The data was processed directly data collection source, with all identifying information either removed or anonymised to comply to the data sharing agreement, funder contract requirements and data protection laws. We include variables (e.g. scores) directly formed from other variables in the original dataset.