paper_ai_pharmacotherapy
Science Score: 44.0%
This score indicates how likely this project is to be science-related based on various indicators:
-
✓CITATION.cff file
Found CITATION.cff file -
✓codemeta.json file
Found codemeta.json file -
✓.zenodo.json file
Found .zenodo.json file -
○DOI references
-
○Academic publication links
-
○Academic email domains
-
○Institutional organization owner
-
○JOSS paper metadata
-
○Scientific vocabulary similarity
Low similarity (4.9%) to scientific vocabulary
Repository
Basic Info
- Host: GitHub
- Owner: mchlbckr
- Language: R
- Default Branch: main
- Size: 1.17 MB
Statistics
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
- Releases: 0
Metadata Files
README.md
Artificial intelligence in pharmacotherapy – mission (im-)possible?
This is the git repository contaning code and data used for the results of the paper Artificial intelligence in pharmacotherapy – mission (im-)possible? by Michael Bücker, Kreshnik Hoti, and Olaf Rose.
Background
Artificial intelligence (AI) has hardly been used in optimizing pharmacotherapy. This study aimed to explore barriers and limits and discuss approaches to overcome them.
Methods
Data of a previous study on medication therapy optimization was updated and adapted for the purpose of this study: predicting medication based on multiple diagonses. 74% of the data was being used for training and 26% for testing. Decision trees were chosen as the underlying model due to their simplicity and interpretability. Overfitting was controlled by bootstrapping, hyperparameters were optimized. Areas under the curve and accuracies were calculated.
Results
The cohort consisted of 101 elderly patients with polymedication and multiple diagnoses. High prediction accuracy was achieved for the cardiovascular drug classes of ACE-inhibitors/angiotensin receptor blockers, mineralocorticoid-receptor antagonists, and nitroglycerin.
Conclusion
The model showed promising results and left space for potential overwriting, in case of sudden therapy changes on safety alerts or new guidelines. Laboratory data and vital signs could not be used for decision-making, as they were influenced by the used drugs and not measured for all patients. An identified problem is the limited availability of optimized therapy plans.
Owner
- Name: Michael Bücker
- Login: mchlbckr
- Kind: user
- Repositories: 1
- Profile: https://github.com/mchlbckr
Citation (CITATION.cff)
cff-version: 1.2.0 message: "If you use this software, please cite it as below." authors: - family-names: "Bücker" given-names: "Michael" orcid: "https://orcid.org/0000-0003-0045-8460" - family-names: "Hoti" given-names: "Kreshnik" orcid: "https://orcid.org/0000-0001-7652-6074" - family-names: "Rose" given-names: "Olaf" orcid: "https://orcid.org/0000-0003-1681-4909" title: "Artificial intelligence in pharmacotherapy – mission (im-)possible?" version: 1.0.0 date-released: 2024-02-05 url: "https://github.com/mchlbckr/paper_ai_pharmacotherapy/tree/main"