This study aims to assess the feasibility of describing the clinical (health) and economic (cost) burden of haemophilia in the UK using CPRD-HES linked data.
Haemophilia is an inherited disorder in which the blood does not clot due to insufficient clotting factors. Haemophilia types A and B affect only males.
We wish to look at the data recorded in GP appointments and hospital visits for people who have haemophilia (type A or B) to see whether we can get a good idea of how haemophilia affects people’s health and how often they have to see a GP, consultant, nurse or other healthcare professional (e.g. physiotherapist) or have tests or admissions to hospital. Sometimes in databases, this information is not well-recorded, which means we would not be able to use the data to get a real understanding of haemophilia. We want to be able to know more about the condition and how it affects people, how many people it affects, and how it impacts on the NHS. We want to know if there is information in this database which is important for showing that a treatment has an impact on a person’s health and on the NHS. To do this, we will show whether important information is recorded and if so, how much is being recorded, to give us an idea of whether the database could show the full picture of living with haemophilia.
Future study: To understand the clinical and economic burden of haemophilia A and B in England.
Feasibility study: To determine the completeness of recording of clinical outcomes and healthcare resource use (HCRU) in haemophilia (A and B).
Aim: To understand whether key clinical and economic outcomes are captured in UK primary and secondary care data to enable future use of such data for haemophilia research
Objective: Explore the recording of haemophilia, its treatments, and specific clinical and economic outcomes
Exposure: Patients with haemophilia A or B
Methods:
Patients with a record of haemophilia A or B with follow-up between 1 Dec 2009 and 31 Dec 2019 will be included. Index date is the later of date of first diagnosis and 1 Dec 2009.
The feasibility cohort will be described in terms of patient characteristics, key clinical outcomes and HCRU.
• Patient characteristics at index date will include age, Index of Multiple Deprivation (via the linked patient-level IMD dataset), age at diagnosis, time since diagnosis and length of follow-up, and will be summarised with descriptive statistics.
• Key clinical outcomes will be largely obtained through HES Admitted Patient Care and Outpatient data (see next section). Indicator variables will be created to denote whether such data are recorded at any point in a patient’s follow-up; this data will be summarised with descriptive statistics on frequency of recording.
• HCRU will be calculated from CPRD by estimating GP visits and medication use, and (from HES), inpatient, outpatient and A&E attendances.
• The top 20 most commonly recorded clinical events (by ICD-10, OPCS and Read code) and prescriptions will be generated.
The results will be reviewed by a clinician to gauge whether the estimates align with expectations. The future study would replicate the feasibility analyses (adjusted following learnings from the feasibility study) using the full CPRD-HES sample.
• Patient population (to understand patient characteristics in this dataset):
o Number with at least 1 year’s data pre-diagnosis
o Type of haemophilia (A or B)
o IMD
o Length of follow-up
o Age at diagnosis
o Time since diagnosis
o Commonly recorded clinical events
• Key Outcomes (to explore the presence and completeness of data):
o Annualised infusion rate
o Factor replacement usage (VIII for Haem A and IX for Haem B)
o Inhibitor to factor replacement
o Factor levels
o Bleeding events
o Imaging (joint and muscle MRIs, joint and muscle ultrasounds, skeletal X-rays, and Peterson scores; CT and MRI head scans)
o Surgical procedures on joints (arthroscopy, arthrodesis/joint fusion, joint replacement)
o Joint Health (Hemophilia Joint Health Score instrument, HJHS)
o Patient Reported Outcome measures (Haem-A-QoL, Haemophilia Activities List [HAL])
• HCRU (to compare rates with external registry):
o Number of GP visits
o Number of outpatient appointments (all-cause and haematology)
o Number and type of inpatient stays
o Number of A&E attendances
o Common prescriptions
Linked datasets:
HES Admitted Patient Care: Will be used to understand clinical outcomes such as major bleeds and HCRU for inpatient stays
HES Accident and Emergency: Will be used to understand HCRU related to A&E
HES Outpatient: Will be used to understand clinical outcomes such as infusion in outpatient clinics
HES Diagnostic Imaging Dataset: Will be used to understand joint outcomes
Patient Level Index of Multiple Deprivation: Will be used to characterise patients’ socio-demographic status
This use of linked health care data will be of benefit to patients with haemophilia to be able to accurately capture the impact of their disease without the difficulty and bias that can be associated with self-reported information which is often the method of data collection for these patients.
Note that to minimise the risk of a patient being re-identified due to small numbers resulting from the outcomes, the results will be obscured (converted to N/A) if the number of patients in a table cell is less than 5. Furthermore, we aim to minimise this risk by requesting data from both CPRD GOLD and Aurum, to attain higher patient counts in this feasibility study.
Shuk-Li Collings - Chief Investigator - Pfizer Ltd - UK
Shuk-Li Collings - Corresponding Applicant - Pfizer Ltd - UK
Graciela Sainz de la Fuente - Collaborator - Pfizer Ltd - UK
Ione Woollacott - Collaborator - Pfizer Ltd - UK
Jim Thomson - Collaborator - Pfizer Ltd - UK
Kate Halsby - Collaborator - Pfizer Ltd - UK
Lewis Carpenter - Collaborator - Not from an Organisation
Lucy Richardson - Collaborator - Pfizer Ltd - UK
HES Accident and Emergency;HES Admitted Patient Care;HES Diagnostic Imaging Dataset;HES Outpatient;Patient Level Index of Multiple Deprivation