The work entitled Tools for Benchmarking from Value-Based Medicine: Archetypes and Patient Classification was defended at the Faculty of Mathematics and Statistics of the Universitat Politècnica de Catalunya-Universitat de Barcelona.
On the 16th of June Maialen Otamendi Garitano, graduate in Mathematics (University of the Basque Country, UPV/EHU) and student of the Master in Statistics and Operations Research at the Universitat Politècnica de Catalunya (UPC)-Universitat de Barcelona, defended her Master’s thesis entitled Tools for Benchmarking from Value-Based Healthcare: Archetypes and Patient Classification at the Faculty of Mathematics and Statistics of the UPC.
The work has been developed within the framework of the international VOICE Community and has been directed by Borja García-Lorenzo, Health Economist at the Biosistemak Institute (formerly Kronikgune). Researchers from the institute, Ania Gorostiza and Itxaso Alayo, also collaborated in the development of the study.
The work presented is framed in the line of medical research on breast cancer, which despite its advances continues to have a clinical and economic impact on patients and the health system. In this context, Value-Based Care (VBA), which puts the patient at the centre and links health outcomes to costs, is emerging as a new paradigm in health services. VBA proposes a step forward, as in addition to looking at traditional clinical and process outcomes, it also takes into account outcomes reported directly by the patient, as well as the use of healthcare resources and associated costs; all of this to determine the value of healthcare. In the published literature, there is no identification of patient archetypes and their corresponding classification based on their care pathway that can be related to patient-reported outcomes. This aspect allows adequate benchmarking between patients and centres from the perspective of VBM, which enables the identification of patterns and anticipation of patient response, as well as determining the most efficient strategies for each group of patients.
In the framework of the European community VOICE, the clinical team raised the need not only to identify archetypes but also to have a tool to classify new patients. The work led by the Biosistemak (formerly Kronikgune) team has a dual objective: the identification of archetypes of patients diagnosed with breast cancer, and the design of a classifier of new patients, both based on their care history.
Latent Class Analysis (LCA) was used to identify patient archetypes and classify them. The results of the objectives were validated with the application of other methods from the literature. Archetype identification was validated with traditional Clustering methods, while classification was validated with Random Forests (AF) and Multinomial Regression (MR). The use of a single methodology to achieve both objectives provides scientific evidence for the use of ACL as an efficient and robust tool for benchmarking in the field of VBM.
The results of this study not only provide the identification of archetypes of patients diagnosed with breast cancer in the context of VBA but also classify the same patients and allow benchmarking between patients and healthcare centres from the perspective of VBA, addressing the needs of the VOICE community care team. Furthermore, the proposed methodological strategy can be transferable to other medical conditions.
If you want to know more about the VOICE community, click on the following link.
