Artikel
Providing an R-package for sampling cluster randomized trails data to show the impact of implementation errors on effect estimation within stepped wedge design studies
Suche in Medline nach
Autoren
-
Diana Trutschel
- Deutsches Zentrum für Neurodegenerative Erkrankungen e.V. (DZNE), Standort Witten, Witten, Germany
-
Sven Reuther
- Deutsches Zentrum für Neurodegenerative Erkrankungen e.V. (DZNE), Standort Witten, Witten, Germany
-
Daniela Holle
- Deutsches Zentrum für Neurodegenerative Erkrankungen e.V. (DZNE), Standort Witten, Witten, Germany
-
Pablo Emilio Verde
- Universität Düsseldorf, Düsseldorf, Germany
| Veröffentlicht: | 26. September 2017 |
|---|
Gliederung
Text
Background: A simulation can be a tool to explore methodological challenges of used study designs or data analysing methods.
Here, we provide an R-package for sampling multidimensional distributed data within cluster randomized trails, for stepped wedge design (SWD) trails as well as cross-over and parallel designs. The SWD is an alternative study designs when a simple parallel design is not useful or not be feasible.
Aim: This design is relative new and for health care researchers in practice several methodological pitfalls are possible. The aim is to give an orientation before beginning a study to determine how sensitive their study is against common scenarios in research practice.
Method: A simulation experiment is performed investigating three factors: the intervention reach not the 100\'25 assumed implementation, number of missing clusters and time point at which clusters were lost. The data within an (cross-sectional as well as longitudinal) SWD trial including the deviations from the assumed perfect situation were sampled using the R-package. Then the followed effect estimation were realized using a linear mixed-effects model.
Results: The results of the simulation study show that the SWD was not robust against a lack of implementation, identifying that a delay in implementation had the greatest influence on the estimates. The variance of the effect estimates increased with the number of lost clusters, where the time-point of clusters loss had only a marginal influence.
Discussion: Researcher may consider using simulation studies to quantify the effect of possible practical lacks within studies. Moreover, solutions to encounter such identified lacks should be found before a study is performed. For example, for implementation errors within SWD trials it is suggested to estimate the degree of intervention implemention via process evaluation, which can ten be included into the statistical model.
Practical implication: The provided R-package is useful to sampling data within such studies and the simulation can be adapted for other settings.
