While randomized controlled trials (RCT) have long been considered the gold standard for evaluating the effectiveness of an intervention or a treatment, they are not always feasible, for instance in evaluating interventions targeted at the population level or when RCTs are deemed unethical. Furthermore, there is often a need to retrospectively evaluate interventions which have already been implemented without randomization or without controls. Interrupted Time Series (ITS), a quasi-experimental design, is sometimes used for evaluation of population-level interventions or when RCTs are not appropriate for reasons mentioned above. ITS design has been used widely and was first introduced to healthcare research in 1981 to evaluate the impact of regionalized perinatal care1. ITS generally involves collecting data at multiple points over a long period of time and constructing a time series of measured outcome, with an interruption by a treatment or implementation of a new policy/program. The intervention is deemed effective if the change in the trend of outcome before and after the interruption is statistically significant2.
There are several advantages to using ITS. Due to the nature of its design, ITS controls for the effect of secular trends, which are fluctuations of data over time independent of intervention. In addition, the large population-level datasets used in ITS studies account for individual bias as well as outliers. The large sample size also allows for stratified analysis of subpopulations and deriving different causal effects of the intervention on different subpopulations3. However, there are still a few caveats to interpreting results from studies that use ITS design. Since ITS is not a controlled study, one of the biggest problems in drawing conclusions from ITS studies is in determining whether a change in outcome is due to the intervention or due to some other factors. Changes independent of the intervention may also be overlooked if the period of the study is not long enough. While researchers can overcome these problems by conducting a control study in parallel, this often increases the cost and time required. Moreover, the power of ITS lies in large sample size, and therefore it is not suitable for studies on interventions targeted at small populations4.
ITS has been increasingly used in clinical research in recent years. Systematic search of MEDLINE for publications that used ITS in drug utilization research shows that only 8% were published before 2000, and 41% were published since 2010 5. More recent use of ITS in clinical setting includes studies on effectiveness of vaccines in certain populations6,7. Despite the increasing popularity of ITS design, there are still inconsistencies and gaps in reported design features and results across a range of healthcare studies, according to a recent investigation in 20198. In addition to differences in reporting, the study also highlighted that there are numerous ways to analyze ITS studies, which can make interpretation of results difficult. It will be interesting to see if there is a move towards standardizing the use of ITS in research and an expansion in application of ITS design in clinical studies.
Gillings D, Makuc D, Siegel E. Analysis of interrupted time series mortaility trends: an example to evaluate regionalized perinatal care. American Journal of Public Health. (1981); 71(1):38-46.
Bernal JL, Cummins S, Gasparrini A. Interrupted Time Series Regression for the Evaluation of Public Health Interventions: A Tutorial. Int J Epidemiol. 2017; 46 (1):348-355.
Penfold RB, Zhang F. Use of Interrupted Time Series Analysis in Evaluating Health Care Quality Improvements. Acad Pediatr. (2013); 13(6 Suppl): S38-44.
Hawley S, Ali MS, Berensci K, Judge A, Preto-Alhambra D. Samples size and power considerations for orginary least squares interrupted time series analysis: a simulation study. Clin Epidemiol. 2019; 11:197-205.
Jandoc R, Burden AM, Mamdani M, Levesque LE, Cadarette SM. Interrupted Time Series Analysis IN Drug Utilization Research Is Increasing: Systematic Review and Recommendations. J Clin Epidemiol 2015; 68 (8): 950-956.
Kabore et. al. Impact of 13-valent Pneumococcal Conjufate Vaccine on the Incidence of Hospitalizations for All-Cause Pneumonia Among Children Ages Less than 5 Years in Burkina Faso: An Interrupted Time-Series Analysis. Int J Infect Dis. 2020; 96 :31-38.
Silaba et. al. Effect of 10-valent pneumococcal conjugate vaccine on the incidence of radiologically-confirmed pneumonia and clinically-defined pneumonia in Kenyan children: an interrupted time-series analysis. The Lancet Global Health. 2019; 7(3): 337-346.
Husdon J, Fielding S, Ramsay RC. Methodology and reporting characteristics of studies using interrupted time series design in healthcare. BMC Medical Research Methodology. 2019; 19:137.