No abstract

Diabetic sensorimotor polyneuropathy (DSPN) is a major complication in patients with diabetes mellitus (DM), and early detection or prediction of DSPN is important for preventing or managing neuropathic pain and foot ulcer. Our aim is to delineate whether machine learning techniques are more useful than traditional statistical methods for predicting DSPN in DM patients. Four hundred seventy DM patients were classified into four groups (normal, possible, probable, and confirmed) based on clinical and electrophysiological findings of suspected DSPN. Three ML methods, XGBoost (XGB), support vector machine (SVM), and random forest (RF), and their combinations were used for analysis. RF showed the best area under the receiver operator characteristic curve (AUC, 0.8250) for differentiating between two categories—criteria by clinical findings (normal, possible, and probable groups) and those by electrophysiological findings (confirmed group)—and the result was superior to that of linear regression analysis (AUC = 0.6620). Average values of serum glucose, International Federation of Clinical Chemistry (IFCC), HbA1c, and albumin levels were identified as the four most important predictors of DSPN. In conclusion, machine learning techniques, especially RF, can predict DSPN in DM patients effectively, and electrophysiological analysis is important for identifying DSPN.

Background and purpose Previous studies on the weekend effect—a phenomenon where stroke outcomes differ depending on whether the stroke occurred on a weekend—mostly targeted ischemic stroke and showed inconsistent results. Thus, we investigated the weekend effect on 30-day mortality in patients with ischemic or hemorrhagic stroke considering the confounding effect of stroke severity and staffing level. Methods We retrospectively analyzed data of patients hospitalized for ischemic or hemorrhagic stroke between January 1, 2015, and December 31, 2018, which were extracted from the claims database of the National Health Insurance System and the Medical Resource Report by the Health Insurance Review & Assessment Service. The primary outcome measure was 30-day all-cause mortality. Results In total, 278,632 patients were included, among whom 84,240 and 194,392 had a hemorrhagic and ischemic stroke, respectively, with 25.8% and 25.1% of patients, respectively, being hospitalized during the weekend. Patients admitted on weekends had significantly higher 30-day mortality rates (hemorrhagic stroke 16.84%>15.55%, p<0.0001; ischemic stroke 5.06%>4.92%, p<0.0001). However, in the multi-level logistic regression analysis adjusted for case-mix, pre-hospital, and hospital level factors, the weekend effect remained consistent in patients with hemorrhagic stroke (odds ratio [OR] 1.05, 95% confidence interval [CI] 1.00–1.10), while the association was no longer evident in patients with ischemic stroke (OR 1.01, 95% CI 0.96–1.06). Conclusions Weekend admission for hemorrhagic stroke was significantly associated with a higher mortality rate after adjusting for confounding factors. Further studies are required to understand factors contributing to mortality during weekend admission.