There is evidence that people with severe mental illness (SMI) have a higher chance of dying from infection than people without SMI. However, very little is known about what types of infection drive this risk. We also don’t know much about how and why risk of death from infection is increased in SMI. This project will use largescale electronic health records (CPRD) to help understand infectious disease mortality in people with SMI. Moreover, it will also try to understand why this risk is increased by looking at the following factors:
(i) Physical health – are there certain combinations of physical conditions (e.g., diabetes, heart disease) and health behaviours (e.g. smoking) which increase risk of death from infection in people with SMI?
(ii) How do sociodemographic factors (e.g., age, gender, ethnicity), socioeconomic factors (e.g. neighbourhood deprivation), and indicators of barriers to healthcare often experienced by people with SMI interact to increase risk of death from infection?
(iii) Does the use of antipsychotic medication affect infection mortality rates in SMI?
Answering these questions will help guide the development and implementation of effective interventions, will inform health policy, and will improve the delivery of health services. Together, this will contribute to a reduction in premature death from infection in people with SMI.
Research questions (RQs)
1. Does SMI selectively increase mortality risk for specific infection types?
2. What SMI-related factors associate most strongly with infectious disease mortality?
3. Do certain combinations of physical long-term conditions and health behaviours increase risk of death from infection in SMI?
4. To what extent do sociodemographic factors and indicators of barriers to healthcare affect infectious disease mortality in SMI?
5. How do antipsychotics affect infection mortality rates in SMI?
We will carry out a retrospective matched cohort study where patients with SMI will be matched to patients without SMI at a ratio of 1:4 using propensity score matching. Patients will be matched on age, sex, primary care practice, and year of primary care practice registration.
The primary outcome of the study will be death from infectious disease. Linkage with Office of National Statistics (ONS) mortality data will provide date and cause of death. Linkage with Hospital Episode Statistics (HES) will allow us to include hospitalisations for infection as a secondary outcome.
We will use Cox proportional hazards regression to examine whether SMI increases mortality risk for specific infection types (RQ1), and to assess what SMI-related factors (e.g., diagnosis) associate most strongly with infectious deaths (RQ2).
We will adjust for covariates not used for control matching. These will include ethnicity, marital status, Index of Multiple Deprivation, smoking status, body mass index (BMI), alcohol and substance misuse, and number of physical long-term conditions (LTCs) (e.g., cancer, diabetes).
Using advanced statistical techniques (e.g., cluster analysis), we will ascertain whether certain combinations of LTCs and health behaviours (e.g., smoking) increase risk of death from infection among people with SMI (RQ3). We will use Cox regressions to estimate associations between these groups of LTCs and infectious disease mortality.
Please note CPRD will not be used to answer RQs 4 and 5.
Infectious disease mortality from linkage with ONS mortality data (primary outcome); hospitalisations for infectious disease from linkage with HES hospital admission data (secondary outcome). (RQs 1, 2, and 3)
Deaths and hospitalisations from infectious disease will be identified using the following ICD-10 codes: central nervous system (CNS) infections (A17, A80-A81, A85-A89, B00.3-B00.4, B01.0-B01.1, B02.0-B02.2, B05.0-B05.1, B06.0, B2.61-B26.2, G00-G01, G02.0, G03, G04.2, G05.0-G05.1, e.g., meningitis, viral encephalitis), gastrointestinal infections (A00-A05, A08, e.g., salmonella, shigellosis), liver infections (B15-B19, e.g., hepatitis A), respiratory infections (A15-A16, A36-A38, J00-J06, J09-J18, J20-J22, e.g., pneumonia, laryngitis), sepsis (A40-A41, e.g., streptococcal sepsis), skin infections (A46, B00-B09, L00-L05, L08, e.g., cellulitis, measles), urogenital infections (N30.0, N39.0, N41.0–41.1, N71-N72, e.g., cystitis, prostatitis), and other infections (A18-A19, A31-A32, A39, A42-A44, A48-A49, B25-B27, B30, B33-B34, B95-B99, H62.0-H62.1, H67.0-H67.1, M00, M01.0-M01.5, N61, e.g., bone infection, mastitis). We also plan to look at COVID-19 deaths and hospitalisations (U07.1, U07.2).
We hypothesise that certain infectious diseases might emerge as primary endpoints in this study. These include influenza (J09-J11), pneumonia (J12-J18), and sepsis (A40, A41).
SMI subtype (bipolar disorder; schizophrenia; other psychoses) will be determined using Read and SNOMED codes (see ‘SMI’ code list in Appendix). (RQ2)
The presence of mental health comorbidities (depression, anxiety, eating disorders, personality disorders, learning disability) will also be determined using Read and SNOMED codes (see example ‘Depression’ code list in Appendix (Please note we were unable to amend the Appendix section at this time, but can forward these lists on request)) (secondary outcome). (RQ2)
SMI duration will be determined using the date the primary SMI diagnosis was first recorded in primary care records (secondary outcome). (RQ2)
Physical long-term conditions and health behaviours (smoking status, body mass index, alcohol misuse, substance misuse) will be determine using Read and SNOMED codes (see example ‘LTC’ code list in appendix (Please note we were unable to amend the Appendix section at this time, but can forward these lists on request)). (RQ3)
Please note that CPRD will not be used to answer RQs 4 and 5.
Amy Ronaldson - Chief Investigator - King's College London (KCL)
Amy Ronaldson - Corresponding Applicant - King's College London (KCL)
Alex Dregan - Collaborator - King's College London (KCL)
Ioannis Bakolis - Collaborator - King's College London (KCL)
Jayati Das-Munshi - Collaborator - King's College London (KCL)
Temi Lampejo - Collaborator - King's College London (KCL)
HES Accident and Emergency;HES Admitted Patient Care;ONS Death Registration Data;Patient Level Index of Multiple Deprivation