Non-alcoholic steatohepatitis (NASH) is a severe chronic liver disease that affects many people. The correct method to diagnose NASH is a liver biopsy, an invasive method where a small piece of the liver is removed with a needle. A liver biopsy is painful, risky and costly. Therefore, NASH is not diagnosed in many patients. Consequently our knowledge on NASH is limited. To better help patients with NASH, we need more knowledge about the disease. An important way to increase our knowledge, is research based on data from NASH patients. However, good quality data is essential for such studies.
We will study the quality of the NASH diagnosis in a general practice database (CPRD Aurum) by comparing it with NASH diagnoses in a hospital database (the Hospital Episode Statistics and the Office for National Statistics Death Registration Data).
We hope that the study can tell us how good NASH diagnoses are in the general practice database. Such knowledge is important to better help patients with NASH.
Non-alcoholic steatohepatitis (NASH) is a growing public health problem. NASH is diagnosed from a liver biopsy (gold standard), which is invasive and costly. Many NASH patients are therefore not diagnosed. Accordingly, data and knowledge about NASH are limited. This hampers medical and societal efforts to reduce the increasing burden of NASH. Therefore, it is essential to identify high quality data for studies of NASH. NASH is often managed by general practitioners and primary care data could be a valuable data source for NASH studies. The quality of such data is largely unknown and also questionable as biopsies needed for diagnosis is performed in secondary care.
The study objective is to describe the quality of the NASH diagnosis recorded in a primary care database (Clinical Practice Research Datalink (CPRD) Aurum) as compared to diagnoses registered in a secondary care database.
The study population is adult patients with a NASH diagnosis in one of the data sources CPRD Aurum, the Hospital Episode Statistics (HES) or the Office for National Statistics Death Registration Data (ONS) between 2010-2021
Aims are to investigate:
- Correctness of the NASH diagnosis in CPRD Aurum as compared with diagnoses in HES and ONS. Correctness is calculated as the proportion of patients with NASH in CPRD Aurum who also has NASH in HES or ONS (primary aim)
- Completeness of the NASH diagnosis in CPRD Aurum as compared with diagnoses in HES and ONS. Completeness is calculated as the proportion of patients with NASH in HES or ONS who also has NASH in CPRD Aurum (secondary aim)
-Potential reasons for discordance between the data sources by describing the NASH populations according to demographic and clinical characteristics (exploratory aim)
The correctness (validity), completeness (presence) and reasons for discordance will be suggestive of the quality of NASH diagnoses in CPRD Aurum(1)
The first outcome to be measured is the correctness of the NASH diagnosis recorded in CPRD Aurum. It will be calculated as the proportion of patients with a diagnosis of NASH recorded in CPRD Aurum, who also has NASH recorded in HES (hospital diagnosis) or ONS (cause of death).
The second outcome to be measured is the completeness of the NASH diagnosis in CPRD Aurum. It will be calculated as the proportion of patients with a NASH diagnosis recorded in HES (diagnosis) or ONS (cause of death), who also has NASH recorded in CPRD Aurum.
Tina Landsvig Berentzen - Chief Investigator - Novo Nordisk A/S
Tina Landsvig Berentzen - Corresponding Applicant - Novo Nordisk A/S
Miriam Martello - Collaborator - Novo Nordisk A/S
Rikke Baastrup Nordsborg - Collaborator - Novo Nordisk A/S
HES Admitted Patient Care;HES Outpatient;ONS Death Registration Data