Premium
Interaction trees: exploring the differential effects of an intervention programme for breast cancer survivors
Author(s) -
Su Xiaogang,
Meneses Karen,
McNees Patrick,
Johnson Wesley O.
Publication year - 2011
Publication title -
journal of the royal statistical society: series c (applied statistics)
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.205
H-Index - 72
eISSN - 1467-9876
pISSN - 0035-9254
DOI - 10.1111/j.1467-9876.2010.00754.x
Subject(s) - recursive partitioning , covariate , breast cancer , intervention (counseling) , permutation (music) , random forest , ranking (information retrieval) , quality of life (healthcare) , tree (set theory) , machine learning , psychology , computer science , clinical psychology , statistics , mathematics , artificial intelligence , cancer , medicine , psychotherapist , nursing , mathematical analysis , physics , acoustics
Summary. The breast cancer education intervention (BCEI) is a tailored psychoeducational intervention programme aiming to improve the quality of life of breast cancer survivors. Borrowing the idea of recursive partitioning and following the convention of classification and regression trees, an exploratory procedure, termed interaction trees, is proposed to understand better the differential effects of the BCEI on longitudinal quality‐of‐life data. The resultant tree model identifies several objectively defined subgroups: in some groups the BCEI is quite effective whereas in others it may not be. Based on the final tree structure, a permutation test is used to assess the overall treatment‐by‐covariate interaction. In addition, a variable importance ranking feature is facilitated via random forests of interaction trees to help to determine important effect modifiers of the BCEI.