A model for identifying potentially inappropriate medication used in older people with dementia: a machine learning study.

BACKGROUND: Older adults with dementia often face the risk of potentially inappropriate medication (PIM) use. The quality of PIM evaluation is hindered by researchers' unfamiliarity with evaluation criteria for inappropriate drug use. While traditional machine learning algorithms can enhance evaluation quality, they struggle with the multilabel nature of prescription data. AIM: This study aimed to combine six machine learning algorithms and three multilabel classification models to identify correlations in prescription information and develop an optimal model to identify PIMs in older adults with dementia. METHOD: This study was conducted from January 1, 2020, to December 31, 2020. We used cluster sampling to obtain prescription data from patients 65 years and older with dementia. We assessed PIMs using the 2019 Beers criteria, the most authoritative and widely recognized standard for PIM detection. Our modeling process used three problem transformation methods (binary relevance, label powerset, and classifier chain) and six classification algorithms. RESULTS: We identified 18,338 older dementia patients and 36 PIMs types. The classifier chain + categorical boosting (CatBoost) model demonstrated superior performance, with the highest accuracy (97.93%), precision (95.39%), recall (94.07%), F1 score (95.69%), and subset accuracy values (97.41%), along with the lowest Hamming loss value (0.0011) and an acceptable duration of the operation (371s). CONCLUSION: This research introduces a pioneering CC + CatBoost warning model for PIMs in older dementia patients, utilizing machine-learning techniques. This model enables a quick and precise identification of PIMs, simplifying the manual evaluation process.

[Knowledge Graph-Based Prediction of Potentially Inappropriate Medication].

Objective: To improve the accuracy of potentially inappropriate medication (PIM) prediction, a PIM prediction model that combines knowledge graph and machine learning was proposed. Methods: Firstly, based on Beers criteria 2019 and using the knowledge graph as the basic structure, a PIM knowledge representation framework with logical expression capabilities was constructed, and a PIM inference process was implemented from patient information nodes to PIM nodes. Secondly, a machine learning prediction model for each PIM label was established based on the classifier chain algorithm, to learn the potential feature associations from the data. Finally, based on a threshold of sample size, a portion of reasoning results from the knowledge graph was used as output labels on the classifier chain to enhance the reliability of the prediction results of low-frequency PIMs. Results: 11 741 prescriptions from 9 medical institutions in Chengdu were used to evaluate the effectiveness of the model. Experimental results show that the accuracy of the model for PIM quantity prediction is 98.10%, the F1 is 93.66%, the Hamming loss for PIM multi-label prediction is 0.06%, and the macroF1 is 66.09%, which has higher prediction accuracy than the existing models. Conclusion: The method proposed has better prediction performance for potentially inappropriate medication and significantly improves the recognition of low-frequency PIM labels.

Developing a Warning Model of Potentially Inappropriate Medications in Older Chinese Outpatients in Tertiary Hospitals: A Machine-Learning Study.

Due to multiple comorbid illnesses, polypharmacy, and age-related changes in pharmacokinetics and pharmacodynamics in older adults, the prevalence of potentially inappropriate medications (PIMs) is high, which affects the quality of life of older adults. Building an effective warning model is necessary for the early identification of PIMs to prevent harm caused by medication in geriatric patients. The purpose of this study was to develop a machine learning-based model for the warning of PIMs in older Chinese outpatients. This retrospective study was conducted among geriatric outpatients in nine tertiary hospitals in Chengdu from January 2018 to December 2018. The Beers criteria 2019 were used to assess PIMs in geriatric outpatients. Three problem transformation methods were used to tackle the multilabel classification problem in prescriptions. After the division of patient prescriptions into the training and test sets (8:2), we adopted six widely used classification algorithms to conduct the classification task and assessed the discriminative performance by the accuracy, precision, recall, F1 scores, subset accuracy (ss Acc), and Hamming loss (hm) of each model. The results showed that among 11,741 older patient prescriptions, 5816 PIMs were identified in 4038 (34.39%) patient prescriptions. A total of 41 types of PIMs were identified in these prescriptions. The three-problem transformation methods included label power set (LP), classifier chains (CC), and binary relevance (BR). Six classification algorithms were used to establish the warning models, including Random Forest (RF), Light Gradient Boosting Machine (LightGBM), eXtreme Gradient Boosting (XGBoost), CatBoost, Deep Forest (DF), and TabNet. The CC + CatBoost model had the highest accuracy value (97.83%), recall value (89.34%), F1 value (90.69%), and ss Acc value (97.79%) with a good precision value (92.18%) and the lowest hm value (0.0006). Therefore, the CC + CatBoost model was selected to predict the occurrence of PIM in geriatric Chinese patients. This study's novelty establishes a warning model for PIMs in geriatric patients by using machine learning. With the popularity of electronic patient record systems, sophisticated computer algorithms can be implemented at the bedside to improve medication use safety in geriatric patients in the future.