Although suicide is a leading cause of death and recognized to be a major and increasing public health problem, it is still a relatively rare event which occurs among 13 per 100,000 people, on average, in the United States each year. Frequently, the rarity of suicide events is compounded by the need to focus research efforts on homogenous populations and sub-populations that are at high risk, but are relatively small in size (e.g. Women Veterans, LGBT Youth, American Indian/Alaska Natives, etc.). In these cases, even when risk is higher than average, the number of events in any given study can be very small. The design, analytic approach, and potential for bias in such studies must be carefully considered upfront and when interpreting research findings. Join us for an overview of common challenges and biases that occur when working with rare outcomes, such as suicide, and learn about some of the analytic approaches for overcoming these challenges. Topics to be covered include, but are not limited to, computing and comparing rare event rates, optimal methods for covariate adjustment in studies with a small number of outcomes, pros and cons of using ‘intermediate’ outcomes (e.g. non-fatal self-harm events), and preferred study designs for working with rare outcomes.
At the conclusion of this webinar, attendees will be able to:
1. Explain at least three common challenges faced by investigators conducting research on rare outcomes
2. Determine when it is and is not appropriate to compute and report event rates for rare outcomes
3. Describe alternatives to standard logistic regression (maximum likelihood estimation) for analyzing studies where the number of observed events is small
Presented by: Claire Hoffmire, PhD
*A link to the Webinar will be sent via email once purchase information is confirmed