Background: Machine learning (ML) can keep improving predictions and generating automated knowledge via data-driven predictors or decisions. Objective: The purpose of this study was to compare different ML methods including random forest, logistics regression, linear support vector machine (SVM), polynomial SVM, radial SVM, and sigmoid SVM in terms of their accuracy, sensitivity, specificity, negative predictor values, and positive predictive values by validating real datasets to predict factors for pressure ulcers (PUs). Methods: We applied representative ML algorithms (random forest, logistic regression, linear SVM, polynomial SVM, radial SVM, and sigmoid SVM) to develop a prediction model (N = 60). Results: The random forest model showed the greatest accuracy (0.814), followed by logistic regression (0.782), polynomial SVM (0.779), radial SVM (0.770), linear SVM (0.767), and sigmoid SVM (0.674). Conclusions: The random forest model showed the greatest accuracy for predicting PUs in nursing homes (NHs). Diverse factors that predict PUs in NHs including NH characteristics and residents’ characteristics were identified according to diverse ML methods. These factors should be considered to decrease PUs in NH residents.

The incidence rate of middle-aged ischemic stroke patients has been increasing. Because of their high risk of recurrence, it is important for them to adhere to their self-care performance. The purpose of this study was to investigate predictors of self-care performance among middle-aged ischemic stroke patients. A descriptive correlational and cross-sectional survey design was used. About 140 participants who had been diagnosed with ischemic stroke within a year and visited neurological outpatients were investigated. Participants answered self-report questionnaires, the Secondary Stroke Prevention Knowledge Scale, the Health-Related Hardiness Scale, and the Subjective Self-Care Performance Scale. Hardiness, secondary-stroke-prevention knowledge, age, and stroke-diagnosis duration had statistically significant influences on self-care performance; hardiness had the largest effect. This study highlighted hardiness as the major predictor of self-care performance. To improve middle-aged ischemic stroke patients’ self-care performance, nurses need an integrated approach that considers a patient’s age, diagnosis duration, secondary-stroke-prevention knowledge, and, especially, hardiness.

Objective: An investigation of the prevalence of medication persistence and associated factors in order to inform effective strategies for improving medication persistence. Methods: A systematic review of the literature from 2010 to the present was performed, using the PRISMA protocol. Primary and empirical observational studies of adult ischemic stroke or transient ischemic attack patients were included. PubMed, CINAHL, Web of Science, Cochrane Library, and PsycInfo databases were searched using the key terms stroke, ischemic stroke, medication persistence, medication adherence, and patient compliance. Results: Of four hundred twenty-eight journal articles retrieved, a final 18 articles were included. Short-term medication persistence was 46.2–96.7%, and long-term medication persistence was 41.7–93.0%. Identified hospital-related factors for medication persistence were stroke unit care, in-hospital medical complications, and early follow-up visit. Demographic factors for medication persistence were older age, and high/adequate financial status; disease-related factors were disease history, stroke subtype, and symptom severity. Age less than 75, female sex, comorbidity, antiplatelet medication switch, and polypharmacy were identified as factors of medication nonpersistence. Conclusions: Stroke patients’ medication persistence decreases over time, and persistence on antiplatelets, anticoagulants, and statin was poor. Several factors were associated with medication persistence, and these factors should be considered in future secondary preventative strategies.

The objective of this study was to test a self-management model for self-management in people living with HIV and type 2 diabetes (PLWH + T2DM). We conducted a predictive, longitudinal study of data from a national research cohort of PLWH using lag analysis to test short- and long-term health outcomes for PLWH + T2DM. We used a dataset from the Center for AIDS Research (CFAR) Network of Integrated Clinic Systems (CNICS), a nation-wide research network of 8 clinics that serves PLWH. Patient-reported outcomes, collected at clinic visit, included depression, adherence, CD4 cell count, and health-related quality of life (HRQoL). We computed summary statistics to describe the sample. Using lag analysis, we then modeled the three variables of adherence, CD4 count, and HRQoL as a function of their predecessors in our conceptual model. In the final model, an increase of in medication adherence corresponded to a small increase in HRQoL. An increase in CD4 count corresponded to a small increase in HRQoL. An increase in lagged depression was associated with a small decrease in HRQoL. The model was not sufficient to predict short- or long-term outcomes in PLWH + T2DM. Although depression had a moderate impact, the final model was not clinically significant. For people with a dual diagnosis of HIV and T2DM, variables other than those traditionally addressed in self-management interventions may be more important.

(1) Background: The nursing home (NH) research field lacks quality reporting about meta-analyses (MAs), and most gradings of MA evidence are biased on analyzing the effectiveness of independent variables in randomized control trials. (2) Objectives: This study aimed to perform a critical methodological review of MAs in the NH research field. (3) Methods: We searched the articles from four databases (PubMed, MEDLINE, CINAHL, and PsycINFO) until 15th January 2021. We reviewed a total of 41 published review articles in the NH research field. (4) Results: The studies primarily fell into the following categories: medicine (17/41), nursing (7/41), and psychiatry or psychology (6/41); 36.6% of the reviewed studies did not use any validated MA guidelines. The lowest correctly reported PRISMA 2000 guideline item was protocol and registration (14.6%), and more than 50% of articles did not report risk of bias. Moreover, 78.0% of studies did not describe missing reports of effect size formula. (5) Discussion: NH researchers must follow appropriate and updated guidelines for their MAs in order to provide validated reviews, as well as consider statistical issues such as the complexity of interventions, proper grouping, and scientific effect-size calculations to improve the quality of their study. Future quality review studies should investigate more diverse studies.