Classifier calibration using splined empirical probabilities in clinical risk prediction

BackgroundSeptic arthritis is an uncommon but potentially significant diagnosis to be considered when a child presents to the emergency department (ED) with non-traumatic limp. Our objective was to determine the diagnostic accuracy of clinical findings (history and examination) and investigation results (pathology tests and imaging) for the diagnosis of septic arthritis among children presenting with acute non-traumatic limp to the ED.MethodsSystematic review of the literature published between 1966 and June 2019 on MEDLINE and EMBASE databases. Studies were included if they evaluated children presenting with lower limb complaints and evaluated diagnostic performance of items from history, physical examination, laboratory testing or radiological examination. Data were independently extracted by two authors, and quality assessment was performed using the Quality Assessment Tool for Diagnostic Accuracy Studies 2 tool.Results18 studies were identified, and included 2672 children (560 with a final diagnosis of septic arthritis). There was substantial heterogeneity in inclusion criteria, study setting, definitions of specific variables and the gold standard used to confirm septic arthritis. Clinical and investigation findings were reported using varying definitions and cut-offs, and applied to differing study populations. Spectrum bias and poor-to-moderate study design quality limit their applicability to the ED setting.Single studies suggest that the presence of joint tenderness (n=189; positive likelihood ratio 11.4 (95% CI 5.9 to 22.0); negative likelihood ratio 0.2 (95% CI 0.0 to 1.2)) and joint effusion on ultrasound (n=127; positive likelihood ratio 8.4 (95% CI 4.1 to 17.1); negative likelihood ratio 0.2 (95% CI 0.1 to 0.3)) appear to be useful. Two promising clinical risk prediction tools were identified, however, their performance was notably lower when tested in external validation studies.DiscussionDifferentiating children with septic arthritis from non-emergent disorders of non-traumatic limp remains a key diagnostic challenge for emergency physicians. There is a need for prospectively derived and validated ED-based clinical risk prediction tools.

Download Full-text

A Novel Method for Clinical Risk Prediction with Low-Quality Data

Artificial Intelligence in Medicine ◽

10.1016/j.artmed.2021.102052 ◽

2021 ◽

pp. 102052

Author(s):

Zeyuan Wang ◽

Josiah Poon ◽

Shuze Wang ◽

Shiding Sun ◽

Simon Poon

Keyword(s):

Risk Prediction ◽

Quality Data ◽

Clinical Risk ◽

Novel Method

Download Full-text

Clinical Risk Prediction in Patients With Left Ventricular Myocardial Noncompaction

Journal of the American College of Cardiology ◽

10.1016/j.jacc.2021.06.016 ◽

2021 ◽

Vol 78 (7) ◽

pp. 643-662

Author(s):

Guillem Casas ◽

Javier Limeres ◽

Gerard Oristrell ◽

Laura Gutierrez-Garcia ◽

Daniele Andreini ◽

...

Keyword(s):

Risk Prediction ◽

Left Ventricular ◽

Clinical Risk

Download Full-text

Performance Metrics for the Comparative Analysis of Clinical Risk Prediction Models Employing Machine Learning

Circulation Cardiovascular Quality and Outcomes ◽

10.1161/circoutcomes.120.007526 ◽

2021 ◽

Author(s):

Chenxi Huang ◽

Shu-Xia Li ◽

César Caraballo ◽

Frederick A. Masoudi ◽

John S. Rumsfeld ◽

...

Keyword(s):

Machine Learning ◽

Risk Prediction ◽

Health Care Professionals ◽

Clinical Decision Making ◽

Performance Metrics ◽

Prediction Models ◽

Learning Models ◽

Risk Prediction Models ◽

Clinical Risk ◽

Machine Learning Models

Background: New methods such as machine learning techniques have been increasingly used to enhance the performance of risk predictions for clinical decision-making. However, commonly reported performance metrics may not be sufficient to capture the advantages of these newly proposed models for their adoption by health care professionals to improve care. Machine learning models often improve risk estimation for certain subpopulations that may be missed by these metrics. Methods and Results: This article addresses the limitations of commonly reported metrics for performance comparison and proposes additional metrics. Our discussions cover metrics related to overall performance, discrimination, calibration, resolution, reclassification, and model implementation. Models for predicting acute kidney injury after percutaneous coronary intervention are used to illustrate the use of these metrics. Conclusions: We demonstrate that commonly reported metrics may not have sufficient sensitivity to identify improvement of machine learning models and propose the use of a comprehensive list of performance metrics for reporting and comparing clinical risk prediction models.

Download Full-text

Standardised measurement of coronary inflammation using cardiovascular CT: integration in clinical care as a prognostic medical device

Cardiovascular Research ◽

10.1093/cvr/cvab286 ◽

2021 ◽

Author(s):

Evangelos K Oikonomou ◽

Alexios S Antonopoulos ◽

David Schottlander ◽

Mohammad Marwan ◽

Chris Mathers ◽

...

Keyword(s):

Risk Factors ◽

Cardiovascular Risk ◽

Coronary Artery ◽

Medical Device ◽

Risk Prediction ◽

Clinical Care ◽

Clinical Risk Factors ◽

Clinical Risk ◽

Cardiovascular Risk Prediction ◽

Perivascular Fat

Abstract Aims Coronary CT angiography (CCTA) is a first-line modality in the investigation of suspected coronary artery disease (CAD). Mapping of perivascular Fat Attenuation Index (FAI) on routine CCTA enables the non-invasive detection of coronary artery inflammation by quantifying spatial changes in perivascular fat composition. We now report the performance of a new medical device, CaRi-Heart®, which integrates standardised FAI mapping together with clinical risk factors and plaque metrics to provide individualised cardiovascular risk prediction. Methods and Results The study included 3912 consecutive patients undergoing CCTA as part of clinical care in the United States (n = 2040) and Europe (n = 1872). These cohorts were used to generate age-specific nomograms and percentile curves as reference maps for the standardised interpretation of FAI. The first output of CaRi-Heart® is the FAI-Score of each coronary artery, which provides a measure of coronary inflammation adjusted for technical, biological and anatomical characteristics. FAI-Score is then incorporated into a risk prediction algorithm together with clinical risk factors and CCTA-derived coronary plaque metrics to generate the CaRi-Heart® Risk that predicts the likelihood of a fatal cardiac event at 8 years. CaRi-Heart® Risk was trained in the US population and its performance was validated externally in the European population. It improved risk discrimination over a clinical risk factor-based model (Δ[C-statistic] of 0.085, P = 0.01 in the US Cohort and 0.149, P < 0.001 in the European cohort) and had a consistent net clinical benefit on decision curve analysis above a baseline traditional risk factor-based model across the spectrum of cardiac risk. Conclusion CaRi-Heart® reliably improves cardiovascular risk prediction by incorporating traditional cardiovascular risk factors along with comprehensive CCTA coronary plaque and perivascular adipose tissue phenotyping. This integration advances the prognostic utility of CCTA for individual patients and paves the way for its use as a screening tool among patients referred for CCTA. Translational Perspective Mapping of perivascular Fat Attenuation Index (FAI) on coronary computed tomography angiography (CCTA) enables the non-invasive detection of coronary artery inflammation by quantifying spatial changes in perivascular fat composition. We now report the performance of a new medical device, CaRi-Heart®, which integrates standardised FAI mapping together with clinical risk factors and plaque metrics to provide age-standardised reference maps and individualised cardiovascular risk prediction. This integration advances the prognostic value of CCTA and paves the way for its use as a screening tool among patients referred for CCTA.

Download Full-text

An empirical characterization of fair machine learning for clinical risk prediction

Journal of Biomedical Informatics ◽

10.1016/j.jbi.2020.103621 ◽

2020 ◽

pp. 103621

Author(s):

Stephen R. Pfohl ◽

Agata Foryciarz ◽

Nigam H. Shah

Keyword(s):

Machine Learning ◽

Risk Prediction ◽

Clinical Risk

Download Full-text

Genetic risk score for risk prediction of diabetic nephropathy in Han Chinese type 2 diabetes patients

Scientific Reports ◽

10.1038/s41598-019-56400-3 ◽

2019 ◽

Vol 9 (1) ◽

Author(s):

Li-Na Liao ◽

Tsai-Chung Li ◽

Chia-Ing Li ◽

Chiu-Shong Liu ◽

Wen-Yuan Lin ◽

...

Keyword(s):

Risk Factors ◽

Type 2 Diabetes ◽

Prediction Model ◽

Risk Prediction ◽

Genetic Risk Score ◽

Clinical Risk Factors ◽

Clinical Risk ◽

Conventional Risk Factors ◽

Diabetes Patients

AbstractWe evaluated whether genetic information could offer improvement on risk prediction of diabetic nephropathy (DN) while adding susceptibility variants into a risk prediction model with conventional risk factors in Han Chinese type 2 diabetes patients. A total of 995 (including 246 DN cases) and 519 (including 179 DN cases) type 2 diabetes patients were included in derivation and validation sets, respectively. A genetic risk score (GRS) was constructed with DN susceptibility variants based on findings of our previous genome-wide association study. In derivation set, areas under the receiver operating characteristics (AUROC) curve (95% CI) for model with clinical risk factors only, model with GRS only, and model with clinical risk factors and GRS were 0.75 (0.72–0.78), 0.64 (0.60–0.68), and 0.78 (0.75–0.81), respectively. In external validation sample, AUROC for model combining conventional risk factors and GRS was 0.70 (0.65–0.74). Additionally, the net reclassification improvement was 9.98% (P = 0.001) when the GRS was added to the prediction model of a set of clinical risk factors. This prediction model enabled us to confirm the importance of GRS combined with clinical factors in predicting the risk of DN and enhanced identification of high-risk individuals for appropriate management of DN for intervention.

Download Full-text