference PB-PG-0906-11098. Some authors were supported by the NIHR Collaboration for Leadership in Applied Health Research and Care East of England, and one author was supported by the Medical Research Council, Unit Programme number U105292687.

Do not use these models as validated risk scores. The authors did not validate them because the sample was small, and only 142 of 726 invited people agreed to take part, so selection bias may affect the results.