Most people with a health condition are excluded from research trials that examine how well a medicine works to treat their condition. Common reasons to be excluded from such research trials are often based upon age, gender, people having other health conditions (comorbidity), and people prescribed other medicines (co-prescribing). Clinical guidelines are commonly used by doctors and other clinicians to guide treatment decisions, but most guideline development is focused on evidence for a single disease, and does not usually explicitly consider whether or not patients have other conditions (comorbidity) or are already being prescribed other medicines (co-prescribing). A key reason for this is that stakeholders do not have access to important information about the clinical population they are making recommendations for. We will examine how treatment effects potentially differ in people that have different numbers of chronic conditions and are prescribed different numbers of medicines including how this varies. We will then examine how people excluded from key research trials differ from people included in trials in terms of the effects of treatments used to manage their condition. We will work stakeholders to explore how summaries of this information could be used for decision making.
Trial eligible and trial ineligible patients have different characteristics that might impact on the effects of treatment. We will perform cohort studies among patients using specific treatments identified in CPRD. Patients will be identified as being eligible or ineligible for the trial using that treatment. A selection of trials will be chosen to cover a range of common chronic diseases involving arthritis, pulmonary disease, diabetes, gout, and cardiovascular disease. Patients in CPRD will be new users of the treatments studied in their corresponding trial. The trials will cover several anticoagulants, anti-hyperglycaemic agents, anti-hypertensives, bronchodilators and antiplatelet medications. We will estimate treatment effects in the cohorts for outcomes studied in the trial. We will use propensity score methods to address confounding control and survival analysis to estimate hazard ratios for each outcome of interest. This will be done in the trial eligible population and the trial ineligible population separately. We will then compare the estimates from trial eligible and ineligible patients to determine which align and which differ. Summary information will then be produced and shared with stakeholders through reports.
The key measure is understanding whether treatment effects, as measured through cohort studies using CPRD data, differ between trial eligible and ineligible populations. We have therefore focused the outcomes as below.
Outcomes related to each clinical trial are fully listed in the Appendix Table 3 but will include:
a) all-cause mortality;
b) fatal disease-specific events (e.g. cardiovascular disease);
c) non-fatal disease-specific events (e.g. acute myocardial infarction, stroke, TIA, hospitalisation);
d) disease exacerbations (e.g. gout, COPD, asthma);
e) adverse events of treatment (e.g. hospitalisation rates, pneumonia, bleeding, acute kidney injury).
Please see Appendix for further details.
Daniel Morales - Chief Investigator - University of Dundee
Daniel Morales - Corresponding Applicant - University of Dundee
Bruce Guthrie - Collaborator - University of Edinburgh
David Moreno Martos - Collaborator - University of Dundee
Huan Wang - Collaborator - University of Dundee
SHONA LIVINGSTONE - Collaborator - University of Dundee
HES Admitted Patient Care;ONS Death Registration Data