Sepsis is caused by an overreaction of a patients body to an infection. It can result in widespread damage to organs in the body and in some cases can be fatal. Around 250,000 people develop sepsis every year in the UK, causing over 50,000 deaths. There are a number of different factors that make people more or less likely to develop sepsis, including age, gender and whether they already have conditions such as type 2 diabetes or heart disease. Most people develop sepsis in the community, however, there are a number of patients who contract it when they are already in hospital being treated for another condition. The project will use an anonymised dataset of patients in the UK who have a diagnosis of sepsis in their hospital records. We will access patients anonymous GP medical records to look at their medical history, including what other health conditions they have and what medications they have been prescribed. The group of sepsis patients will be compared to healthy patients who have not been diagnosed with sepsis. The aim of this project is to understand further the factors associated with developing sepsis, to see if there are differences between patients who develop sepsis in the community and patients who acquire it in hospital. We will assess whether patients receiving more prescriptions of antibiotics from their GP are more likely to develop sepsis. Additionally, we will look at the survival rate of sepsis patients and how many are readmitted to hospital longer term.
Around 250,000 people develop sepsis each year in the UK, with the majority contracting it in the community. It is a complex condition with a wide range of symptoms, making it difficult to recognise clinically. There are many risk factors associated with developing sepsis such as age, gender and multiple comorbidities. Much of the research into sepsis has involved single hospitals or ICUs, with few large population-based cohort studies performed.
The aims of this study are to use linked primary and secondary care data to look at risk factors associated with developing community-acquired and hospital-acquired sepsis, and to see whether increased antibiotic prescribing in primary care increases a patients risk of developing sepsis. The study will be a retrospective cohort study of patients with a hospital diagnosis of sepsis between 2000 and 2018. Patients diagnosed within 48 hours of admission to hospital will be defined as community-acquired, with those diagnosed more than 48 hours after admission classed as hospital-acquired sepsis. ICD-10 codes for sepsis will be used to extract the cohort in the Hospital Episode Statistics (HES) data and the primary care records obtained from the CPRD GOLD dataset. These will also be linked with Office for National Statistics mortality data. Comparator cohorts of patients without a sepsis diagnosis will be created, matched by age, gender, GP practice or duration of hospital stay prior to index date.
Outcomes assessed will be mortality rates, hospital length of stay, critical care length of stay and readmission rates, comparing the two groups of patients. Survival rates will be compared using Kaplan-Meier plots and modelled using Cox proportional hazards regression. Logistic regression modelling will be used to characterise risk factors for the development of sepsis, including age, gender, comorbidity and prescription data. Analyses will be conducted separately for children (age < 18) and adults.
In-hospital mortality; admission length of stay; critical care length of stay; all-cause mortality (1-year and longer term); mortality with sepsis as one of the causes of death; all-cause readmission rates (1-year and longer term); readmission with hospital diagnosis of sepsis.
Tjeerd van Staa - Chief Investigator - University of Manchester
Sian Bladon - Corresponding Applicant - University of Manchester
Darren Ashcroft - Collaborator - University of Manchester
HES Admitted Patient Care;ONS Death Registration Data;Patient Level Index of Multiple Deprivation