Article type
Abstract
Background: Subgroup analysis is a commonly used method to study whether treatment effects differ based on different covariates.
Objectives: This study aimed to introduce the concepts, advantages, and subpopulation delineation methods of the Subpopulation Treatment Effect Pattern Plot (STEPP) and a real example for its application.
Method: This article introduced the origin and development of the STEPP method and its advantages compared with traditional subgroup analysis and explained how the STEPP method divides overlapping subpopulations. Based on the data of a randomized controlled trial of Tongxie Yaofang in the treatment of diarrhea-predominant irritable bowel syndrome (IBS-D), R software was used to demonstrate the application effect of the STEPP method.
Results: Compared with traditional subgroup analysis, the STEPP method (1) does not need to consider multiplicity to adjust the P value, (2) can determine whether there is an interaction between the treatment effect and the covariates, and (3) shows whether there is a functional relationship between the treatment effect and the covariates without indicating whether the divided subpopulations have clinical significance. The example of this study was based on the patients' pretreatment Self-Rating Anxiety Scale (SAS) standard scores to divide subgroups and analyzed the IBS symptom severity scale scores of IBS-D patients after the end of the trial. There was a statistical difference in the treatment effect difference between each subpopulation (P = 0.0485). According to plot analysis, the effect difference of SAS standard score showed a "U" shape in the area <46. The effect difference had a larger decrease when the mean SAS standard score reached 49.It is speculated that as the level of anxiety further increases, the difference in efficacy between the Tongxie Yaofang and the placebo may further increase.
Conclusions: The STEPP method can be widely used to explore the relationship between continuous covariates and treatment effects so as to explore the influencing factors of treatment effects, and thus certain predictions of treatment effects can be made based on the changes of covariates, which can provide references to the individualized treatment and improve the patients’ clinical outcomes.
Objectives: This study aimed to introduce the concepts, advantages, and subpopulation delineation methods of the Subpopulation Treatment Effect Pattern Plot (STEPP) and a real example for its application.
Method: This article introduced the origin and development of the STEPP method and its advantages compared with traditional subgroup analysis and explained how the STEPP method divides overlapping subpopulations. Based on the data of a randomized controlled trial of Tongxie Yaofang in the treatment of diarrhea-predominant irritable bowel syndrome (IBS-D), R software was used to demonstrate the application effect of the STEPP method.
Results: Compared with traditional subgroup analysis, the STEPP method (1) does not need to consider multiplicity to adjust the P value, (2) can determine whether there is an interaction between the treatment effect and the covariates, and (3) shows whether there is a functional relationship between the treatment effect and the covariates without indicating whether the divided subpopulations have clinical significance. The example of this study was based on the patients' pretreatment Self-Rating Anxiety Scale (SAS) standard scores to divide subgroups and analyzed the IBS symptom severity scale scores of IBS-D patients after the end of the trial. There was a statistical difference in the treatment effect difference between each subpopulation (P = 0.0485). According to plot analysis, the effect difference of SAS standard score showed a "U" shape in the area <46. The effect difference had a larger decrease when the mean SAS standard score reached 49.It is speculated that as the level of anxiety further increases, the difference in efficacy between the Tongxie Yaofang and the placebo may further increase.
Conclusions: The STEPP method can be widely used to explore the relationship between continuous covariates and treatment effects so as to explore the influencing factors of treatment effects, and thus certain predictions of treatment effects can be made based on the changes of covariates, which can provide references to the individualized treatment and improve the patients’ clinical outcomes.