Investigating the real-world safety and effectiveness of medications is an important, post-licensing, stage of drug development. These studies give us more in-depth knowledge, especially surrounding the long-term and rare effects of medications, and help us to characterise patient experiences outside the controlled setting of a clinical trial. Frequently, we will want to compare users of a newly licensed drug to users of an older drug which is prescribed for similar reasons.
Traditionally, studies of this type are designed as follows. We identify a population of patients who have no prior usage of the two drugs and compare initiators of the newer drug to initiators of the older drug. Patients are then followed up and the risk of a particular outcome compared between the two groups.
However, this approach excludes a large number of initiators of the newer drug that previously received the older drug. This can lead to concerns surrounding the representativeness of the study population compared to the usage of the drugs in practice. A recently proposed study design aims to address this limitation and explicitly accounts for patients who switch from the older drug to the newer drug.
In this study, we will provide an assessment of the results obtained from the traditional design against variations of the newly developed study design. This will help to inform us of the potential for this design to be used in future studies.
The active comparator new user (ACNU) study design has become the gold-standard for conducting cohort studies to assess the real-world safety and effectiveness of medications 1. The attractiveness of the ACNU approach is largely due to the baseline washout period (mimicking a clinical trial) and using an active comparator to reduce confounding by indication.
However, one issue with the ACNU is the selection of a suitable comparator drug. Often the comparator is an older drug that has been on the market for a long time. In the patient population, this means that many new users of the study drug are not, in fact, treatment naïve but have instead switched to the newer drug from the old comparator. The ACNU would typically exclude those who switched from the comparator drug to the study drug and this can result in investigators mischaracterising the real world patient population. The newly proposed Prevalent New User (PNU) design aims to address this limitation by incorporating patients who have switched from the older drug to the newer drug.
In this study, we aim to provide an assessment of the PNU design and proposed variations of defining exposure sets compared to the existing ACNU design. Furthermore, we will investigate the use of high-dimensional propensity scores (hd-PS) for confounder adjustment in the context of PNU designs and provide initial guidance to investigators planning to combine these approaches.
This is a methodological study. We assess several study designs in the context of a cohort study, with the following outcome:
Upper gastrointestinal bleeding leading to hospitalisation or death. .
Daniel Gibbons - Chief Investigator - GlaxoSmithKline Research & Development Limited (UK)
John Tazare - Corresponding Applicant - GSK
Elizabeth Williamson - Collaborator - London School of Hygiene & Tropical Medicine ( LSHTM )
Ian Douglas - Collaborator - London School of Hygiene & Tropical Medicine ( LSHTM )
John Logie - Collaborator - GlaxoSmithKline Research & Development Limited (UK)
Liam Smeeth - Collaborator - London School of Hygiene & Tropical Medicine ( LSHTM )
Sanni Ali - Collaborator - London School of Hygiene & Tropical Medicine ( LSHTM )
HES Accident and Emergency;HES Admitted Patient Care;ONS Death Registration Data;Patient Level Index of Multiple Deprivation