Identifying trends and causes of hospitalisations in people with and without the top 5 chronic conditions: a population-based study

Date of Approval: 
2021-05-26 00:00:00
Lay Summary: 
The number of hospitalisations has increased substantially over the last decade, where nearly four out of ten UK hospital admissions are unplanned. Hospitalisations have social and health implications and are an expensive part of medical care, placing high burden on national resources. People with cancer, heart disease, stroke, lung disease, and liver disease are more likely to die than people without these conditions. It is estimated these five conditions are responsible for more than 115,000 deaths a year in England. Studies examining the trends of hospital admissions in people with these conditions are lacking. In our study, we will use anonymised data collected in general practices and hospitals in England to: firstly, describe the full characteristics (such as age, gender, ethnicity, residence region, other illnesses) of people with these five conditions and describe the trends of their hospitalisations (planned and unplanned admissions including COVID-19 admissions, A&E, and outpatient visits) by numbers, types, and causes, between 2006 and 2020; secondly, investigate if hospitalisations are linked to death. In all analyses, people with these five conditions registered in English general practices (cases), will be compared to a group of people without these conditions (comparators) with the same age, gender, and general practice. The expected results of our study will help find out which patient groups are at higher risks for first and recurrent hospitalisations and death. This can support primary care doctors and healthcare teams provide better care to people at greater need and help reduce future hospitalisations and unplanned admissions.
Technical Summary: 
Background/Aims Hospitalisations have increased by up to 79% from the last decade in England and are a considerable NHS strain, placing unprecedented pressures on healthcare resources. Nearly 40% of admissions are unplanned. The top five chronic conditions responsible for premature mortality in England (>115,000 deaths/year) are cancer, heart disease, stroke, lung, and liver diseases. However, studies examining the hospitalisation trends of people with these conditions are lacking. We aim to identify the socio-demographic and clinical characteristics of patients with and without these conditions, identify the trends and predictors of first and recurrent hospitalisations and mortality. Methods Using cohorts of people with each of these top five conditions between 2006-2020, matched to up to five comparators without any of these conditions on age, gender, and general practice. Descriptive analyses will be used to identify the socio-demographic (including ethnicity, social-deprivation) and clinical characteristics of the matched cohorts and the longitudinal trends, primary cause of admission, and type (planned or unplanned) of hospitalisations. These trends will be categorised by patient age; gender, history of past admissions, deprivation, geographic location, comorbidities, and ambulatory care sensitive conditions. Competing risk analysis will assess the time to first all-cause and condition-related hospital admission, A&E events and OP visits. Cox proportional hazards regressions will estimate the risk for recurrent hospitalisations and the association between hospitalisation (by type, cause, length of stay) and the risk of all-cause and condition-specific (including COVID-19) mortality. Expected impact The findings can be a proxy for primary care quality to inform a more targeted and effective pre-hospitalisation care delivery (at primary care level) to the high-risk patient groups with potential benefits towards providing focus on what actions are needed to reduce premature mortality and admissions in people with these five conditions to assist better use of limited NHS resources.
Health Outcomes to be Measured: 
• Annual trends of i) hospital admissions by count, causes of admission, and type (elective vs. non-elective), by the conditions reportedly accounting for three-quarters of Ambulatory care sensitive conditions (ACSCs) spells (as appropriate per each of the five conditions), and ii) A&E attendances and outpatient visits (in years where HES A&E and HES OP data are available), categorised by patient characteristics of the matched cohorts. • Risk for first and recurrent (readmissions within 30 days of hospital discharge) all-cause, condition-specific, and COVID-19-related hospital admissions, A&E attendances and outpatient visits. • Risk for all-cause and condition-specific mortality (including COVID-19-related deaths).
Application Number: 

Salwa Zghebi - Chief Investigator - University of Manchester
Salwa Zghebi - Corresponding Applicant - University of Manchester
Darren Ashcroft - Collaborator - University of Manchester
Dimitri Varsamis - Collaborator - NHS England
Evangelos Kontopantelis - Collaborator - University of Manchester
Mamas Mamas - Collaborator - Keele University
Martin Rutter - Collaborator - University of Manchester

2011 Rural-Urban Classification at LSOA level;HES Accident and Emergency;HES Admitted Patient Care;HES Outpatient;ONS Death Registration Data;Patient Level Index of Multiple Deprivation;Practice Level Index of Multiple Deprivation