Length of Stay on Psychiatric Wards
Lead Research Organisation:
King's College London
Department Name: Institute of Psychiatry
Abstract
The length of admissions to psychiatric hospitals is very variable but we don‘t really know why. I believe that being excluded from housing by your landlord or family makes an important contribution to length of stay (LOS), separate from those other individual characteristics that predict longer admissions.
First, I will test a simplified version of this idea by re-examining results from some existing studies. I will use being discharged to an address other than the one before admission as an indirect indication that the person has been excluded from housing. Then I will study housing changes in a group admitted to psychiatric hospitals in order to see if the theory above describes what actually happens and work out a simple way of measuring and recording exclusion from housing. Next I will try to summarise all previous studies of LOS in order to check that I will be recording all other relevant information in my own study. Finally, based on these results, I will collect information from a large group of people who have been admitted to hospital and investigate the relationships between individual characteristics, loss of housing and length of stay, again testing the idea above.
First, I will test a simplified version of this idea by re-examining results from some existing studies. I will use being discharged to an address other than the one before admission as an indirect indication that the person has been excluded from housing. Then I will study housing changes in a group admitted to psychiatric hospitals in order to see if the theory above describes what actually happens and work out a simple way of measuring and recording exclusion from housing. Next I will try to summarise all previous studies of LOS in order to check that I will be recording all other relevant information in my own study. Finally, based on these results, I will collect information from a large group of people who have been admitted to hospital and investigate the relationships between individual characteristics, loss of housing and length of stay, again testing the idea above.
Technical Summary
Aims and opportunities
Long psychiatric admissions are common, costly, unpleasant and conflict with the ethos of community care. However, previous research does not satisfactorily explain why they occur. My current research suggests that loss of housing during admission has been neglected as a potential determinant. But rather than simply demonstrate that this has a degree of independence of other variables, I aim to test a model in which the decision to exclude an individual from housing is treated as an intervening variable on an indirect path between symptoms and behaviour at admission and length of stay (LOS), these also being linked by a direct path. The fellowship is an opportunity to set a high standard for study design while addressing a clinically important question in a novel way.
Objectives
Year 1: Study MSc in Epidemiology and apply the principles above to the analysis of data on admissions from the AESOP (Aetiology and Ethnicity in Schizophrenia and Other Psychoses) and UK700 studies of psychosis.
Year 2: Study of housing change among inpatients and systematic review of studies of LOS.
Year 3/4: Prospective study of admissions.
Design and methods
In both initial and final phases, I will test the ‘dual-path‘ hypothesis above, using Cox regression for hazard of discharge (LOS) and logistic regression for change of/ exclusion from housing. Change of housing will be a proxy for exclusion from housing in the analysis of the UK700/AESOP data. Housing type (living alone, in supported accommodation or with the family) will be treated as a moderating variable.
In year 2 I will ask professionals, patients and carers open questions about decisions regarding housing change in a group of admissions, and also examine their casenotes. I will clarify when and by whom decisions regarding housing change are made. A simple instrument will be designed and validated with a second sample so that it can be used in a prospective study.
In years 3 and 4 I will study 200 admissions in Southwark, oversampling subjects with longer previous admissions. (A power calculation dictates at least 135 subjects). I will record demographic, symptom, diagnostic and behavioural measures, IQ, cognitive state, measures of previous admissions, and additional variables suggested by the systematic review. I will record housing exclusion and eventual housing change using the new instrument. Measurements of ward atmosphere will be made in case this should be controlled for. Analysis will be as above.
Long psychiatric admissions are common, costly, unpleasant and conflict with the ethos of community care. However, previous research does not satisfactorily explain why they occur. My current research suggests that loss of housing during admission has been neglected as a potential determinant. But rather than simply demonstrate that this has a degree of independence of other variables, I aim to test a model in which the decision to exclude an individual from housing is treated as an intervening variable on an indirect path between symptoms and behaviour at admission and length of stay (LOS), these also being linked by a direct path. The fellowship is an opportunity to set a high standard for study design while addressing a clinically important question in a novel way.
Objectives
Year 1: Study MSc in Epidemiology and apply the principles above to the analysis of data on admissions from the AESOP (Aetiology and Ethnicity in Schizophrenia and Other Psychoses) and UK700 studies of psychosis.
Year 2: Study of housing change among inpatients and systematic review of studies of LOS.
Year 3/4: Prospective study of admissions.
Design and methods
In both initial and final phases, I will test the ‘dual-path‘ hypothesis above, using Cox regression for hazard of discharge (LOS) and logistic regression for change of/ exclusion from housing. Change of housing will be a proxy for exclusion from housing in the analysis of the UK700/AESOP data. Housing type (living alone, in supported accommodation or with the family) will be treated as a moderating variable.
In year 2 I will ask professionals, patients and carers open questions about decisions regarding housing change in a group of admissions, and also examine their casenotes. I will clarify when and by whom decisions regarding housing change are made. A simple instrument will be designed and validated with a second sample so that it can be used in a prospective study.
In years 3 and 4 I will study 200 admissions in Southwark, oversampling subjects with longer previous admissions. (A power calculation dictates at least 135 subjects). I will record demographic, symptom, diagnostic and behavioural measures, IQ, cognitive state, measures of previous admissions, and additional variables suggested by the systematic review. I will record housing exclusion and eventual housing change using the new instrument. Measurements of ward atmosphere will be made in case this should be controlled for. Analysis will be as above.