Development and validation of a nomogram for predicting stroke risk in rheumatoid arthritis patients.

We developed and validated a nomogram to predict the risk of stroke in patients with rheumatoid arthritis (RA) in northern China. Out of six machine learning algorithms studied to improve diagnostic and prognostic accuracy of the prediction model, the logistic regression algorithm showed high performance in terms of calibration and decision curve analysis. The nomogram included stratifications of sex, age, systolic blood pressure, C-reactive protein, erythrocyte sedimentation rate, total cholesterol, and low-density lipoprotein cholesterol along with the history of traditional risk factors such as hypertensive, diabetes, atrial fibrillation, and coronary heart disease. The nomogram exhibited a high Hosmer-Lemeshow goodness-for-fit and good calibration (P > 0.05). The analysis, including the area under the receiver operating characteristic curve, the net reclassification index, the integrated discrimination improvement, and clinical use, showed that our prediction model was more accurate than the Framingham risk model in predicting stroke risk in RA patients. In conclusion, the nomogram can be used for individualized preoperative prediction of stroke risk in RA patients.

Risk prediction of in-hospital mortality among patients with type 2 diabetes mellitus and concomitant community-acquired pneumonia.

BACKGROUND: The aim of the present study was to investigate the risk factors for in-hospital mortality among patients with type 2 diabetes mellitus (T2DM) and concomitant community-acquired pneumonia (CAP) and establish a risk prediction score. METHODS: Data from 1,360 adult patients with T2DM and concomitant CAP hospitalized in two grade 3A hospitals between 2009 and 2019 were collected through electronic medical records. Data obtained included the status of diabetes mellitus, comorbidities, laboratory and imaging findings, and treatment outcomes. Statistical analysis was conducted to investigate the risk factors affecting prognosis, and a clinical risk prediction score was designed. RESULTS: Based on the patients' treatment outcomes (deceased, improved and cured), the cohort was divided into two groups: deceased and improved; 16 parameters were significant after segmentation. However, the following nine parameters were independent predictors of mortality: neutrophil-lymphocyte ratio (NLR) ≥4, pulse rate ≥125 bpm, change in state of consciousness, arterial blood pH ≤7.35, age ≥65 years, serum sodium ≤130 mmol/L, initial fasting blood glucose ≥9 mmol/L, multilobar involvement, and diabetic nephropathy. Based on these findings, a risk prediction score was established, and bootstrap validation was performed. The risk prediction score was significantly superior to CURB-65 [confusion, urea >7 mmol/L, respiratory rate >30/min, low blood pressure (systolic <90 mmHg or diastolic <60 mmHg), age >65 years] and slightly superior than the pneumonia severity index (PSI). CONCLUSIONS: The influencing factors for in-hospital mortality among patients with T2DM and concomitant CAP included advanced age, change in state of consciousness, increased pulse rate, acidosis, high NLR, high platelet-lymphocyte ratio, hyponatremia, hyperglycemia, and diabetic nephropathy. These parameters should be recognized in clinical practice, with active interventions to improve the treatment success rate. The risk prediction score effectively differentiated the mortality risk of inpatients, thereby providing guidance on clinical decision-making.

Identifying interactions in the time and frequency domains in local and global networks - A Granger Causality Approach.

BACKGROUND: Reverse-engineering approaches such as Bayesian network inference, ordinary differential equations (ODEs) and information theory are widely applied to deriving causal relationships among different elements such as genes, proteins, metabolites, neurons, brain areas and so on, based upon multi-dimensional spatial and temporal data. There are several well-established reverse-engineering approaches to explore causal relationships in a dynamic network, such as ordinary differential equations (ODE), Bayesian networks, information theory and Granger Causality. RESULTS: Here we focused on Granger causality both in the time and frequency domain and in local and global networks, and applied our approach to experimental data (genes and proteins). For a small gene network, Granger causality outperformed all the other three approaches mentioned above. A global protein network of 812 proteins was reconstructed, using a novel approach. The obtained results fitted well with known experimental findings and predicted many experimentally testable results. In addition to interactions in the time domain, interactions in the frequency domain were also recovered. CONCLUSIONS: The results on the proteomic data and gene data confirm that Granger causality is a simple and accurate approach to recover the network structure. Our approach is general and can be easily applied to other types of temporal data.

Granger causality vs. dynamic Bayesian network inference: a comparative study.

BACKGROUND: In computational biology, one often faces the problem of deriving the causal relationship among different elements such as genes, proteins, metabolites, neurons and so on, based upon multi-dimensional temporal data. Currently, there are two common approaches used to explore the network structure among elements. One is the Granger causality approach, and the other is the dynamic Bayesian network inference approach. Both have at least a few thousand publications reported in the literature. A key issue is to choose which approach is used to tackle the data, in particular when they give rise to contradictory results. RESULTS: In this paper, we provide an answer by focusing on a systematic and computationally intensive comparison between the two approaches on both synthesized and experimental data. For synthesized data, a critical point of the data length is found: the dynamic Bayesian network outperforms the Granger causality approach when the data length is short, and vice versa. We then test our results in experimental data of short length which is a common scenario in current biological experiments: it is again confirmed that the dynamic Bayesian network works better. CONCLUSION: When the data size is short, the dynamic Bayesian network inference performs better than the Granger causality approach; otherwise the Granger causality approach is better.