BACKGROUND AND OBJECTIVE

Access to healthcare remains a challenge in most areas in the Philippines. Fifty-three percent (53%) of the Philippine population do not have access to a rural health unit (RHU) within a 30-minute travel t ime. As a response, the Department of Health (DOH) needs to construct an additional 2400 RHUs by 2025. This paper uses the Philippine Health Facility Development Plan 2020-2040 (PHFDP) as a reference to present a solution for locating sites for RHU placement in under-served areas using neural networks to meet the 30-minute travel time by maximizing population accessibility.

RHU accessibility was measured using geographic attributes as inputs to a back propagation neural network (BPNN) and a recurrent neural network (RNN): (1) land coverage and hazard data, representing geographical limitations; (2) population density and distribution, indicating demand for healthcare services; and (3) infrastructure-related features, such as road networks, points of interest, and the locations of existing RHUs, which influence healthcare accessibility. The models were trained to identify underserved areas and were implemented on a nationwide scale, excluding NCR, to locate candidate areas to increase population access to the new RHUs. The models were validated using a healthcare facility accessibility index (HCFAI) to assess RHU coverage improvement.

The BPNN showed stronger generalization across regions, achieving 79.1% average accuracy in distinguishing low from high accessible areas on Region 1 and identifying 1668 out of 3305 locations in the region as candidate sites. The RNN, better capturing unique regional characteristics, required separate training: 77.2% average accuracy on Region 1, identifying 1593 candidate sites. Our findings suggest expanding the use of land improves population access to healthcare facilities. Both models found more than the needed number of RHUs by 2040. The BPNN was more consistent than RNN to improve a region’s overall accessibility by increasing the HCFAI. The BPNN can increase population access to an RHU from 2.5-98.5% from its original population with access to an RHU.

The study demonstrates the usage of geographic attributes and neural networks to improve healthcare accessibility. The BPNN and RNN are adequate algorithms to find under-served areas and candidate sites for RHU construction to maximize population accessibility. The HCFAI metric validates the locations to highlight which neural network maximizes more of the region’s populat ion. The study contributes to ongoing efforts to improve healthcare infrastructure and accessibility, offering datadriven recommendations for RHU locations.

Human ; Humans ; Humanities ; Health ; Forecasting ; Algorithms

OBJECTIVE: To identify the optimal machine learning algorithm for predicting post-stroke epilepsy (PSE) within one year following acute stroke, establish a nomogram model based on this algorithm, and perform external validation to achieve accurate prediction of secondary epilepsy. METHODS: A total of 870 acute stroke patients admitted to the emergency department of Xiang'an Hospital of Xiamen University from June 2019 to June 2023 were enrolled for model development (model group). An external validation cohort of 435 acute stroke patients admitted to the Fifth Hospital of Xiamen during the same period was used to validate the machine learning algorithms and nomogram model. Patients were classified into control and epilepsy groups based on the development of PSE within one year. Clinical and laboratory data, including baseline characteristics, stroke location, vascular status, complications, hematologic parameters, and National Institutes of Health Stroke Scale (NIHSS) score, were collected for analysis. Nine machine learning algorithms such as logistic regression, CN2 rule induction, K-nearest neighbors, adaptive boosting, random forest, gradient boosting, support vector machine, naive Bayes, and neural network were applied to evaluate predictive performance. The area under the curve (AUC) of receiver operator characteristic curve (ROC curve) was used to identify the optimal algorithm. Logistic regression was used to screen risk factors for PSE, and the top 10 predictors were selected to construct the nomogram model. The predictive performance of the model was evaluated using the ROC curve in both the model and validation groups. RESULTS: Among the 870 patients in the model group, 29 developed PSE within one year. Among the nine algorithms tested, logistic regression demonstrated the best performance and generalizability, with an AUC of 0.923. Univariate logistic regression identified several risk factors for PSE, including platelet count, white blood cell count, red blood cell count, glycated hemoglobin (HbA1c), C-reactive protein (CRP), triglycerides, high-density lipoprotein (HDL), aspartate aminotransferase (AST), alanine aminotransferase (ALT), activated partial thromboplastin time (APTT), thrombin time, D-dimer, fibrinogen, creatine kinase (CK), creatine kinase-MB (CK-MB), lactate dehydrogenase (LDH), serum sodium, lactic acid, anion gap, NIHSS score, brain herniation, periventricular stroke, and carotid artery plaque. Further multivariate logistic regression analysis showed that white blood cell count, HDL, fibrinogen, lactic acid and brain herniation were independent risk factors [odds ratio (OR) were 1.837, 198.039, 47.025, 11.559, 70.722, respectively, all P < 0.05]. In the external validation group, univariate logistic regression analysis showed that platelet count, white blood cell count, CRP, triacylglycerol, APTT, D-dimer, fibrinogen, CK, CK-MB, LDH, NIHSS score, and cerebral herniation were risk factors for PSE one year after acute stroke. Further multiple logistic regression analysis showed that APTT and cerebral herniation were independent predictors (OR were 0.587 and 116.193, respectively, both P < 0.05). The nomogram model, constructed using 10 key variables-brain herniation, periventricular stroke, carotid artery plaque, white blood cell count, triglycerides, thrombin time, D-dimer, serum sodium, lactic acid, and NIHSS score-achieved an AUC of 0.908 in the model group and 0.864 in the external validation group. CONCLUSIONS The logistic regression-based prediction model for epilepsy one year after acute stroke, developed using machine learning algorithms, showed optimal predictive performance. The nomogram model based on the logistic regression-derived predictors showed strong discriminative power and was successfully validated externally, suggesting favorable clinical applicability and generalizability.
Humans ; Machine Learning ; Stroke/complications* ; Nomograms ; Epilepsy/etiology* ; Algorithms ; Male ; Female ; Logistic Models ; Middle Aged ; Aged ; Risk Factors ; Bayes Theorem

OBJECTIVE: To develop and compare risk prediction models for in-hospital post-cardiac arrest brain injury (PCABI) in critically ill patients using nomograms and random forest algorithms, aiming to identify the optimal model for early identification of high-risk PCABI patients and providing evidence for precise treatment. METHODS: A retrospective cohort study was used to collect the first-time in-hospital cardiac arrest (IHCA) patients admitted to the intensive care unit (ICU) from 2008 to 2019 in the Medical Information Mart for Intensive Care-IV (MIMIC-IV) as the study population, and the patients' age, gender, body mass, health insurance utilization, first vital signs and laboratory tests within 24 hours of ICU admission, mechanical ventilation, and critical care scores were extracted. Independent influencing factors of PCABI were identified through univariate and multivariate Logistic regression analyses. The included patients were randomly divided into a training cohort and an internal validation cohort in a 7:3 ratio, and the PCABI risk prediction model was constructed by the nomogram and random forest algorithm, respectively, and the model was evaluated by receiver operator characteristic curve (ROC curve), the calibration curve, and the decision curve analysis (DCA), and after the better model was selected, 179 patients admitted to Tianjin Medical University General Hospital as the external validation cohort for external evaluation were collected by using the same inclusion and exclusion criteria. RESULTS: A total of 1 419 patients with without traumatic brain injury who had their first-time IHCA were enrolled, including 995 in the training cohort (including 176 PCABI and 819 non-PCABI) and 424 in the internal validation cohort (including 74 PCABI and 350 non-PCABI). Univariate and multivariate analysis showed that age, potassium, urea nitrogen, sequential organ failure assessment (SOFA), acute physiology and chronic health evaluation III (APACHE III), and mechanical ventilation were independent influences on the occurrence of PCABI in patients with IHCA (all P < 0.05). Combining the above variables, we constructed a nomogram model and a random forest model for comparison, and the results show that the nomogram model has better predictive efficacy than the random forest model [nomogram model: area under the ROC curve (AUC) of the training cohort = 0.776, with a 95% credible interval (95%CI) of 0.741-0.811; internal validation cohort AUC = 0.776, with a 95%CI of 0.718-0.833; random forest model: AUC = 0.720, with a 95%CI of 0.653-0.787], and they performed similarly in terms of calibration curves, but the nomogram performed better in terms of decision curve analysis (DCA); at the same time, the nomogram model was robust in terms of external validation cohort (external validation cohort AUC = 0.784, 95%CI was 0.692-0.876). CONCLUSIONS A nomogram risk prediction model for the occurrence of PCABI in critically ill patients was successfully constructed, which performs better than the random forest model, helps clinicians to identify the risk of PCABI in critically ill patients at an early stage and provides a theoretical basis for early intervention.
Humans ; Critical Illness ; Retrospective Studies ; Heart Arrest/complications* ; Nomograms ; Brain Injuries/etiology* ; Intensive Care Units ; Algorithms ; Male ; Female ; Middle Aged ; ROC Curve ; Risk Factors ; Risk Assessment ; Logistic Models ; Aged

OBJECTIVE: To construct machine learning prediction model for sepsis-associated encephalopathy (SAE), and analyze the application value of the model on early identification of SAE risk in elderly septic patients. METHODS: Patients aged over 60 years with a primary diagnosis of sepsis admitted to intensive care unit (ICU) from 2008 to 2023 were selected from Medical Information Mart for Intensive Care-IV 2.2 (MIMIC-IV 2.2). Demographic variables, disease severity scores, comorbidities, interventions, laboratory indicators, and hospitalization details were collected. Key factors associated with SAE were identified using univariate Logistic regression analysis. The data were randomly divided into training and validation sets in a 7 : 3 ratio. Multivariable Logistic regression analysis was conducted in the training set and visualized using a nomogram model for prediction of SAE. The discrimination of the model was evaluated in the validation set using the receiver operator characteristic curve (ROC curve), and its calibration was assessed using calibration curve. Furthermore, multiple machine learning algorithms, including multi-layer perceptron (MLP), support vector machine (SVM), naive bayes (NB), gradient boosting machine (GBM), random forest (RF), and extreme gradient boosting (XGB), were constructed in the training set. Their predictive performance was subsequently evaluated on the validation set. Taking the XGB model as an example, the interpretability of the model through the SHapley Additive exPlanations (SHAP) algorithm was enhanced to identify the key predictive factors and their contributions. RESULTS: A total of 2 204 septic patients were finally enrolled, of whom 840 developed SAE (38.1%). A total of 21 variables associated with SAE were screened through univariate Logistic regression analysis. Multivariable Logistic regression analysis showed that endotracheal intubation [odds ratio (OR) = 0.40, 95% confidence interval (95%CI) was 0.19-0.88, P < 0.001], oxygen therapy (OR = 0.76, 95%CI was 0.53-0.95, P = 0.023), tracheotomy (OR = 0.20, 95%CI was 0.07-0.53, P < 0.001), continuous renal replacement therapy (CRRT; OR = 0.32, 95%CI was 0.15-0.70, P < 0.001), cerebrovascular disease (OR = 0.31, 95%CI was 0.16-0.60, P < 0.001), rheumatic disease (OR = 0.44, 95%CI was 0.19-0.99, P < 0.001), male (OR = 0.68, 95%CI was 0.54-0.86, P = 0.001), and maximum anion gap (AG; OR = 0.95, 95%CI was 0.93-0.97, P < 0.001) were associated with an decreased probability of SAE, and age (OR = 1.05, 95%CI was 1.03-1.06, P < 0.001), acute physiology score III (APSIII; OR = 1.02, 95%CI was 1.01-1.02, P < 0.001), Oxford acute severity of illness score (OASIS; OR = 1.04, 95%CI was 1.03-1.06, P < 0.001), and length of hospital stay (OR = 1.01, 95%CI was 1.01-1.02, P < 0.001) were associated with an increased probability of SAE. A nomogram model was constructed based on these variables. In the validation set, ROC curve analysis showed that the model achieved an area under the ROC curve (AUC) of 0.723, and the calibration curve showed good consistency between the predicted probability of the model and the observed probability. Among the machine learning algorithms, including MLP, SVM, NB, GBM, RF, and XGB, the SVM model and RF model demonstrated relatively good predictive performance, with AUC of 0.748 and 0.739, respectively, and the sensitivity was both exceeding 85%. The predictive performance of the XGB model was explained through SHAP analysis, and the results indicated that APSIII score (SHAP value was 0.871), age (SHAP value was 0.521), and OASIS score (SHAP value was 0.443) were important factors affecting the predictive performance of the model. CONCLUSIONS The machine learning-based SAE prediction model exhibits good predictive capability and holds significant application value for the early identification of SAE risk in elderly septic patients.
Humans ; Machine Learning ; Aged ; Sepsis-Associated Encephalopathy ; Sepsis/complications* ; Intensive Care Units ; Logistic Models ; Middle Aged ; Male ; ROC Curve ; Female ; Bayes Theorem ; Nomograms ; Support Vector Machine ; Algorithms

OBJECTIVE: To develop a digital intelligent multimodal shift handover system for the intensive care unit (ICU) and evaluate its application effect in ICU shift handovers. METHODS: A research and development team was established, consisting of 1 department director, 1 head nurse, 3 information technology engineers, 3 nurses, and 2 doctors. Team members were assigned responsibilities including overall coordination and planning, platform design and maintenance, pre-application training, collection and organization of clinical feedback, and research investigation respectively. A digital intelligent multimodal shift handover system was developed for ICU based on the Shannon-Weaver linear transmission model. This innovative system integrated automated data collection, intelligent dynamic monitoring, multidimensional condition analysis and visual reporting functions. A cloud platform was used to gather data from multi-parameter vital signs monitors, infusion pumps, ventilators and other devices. Artificial intelligence algorithms were employed to standardize and analyze the data, providing personalized recommendations for healthcare professionals. A self-controlled before-after method was adopted. Before the application of the ICU digital intelligent multimodal shift handover system (from December 2023 to March 2024), the traditional verbal bedside handover was used; from June 2024 to March 2025, the ICU digital intelligent multimodal shift handover system was applied for shift handovers. Questionnaires before the application of the shift handover system were collected in April 2024, and those after the application were collected in April 2025. The shift handover time, handover quality (scored by the nursing handover evaluation scale), satisfaction with doctor-nurse communication (scored by the ICU doctor-nurse scale) before and after the application of the handover system were compared, and nurses' satisfaction with the shift handover system (scored by the clinical nursing information system effectiveness evaluation scale) was investigated. RESULTS: After the application of the ICU digital intelligent multimodal shift handover system, the shift handover time was significantly shorter than that before the application [minutes: 20 (15, 25) vs. 30 (22, 40)], the handover quality was significantly higher than that before the application [score: 84.0 (78.0, 88.5) vs. 71.0 (55.0, 79.0)], and the satisfaction with doctor-nurse communication was also significantly higher than that before the application (score: 84.58±6.79 vs. 74.50±11.30). All differences were statistically significant (all P < 0.05). In addition, the nurses' system effectiveness evaluation scale score was 102.30±10.56, which indicated that nurses had a very high level of satisfaction with the ICU digital intelligent multimodal shift handover system. CONCLUSIONS The application of the ICU digital intelligent multimodal shift handover system can shorten the shift handover time, improve the handover quality, and enhance the satisfaction with doctor-nurse communication. Nurses have a high level of satisfaction with this system.
Intensive Care Units ; Humans ; Patient Handoff ; Artificial Intelligence ; Algorithms

Accurate prediction of drug-target interactions (DTIs) plays a pivotal role in drug discovery, facilitating optimization of lead compounds, drug repurposing and elucidation of drug side effects. However, traditional DTI prediction methods are often limited by incomplete biological data and insufficient representation of protein features. In this study, we proposed KG-CNNDTI, a novel knowledge graph-enhanced framework for DTI prediction, which integrates heterogeneous biological information to improve model generalizability and predictive performance. The proposed model utilized protein embeddings derived from a biomedical knowledge graph via the Node2Vec algorithm, which were further enriched with contextualized sequence representations obtained from ProteinBERT. For compound representation, multiple molecular fingerprint schemes alongside the Uni-Mol pre-trained model were evaluated. The fused representations served as inputs to both classical machine learning models and a convolutional neural network-based predictor. Experimental evaluations across benchmark datasets demonstrated that KG-CNNDTI achieved superior performance compared to state-of-the-art methods, particularly in terms of Precision, Recall, F1-Score and area under the precision-recall curve (AUPR). Ablation analysis highlighted the substantial contribution of knowledge graph-derived features. Moreover, KG-CNNDTI was employed for virtual screening of natural products against Alzheimer's disease, resulting in 40 candidate compounds. 5 were supported by literature evidence, among which 3 were further validated in vitro assays.
Alzheimer Disease/drug therapy* ; Biological Products/therapeutic use* ; Humans ; Neural Networks, Computer ; Machine Learning ; Drug Discovery/methods* ; Algorithms ; Drug Evaluation, Preclinical/methods*

OBJECTIVES: This study investigates independent risk factors for postoperative internal fixation device infection in patients with maxillofacial fractures and proposes an early warning model based on the synthetic minority over-sampling technique (SMOTE) algorithm. METHODS: A total of 1 104 patients who underwent surgical treatment for maxillofacial fractures at Oral and Maxillofacial Surgery Department, Affiliated Hospital of Nantong University from January 2021 to December 2024 were retrospectively analyzed. The patients were divided into two groups based on the presence of postoperative internal fixation device infection: the infection group (27 cases) and non-infection group (1 077 cases). Clinical data from both groups were collected and subjected to statistical analysis. Univariate and binary Logistic regression analysis were used to identify risk factors for postoperative internal fixation device infection in maxillofacial fractures. Subsequently, a Logistic regression model was established, and the dataset was improved based on the SMOTE algorithm to construct an early warning model with the improved dataset. The prediction performance of the models was compared and validated. RESULTS: Among the 1 104 patients who underwent surgical treatment for maxillofacial fractures, 27 cases of postoperative internal fixation device infections were identified, corresponding to an infection rate of 2.45% (27/1 104). Age, diabetes history, fracture severity, and oral hygiene status were all identified as risk factors for postoperative internal fixation device infections in maxillofacial fractures (all P<0.05). The prediction model based on the original data (P1). The prediction model based on the SMOTE algorithm (P2). Receiver operating characteristic (ROC) curve analysis shows that the area under curve (AUC) for the P2 model was 0.882, the P1 model was 0.861, indicating the superior predictive performance of the P2 model. The DeLong test results show that the difference in AUC between the two models was statistically significant (P<0.05). CONCLUSIONS Age, diabetes history, postoperative fracture severity, and oral hygiene status are all risk factors for infections associated with internal fixation devices after maxillofacial fracture surgery. The proposed early warning model demonstrated good predictive performance. Medical professionals can utilize this model to effectively intervene and anticipate infections related to internal fixation devices after maxillofacial fracture surgery.
Humans ; Algorithms ; Retrospective Studies ; Male ; Female ; Fracture Fixation, Internal/instrumentation* ; Risk Factors ; Middle Aged ; Adult ; Logistic Models ; Surgical Wound Infection/epidemiology* ; Aged ; Internal Fixators/adverse effects* ; Maxillofacial Injuries/surgery* ; Adolescent

OBJECTIVES: This study aimed to use machine learning algorithms to build a prediction model of the first permanent molar caries of 9-year-old children in Suzhou and screen out risk factors. METHODS: Random stratified whole group sampling was applied to randomly select 9-year-old students from 38 primary schools in 14 townships and streets in Wuzhong District for oral examination and questionnaire survey. Multifactor Logistics regression was used to analyze the risk factors of tooth decay. The data set was randomly divided into training sets and verification sets according to 8∶2, and R 4.3.1 was used to build five machine learning algorithms: random forest, decision tree, extreme gradient boosting (XGBoost), Logistics regression, and lightweight gradient enhancement (LightGBM). The predictive effect of these five models was evaluated using the area under the characteristic curve (AUC). The marginal contribution of quantitative characteristics to the caries prediction model was determined through Shapley additive explanations (SHAP). RESULTS: This study included 7 225 samples that met the standard. The caries rate of the first permanent molar was 54.96%. Multifactor Logistic regression analysis showed that sweet drinks, dessert and candy, snack frequency, and snacks before going to bed after brushing teeth were correlated with the occurrence of first permanent molar caries (P<0.05). The AUC values of decision tree, Logistic regression, LightGBM, random forest, and XGBoost were 75.5%, 83.9%, 88.6%, 88.9%, and 90.1%, respectively. Compared with the variables after single heat coding, the SHAP value of high-frequency sweets (such as dessert candy ≥2 times a day, mother's sugary diet ≥2 times a day) and bad oral hygiene habits (such as frequent snacks before going to bed after brushing teeth and irregular brushing teeth) exhibited the highest positive. CONCLUSIONS XGBoost algorithm has a good prediction effect for first permanent molar caries in 9-year-old children. High-frequency sweet factors and bad oral hygiene habits have a strong positive impact on the risk of first permanent molar caries and are key drivers that can be used in the formulation of targeted interventions.
Humans ; Dental Caries/epidemiology* ; Child ; Machine Learning ; China/epidemiology* ; Molar ; Risk Factors ; Female ; Logistic Models ; Male ; Decision Trees ; Algorithms

Nucleic acid elements are essential functional sequences that play critical roles in regulating gene expression, optimizing pathways, and enabling gene editing to enhance the production of target products in biomanufacturing. Therefore, the design and optimization of these elements are crucial in constructing efficient cell factories. Artificial intelligence (AI) provides robust support for biomanufacturing by accurately predicting functional nucleic acid elements, designing and optimizing sequences with quantified functions, and elucidating the operating mechanisms of these elements. In recent years, AI has significantly accelerated the progress in biomanufacturing by reducing experimental workloads through the design and optimization of promoters, ribosome-binding sites, terminators, and their combinations. Despite these advancements, the application of AI in biomanufacturing remains limited due to the complexity of biological systems and the lack of highly quantified training data. This review summarizes the various nucleic acid elements utilized in biomanufacturing, the tools developed for predicting and designing these elements based on AI algorithms, and the case studies showcasing the applications of AI in biomanufacturing. By integrating AI with synthetic biology and high-throughput techniques, we anticipate the development of more efficient tools for designing nucleic acid elements and accelerating the application of AI in biomanufacturing.
Artificial Intelligence ; Synthetic Biology ; Nucleic Acids/genetics* ; Algorithms ; Gene Editing ; Promoter Regions, Genetic ; Biotechnology/methods*