Event

Descriptive health inequality estimands in the presence of competing events

Quantifying social health inequalities (e.g. by socioeconomic position) in outcomes such as cancer fatality is a research priority but is complicated by competing events, such as death from other causes, which may also be subject to inequalities. Young et al. (Stat Med. 2020) used a counterfactual framework to clarify the interpretation, identifiability, and estimation of causal estimands in survival analysis with competing events. Using inequalities in colorectal cancer fatality as an example, this talk reviews three descriptive analogues and their value for describing health inequalities across interpretation, policy relevance, comparability, and identifiability: 1) Direct inequality: inequality in cancer fatality under a hypothetical intervention eliminating the competing event; 2) Total inequality: inequality in the presence of the competing event; and 3) Inequality in survivors: inequality among individuals who survive the competing event. The talk will also introduce a new estimand that treats the competing event as a mediator using the interventional effects framework. This interventional inequality estimand captures inequality under a hypothetical intervention that equalises the distribution of the competing event across socioeconomic groups to a user-specified distribution (e.g. that of the most advantaged group). Finally, the importance of defining the target population as a hypothetical population, in which groups share the same distribution of factors not considered sources of unfair differences, will be discussed to ensure estimands reflect genuine health inequalities.

Dr Ghazaleh Dashti is a Biostatistician within the Clinical Epidemiology and Biostatistics Unit (CEBU) at MCRI and the University of Melbourne. Her primary areas of research interest are developing and evaluating methods to better design and analyse observational studies addressing questions that would inform public health and clinical practice decision-making. Specifically, her research, currently supported by an NHMRC Investigator Grant focuses on causal inference and missing data methods. The substantive areas to which she contributes include cancer, mental health, adolescent health, transgender health, intergenerational trauma, and family violence.

Zoom

This seminar will be held via Zoom.

Please click this URL to start or join. https://monash.zoom.us/j/88922203623?pwd=6MSlUsNEYVaNPrhbOwINqxHuNp2vDI.1
Or, go to https://monash.zoom.us/join and enter meeting ID: 889 2220 3623 and passcode: 204213

Date and time

Thursday 28 May, 9:30am AEST
──────────