Artificial intelligence prediction of In-Hospital mortality in patients with dementia: A multi-center study.

BACKGROUND: Prediction of mortality is very important for care planning in hospitalized patients with dementia and artificial intelligence has the potential to serve as a solution; however, this issue remains unclear. Thus, this study was conducted to elucidate this matter. METHODS: We identified 10,573 hospitalized patients aged ≥ 45 years with dementia from three hospitals between 2010 and 2020 for this study. Utilizing 44 feature variables extracted from electronic medical records, an artificial intelligence (AI) model was constructed to predict death during hospitalization. The data was randomly separated into 70 % training set and 30 % testing set. We compared predictive accuracy among six algorithms including logistic regression, random forest, extreme gradient boosting (XGBoost), Light Gradient Boosting Machine (LightGBM), multilayer perceptron (MLP), and support vector machine (SVM). Additionally, another set of data collected in 2021 was used as the validation set to assess the performance of six algorithms. RESULTS: The average age was 79.8 years, with females constituting 54.5 % of the sample. The in-hospital mortality rate was 6.7 %. LightGBM exhibited the highest area under the curve (0.991) for predicting mortality compared to other algorithms (XGBoost: 0.987, random forest: 0.985, logistic regression: 0.918, MLP: 0.898, SVM: 0.897). The accuracy, sensitivity, positive predictive value, and negative predictive value of LightGBM were 0.943, 0.944, 0.943, 0.542, and 0.996, respectively. Among the features in LightGBM, the three most important variables were the Glasgow Coma Scale, respiratory rate, and blood urea nitrogen. In the validation set, the area under the curve of LightGBM reached 0.753. CONCLUSIONS: The AI prediction model demonstrates strong accuracy in predicting in-hospital mortality among patients with dementia, suggesting its potential implementation to enhance future care quality.

Heat-related illness and dementia: a study integrating epidemiological and experimental evidence.

BACKGROUND: Heat-related illness (HRI) is commonly considered an acute condition, and its potential long-term consequences are not well understood. We conducted a population-based cohort study and an animal experiment to evaluate whether HRI is associated with dementia later in life. METHODS: The Taiwan National Health Insurance Research Database was used in the epidemiological study. We identified newly diagnosed HRI patients between 2001 and 2015, but excluded those with any pre-existing dementia, as the study cohort. Through matching by age, sex, and the index date with the study cohort, we selected individuals without HRI and without any pre-existing dementia as a comparison cohort at a 1:4 ratio. We followed each cohort member until the end of 2018 and compared the risk between the two cohorts using Cox proportional hazards regression models. In the animal experiment, we used a rat model to assess cognitive functions and the histopathological changes in the hippocampus after a heat stroke event. RESULTS: In the epidemiological study, the study cohort consisted of 70,721 HRI patients and the comparison cohort consisted of 282,884 individuals without HRI. After adjusting for potential confounders, the HRI patients had a higher risk of dementia (adjusted hazard ratio [AHR] = 1.24; 95% confidence interval [CI]: 1.19-1.29). Patients with heat stroke had a higher risk of dementia compared with individuals without HRI (AHR = 1.26; 95% CI: 1.18-1.34). In the animal experiment, we found cognitive dysfunction evidenced by animal behavioral tests and observed remarkable neuronal damage, degeneration, apoptosis, and amyloid plaque deposition in the hippocampus after a heat stroke event. CONCLUSIONS: Our epidemiological study indicated that HRI elevated the risk of dementia. This finding was substantiated by the histopathological features observed in the hippocampus, along with the cognitive impairments detected, in the experimental heat stroke rat model.

Higher migraine risk in healthcare professionals than in general population: a nationwide population-based cohort study in Taiwan.

BACKGROUND: High stress levels and shift work probably trigger migraine in healthcare professionals (HCPs). However, the migraine risk differences between HCPs and the general population is unknown. METHODS: This nationwide population-based cohort study used Taiwan's National Health Insurance Research Database. Physicians (50,226), nurses (122,357), and other HCPs (pharmacists, technicians, dietitians, rehabilitation therapists, social workers, etc.) (45,736) were enrolled for the study cohort, and randomly selected non-HCPs (218,319) were enrolled for the comparison cohort. Conditional logistical regression analysis was used to compare the migraine risks. Comparisons between HCPs and between physician specialties were also done. RESULTS: Physicians, nurses, and other HCPs had higher migraine risks than did the general population (adjusted odds ratio [AOR]: 1.672; 95 % confidence interval [CI]: 1.468-1.905, AOR: 1.621; 95 % CI: 1.532-1.714, and AOR: 1.254; 95 % CI: 1.124-1.399, respectively) after stroke, hypertension, epilepsy, anxiety, depression, and insomnia had been adjusted for. Nurses and physicians had higher migraine risks than did other HCPs (AOR: 1.303; 95 % CI: 1.206-1.408, and AOR: 1.193; 95 % CI: 1.069-1.332, respectively). Obstetricians and gynecologists had a lower migraine risk than did other physician specialists (AOR: 0.550; 95 % CI: 0.323-0.937). CONCLUSION: HCPs in Taiwan had a higher migraine risk than did the general population. Heavy workloads, emotional stress, and rotating night shift sleep disturbances appear to be the most important risk factors. These findings should provide an important reference for promoting occupational health in HCPs in Taiwan.

Will computed tomography (CT) miss something? The characteristics and pitfalls of torso CT in evaluating patients with blunt solid organ trauma.

BACKGROUND: Selective nonoperative management has become the Standard care for blunt solid organ trauma patients, and torso computed tomography (CT) provides useful therapeutic clues. We conducted this study to determine the frequency and character of missed diagnoses in blunt solid organ trauma patients. METHODS: We reviewed the medical records of all blunt trauma patients who underwent torso CT and who were admitted for solid organ injuries (liver, spleen and kidney) at the Chi- Mei Medical Center from August 2003 to October 2006. RESULTS: The patients were divided into the Missed Group (24 patients) and the Unaltered Group (262 patients) according to the presence or absence of a missed diagnosis. The overall missed diagnosis rate was 8.4%. Only one unidentified bowel injury was disclosed by follow-up CT, and all of the missed injuries were revealed by laparotomy. The Missed Group had a higher Injury Severity Score, lower Glasgow Coma Scale, more Intensive Care Unit (ICU) care, and longer duration of hospitalization. CONCLUSION: Discovery of missed diagnoses is not uncommon in patients who sustain severe trauma. Laparotomy revealed all of the missed diagnoses, and follow-up CT demonstrated a poor ability to detect unidentified injuries. We suggest laparotomy instead of follow-up CT in the nonoperative management of patients with blunt solid organ injuries if clinical deterioration occurs.