Grant: Banting Postdoctoral Fellowship Program

Network meta-analysis (NMA) is the only approach to formally compare the efficacy of all available treatments. However, many NMAs fail to account for differences in effects within different groups of individuals such as those who are older or those at different stages of illness. A key assumption in NMA is that trials are sufficiently similar in their design (including patient characteristics), called consistency. A different distribution of an effect modifier (e.g. patient age) across treatment comparisons can cause inconsistency questioning NMA validity. For example, while oral anticoagulants are effective in reducing stroke in all patients with non-valvular atrial fibrillation, older patients (i.e. >75 years) are at the highest risk of a stroke and achieve greater benefit than patients <75 years. Similarly, this older age group is at a higher risk of bleeding with these agents. These types of analyses are optimally informed by the use of individual patient data (IPD) where data on each patient enrolled in the included trials is combined. NMA is commonly performed with aggregated data (AD) where the summary results of included trials are combined, and the value of IPD in NMA is currently unknown. The IPD NMA can increase power, and identify interactions not detected with AD. However, further research is needed to establish the benefits and properties of IPD in networks of interventions. My objectives are to: 1) establish and advance IPD NMA methods, and 2) make NMA more useful and applicable for clinical practice. My results will be used to answer whether the same treatment effect is applicable to all patients, or whether the treatment effect for patients in population A (e.g. >75 years) is greater than for patients in population B (e.g. < 75 years), or whether intervention 1 is more effective for patients A and intervention 2 is more effective for patients B. My research will inform decision making by providing contextualized information tailored to the patient.