Retrospective cohort analysis using Clinical Practice Research Datalink (CPRD) linked to Hospital Episode Statistics (HES) to investigate adherence to androgen deprivation therapy (ADT) in patients with prostate cancer

Study type
Date of Approval
Study reference ID
Lay Summary

Prostate cancer is one of the most common cancers diagnosed worldwide and a leading cause of cancer related deaths in men. Men with advanced prostate cancer may be prescribed Androgen Deprivation Therapy (ADT), which lowers levels of the hormone testosterone, aiming to slow or stop growth of prostate tumours. There are two main types of ADT used in the UK for treating prostate cancer, long-acting luteinising hormone-releasing hormone (LHRH) agonists and gonadotropin-releasing hormone (GnRH) antagonists.

This study aims to describe current experiences of patients receiving ADT in the UK by type (LHRHs and GnRHs). Specifically, how frequently people with prostate cancer receive ADT prescriptions, if this is different from the plan for their treatment, and whether they experience delays, interruptions or stop their treatment early. The study will also describe the disease characteristics of people with prostate cancer receiving ADT, and how often they receive medical care after starting their ADT treatment. Finally, the study will also describe patients’ experience on ADT treatment before, during and after the COVID-19 pandemic.

Understanding treatment patterns and associated medical care received by patients on ADT will have a public health benefit in helping to describe any unmet need in the treatment of patients with prostate cancer, including understanding whether there are issues with treatment adherence, preventing patients from taking ADT medication as prescribed.

The study will use electronic health records from Clinical Practice Research Datalink linked to data from Hospital Episode Statistics (HES) and the Office for National Statistics (ONS).

Technical Summary

This study aims to describe treatment patterns, safety outcomes and healthcare resource use (HCRU) in adult patients with prostate cancer, treated with an Androgen Deprivation Therapy (ADT). Our study population will have a diagnosis of prostate cancer at any time and a first prescription for an ADT (injectable LHRH agonist or GnRH antagonist) from 1st April 2015 in CPRD.

Our primary exposure will be treatment with an ADT. CPRD data will be used to assess patient demographics, treatment patterns, safety outcomes and primary-care HCRU. Linked HES data will be used to assess safety outcomes and secondary-care HCRU (hospital admissions, outpatient visits, relevant procedures, and emergency visits).

This is a retrospective cohort study. Our analysis will be presented overall, by ADT agent and formulation and stratified by pre-, during- and post- COVID pandemic periods. Patients will be followed-up from first ADT prescription.

Descriptive statistics will describe:

-Patient characteristics (e.g., prostate-specific antigen [PSA] levels); comorbidities and concomitant treatments
-Treatment patterns, including patient adherence to ADT proxied by proportion of days covered (PDC), treatment delays, interruptions, discontinuation and switching, each defined by pre-specified time gaps within which another ADT treatment is not prescribed. Only CPRD based prescription information will be available for this analysis
-Safety outcomes of interest (e.g. hypertension, bone health)
-All-cause and prostate cancer-specific HCRU

Cox proportional hazard modelling will be used to describe incidence of major adverse cardiovascular (MACE) and other cardiovascular-related safety outcomes of interest.

Using CPRD-HES will provide real-world evidence of the usage and outcomes following treatment with ADT in prostate cancer patients in England. Understanding treatment patterns and associated HCRU of patients on ADT will have a public health benefit in helping to describe any unmet need in the treatment of patients with prostate cancer that may be leading to treatment disruptions or adverse outcomes.

Health Outcomes to be Measured

The below will be measured overall and stratified by COVID-period (patients assigned to COVID-period by date of patients’ first ADT injection prescription [index])

Treatment patterns:
-Length of follow-up time, average number of injections of an LHRH agonist or GnRH antagonist over follow-up time.
-Proportion of days covered by drug formulation,
-Treatment delays (continuous and categorical: <4 days (no delay), 4 – 6 days, 7 – 13 days, 14 - 27 days, ≥28 days) by drug formulation
-Treatment interruptions (% of patients with an interruption by type and frequency per patient, defined as a gap in treatment >6 months before another ADT injection is administered) by drug formulation
-Treatment discontinuation (% of patients with a discontinuation by type, defined as a gap >6 months with no further ADT therapy observed) by drug formulation
-Treatment switching (% that switch ADT) by drug formulation

Demographics and clinical characteristics:
age at diagnosis; age at index; year of index; ethnicity; socio-economic status (IMD quintile); NHS region of GP practice; metastatic status of prostate tumour at diagnosis; PSA levels; comorbidities (including obesity, anxiety depression); concomitant prescriptions to ADT (including anti-androgens, biphosphates, statins, diabetes medications, antiplatelet therapies, alpha-blockers, non-opioid analgesics, angiotensin-converting enzyme inhibitors, beta-adrenoreceptor blocking drugs, calcium channel blockers); death during the study period

Safety events:
Major Adverse Cardiovascular Events (MACE) (including Myocardial infarction, Non-fatal stroke, Cardiovascular death) and MACE-related adverse events (nervous system haemorrhages and cerebrovascular conditions), Composite MACE (first of myocardial infarction, non-fatal stroke or cardiovascular death); Safety events (including hypertension; non-fatal stroke; myocardial infarction; osteoporosis; fractures; urinary tract infections; impotence/erectile dysfunction, type II diabetes);

HCRU from index to follow up, all-cause and prostate-cancer specific (including primary care consultations; primary care prescriptions, primary care prescriptions for ADTs of interest, primary care laboratory tests of interest (including, blood pressure, lipid panels, HbA1c testing), secondary care prescriptions, inpatient admissions, length of inpatient stay, outpatient attendance, referral to cardiology specialist, accident and emergency visits, cost of primary care prescriptions, cost of all care, cost of prostate-cancer specific care.


Craig Currie - Chief Investigator - Pharmatelligence Limited t/a Human Data Sciences
Nick Denholm - Corresponding Applicant - Harvey Walsh Ltd
Emily Wilkes - Collaborator - OPEN Health Group
John Robinson - Collaborator - Harvey Walsh Ltd
Leah Fisher - Collaborator - Pharmatelligence Limited t/a Human Data Sciences
Mark McCormack - Collaborator - Harvey Walsh Ltd
Stephen Boult - Collaborator - Harvey Walsh Ltd


HES Accident and Emergency;HES Admitted Patient Care;HES Outpatient;ONS Death Registration Data;Patient Level Index of Multiple Deprivation