Explore chapters and articles related to this topic
Analysis of Vaccine Studies and Causal Inference
Published in Leonhard Held, Niel Hens, Philip O’Neill, Jacco Wallinga, Handbook of Infectious Disease Data Analysis, 2019
The basic approach is a segmented regression. Let be the outcome at time t, be a dummy variable indicating pre-vaccination ( set to 0) or post-vaccination period ( set to 1), and T is the time since the start of the study. In a simple interrupted time series analysis, the regression model is
Introduction to Time Series Analysis
Published in Mohamed M. Shoukri, Analysis of Correlated Data with SAS and R, 2018
An important design to establish causality attributed to intervention is the interrupted time-series experiment. In Figure 8.21 we show the effect of intervention (in this case launching campaign to advocate regular hand washing) on the incidence of nosocomial infection in a large tertiary care hospital.
Stepped wedge, natural experiments and interrupted time series analysis designs
Published in David A. Richards, Ingalill Rahm Hallberg, Complex Interventions in Health, 2015
The randomized controlled trial is perceived as the gold standard by which effectiveness is measured in a study. However, alternative types of quasi-experimental designs – controlled before and after studies and ITS studies – are recognized as practical approaches for improving the quality of information for decision-makers in the real world. Interrupted time series designs use routine monitoring data collected at equally spaced intervals of time before and after an intervention, with the period before intervention serving as a control group (Grimshaw et al., 2003). ITS can be used to assess change of behavior, practices or outcomes over time. This design can be particularly applied to routine surveillance data that have no obvious control group. The ITS design is a much more practical option in many settings, even though the allocation of subjects is not randomized. Therefore, when experimental study is not possible, an ITS study can provide a robust method of measuring the effect of an intervention (Grimshaw et al., 2003). The ITS approach has been used to assess the effectiveness of a variety of interventions in various fields such as in environmental, financial and health sciences.
Trends in concomitant and single opioid and benzodiazepine exposures reported to the California Poison Control System following the Centers for Disease Control and Prevention release of opioid guidelines in 2016
Published in Clinical Toxicology, 2023
Emily Chu, Gina Cocos, Ho Jun Lee, Jane Go, Justin Lewis, Dorie E. Apollonio
Our primary analysis relied on interrupted time series analyses. We assessed the risk of delayed implementation (using “actest” for autocorrelation) and identified a lag of one month for our analyses of time trends. The regression model, assuming the form of Yt = β0 + β1Tt + β2Xt + β3XtTt + ϵt, tested for several parameters which included: β0, the intercept or initial level; β1, the trend before the intervention; β2, the immediate change following an intervention; and β3, the difference in the trend of pre- and post-intervention periods [15]. The outcome variable was the count of total monthly exposures to our drug(s) of interest: opioid, benzodiazepine, or concomitant opioid and benzodiazepine. Our intervention was the release of the CDC Guideline in March 2016, meaning that exposures occurring before that month were regarded as pre-intervention while exposures occurring after were post-intervention.
The impact of converting a power plant from coal to natural gas on pediatric acute asthma
Published in Journal of Asthma, 2022
Robert Clemons, Maiying Kong, Kahir Jawad, Yana Feygin, Kerry Caperell
The interrupted time series is a quasi-experimental study design. The term quasi-experimental refers to an absence of randomization. Interrupted time series a routine methodology used is for analyzing observational data where randomization is not possible. Its main advantage over alternative approaches is that it can make full use of the longitudinal nature of the data and account for pre-intervention trends. This design is particularly useful when “natural experiments” in real word settings occur—for example, when a power company change fuel sources. However, it is not appropriate when trends have external time varying effects or autocorrelation (e.g., seasonality). While these can be potentially handled through modeling if the relevant information is known. It is difficult to understand all factors that impact a longitudinal outcome.
Living kidney and liver donations and transplantations: an interrupted time series analysis spanning years, 1988–2020
Published in International Journal of Healthcare Management, 2022
Amrita Shenoy, Gopinath N. Shenoy, Gayatri G. Shenoy
There are some strengths to this method. First, ITS is a valuable study design for evaluating the effectiveness at population-level interventions [34]. Second, a segmented regression approach can be used to analyze an interrupted time series study by testing the effect of an intervention on the outcome of interest using an appropriately defined impact model [34]. Third, it is rarely possible to assess the impact of policy changes with randomized controlled trials. Time series designs, therefore, are the strongest, quasi-experimental designs to estimate intervention effects in non-randomized settings [22]. Fourth, segmented regression analysis of ITS data allows analysts to control for prior trends in the outcome and to study the dynamics of change in response to an intervention [22]. Fifth, even without a control group, segmented regression analysis addresses important threats to internal validity by making multiple assessments of the outcome variable both before and after the policy intervention [22]. Finally, segmented regression analysis can estimate the size of the effect at different time points, as well as changes in the trend of the effect over time [22].