Past Students

“Reducing potentially avoidable acute care transfers from long-term care homes: Evaluating intervention effectiveness and improving approaches for their analysis”, Dr. Deniz Cetin-Sahin, PhD Student. 

Background: Potentially avoidable acute care transfers from long-term care homes (CHSLDs) may occur contrary to residents’ advance directives, result in adverse patient outcomes, and represent inefficiencies in our health care system. As such, long-term care physicians and other frontline staff are increasingly being directed to reduce them. However, the current understanding of the effectiveness of interventions aiming to reduce these transfers is incomplete. As well, novel causal inference and data simulation methods have yet to be applies in this area although they offer wide-ranging potential in reducing design bias and increasing precision. 

Aim and Objectives: The aim of my doctoral dissertation is to conduct a thorough investigation of potentially avoidable emergency department transfers and hospitalizations from long-term care homes with a view of designing an observational study using causal inference methods.

Methods: I will conduct four sub-studies, which will yield four thesis manuscripts, to:

1. Conduct a comprehensive systematic review to synthesize the published evidence that assesses the effectiveness/efficacy of interventions that aim to reduce emergency department transfers and/or hospitalizations for long-term care home residents experiencing an acute health-status change.
2. Compare potentially avoidable emergency department transfers and hospitalizations among our CIUSSS CHSLD residents and describe these outcome measures for conditions that are, in theory, ‘clinically manageable’ in the CHSLD setting. In a cross-sectional study using real-world data pertaining to our residents who received care at the Jewish General Hospital emergency department, I aim to prioritize a transfer outcome measure for the proposed observational study.
3. Gain better understanding of the anticipated results and their sensitivity to violations of underlying statistical model assumptions. Using a data simulation approach, I aim to:
4. Assess the potential magnitude of biases and performance of the statistical inference models used to analyze the observational study.
5. Estimate the statistical power and minimum sample size needed to detect clinically relevant effect sizes.
6. Develop an observational study protocol applying a causal inference framework to estimate the average causal effect of a selected exposure existing in real world (e.g., advance care planning) on reducing the prioritized outcome measure, using the findings of the 3 substudies.

Expected Results: My thesis project will be the first to demonstrate designing an observational study informed by a systematic review, a cross-sectional study, and data simulations, through the lens of a causal inference framework. Deliverables emanating from this dissertation will provide new and important insights for long-term care stakeholders, policy makers, and researchers. Future steps will involve executing the study protocol, and moving the results into recommendations for action through deliberative dialogues with long-term care stakeholders.