Identifying Individuals at High Risk for HIV and Sexually Transmitted Infections With an Artificial Intelligence-Based Risk Assessment Tool

Background

We have previously developed an artificial intelligence-based risk assessment tool to identify the individual risk of HIV and sexually transmitted infections (STIs) in a sexual health clinical setting. Based on this tool, this study aims to determine the optimal risk score thresholds to identify individuals at high risk for HIV/STIs.

Methods

Using 2008-2022 data from 216 252 HIV, 227 995 syphilis, 262 599 gonorrhea, and 320 355 chlamydia consultations at a sexual health center, we applied MySTIRisk machine learning models to estimate infection risk scores. Optimal cutoffs for determining high-risk individuals were determined using Youden's index.

Results

The HIV risk score cutoff for high risk was 0.56, with 86.0% sensitivity (95% CI, 82.9%-88.7%) and 65.6% specificity (95% CI, 65.4%-65.8%). Thirty-five percent of participants were classified as high risk, which accounted for 86% of HIV cases. The corresponding cutoffs were 0.49 for syphilis (sensitivity, 77.6%; 95% CI, 76.2%-78.9%; specificity, 78.1%; 95% CI, 77.9%-78.3%), 0.52 for gonorrhea (sensitivity, 78.3%; 95% CI, 77.6%-78.9%; specificity, 71.9%; 95% CI, 71.7%-72.0%), and 0.47 for chlamydia (sensitivity, 68.8%; 95% CI, 68.3%-69.4%; specificity, 63.7%; 95% CI, 63.5%-63.8%). High-risk groups identified using these thresholds accounted for 78% of syphilis, 78% of gonorrhea, and 69% of chlamydia cases. The odds of positivity were significantly higher in the high-risk group than otherwise across all infections: 11.4 (95% CI, 9.3-14.8) times for HIV, 12.3 (95% CI, 11.4-13.3) for syphilis, 9.2 (95% CI, 8.8-9.6) for gonorrhea, and 3.9 (95% CI, 3.8-4.0) for chlamydia.

Conclusions

Risk scores generated by the AI-based risk assessment tool MySTIRisk, together with Youden's index, are effective in determining high-risk subgroups for HIV/STIs. The thresholds can aid targeted HIV/STI screening and prevention.

Optimal cutoff value of the dry eye-related quality-of-life score for diagnosing dry eye disease

This retrospective study aimed to determine the optimal cutoff values of the Dry Eye-Related Quality-of-Life Score (DEQS) questionnaire for diagnosing dry eye disease (DED) and classifying DED severities. Participants completed the DEQS questionnaire, the Japanese version of the Ocular Surface Disease Index (J-OSDI) questionnaire, and DED examinations. DED was diagnosed according to the 2016 Asia Dry Eye Society diagnostic criteria based on DED symptoms (J-OSDI ≥ 13 points) and tear film breakup time ≤ 5 s. Receiver operating characteristic (ROC) analysis was used to calculate the optimal cutoff values of the DEQS summary score for detecting DED and grading its severity. Among 427 patients, 296 (69.3%) and 131 (30.7%) were diagnosed with DED and non-DED, respectively. ROC analysis determined an optimal cutoff value of 15.0 points for DED diagnosis, with 83.5% sensitivity, 87.0% specificity, and an area under the curve of 0.915. The positive and negative predictive values for DEQS ≥ 15.0 points were 93.6% and 69.9%, respectively. DEQS cutoff values of 15.0, 20.0, and 26.8 points could be accepted for severity classification of DED subjective symptoms in clinical use and represent mild, moderate, and severe DED, respectively. Conclusively, the optimal cutoff values of DEQS enable DED detection and subjective symptom severity classification.

Role of fatty liver in the epidemic of advanced chronic liver disease among people with HIV: protocol for the Canadian LIVEHIV multicentre prospective cohort

INTRODUCTION

Advanced chronic liver disease (ACLD) is a major cause of death for people with HIV (PWH). While viral hepatitis coinfections are largely responsible for this trend, metabolic dysfunction-associated steatotic liver disease (MASLD) is an emerging concern for PWH. We aimed to assess the contribution of MASLD to incident ACLD in PWH.

METHODS AND ANALYSIS

This multicentre prospective observational cohort study will enrol 968 consecutive HIV monoinfected patients from four Canadian sites, excluding subjects with alcohol abuse, liver disease other than MASLD, or ACLD at baseline. Participants will be followed annually for 4 years by clinical evaluation, questionnaires, laboratory testing and Fibroscan to measure liver stiffness measurement (LSM) and controlled attenuation parameter (CAP). The primary outcome will be incidence of ACLD, defined as LSM>10 kPa, by MASLD status, defined as CAP≥285 dB/m with at least one metabolic abnormality, and to develop a score to classify PWH according to their risk of ACLD. Secondary outcomes will include health-related quality of life (HRQoL) and healthcare resource usage. Kaplan-Meier survival method and Cox proportional hazards regression will calculate the incidence and predictors of ACLD, respectively. Propensity score methods and marginal structural models will account for time-varying exposures. We will split the cohort into a training set (to develop the risk score) and a validation set (for validation of the score). HRQoL scores and healthcare resource usage will be compared by MASLD status using generalised linear mixed effects model.

ETHICS AND DISSEMINATION

This protocol has been approved by the ethics committees of all participating institutions. Written informed consent will be obtained from all study participants. The results of this study will be shared through scientific publications and public presentations to advocate for the inclusion of PWH in clinical trials of MASLD-targeted therapies and case-finding of ACLD in PWH.

Specificity and sensitivity of the social communication questionnaire lifetime screening tool for autism spectrum disorder in a UK CAMHS service

INTRODUCTION

The Social Communication Questionnaire is used to identify children and young people (CYP) who may require formal ASD assessment. However, there is a paucity of research on its utility in Children and Adolescent Mental Health Services. This evaluation aimed to determine the sensitivity and specificity of the Social Communication Questionnaire (SCQ) in a UK, Midlands CAMHS service.

METHOD

Forty young people (mean age 13.75 years) were screened using the caregiver reported SCQ before completing 'gold standard' assessment.

RESULTS

The SCQ had a sensitivity of 80% and a specificity of 25.7%. ROC curve analysis indicated low diagnostic accuracy. Differences in predictive accuracy of SCQ and diagnostic standard were statistically significant (p < 0.0001).

CONCLUSION

This evaluation builds on previous research suggesting that the SCQ may not be an efficient screening tool in CAMHS settings.

The association between low protein diet and body composition, muscle function, inflammation, and amino acid-based metabolic profile in chronic kidney disease stage 3-5 patients

BACKGROUND

Chronic kidney disease (CKD) is a global burden in the world. Low protein diet (LPD) recommendation is suggested in CKD patients to avoid or defer dialysis initiation and slow down CKD progression. However, nutritional imbalance and protein energy wasting represent key worries. The amino acid-based metabolic profile may provide a sensitive biomarker to evaluate CKD patients' nutrition status with LPD recommendations.

METHODS

We conducted a cross-sectional study in CKD stage 3-5 patients who had received LPD recommendation to evaluate the association between LPD and traditional markers (including plasma levels of albumin, pre-albumin, transferrin, total iron-binding capacity), inflammation markers (including peripheral leukocyte count and plasma levels of high-sensitivity C-reactive protein), body composition, muscle strength, and physical function, and novel nutrition markers (including amino acid-based metabolic profile) in CKD stage 3-5 patients.

RESULTS

In our study CKD stage 3-5 patients with the total number of 73, the mean age was around 71 ± 10 years old. The mean daily protein intake (DPI) was around 0.9 ± 0.3 g/kg/day and 25 (34%) patients met the recommended goal of DPI <0.8 g/kg/day. The mean daily calorie intake (DCI) was around 23 ± 6 kcal/kg/day, with only 11 (15%) patients met the recommend DCI with 30-35 kcal/kg/day. Compared to CKD patients with non-LPD, patients with LPD had significantly lower hemoglobin and albumin levels, shorter 6-min walking distance (6MWD), and lower leucine levels. Multivariable analysis found that lower hemoglobin and leucine levels, and shorter 6MWD were negatively and independently associated with LPD (all p < 0.05). Then ROC curve analysis found that the optimal cut-off value of leucine plasma levels was 95.5 μM with 60% sensitivity and 71% specificity to predict those CKD patients with LPD with the area under the curve of 0.646 (95% CI: 0.512-0.780).

CONCLUSION

LPD attainment was noted in 34% patients and most of CKD stage 3-5 patients (around 85%) had inadequate daily calorie intake although receiving standard dietary counseling routinely. A low protein diet and inadequate daily calorie intake in CKD patients were associated with shorter 6MWD, and lower hemoglobin and leucine levels. Plasma leucine levels lower than 95.5 μM may be a herald for muscle wasting and malnutrition in these CKD stage 3-5 patients with inadequate calorie intake.

Cell free HPV DNA provides an accurate and rapid diagnosis of HPV-associated head and neck cancer

PURPOSE

HPV-associated Head and Neck Squamous Cell Carcinoma(HPV+HNSCC) is the most common HPV-associated malignancy in the United States and continues to increase in incidence. Current diagnostic approaches for HPV+HNSCC rely on tissue biopsy followed by histomorphologic assessment and detection of HPV indirectly by p16 immunohistochemistry. Such approaches are invasive and have variable sensitivity.

EXPERIMENTAL DESIGN

We conducted a prospective observational study in 140 subjects (70 cases and 70 controls) to test the hypothesis that a non-invasive diagnostic approach for HPV+HNSCC would have improved diagnostic accuracy, lower cost, and shorter Diagnostic Interval compared to standard approaches. Blood was collected, processed for circulating tumor HPV DNA(ctHPVDNA) and analyzed with custom ddPCR assays for HPV genotypes 16,18, 33, 35 and 45. Diagnostic performance, cost and Diagnostic Interval were calculated for standard clinical work up and compared to a non-invasive approach using ctHPVDNA combined with cross-sectional imaging and physical exam findings.

RESULTS

Sensitivity and specificity of ctHPVDNA for detecting HPV+HNSCC was 98.4% and 98.6%. Sensitivity and specificity of a composite non-invasive diagnostic using ctHPVDNA and imaging/physical exam were 95.1% and 98.6%. Diagnostic accuracy of this non-invasive approach was significantly higher than standard of care (Youden index 0.937 vs 0.707, p=0.0006). Costs of non-invasive diagnostic were 36-38% less than standard clinical work up and the median Diagnostic Interval was 26 days less.

CONCLUSIONS

A non-invasive diagnostic approach for HPV+HNSCC demonstrated improved accuracy, reduced cost and a shorter time to diagnosis compared to standard clinical workup and could be a viable alternative in the future.

IMPACT OF THORACIC ENDOVASCULAR AORTIC REPAIR TIMING ON AORTIC REMODELING IN ACUTE TYPE B AORTIC INTRAMURAL HEMATOMA

INTRODUCTION

Thoracic endovascular aortic repair (TEVAR) is increasingly utilized in the management of acute type B aortic intramural hematoma (TBIMH). Optimal timing for intervention has not been described. The aim of this study was to evaluate TEVAR timing on post-operative aortic remodeling.

METHODS

A retrospective chart review was performed on patients who underwent TEVAR for TBIMH from January 2008 to September 2018. Imaging was reviewed pre- and postoperatively. Primary data points included true lumen diameter (TLD) and total aortic diameter (TAD) at the site of maximal pathology. Primary endpoint was aortic remodeling evidenced by a TAD/TLD ratio closest to 1.0. Secondary outcome was occurrence of aortic-related adverse events and mortality (AREM): aortic rupture, aortic-related death, progression to dissection or need for aortic re-intervention within 12 months. Patients undergoing emergent TEVAR (within 24 hours, 'eTEVAR') were compared to the remainder - delayed TEVAR ('dTEVAR').

RESULTS

We analyzed 71 patients that underwent TEVAR FOR TBIMH; 25 underwent emergent TEVAR and 46 patients underwent dTEVAR (median: 5.5 days, range: 2-120 days). There were no differences in demographics and comorbidities and patients did not differ in presenting IMH thickness (12.6±3.1 mm vs. 11.3±4.1 mm, p=0.186) nor presenting TAD/TLD ratio (1.535±0.471 vs. 1.525±0.397, p=0.928) for eTEVAR and dTEVAR groups, respectively. eTEVAR patients had larger average presenting maximal descending aortic (DTA) diameter (45.8±14.3 mm vs. 38.2±7.5 mm, p=0.018) and higher incidence of penetrating aortic ulcer (PAU) on presenting CT angiography (52.0% vs 21.7%, p=0.033). Thirty day mortality was 2/25 (8.0%) for eTEVAR and 2/45 (4.4%) for dTEVAR (p=0.602). Postoperative aortic remodeling was more complete in the dTEVAR group (1.23±0.12 vs. 1.33±0.15, p=0.004). Case-control matching (controlling for presenting DTA diameter and PAU) on 30 patients still showed better aortic remodeling in the dTEVAR group (1.125±0.100 vs 1.348±0.42, p<0.001). The incidence of AREM was higher in the eTEVAR - 6/25 (24.0%) - group compared to the dTEVAR group - 2/46 (4.3%). At 12 months, freedom from AREM was higher in the dTEVAR group (95.7% vs. 76.0%, p=0.011). Postoperative TAD/TLD ratio was the best predictor for late aortic-related adverse events (AUROC=0.825, p=0.003).

CONCLUSION

TEVAR for acute type B IMH within 24 hours of admission is associated with lower aortic remodeling and higher occurrence of late aortic related adverse events and mortality. Delaying TEVAR when clinically possible could improve aortic remodeling and aortic-related outcomes.

On the Origin of Sensitivity and Specificity

Although it is commonly said that the notions of sensitivity and specificity were first defined by Jacob Yerushalmy in 1947, the sensitivity and specificity of diagnostic tests have been assessed as far back as the early 1900s. These notions share a common origin with the development of serology. They were originally immunologic concepts, closely associated with the development of complement fixation reactions for syphilis. Here, the authors trace how immunologic sensitivity and specificity were transformed into diagnostic sensitivity and specificity. By relocating the origins of these concepts to the early 20th century, they highlight how these origins were bound to then-commonplace assumptions about specific infectious disease entities.

Joint hypothesis testing of the area under the receiver operating characteristic curve and the Youden index

In the receiver operating characteristic (ROC) analysis, the area under the ROC curve (AUC) serves as an overall measure of diagnostic accuracy. Another popular ROC index is the Youden index (J), which corresponds to the maximum sum of sensitivity and specificity minus one. Since the AUC and J describe different aspects of diagnostic performance, we propose to test if a biomarker beats the pre-specified targeting values of AUC₀ and J₀ simultaneously with H₀  : AUC ≤ AUC₀ or J ≤ J₀ against H_a  : AUC > AUC₀ and J > J₀ . This is a multivariate order restrictive hypothesis with a non-convex space in H_a , and traditional likelihood ratio-based tests cannot apply. The intersection-union test (IUT) and the joint test are proposed for such test. While the IUT test independently tests for the AUC and the Youden index, the joint test is constructed based on the joint confidence region. Findings from the simulation suggest both tests yield similar power estimates. We also illustrated the tests using a real data example and the results of both tests are consistent. In conclusion, testing jointly on AUC and J gives more reliable results than using a single index, and the IUT is easy to apply and have similar power as the joint test.

Predictors of outcomes in patients with mitral regurgitation undergoing percutaneous valve repair

Percutaneous mitral valve repair has been increasingly performed worldwide after approval. We sought to investigate predictors of clinical outcome in patients with mitral regurgitation undergoing percutaneous valve repair. The MITRA-UMG study, a single-centre registry, retrospectively collected consecutive patients with symptomatic moderate-to-severe or severe MR undergoing MitraClip therapy. The primary endpoint was the composite of cardiovascular death or rehospitalization for heart failure. Between March 2012 and July 2018, a total of 150 consecutive patients admitted to our institution were included. Procedural success was obtained in 95.3% of patients. The composite primary endpoint of cardiovascular death or rehospitalization for HF was met in 55 patients (37.9%) with cumulative incidences of 7.6%, 26.2%, at 30 days and 1-year, respectively. In the Cox multivariate model, NYHA functional class and left ventricular end-diastolic volume index (LVEDVi), independently increased the risk of the primary endpoint at long-term follow-up. At Kaplan–Meier analysis, a LVEDVi > 92 ml/m² was associated with an increased incidence of the primary endpoint. In this study, patients presenting with dilated ventricles (LVEDVi > 92 ml/m²) and advanced heart failure symptoms (NYHA IV) at baseline carried the worst prognosis after percutaneous mitral valve repair.

Diagnostic accuracy of FeNO [fractional exhaled nitric oxide] and asthma symptoms increased when evaluated with a superior reference standard

OBJECTIVES

The objective of the study is to determine the impact of changing reference standards (RS), namely spirometry vs. whole-body plethysmography (WBP), on estimation of the diagnostic accuracy of fractional exhaled nitric oxide (FeNO) and clinical signs and symptoms (CSS) as index tests regarding asthma diagnosis.

STUDY DESIGN AND SETTING

This was a diagnostic study conducted in 393 patients attending a private practice of pneumologists with complaints suspicious of asthma. First, the index tests were compared with the diagnostic results of spirometry in terms of forced expiratory volume in the first second (FEV₁) responsiveness. Second, the index tests were compared with the results of WBP in terms of specific airway resistance and FEV₁ responsiveness. Areas under the curve (AUC) were compared with a generalized estimating equation approach based on binary logistic regression.

RESULTS

FeNO values and CSS 'wheezing' and 'allergic rhinitis' showed higher specificities (P < 0.001) and sensitivities (not significant) when evaluated with WBP; also, Youden indices increased in these CSS (P < 0.05). AUC of FeNO in combination with 'wheezing' and 'allergic rhinitis' when WBP was used as RS (AUC = 0.724; 95% confidence interval 0.672 to 0.776) was higher compared with spirometry as RS (AUC = 0.654; 95% confidence interval 0.585 to 0.722) (P < 0.001).

CONCLUSION

In case of asthma, superior RS led to more favorable assessment of index tests. FeNO measurement might have been underestimated in some previous studies.

Statistical inference for decision curve analysis, with applications to cataract diagnosis

Statistical learning methods are widely used in medical literature for the purpose of diagnosis or prediction. Conventional accuracy assessment via sensitivity, specificity, and ROC curves does not fully account for clinical utility of a specific model. Decision curve analysis (DCA) becomes a novel complement as it incorporates a clinical judgment of the relative value of benefits (treating a true positive case) and harms (treating a false positive case) associated with prediction models. The preference of a patient or a policy-maker is formulated statistically as the underlying threshold probability, above which the patient would choose to be treated. Net benefit is then calculated for possible threshold probability, which places benefits and harms on the same scale. We consider the inference problems for DCA in this paper. Interval estimation procedure and inference methodology are provided after we derive the relevant asymptotic properties. Our formulation can accommodate the classification problems with multiple categories. We carry out numerical studies to assess the performance of the proposed methods. An eye disease dataset is analyzed to illustrate our proposals.

Machine Learning For Tuning, Selection, And Ensemble Of Multiple Risk Scores For Predicting Type 2 Diabetes

Background

This study proposes the use of machine learning algorithms to improve the accuracy of type 2 diabetes predictions using non-invasive risk score systems.

Methods

We evaluated and compared the prediction accuracies of existing non-invasive risk score systems using the data from the REACTION study (Risk Evaluation of Cancers in Chinese Diabetic Individuals: A Longitudinal Study). Two simple risk scores were established on the bases of logistic regression. Machine learning techniques (ensemble methods) were used to improve prediction accuracies by combining the individual score systems.

Results

Existing score systems from Western populations performed worse than the scores from Eastern populations in general. The two newly established score systems performed better than most existing scores systems but a little worse than the Chinese score system. Using ensemble methods with model selection algorithms yielded better prediction accuracy than all the simple score systems.

Conclusion

Our proposed machine learning methods can be used to improve the accuracy of screening the undiagnosed type 2 diabetes and identifying the high-risk patients.

Comparison between the Bouchama and Japanese Association for Acute Medicine Heatstroke Criteria with Regard to the Diagnosis and Prediction of Mortality of Heatstroke Patients: A Multicenter Observational Study

Background: This study aims to compare the Bouchama heatstroke (B-HS) and Japanese Association for Acute Medicine heatstroke (JAAM-HS) criteria with regard to the diagnosis and prediction of mortality and neurological status of heatstroke patients. Methods: This multicenter observational study recruited eligible patients from the emergency departments of 110 major hospitals in Japan from 1 July to 30 September, 2014. Results: A total of 317 patients (median age, 65 years; interquartile range, 39–80 years) were included and divided into the B-HS, JAAM-HS, and non-HS groups, with each group consisting of 97, 302, and 15 patients, respectively. The JAAM-HS (1.0; 95% confidence interval [CI], 0.87–1.0) and B-HS (0.29; 95% CI, 0.14–0.49) criteria showed high and low sensitivity to mortality, respectively. Similarly, the JAAM-HS (1.0; 95% CI, 0.93–1.0) and B-HS (0.35; 95% CI, 0.23–0.49) criteria showed high and low sensitivity to poor neurological status, respectively. Meanwhile, the sequential organ failure assessment (SOFA) scores demonstrated good accuracy in predicting mortality among heat-related illness (HRI) patients. However, both JAAM-HS and B-HS criteria could not predict in-hospital mortality. The AUC of the SOFA score for mortality was 0.83 (day 3) among the HRI patients. The patients’ neurological status was difficult to predict using the JAAM-HS and B-HS criteria. Concurrently, the total bilirubin level could relatively predict the central nervous system function at discharge. Conclusions: The JAAM-HS criteria showed high sensitivity to mortality and could include all HRI patients who died. The JAAM-HS criterion was considered a useful tool for judgement of admission at ED. Further investigations are necessary to determine the accuracy of both B-HS and JAAM-HS criteria in predicting mortality and neurological status at discharge.

Mining latent information in PTSD psychometrics with fuzziness for effective diagnoses

The options of traditional self-report rating-scale, like the PTSD Checklist Civilian (PCL-C) scale, have no clear boundaries which might cause considerable biases and low effectiveness. This research aimed to explore the feasibility of using fuzzy set in the data processing to promote the screening effectiveness of PCL-C in real-life practical settings. The sensitivity, specificity, Youden's index etc., of PCL-C at different cutoff lines (38, 44 and 50 respectively) were analyzed and compared with those of fuzzy set approach processing. In practice, no matter the cutoff line of the PCL-C was set at 50, 44 or 38, the PCL-C showed good specificity, but failed to exhibit good sensitivity and screening effectiveness. The highest sensitivity was at 65.22%, with Youden's index being 0.64. After fuzzy processing, the fuzzy-PCL-C's sensitivity increased to 91.30%, Youden's index rose to 0.91, having seen marked augmentation. In conclusion, this study indicates that fuzzy set can be used in the data processing of psychiatric scales which have no clear definition standard of the options to improve the effectiveness of the scales.

A Predictive Model Facilitates Early Recognition of Spinal Epidural Abscess in Adults

Introduction

Spinal epidural abscess (SEA), a highly morbid and potentially lethal deep tissue infection of the central nervous system has more than tripled in incidence over the past decade. Early recognition at the point of initial clinical presentation may prevent irreversible neurologic injury or other serious, adverse outcomes. To facilitate early recognition of SEA, we developed a predictive scoring model.

Methods

Using data from a 10-year, retrospective, case-control study of adults presenting for care at a tertiary-care, regional, academic medical center, we used the Integrated Discrimination Improvement Index (IDI) to identify candidate discriminators and created a multivariable logistic regression model, refined based on p-value significance. We selected a cutpoint that optimized sensitivity and specificity.

Results

The final multivariable logistic regression model based on five characteristics –patient age, fever and/or rigor, antimicrobial use within 30 days, back/neck pain, and injection drug use – shows excellent discrimination (AUC 0.88 [95% confidence interval {0.84, 0.92}]). We used the model’s β coefficients to develop a scoring system in which a cutpoint of six correctly identifies cases 89% of the time. Bootstrapped validation measures suggest this model will perform well across samples drawn from this population.

Conclusion

Our predictive scoring model appears to reliably discriminate patients who require emergent spinal imaging upon clinical presentation to rule out SEA and should be used in conjunction with clinical judgment.

Feasibility of anthropometric indices to identify dyslipidemia among adults in Jilin Province: a cross-sectional study

Background

Dyslipidemia and other cardiovascular disease (CVD) risk factors have a strong association with obesity. Anthropometric indices have been widely used to evaluate obesity in clinical and epidemiological studies. We aim to investigate association between serum lipid levels and different anthropometric indices.

Methods

Our study included 17,554 participants. We mainly investigated area under the receiver operating characteristic (AUROC) curves and optimal operating points (OOPs) between the anthropometric indices and serum lipid levels or categories of abnormal serum lipid indices.

Results

For predicting one/two categories of abnormal serum lipid indices among the anthropometric indices, AUROC value of WC was the highest in men (0.718), and AUROC values of BRI and WHtR were the highest in women (0.700 and 0.700) (all P < 0.001); OOP of WC was 82.450 in men; OOPs of BRI and WHtR were 3.435 and 0.504 in women. For predicting three/more categories of abnormal serum lipid indices among the anthropometric indices, AUROC value of WC was the highest in men (0.806), and AUROC values of BRI and WHtR were the highest in women (0.783 and 0.783) (all P < 0.001); OOP of WC was 84.150 in men; OOPs of BRI and WHtR were 3.926 and 0.529 in women.

Conclusions

WC was a good predictor for one/two or three/more categories of abnormal serum lipid indices in men. However, BRI and WHtR were good predictors for one/two or three/more categories of abnormal serum lipid indices in women. ABSI showed the weakest predictive power.

Electronic supplementary material

The online version of this article (10.1186/s12944-017-0648-6) contains supplementary material, which is available to authorized users.

Performance evaluation of existing immunoassays for Clonorchis sinensis infection in China

BACKGROUND

Clonorchiasis ranks among the most important food-borne parasitic diseases in China. However, due to low compliance to traditional fecal examination techniques in the general population and medical personnel, immunodiagnosis is expected. This study evaluated, in parallel, the performance of four immunodiagnostic kits detecting clonorchiasis in China.

RESULTS

A bank with 475 sera was established in this study. Except for the low performance of the kit detecting IgM, the other three kits detecting IgG showed sensitivities ranging from 81.51% (194/238) to 99.16% (236/238). Higher sensitivity was presented in heavy infection intensity [89.47% (68/76) to 100% (76/76)]. Among the four kits, the overall specificity varied from 73.42% (174/237) to 87.34% (207/237). It was observed that the specificity was lower in the sera of the participants living in clonorchiasis-endemic areas but without any parasite infection [67.5% (81/120) to 90% (108/120)], as compared to those from the non-endemic area [94% (47/50) to 98% (49/50)]. The cross-reaction rate varied from 14.93% (10/67) to 31.34% (21/67). Youden's index was -0.022, 0.689, 0.726, and 0.802 for kits T1, T2, T3 and T4, respectively. Repeatability was high in all four kits.

CONCLUSIONS

Three immunodiagnosis kits targeting IgG antibody had high performance on detecting chronic Clonorchis sinensis infection, but that detecting IgM antibody had not. The kits detecting IgG antibody also showed high sensitivity in heavy infection intensity. Research on immunological diagnosis of clonorchiasis is expected to be strengthened to improve the sensitivity in light infection and specificity.

Development of the Sydney Falls Risk Screening Tool in brain injury rehabilitation: A multisite prospective cohort study

AIMS AND OBJECTIVES

To develop a falls risk screening tool (FRST) sensitive to the traumatic brain injury rehabilitation population.

BACKGROUND

Falls are the most frequently recorded patient safety incident within the hospital context. The inpatient traumatic brain injury rehabilitation population is one particular population that has been identified as at high risk of falls. However, no FRST has been developed for this patient population. Consequently in the traumatic brain injury rehabilitation population, there is the real possibility that nurses are using falls risk screening tools that have a poor clinical utility.

DESIGN

Multisite prospective cohort study.

METHODS

Univariate and multiple logistic regression modelling techniques (backward elimination, elastic net and hierarchical) were used to examine each variable's association with patients who fell. The resulting FRST's clinical validity was examined.

RESULTS

Of the 140 patients in the study, 41 (29%) fell. Through multiple logistic regression modelling, 11 variables were identified as predictors for falls. Using hierarchical logistic regression, five of these were identified for inclusion in the resulting falls risk screening tool: prescribed mobility aid (such as, wheelchair or frame), a fall since admission to hospital, impulsive behaviour, impaired orientation and bladder and/or bowel incontinence. The resulting FRST has good clinical validity (sensitivity = 0.9; specificity = 0.62; area under the curve = 0.87; Youden index = 0.54). The tool was significantly more accurate (p = .037 on DeLong test) in discriminating fallers from nonfallers than the Ontario Modified STRATIFY FRST.

CONCLUSION

A FRST has been developed using a comprehensive statistical framework, and evidence has been provided of this tool's clinical validity.

RELEVANCE TO CLINICAL PRACTICE

The developed tool, the Sydney Falls Risk Screening Tool, should be considered for use in brain injury rehabilitation populations.

Optimal cut-off of obesity indices to predict cardiovascular disease risk factors and metabolic syndrome among adults in Northeast China

Background

CVD risk factors (hypertension, dyslipidemia and diabetes) and MetS are closely related to obesity. The selection of an optimal cut-off for various obesity indices is particularly important to predict CVD risk factors and MetS.

Methods

Sixteen thousand seven hundred sixty-six participants aged 18–79 were recruited in Jilin Province in 2012. Five obesity indices, including BMI, WC, WHR, WHtR and BAI were investigated. ROC analyses were used to evaluate the predictive ability and determine the optimal cut-off values of the obesity indices for CVD risk factors and MetS.

Results

BMI had the highest adjusted ORs, and the adjusted ORs for hypertension, dyslipidemia, diabetes and MetS were 1.19 (95 % CI, 1.17 to 1.20), 1.20 (95 % CI, 1.19 to 1.22), 1.12 (95 % CI, 1.10 to 1.13), and 1.40 (95 % CI, 1.38 to 1.41), respectively. However, BMI did not always have the largest adjusted AUROC. In general, the young age group (18 ~ 44) had higher ORs and AUROCs for CVD risk factors and MetS than those of the other age groups. In addition, the optimal cut-off values for WC and WHR in males were relatively higher than those in females, whereas the BAI in males was comparatively lower than that in females.

Conclusions

The appropriate obesity index, with the corresponding optimal cut-off values, should be selected in different research studies and populations. Generally, the obesity indices and their optimal cut-off values are: BMI (24 kg/m²), WC (male: 85 cm; female: 80 cm), WHR (male: 0.88; female: 0.85), WHtR (0.50), and BAI (male: 25 cm; female: 30 cm). Moreover, WC is superior to other obesity indices in predicting CVD risk factors and MetS in males, whereas, WHtR is superior to other obesity indices in predicting CVD risk factors and MetS in females.

Electronic supplementary material

The online version of this article (doi:10.1186/s12889-016-3694-5) contains supplementary material, which is available to authorized users.

Cancers Screening in an Asymptomatic Population by Using Multiple Tumour Markers

BACKGROUND

Analytic measurement of serum tumour markers is one of commonly used methods for cancer risk management in certain areas of the world (e.g. Taiwan). Recently, cancer screening based on multiple serum tumour markers has been frequently discussed. However, the risk-benefit outcomes appear to be unfavourable for patients because of the low sensitivity and specificity. In this study, cancer screening models based on multiple serum tumour markers were designed using machine learning methods, namely support vector machine (SVM), k-nearest neighbour (KNN), and logistic regression, to improve the screening performance for multiple cancers in a large asymptomatic population.

METHODS

AFP, CEA, CA19-9, CYFRA21-1, and SCC were determined for 20 696 eligible individuals. PSA was measured in men and CA15-3 and CA125 in women. A variable selection process was applied to select robust variables from these serum tumour markers to design cancer detection models. The sensitivity, specificity, positive predictive value (PPV), negative predictive value, area under the curve, and Youden index of the models based on single tumour markers, combined test, and machine learning methods were compared. Moreover, relative risk reduction, absolute risk reduction (ARR), and absolute risk increase (ARI) were evaluated.

RESULTS

To design cancer detection models using machine learning methods, CYFRA21-1 and SCC were selected for women, and all tumour markers were selected for men. SVM and KNN models significantly outperformed the single tumour markers and the combined test for men. All 3 studied machine learning methods outperformed single tumour markers and the combined test for women. For either men or women, the ARRs were between 0.003-0.008; the ARIs were between 0.119-0.306.

CONCLUSION

Machine learning methods outperformed the combined test in analysing multiple tumour markers for cancer detection. However, cancer screening based solely on the application of multiple tumour markers remains unfavourable because of the inadequate PPV, ARR, and ARI, even when machine learning methods were incorporated into the analysis.

Troponin I Assay for Identification of a Significant Coronary Stenosis in Patients with Suspected Acute Myocardial Infarction and Wide QRS Complex

Background

Common ECG criteria such as ST-segment changes are of limited value in patients with suspected acute myocardial infarction (AMI) and bundle branch block or wide QRS complex. A large proportion of these patients do not suffer from an AMI, whereas those with ST-elevation myocardial infarction (STEMI) equivalent AMI benefit from an aggressive treatment. Aim of the present study was to evaluate the diagnostic information of cardiac troponin I (cTnI) in hemodynamically stable patients with wide QRS complex and suspected AMI.

Methods

In 417 out of 1818 patients presenting consecutively between 01/2007 and 12/2008 in a prospective multicenter observational study with suspected AMI a prolonged QRS duration was observed. Of these, n = 117 showed significant obstructive coronary artery disease (CAD) used as diagnostic outcome variable. cTnI was determined at admission.

Results

Patients with significant CAD had higher cTnI levels compared to individuals without (median 250ng/L vs. 11ng/L; p<0.01). To identify patients needing a coronary intervention, cTnI yielded an area under the receiver operator characteristics curve of 0.849. Optimized cut-offs with respect to a sensitivity driven rule-out and specificity driven rule-in strategy were established (40ng/L/96ng/L). Application of the specificity optimized cut-off value led to a positive predictive value of 71% compared to 59% if using the 99^th percentile cut-off. The sensitivity optimized cut-off value was associated with a negative predictive value of 93% compared to 89% provided by application of the 99^th percentile threshold.

Conclusion

cTnI determined in hemodynamically stable patients with suspected AMI and wide QRS complex using optimized diagnostic thresholds improves rule-in and rule-out with respect to presence of a significant obstructive CAD.

Real-time prediction of inpatient length of stay for discharge prioritization

OBJECTIVE

Hospitals are challenged to provide timely patient care while maintaining high resource utilization. This has prompted hospital initiatives to increase patient flow and minimize nonvalue added care time. Real-time demand capacity management (RTDC) is one such initiative whereby clinicians convene each morning to predict patients able to leave the same day and prioritize their remaining tasks for early discharge. Our objective is to automate and improve these discharge predictions by applying supervised machine learning methods to readily available health information.

MATERIALS AND METHODS

The authors use supervised machine learning methods to predict patients' likelihood of discharge by 2 p.m. and by midnight each day for an inpatient medical unit. Using data collected over 8000 patient stays and 20 000 patient days, the predictive performance of the model is compared to clinicians using sensitivity, specificity, Youden's Index (i.e., sensitivity + specificity - 1), and aggregate accuracy measures.

RESULTS

The model compared to clinician predictions demonstrated significantly higher sensitivity (P < .01), lower specificity (P < .01), and a comparable Youden Index (P > .10). Early discharges were less predictable than midnight discharges. The model was more accurate than clinicians in predicting the total number of daily discharges and capable of ranking patients closest to future discharge.

CONCLUSIONS

There is potential to use readily available health information to predict daily patient discharges with accuracies comparable to clinician predictions. This approach may be used to automate and support daily RTDC predictions aimed at improving patient flow.