Fig. 3From: Comparison of Bayesian Networks, G-estimation and linear models to estimate causal treatment effects in aggregated N-of-1 trials with carry-over effectsThe DAG shows the assumed network structure between the nodes Activity (A), Treatment (T) and Nonspecific Low Back Pain (O) at time point t that is modeled in the Bayesian Network with time adjustment with lags. With that, the model is accounting for Activity as a confounding variable, although we do not simulate it. This is a snippet of the relationships between these three variables, and embedded in the bigger network in Fig. 1Back to article page