Estimating the effects of early intervention services for first-episode psychosis and model fidelity on hospitalisations: a quasi-experimental study

Date of Approval: 
2020-09-22 00:00:00
Lay Summary: 
Psychotic disorders are serious mental conditions that can lead to a great deal of suffering for the individual and their family and can place a large burden on society. Early intervention for psychosis (EIP) services aim to improve patient outcomes by providing a variety of wide-ranging treatments, as early as possible, to people experiencing psychosis for the first time. Among many goals, EIP services try to reduce the number of times a patient is hospitalised for psychosis. There are many EIP services in England; however, there are no reported research findings on whether EIP services improve patient health from a population level perspective. Most studies have only looked at smaller groups of patients chosen from a small number of EIP teams, and the research has some shortcomings. Even though there are guidelines describing how the EIP model of care should be delivered, services are still provided in ways that differ from the ideal model of EIP care. There is little research on how the different ways of delivering care can affect the health of patients. Therefore, this study will look at whether EIP services decrease the number of patient hospitalisations and time spent in the hospital among people with early psychosis in England. It will also investigate whether there are differences in hospitalisations between regions which more (or less) closely follow the “ideal” model of care. The findings will be useful in planning EIP policy and improving delivery of services, which will improve patient care and health.
Technical Summary: 
The objective of this research is to estimate the effects of early intervention for psychosis (EIP) services and EIP services ‘fidelity’ (i.e. the degree to which an EIP service is delivering care as intended) on two outcomes: the rate of and length of stay in inpatient psychiatric hospitalisations among people experiencing their first psychosis episode in England. Using a quasi-experimental study design, we will compare hospitalisation outcomes in Strategic Health Authority (SHA) regions that vary in exposure to EIP services over time, specifically, the number of EIP teams in each region. Whether the level of EIP ‘fidelity’ in the region is an effect measure modifier of these relationships will also be examined. We will define a cohort of people who experienced their first episode of psychosis between 2002 and 2013 in the Clinical Practice Research Datalink (CPRD) and the Hospital Episodes Statistics (HES) database. The HES will also provide data on the hospitalisation outcomes of interest. Using data from a national survey of EIP teams, we will derive a series of aggregate, time-varying exposures (number of EIP teams) and effect measure modifier (fidelity) variables, at the level of the SHA, each year. We will use fixed-effects Poisson and linear regression models to estimate the effects of EIP on the two hospitalisation outcomes and examine effect measure modification by EIP fidelity level in the SHAs. The fixed-effects approach allows us to control confounding by often difficult or impossible to measure stable, region-level characteristics. This research will estimate the impact of EIP services on patient hospitalisations across England and provide insight into how fidelity of EIP services influences these outcomes. This in turn, will be useful in informing service delivery and policy at a national level.
Health Outcomes to be Measured: 
Absolute difference in hospitalisation rate; mean days (or weeks) of hospitalisation
Application Number: 

Samy Suissa - Chief Investigator - McGill University
Rebecca Fuhrer - Corresponding Applicant - McGill University
Ashok Malla - Collaborator - McGill University
David Stephens - Collaborator - McGill University
Miriam Kinkaid - Collaborator - McGill University
Sam Harper - Collaborator - McGill University
STEPHEN MCGOWAN - Collaborator - NHS England

HES Admitted Patient Care