Despite hospitals capturing vast amounts of data on their patients, healthcare systems in England know remarkably little about which patients have which diseases, in which combinations, and what the outcomes are for patients. When a doctor sees a patient, (s)he will record information about the patient’s disease in a computer system using a series of codes. However, diseases are not coded in the same way in general practice and hospitals, making research challenging. Recent opportunities for accessing national hospital data for patients stored electronically, in a safe and secure way, has allowed researchers to generate entirely new information about all diseases. We will build on research carried out with national data by using data collected on patients with one or more diseases in general practice, linked to hospital data. Our research aims are to use the coding system adopted in general practice to define diseases to help researchers understand whether this form of coding captures additional diseases compared to codes used in hospitals; and analyse data for diseases defined used the GP coding system to provide novel information about how often diseases occur, which diseases occur together, what outcomes are for patients, and use of medications and procedures. This new body of knowledge will be made available to patients, doctors, researchers and people involved in guiding healthcare decisions in a ‘Disease Atlas’ to aid understanding of what diseases patients have, or are at risk of, so healthcare needs can be addressed and adopted into policy to guide treatment.
Background: Systematic approaches to studying diseases in populations have focused on common diseases (eg Global Burden of Diseases) but have not considered less common and rare diseases, nor considered how diseases co-occur in people or polypharmacy.
Objectives: To develop a Disease Atlas across linked primary–secondary care data.
Methods: as in previous CPRD approved protocols (12_153, 16_022 and 20_074), we will use all patients, all ages with up to standard linked data. A participant will enter a disease cohort on the date of first occurrence of a SNOMED-CT disorder term or an ICD-10 disease term. From that date and within this period, we will follow back to a patients first record (for co-occurrence), and follow forward for 365 days for death and hospitalisations. We will estimate prescribed drugs within 12 months of the first record of disease. Specifically we will define diseases using SNOMED-CT disorder terms in UK primary care alone or in combination with ICD-10 in HES APC and A&E. We will estimate estimates from primary care data mirroring those previously generated in secondary care data in 56m population. For each disease these estimates include frequency, co-occurrence, prognosis and use of medications and procedures. For diseases which are sufficiently common we will develop and validate parsimonious prognostic models. We will determine the extent to which SNOMED-CT data offers differences in understanding beyond ICD-10 comparing for example in disease resolution: to what extent does SNOMED-CT capture additional diseases, beyond those in ICD-10? We will also investigate patterns of disease co-occurrence: how do they differ among people who have and do not have hospital data?
Outputs: As well as publishing papers in peer-reviewed journals, we will work with stakeholders (patients, clinicians, policymakers) on the development, testing and launch of a public-facing website to communicating relevant information from the Atlas.
For each disease we will estimate outcomes at 1 year:
Mortality, all cause and cause-specific
Hospital admissions, all cause and cause specific
Harry Hemingway - Chief Investigator - University College London ( UCL )
Natalie Fitzpatrick - Corresponding Applicant - University College London ( UCL )
Alvina Lai - Collaborator - University College London ( UCL )
Amitava Banerjee - Collaborator - University College London ( UCL )
Ana Torralbo - Collaborator - University College London ( UCL )
Andrej Ivanovic - Collaborator - University College London ( UCL )
Arturo Gonzalez-Izquierdo - Collaborator - University College London ( UCL )
Christopher Tomlinson - Collaborator - University College London ( UCL )
Johan Thygesen - Collaborator - University College London ( UCL )
Laura Pasea - Collaborator - University College London ( UCL )
Muhammad (Ashkan) Dashtban - Collaborator - University College London ( UCL )
Muhammad Qummer ul Arfeen - Collaborator - University College London ( UCL )
Spiros Denaxas - Collaborator - University College London ( UCL )
HES Accident and Emergency;HES Admitted Patient Care;ONS Death Registration Data;Patient Level Index of Multiple Deprivation