Skip to main content

Table 1 Model for a controlled interrupted time series analysis

From: Assessing the impact of a national clinical guideline for the management of chronic pain on opioid prescribing rates: a controlled interrupted time series analysis

Y = β0 + β1 * Time + β2 * Intervention status + β3 * Intervention status * Time + β4 * Cohort status + β5 * Cohort status * Time + β6 * Cohort status * Intervention status + β7 * Cohort status * Intervention status * Time

Y is the outcome (prescribing rate).

β0–3 are coefficients representing the control cohort (e.g. gabapentinoid series) where β0 is the intercept or value of the outcome at the start of the study period, β1 is the change in outcome per unit time (trend) before the intervention, β2 is the immediate step change in level following the intervention and β3 is the change in trend following the intervention (relative to the trend before the intervention – β1). β1 and β3 can therefore be summed to provide the trend following the intervention. β0–3 can also be used in isolation as a standalone model for a single interrupted time series analysis.

β4–7 are coefficients representing the difference between the case series (e.g. opioid series) and the control series. β4 is the difference in intercept level, β5 is the difference in trend before the intervention, β6 is the difference in immediate change in level following the intervention and β7 is the difference in change in trend following the intervention. β5 and β7 can be summed to provide the difference in trend following the intervention.

Time, intervention status and cohort status relate to variables in the dataset representing the time elapsed since the start of the study period, the pre- or post-intervention period and the time series assignment (e.g. opioids or gabapentinoids).