@phdthesis{,
    author = {Hinnenthal, Marie},
    title = {Endpoint-oriented clustering of individual patient data : A new approach for the health economic evaluation of medical interventions},
    editor = {},
    booktitle = {},
    series = {},
    journal = {},
    address = {},
    publisher = {},
    edition = {},
    year = {2017},
    isbn = {},
    volume = {},
    number =  {},
    pages = {},
    url = {},
    doi = {},
    keywords = {clustering, meta-analysis, heterogeneity},
    abstract = {In the age of big data, the collection of individual patient information, with the help of clinical trials, is becoming an increasingly important area of healthcare. These large amounts of data have the potential to provide an improved medical care for patients. Especially with regard to the health economic evaluation of medical interventions, the analysis of this data can lead to a better patient-oriented medication in terms of evidence-based medicine. Meta-analytic approaches for the evaluation of clinical drug studies only estimate a weighted mean value of the measured endpoints. The whole potential of the big amounts of individual patient data is therefore not nearly exploited. The collected patients' information, e.g. the differences in the socio-economic parameters of different patient groups and the associated heterogeneity in the efficacy of a drug, are not sufficiently considered in the common analysis. For approaches like the subgroup analysis, which uses such information, the considered groups are often too small to provide statistically well-founded results. Therefore, new methods are needed for the analysis of the extensive patient data. The new approaches presented in this thesis are all based on an innovative endpoint-oriented clustering, developed by Brieden and Gritzmann. The algorithm identifies hidden multidimensional structures and forms sufficiently large clusters in which patients are grouped with similar combinations of their characteristic values. The new invented methods, which are applied on the identified patient collectives, deal with the evaluation and prediction of the efficacy of medical interventions and the identification of clinical heterogeneity in the treatment effects.},
    note = {},
    
    school = {Universität der Bundeswehr München},
    
}