Figure 4: Possible relationship between a measured confounder (years of opioid dependence), prenatal opioid agonist therapy, risky behaviour, gestational age at birth and neonatal abstinence syndrome (NAS) severity.