Early prediction of in-hospital mortality utilizing multivariate predictive modelling of electronic medical records and socio-determinants of health of the first day of hospitalization.

BACKGROUND: In France an average of 4% of hospitalized patients die during their hospital stay. To aid medical decision making and the attribution of resources, within a few days of admission the identification of patients at high risk of dying in hospital is essential. METHODS: We used de-identified routine patient data available in the first 2 days of hospitalization in a French University Hospital (between 2016 and 2018) to build models predicting in-hospital mortality (at ≥ 2 and ≤ 30 days after admission). We tested nine different machine learning algorithms with repeated 10-fold cross-validation. Models were trained with 283 variables including age, sex, socio-determinants of health, laboratory test results, procedures (Classification of Medical Acts), medications (Anatomical Therapeutic Chemical code), hospital department/unit and home address (urban, rural etc.). The models were evaluated using various performance metrics. The dataset contained 123,729 admissions, of which the outcome for 3542 was all-cause in-hospital mortality and 120,187 admissions (no death reported within 30 days) were controls. RESULTS: The support vector machine, logistic regression and Xgboost algorithms demonstrated high discrimination with a balanced accuracy of 0.81 (95%CI 0.80-0.82), 0.82 (95%CI 0.80-0.83) and 0.83 (95%CI 0.80-0.83) and AUC of 0.90 (95%CI 0.88-0.91), 0.90 (95%CI 0.89-0.91) and 0.90 (95%CI 0.89-0.91) respectively. The most predictive variables for in-hospital mortality in all three models were older age (greater risk), and admission with a confirmed appointment (reduced risk). CONCLUSION: We propose three highly discriminating machine-learning models that could improve clinical and organizational decision making for adult patients at hospital admission.

Cohort profile for development of machine learning models to predict healthcare-related adverse events (Demeter): clinical objectives, data requirements for modelling and overview of data set for 2016-2018.

PURPOSE: In-hospital health-related adverse events (HAEs) are a major concern for hospitals worldwide. In high-income countries, approximately 1 in 10 patients experience HAEs associated with their hospital stay. Estimating the risk of an HAE at the individual patient level as accurately as possible is one of the first steps towards improving patient outcomes. Risk assessment can enable healthcare providers to target resources to patients in greatest need through adaptations in processes and procedures. Electronic health data facilitates the application of machine-learning methods for risk analysis. We aim, first to reveal correlations between HAE occurrence and patients' characteristics and/or the procedures they undergo during their hospitalisation, and second, to build models that allow the early identification of patients at an elevated risk of HAE. PARTICIPANTS: 143 865 adult patients hospitalised at Grenoble Alpes University Hospital (France) between 1 January 2016 and 31 December 2018. FINDINGS TO DATE: In this set-up phase of the project, we describe the preconditions for big data analysis using machine-learning methods. We present an overview of the retrospective de-identified multisource data for a 2-year period extracted from the hospital's Clinical Data Warehouse, along with social determinants of health data from the National Institute of Statistics and Economic Studies, to be used in machine learning (artificial intelligence) training and validation. No supplementary information or evaluation on the part of medical staff will be required by the information system for risk assessment. FUTURE PLANS: We are using this data set to develop predictive models for several general HAEs including secondary intensive care admission, prolonged hospital stay, 7-day and 30-day re-hospitalisation, nosocomial bacterial infection, hospital-acquired venous thromboembolism, and in-hospital mortality.

Psoriatic arthritis epidemiology, comorbid disease profiles and risk factors: results from a claims database analysis.

OBJECTIVE: Psoriasis is a systemic inflammatory disease often accompanied by comorbidities, including metabolic syndrome, cardiovascular diseases and depression. Up to 41% of psoriasis patients develop psoriatic arthritis (PsA), making it one of the most relevant manifestations. A large health claims data set was analysed to determine the rate of PsA development in psoriasis patients. Furthermore, comorbid disease profiles of psoriasis patients with or without PsA were compared, and potential risk factors for the development of PsA were identified. METHODS: This was a non-interventional, retrospective analysis of anonymized insurance health claims data using a subset of the Institute of Applied Health Research Berlin (InGef) database. The primary outcome was the prevalence and incidence of diagnosed PsA among psoriasis patients in Germany. Risk factors for the development of PsA in psoriasis patients were determined by conditional logistic regression analysis. RESULTS: The cumulative percentage of patients with existing psoriasis developing concomitant PsA over 4 years was 3.44%, with a mean time to diagnosis of PsA of 1.5 years. Psoriasis patients diagnosed with acute rheumatism (odds ratio: 2.93, 95% CI = 1.76, 4.86; P < 0.001) or pain in unspecific joints (odds ratio: 1.74, 95% CI = 1.01, 2.99; P = 0.047) showed an increased risk for development of PsA later on. Interestingly, fewer than half of the patients with concomitant PsA consulted a rheumatologist. CONCLUSIONS: Unspecific arthritic symptoms are likely to precede PsA diagnoses and can develop soon after onset of psoriasis, with accumulating risk over time. There is a high unmet need for early rheumatological assessment of psoriasis patients.

Involvement of Lactate and Pyruvate in the Anti-Inflammatory Effects Exerted by Voluntary Activation of the Sympathetic Nervous System.

We recently demonstrated that the sympathetic nervous system can be voluntarily activated following a training program consisting of cold exposure, breathing exercises, and meditation. This resulted in profound attenuation of the systemic inflammatory response elicited by lipopolysaccharide (LPS) administration. Herein, we assessed whether this training program affects the plasma metabolome and if these changes are linked to the immunomodulatory effects observed. A total of 224 metabolites were identified in plasma obtained from 24 healthy male volunteers at six timepoints, of which 98 were significantly altered following LPS administration. Effects of the training program were most prominent shortly after initiation of the acquired breathing exercises but prior to LPS administration, and point towards increased activation of the Cori cycle. Elevated concentrations of lactate and pyruvate in trained individuals correlated with enhanced levels of anti-inflammatory interleukin (IL)-10. In vitro validation experiments revealed that co-incubation with lactate and pyruvate enhances IL-10 production and attenuates the release of pro-inflammatory IL-1ß and IL-6 by LPS-stimulated leukocytes. Our results demonstrate that practicing the breathing exercises acquired during the training program results in increased activity of the Cori cycle. Furthermore, this work uncovers an important role of lactate and pyruvate in the anti-inflammatory phenotype observed in trained subjects.

Promising Metabolite Profiles in the Plasma and CSF of Early Clinical Parkinson's Disease.

Parkinson's disease (PD) shows high heterogeneity with regard to the underlying molecular pathogenesis involving multiple pathways and mechanisms. Diagnosis is still challenging and rests entirely on clinical features. Thus, there is an urgent need for robust diagnostic biofluid markers. Untargeted metabolomics allows establishing low-molecular compound biomarkers in a wide range of complex diseases by the measurement of various molecular classes in biofluids such as blood plasma, serum, and cerebrospinal fluid (CSF). Here, we applied untargeted high-resolution mass spectrometry to determine plasma and CSF metabolite profiles. We semiquantitatively determined small-molecule levels (≤1.5 kDa) in the plasma and CSF from early PD patients (disease duration 0-4 years; n = 80 and 40, respectively), and sex- and age-matched controls (n = 76 and 38, respectively). We performed statistical analyses utilizing partial least square and random forest analysis with a 70/30 training and testing split approach, leading to the identification of 20 promising plasma and 14 CSF metabolites. These metabolites differentiated the test set with an AUC of 0.8 (plasma) and 0.9 (CSF). Characteristics of the metabolites indicate perturbations in the glycerophospholipid, sphingolipid, and amino acid metabolism in PD, which underscores the high power of metabolomic approaches. Further studies will enable to develop a potential metabolite-based biomarker panel specific for PD.

Metabolomic Profiles for Primary Progressive Multiple Sclerosis Stratification and Disease Course Monitoring.

Primary progressive multiple sclerosis (PPMS) shows a highly variable disease progression with poor prognosis and a characteristic accumulation of disabilities in patients. These hallmarks of PPMS make it difficult to diagnose and currently impossible to efficiently treat. This study aimed to identify plasma metabolite profiles that allow diagnosis of PPMS and its differentiation from the relapsing-remitting subtype (RRMS), primary neurodegenerative disease (Parkinson's disease, PD), and healthy controls (HCs) and that significantly change during the disease course and could serve as surrogate markers of multiple sclerosis (MS)-associated neurodegeneration over time. We applied untargeted high-resolution metabolomics to plasma samples to identify PPMS-specific signatures, validated our findings in independent sex- and age-matched PPMS and HC cohorts and built discriminatory models by partial least square discriminant analysis (PLS-DA). This signature was compared to sex- and age-matched RRMS patients, to patients with PD and HC. Finally, we investigated these metabolites in a longitudinal cohort of PPMS patients over a 24-month period. PLS-DA yielded predictive models for classification along with a set of 20 PPMS-specific informative metabolite markers. These metabolites suggest disease-specific alterations in glycerophospholipid and linoleic acid pathways. Notably, the glycerophospholipid LysoPC(20:0) significantly decreased during the observation period. These findings show potential for diagnosis and disease course monitoring, and might serve as biomarkers to assess treatment efficacy in future clinical trials for neuroprotective MS therapies.