Zde se nacházíte:
Informace o publikaci
CLINICAL PHENOTYPES THROUGH MACHINE LEARNING OF FIRST-LINE TREATED PATIENTS DURING THE PERIOD 2013-2022: DATA FROM THE EUROPEAN REGISTRY ON HELICOBACTER PYLORI MANAGEMENT (HPEUREG)
Autoři | |
---|---|
Rok publikování | 2023 |
Druh | Konferenční abstrakty |
Fakulta / Pracoviště MU | |
Citace | |
Popis | Introduction: The segmentation of patients in homogeneous groups, according to their clinical variables and treatments, could help to improve the effectiveness of current eradication therapies. Aims & Methods: 1. To determine the most important characteristics of the treatments used in the European Registry on H. pylori management (Hp-EuReg), using machine learning techniques. 2. To evaluate the effectiveness of the treatments according to the year of the visit and the country using a cluster decomposition. Sub-study of the Hp-EuReg, a systematic, prospective, registry of the routine clinical practice of European gastroenterologists on the management of H. pylori infection. All cases with a first-line empirical eradication treatment registered from June 2013 to December 2022, were included in the current analysis. Boruta, a random-forest-like method was used to determine the following ‘most important’ variables: compliance, duration of treatment, PPI dosage, patient’s country, and treatment scheme. Results: In total, 35,852 European patients were analysed. Table 1 shows the increasing trend in the effectiveness of treatments, from an average 87% in 2013 to 93% in 2022 (more than 100 patients/clusters). The cluster 3 in 2016 (lowest effectiveness) was composed of 97.5% triple therapy with clarithromycin-amoxicillin/metronidazole, mainly in Slovenia (54%), with a majority (85%) of 7-day prescriptions, and 99% compliance. The highest effectiveness was obtained in cluster #1 in 2022, with 81% of Spanish cases, 32% of concomitant therapy with clarithromycin-amoxicillinmetronidazole/tinidazole and 63% of bismuth quadruple therapy with tetracycline-metronidazole (prescribed as single capsule), 69% of 10 days prescriptions, and 32% of 14 days treatment length. |