From the Guidelines
Controlled interrupted time series is a quasi-experimental study design that evaluates the impact of an intervention by comparing trends before and after implementation, using a control group to account for external factors and time trends. This approach is particularly useful in situations where randomization is not feasible or ethically not acceptable, as noted in a 2020 study published in Clinical Microbiology and Infection 1. The design involves collecting multiple observations over time, with clear pre-intervention and post-intervention periods, allowing researchers to assess whether the intervention caused a significant change in the outcome of interest beyond what would have occurred naturally.
Key characteristics of controlled interrupted time series include:
- Use of a control group that did not receive the intervention to account for external factors
- Collection of multiple observations over time, with clear pre-intervention and post-intervention periods
- Analysis that accounts for time trends, such as segmented regression analysis and adjustment for autocorrelation
- Reporting of immediate effects on outcome and trends before and after the implementation, and assessment of whether trends are non-linear, as recommended by the working group in the 2020 study 1
The controlled interrupted time series design is valuable in health services research and policy evaluation because it can establish causality more convincingly than simple pre-post comparisons by controlling for secular trends, seasonal patterns, and other confounding factors that might affect outcomes over time. For example, if studying the effect of a new hospital policy on infection rates, researchers would collect infection data for several time points before and after implementation at both the hospital implementing the change (intervention group) and similar hospitals not implementing it (control group), as described in the 2020 study 1.
From the Research
Definition of Controlled Interrupted Time Series
- Controlled interrupted time series is a quasi-experimental design that involves a before-after comparison within a single population, as well as a comparison with a control group 2.
- This design is used to evaluate the impact of interventions or programs implemented in healthcare settings, and can help to minimize potential confounding from simultaneous events 2.
Strengths and Limitations
- The controlled interrupted time series design has the advantage of limiting selection bias and confounding due to between-group differences 2.
- However, it requires careful consideration of potential confounding events and the choice of control group, as different controls can have associated strengths and limitations 2.
- The design can be used to evaluate outcomes using population-level data, and can provide clear graphical presentation of results 3.
- Limitations include the need for a minimum of 8 time periods before and 8 after an intervention to evaluate changes statistically, and difficulty in analyzing the independent impact of separate components of a program that are implemented close together in time 3.
Types of Controls
- A range of different types of controls can be used with interrupted time series designs, including synthetic controls 4.
- Synthetic control methodology is a data-driven technique for deriving a control series from a pool of unexposed populations, and can be used to strengthen an interrupted time series design 4.
- However, the use of synthetic controls may not always nullify important threats to validity nor improve causal inference 4.
Statistical Methods
- Multiple statistical methods are available to analyze data from interrupted time series studies, including segmented linear regression 5.
- The choice of statistical method can importantly affect the level and slope change point estimates, their standard errors, width of confidence intervals and p-values 6.
- Pre-specification of the statistical method is encouraged, and naive conclusions based on statistical significance should be avoided 6.