We are studying the risk factors for healthcare-associated infections (HAI), considering both patient and hospital-level variables.
To achieve as more as possible a causal interpretation of the results, we are wondering how should we manage variables that can be both a cause and consequence of the outcome, both when analyzing the risk associated with those variables and when considering them as confounders or moderators.
Our data is cross-sectional, from a prevalence study.
For example, let's say we want to evaluate the contribution of some HAI control programs on the patient HAI risk: - should we include it as a confounder/moderator when assessing other patient-level risk factors? - how should we do assess the program causal effect on the HAI risk?