Many patients do not participate in research. Can results still be generalised?

As in all studies of this kind, not all eligible patients participate. Out of more than 1,650 patients who were eligible to participate in the study, more than 800 (ranging between 40 – 59% at 4 sites) were not interviewed for varying reasons. At the time of the first interview within one week after admission many couldn't be contacted for organizational reasons, did not want to take part in the study, were too ill to give informed consent, were already discharged, transferred to another hospital, transferred to prison, hospital staff said they were too ill to approach, hospital staff refused contact because patients were too violent/dangerous/unpredictable or sedated, or the patient simply refused to see anyone. This significant group of non-recruited patients required us to have systematic data describing the eligible sample and the sample actually interviewed. Exclusion of non-participants could bias data collection. For all patients who met the inclusion criteria within the study period and were therefore eligible to participate but were not interviewed for the above reasons data on age, gender, ethnicity, legal status of admission, discharge diagnosis (ICD-10), length of index admission, time until readmission, number and length of further admissions to psychiatric hospital care within a one year of follow-up were collected from records or obtained in aggregate form. This paper will present findings of the comparison of baseline characteristic and outcomes between the investigated sample and non-participants. Presented findings will show how representative the recruited sample (N = 843) is for all involuntarily admitted patients and estimate in what way the findings of the investigated sample can be generalised to all involuntarily admitted patients in England.