Moving beyond prevalent exposure designs for causal inference in dementia research.

We found that most (72\%) recent studies of dementia prevention interventions used a study design that is ill-suited for causal inference. The target trial framework provides an alternative approach that can improve the robustness and relevance of dementia prevention evidence.

What is the prevalent exposure design?
Exposure is initiated before the start of follow-up, and no design or statistical adjustment is made for past exposure.

How common is it?
We systematically searched Embase and MEDLINE for cohort studies:
published from 2018 to 2023
investigating selected risk or protective factors (physical activity, hearing aids, antihypertensives, antidiabetics, antidepressants)
using dementia or cognitive decline as the outcome
72% of identified studies (84 of 116) used this design approach.

Why does it matter?
Prevalent exposure designs are vulnerable to several problems.
Ill-defined intervention: no clear start time or duration.
Confounding adjustment difficult: baseline covariates may be confounders, mediators, or both.
Selection bias: pre-study exposure can influence who is selected into the study.
The sedentary group is ``depleted of susceptibles'' before follow-up begins.

A better approach: the target trial framework
specify the protocol for the hypothetical randomised trial that would answer your research question (the ``target trial'').
use observational data to emulate the target trial.

Employing the target trial approach aligns observational analyses with the design and analysis principles of randomised trials:
Explicit causal question
Transparency around required assumptions
Clear timing and duration of strategy under study
Start of follow-up aligned with initiation of that strategy
Avoid common biases and more relevant evidence for decision-makers.