Assessing the capacity of symptom scores to predict COVID-19 positivity in Nigeria: a national derivation and validation cohort study

ObjectivesThis study aimed to develop and validate a symptom prediction tool for COVID-19 test positivity in Nigeria.DesignPredictive modelling study.SettingAll Nigeria States and the Federal Capital Territory.ParticipantsA cohort of 43 221 individuals within the national COVID-19 surveillance dataset from 27 February to 27 August 2020. Complete dataset was randomly split into two equal halves: derivation and validation datasets. Using the derivation dataset (n=21 477), backward multivariable logistic regression approach was used to identify symptoms positively associated with COVID-19 positivity (by real-time PCR) in children (≤17 years), adults (18–64 years) and elderly (≥65 years) patients separately.Outcome measuresWeighted statistical and clinical scores based on beta regression coefficients and clinicians’ judgements, respectively. Using the validation dataset (n=21 744), area under the receiver operating characteristic curve (AUROC) values were used to assess the predictive capacity of individual symptoms, unweighted score and the two weighted scores.ResultsOverall, 27.6% of children (4415/15 988), 34.6% of adults (9154/26 441) and 40.0% of elderly (317/792) that had been tested were positive for COVID-19. Best individual symptom predictor of COVID-19 positivity was loss of smell in children (AUROC 0.56, 95% CI 0.55 to 0.56), either fever or cough in adults (AUROC 0.57, 95% CI 0.56 to 0.58) and difficulty in breathing in the elderly (AUROC 0.53, 95% CI 0.48 to 0.58) patients. In children, adults and the elderly patients, all scoring approaches showed similar predictive performance.ConclusionsThe predictive capacity of various symptom scores for COVID-19 positivity was poor overall. However, the findings could serve as an advocacy tool for more investments in resources for capacity strengthening of molecular testing for COVID-19 in Nigeria.

Download Full-text

Easy Screening for Mild Alzheimer's Disease and Mild Cognitive Impairment from Elderly Speech

Current Alzheimer Research ◽

10.2174/1567205014666171120144343 ◽

2018 ◽

Vol 15 (2) ◽

pp. 104-110 ◽

Cited By ~ 3

Author(s):

Shohei Kato ◽

Akira Homma ◽

Takuto Sakuma

Keyword(s):

Alzheimer’S Disease ◽

Alzheimer's Disease ◽

Cognitive Impairment ◽

Mild Cognitive Impairment ◽

Characteristic Curve ◽

Linear Regression Analysis ◽

Speech Sound ◽

The Elderly ◽

Speech Sounds ◽

Mel Frequency Cepstral Coefficients

Objective: This study presents a novel approach for early detection of cognitive impairment in the elderly. The approach incorporates the use of speech sound analysis, multivariate statistics, and data-mining techniques. We have developed a speech prosody-based cognitive impairment rating (SPCIR) that can distinguish between cognitively normal controls and elderly people with mild Alzheimer's disease (mAD) or mild cognitive impairment (MCI) using prosodic signals extracted from elderly speech while administering a questionnaire. Two hundred and seventy-three Japanese subjects (73 males and 200 females between the ages of 65 and 96) participated in this study. The authors collected speech sounds from segments of dialogue during a revised Hasegawa's dementia scale (HDS-R) examination and talking about topics related to hometown, childhood, and school. The segments correspond to speech sounds from answers to questions regarding birthdate (T1), the name of the subject's elementary school (T2), time orientation (Q2), and repetition of three-digit numbers backward (Q6). As many prosodic features as possible were extracted from each of the speech sounds, including fundamental frequency, formant, and intensity features and mel-frequency cepstral coefficients. They were refined using principal component analysis and/or feature selection. The authors calculated an SPCIR using multiple linear regression analysis. Conclusion: In addition, this study proposes a binary discrimination model of SPCIR using multivariate logistic regression and model selection with receiver operating characteristic curve analysis and reports on the sensitivity and specificity of SPCIR for diagnosis (control vs. MCI/mAD). The study also reports discriminative performances well, thereby suggesting that the proposed approach might be an effective tool for screening the elderly for mAD and MCI.

Download Full-text

Development of Machine Learning Models to Predict Probabilities and Types of Stroke at Prehospital Stage: the Japan Urgent Stroke Triage Score Using Machine Learning (JUST-ML)

Translational Stroke Research ◽

10.1007/s12975-021-00937-x ◽

2021 ◽

Author(s):

Kazutaka Uchida ◽

Junichi Kouno ◽

Shinichi Yoshimura ◽

Norito Kinjo ◽

Fumihiro Sakakibara ◽

...

Keyword(s):

Machine Learning ◽

Logistic Regression ◽

Random Forests ◽

Prediction Models ◽

Characteristic Curve ◽

Predictive Performance ◽

Vessel Occlusion ◽

Predictive Values ◽

Training Cohort ◽

Sensitivity Specificity

AbstractIn conjunction with recent advancements in machine learning (ML), such technologies have been applied in various fields owing to their high predictive performance. We tried to develop prehospital stroke scale with ML. We conducted multi-center retrospective and prospective cohort study. The training cohort had eight centers in Japan from June 2015 to March 2018, and the test cohort had 13 centers from April 2019 to March 2020. We use the three different ML algorithms (logistic regression, random forests, XGBoost) to develop models. Main outcomes were large vessel occlusion (LVO), intracranial hemorrhage (ICH), subarachnoid hemorrhage (SAH), and cerebral infarction (CI) other than LVO. The predictive abilities were validated in the test cohort with accuracy, positive predictive value, sensitivity, specificity, area under the receiver operating characteristic curve (AUC), and F score. The training cohort included 3178 patients with 337 LVO, 487 ICH, 131 SAH, and 676 CI cases, and the test cohort included 3127 patients with 183 LVO, 372 ICH, 90 SAH, and 577 CI cases. The overall accuracies were 0.65, and the positive predictive values, sensitivities, specificities, AUCs, and F scores were stable in the test cohort. The classification abilities were also fair for all ML models. The AUCs for LVO of logistic regression, random forests, and XGBoost were 0.89, 0.89, and 0.88, respectively, in the test cohort, and these values were higher than the previously reported prediction models for LVO. The ML models developed to predict the probability and types of stroke at the prehospital stage had superior predictive abilities.

Download Full-text

Systemic Immune-Inflammation Index May Predict the Development of Contrast-Induced Nephropathy in Patients With ST-Segment Elevation Myocardial Infarction

Angiology ◽

10.1177/00033197211030053 ◽

2021 ◽

pp. 000331972110300

Author(s):

Ali Bağcı ◽

Fatih Aksoy ◽

Hasan Aydin Baş

Keyword(s):

Myocardial Infarction ◽

Characteristic Curve ◽

Contrast Induced Nephropathy ◽

Predictive Capacity ◽

St Segment Elevation ◽

Neutrophil Lymphocyte Ratio ◽

Lymphocyte Ratio ◽

Segment Elevation Myocardial Infarction ◽

St Segment ◽

Immune Inflammation

The aim of this study was to investigate the predictive capacity of a systemic immune-inflammation index (SII) in the detection of contrast-induced nephropathy (CIN) following ST-segment elevation myocardial infarction (STEMI). A total of 477 STEMI patients were enrolled in the study. The patients were divided into 2 groups according to CIN development. A cutoff point of 5.91 for logarithm-transformed SII was identified with 73.0% sensitivity and 57.5% specificity to predict CIN following STEMI. According to a pairwise analysis of receiver operating characteristic curve analysis, the predictive power of SII in detecting CIN following STEMI was similar to that of high-sensitivity C-reactive protein and better than the neutrophil/lymphocyte ratio or platelet/lymphocyte ratio. As a result, SII can be used as one of the independent predictors of CIN after STEMI.

Download Full-text

Optimisation of microalgal cultivation via nutrient-enhanced strategies: the biorefinery paradigm

Biotechnology for Biofuels ◽

10.1186/s13068-021-01912-2 ◽

2021 ◽

Vol 14 (1) ◽

Author(s):

Gonzalo M. Figueroa-Torres ◽

Jon K. Pittman ◽

Constantinos Theodoropoulos

Keyword(s):

Kinetic Model ◽

Growth Dynamics ◽

Value Added ◽

Cost Effective ◽

Predictive Modelling ◽

Nitrogen And Phosphorus ◽

Predictive Capacity ◽

Microalgal Biomass ◽

Microalgal Cultivation ◽

Lipid Formation

Abstract Background The production of microalgal biofuels, despite their sustainable and renowned potential, is not yet cost-effective compared to current conventional fuel technologies. However, the biorefinery concept increases the prospects of microalgal biomass as an economically viable feedstock suitable for the co-production of multiple biofuels along with value-added chemicals. To integrate biofuels production within the framework of a microalgae biorefinery, it is not only necessary to exploit multi-product platforms, but also to identify optimal microalgal cultivation strategies maximising the microalgal metabolites from which biofuels are obtained: starch and lipids. Whilst nutrient limitation is widely known for increasing starch and lipid formation, this cultivation strategy can greatly reduce microalgal growth. This work presents an optimisation framework combining predictive modelling and experimental methodologies to effectively simulate and predict microalgal growth dynamics and identify optimal cultivation strategies. Results Microalgal cultivation strategies for maximised starch and lipid formation were successfully established by developing a multi-parametric kinetic model suitable for the prediction of mixotrophic microalgal growth dynamics co-limited by nitrogen and phosphorus. The model’s high predictive capacity was experimentally validated against various datasets obtained from laboratory-scale cultures of Chlamydomonas reinhardtii CCAP 11/32C subject to different initial nutrient regimes. The identified model-based optimal cultivation strategies were further validated experimentally and yielded significant increases in starch (+ 270%) and lipid (+ 74%) production against a non-optimised strategy. Conclusions The optimised microalgal cultivation scenarios for maximised starch and lipids, as identified by the kinetic model presented here, highlight the benefits of exploiting modelling frameworks as optimisation tools that facilitate the development and commercialisation of microalgae-to-fuel technologies.

Download Full-text

Review and comparative assessment of similarity-based methods for prediction of drug–protein interactions in the druggable human proteome

Briefings in Bioinformatics ◽

10.1093/bib/bby069 ◽

2018 ◽

Vol 20 (6) ◽

pp. 2066-2087 ◽

Cited By ~ 8

Author(s):

Chen Wang ◽

Lukasz Kurgan

Keyword(s):

Protein Interactions ◽

Drug Targets ◽

Characteristic Curve ◽

Protein Structures ◽

Predictive Performance ◽

Computational Prediction ◽

Comprehensive Analysis ◽

Model Combining ◽

Benchmark Database ◽

Key Aspects

AbstractDrug–protein interactions (DPIs) underlie the desired therapeutic actions and the adverse side effects of a significant majority of drugs. Computational prediction of DPIs facilitates research in drug discovery, characterization and repurposing. Similarity-based methods that do not require knowledge of protein structures are particularly suitable for druggable genome-wide predictions of DPIs. We review 35 high-impact similarity-based predictors that were published in the past decade. We group them based on three types of similarities and their combinations that they use. We discuss and compare key aspects of these methods including source databases, internal databases and their predictive models. Using our novel benchmark database, we perform comparative empirical analysis of predictive performance of seven types of representative predictors that utilize each type of similarity individually and all possible combinations of similarities. We assess predictive quality at the database-wide DPI level and we are the first to also include evaluation over individual drugs. Our comprehensive analysis shows that predictors that use more similarity types outperform methods that employ fewer similarities, and that the model combining all three types of similarities secures area under the receiver operating characteristic curve of 0.93. We offer a comprehensive analysis of sensitivity of predictive performance to intrinsic and extrinsic characteristics of the considered predictors. We find that predictive performance is sensitive to low levels of similarities between sequences of the drug targets and several extrinsic properties of the input drug structures, drug profiles and drug targets. The benchmark database and a webserver for the seven predictors are freely available at http://biomine.cs.vcu.edu/servers/CONNECTOR/.

Download Full-text

Development and validation of a clinical and genetic model for predicting risk of severe COVID-19

10.1101/2021.03.09.21253237 ◽

2021 ◽

Author(s):

Gillian S. Dite ◽

Nicholas M. Murphy ◽

Richard Allman

Keyword(s):

Logistic Regression ◽

Genetic Model ◽

Characteristic Curve ◽

Accurate Prediction ◽

Training Dataset ◽

Validation Dataset ◽

Individual Risk ◽

Prototype Model ◽

List Type ◽

New Model

SummaryClinical and genetic risk factors for severe COVID-19 are often considered independently and without knowledge of the magnitudes of their effects on risk. Using SARS-CoV-2 positive participants from the UK Biobank, we developed and validated a clinical and genetic model to predict risk of severe COVID-19. We used multivariable logistic regression on a 70% training dataset and used the remaining 30% for validation. We also validated a previously published prototype model. In the validation dataset, our new model was associated with severe COVID-19 (odds ratio per quintile of risk=1.77, 95% confidence interval [CI]=1.64, 1.90) and had excellent discrimination (area under the receiver operating characteristic curve=0.732, 95% CI=0.708, 0.756). We assessed calibration using logistic regression of the log odds of the risk score, and the new model showed no evidence of over- or under-estimation of risk (α=−0.08; 95% CI=−0.21, 0.05) and no evidence or over- or under-dispersion of risk (β=0.90, 95% CI=0.80, 1.00). Accurate prediction of individual risk is possible and will be important in regions where vaccines are not widely available or where people refuse or are disqualified from vaccination, especially given uncertainty about the extent of infection transmission among vaccinated people and the emergence of SARS-CoV-2 variants of concern.Key resultsAccurate prediction of the risk of severe COVID-19 can inform public heath interventions and empower individuals to make informed choices about their day-to-day activities.Age and sex alone do not accurately predict risk of severe COVID-19.Our clinical and genetic model to predict risk of severe COVID-19 performs extremely well in terms of discrimination and calibration.

Download Full-text

External validation of the predictive performance of population pharmacokinetic models for phenobarbital in pediatric patients

10.1101/2020.09.10.20192005 ◽

2020 ◽

Author(s):

Sunae Ryu ◽

Woo Jin Jung ◽

Zheng Jiao ◽

Jung Woo Chae ◽

Hwi-yeol Yun

Keyword(s):

External Validation ◽

Predictive Performance ◽

Validation Dataset ◽

Therapeutic Drug ◽

Population Pharmacokinetic ◽

Pharmacokinetic Models ◽

Bayesian Forecasting ◽

Predictive Capability ◽

Young Infants ◽

Simulation Based

Aim: Several studies have reported population pharmacokinetic models for phenobarbital (PB), but the predictive performance of these models has not been well documented. This study aims to do external validation of the predictive performance in published pharmacokinetic models. Methods: Therapeutic drug monitoring data collected in neonates and young infants treated with PB for seizure control, was used for external validation. A literature review was conducted through PubMed to identify population pharmacokinetic models. Prediction- and simulation-based diagnostics, and Bayesian forecasting were performed for external validation. The incorporation of size or maturity functions into the published models was also tested for prediction improvement. Results: A total of 79 serum concentrations from 28 subjects were included in the external validation dataset. Seven population pharmacokinetic studies of PB were selected for evaluation. The model by Voller et al. [27] showed the best performance concerning prediction-based evaluation. In simulation-based analyses, the normalized prediction distribution error of two models (those of Shellhaas et al. [24] and Marsot et al. [25]) obeyed a normal distribution. Bayesian forecasting with more than one observation improved predictive capability. Incorporation of both allometric size scaling and maturation function generally enhanced the predictive performance, but with marked improvement for the adult pharmacokinetic model. Conclusion: The predictive performance of published pharmacokinetic models of PB was diverse, and validation may be necessary to extrapolate to different clinical settings. Our findings suggest that Bayesian forecasting improves the predictive capability of individual concentrations for pediatrics.

Download Full-text

Machine learning approaches revealed metabolic signatures of incident chronic kidney disease in persons with pre-and type 2 diabetes

10.2337/figshare.13022624.v2 ◽

2020 ◽

Author(s):

Ada Admin ◽

Jialing Huang ◽

Cornelia Huth ◽

Marcela Covic ◽

Martina Troll ◽

...

Keyword(s):

Machine Learning ◽

Type 2 Diabetes ◽

Chronic Kidney Disease ◽

Kidney Disease ◽

Characteristic Curve ◽

Predictive Performance ◽

Population Based ◽

Learning Approaches ◽

Clinical Variables

Early and precise identification of individuals with pre-diabetes and type 2 diabetes (T2D) at risk of progressing to chronic kidney disease (CKD) is essential to prevent complications of diabetes. Here, we identify and evaluate prospective metabolite biomarkers and the best set of predictors of CKD in the longitudinal, population-based Cooperative Health Research in the Region of Augsburg (KORA) cohort by targeted metabolomics and machine learning approaches. Out of 125 targeted metabolites, sphingomyelin (SM) C18:1 and phosphatidylcholine diacyl (PC aa) C38:0 were identified as candidate metabolite biomarkers of incident CKD specifically in hyperglycemic individuals followed during 6.5 years. Sets of predictors for incident CKD developed from 125 metabolites and 14 clinical variables showed highly stable performances in all three machine learning approaches and outperformed the currently established clinical algorithm for CKD. The two metabolites in combination with five clinical variables were identified as the best set of predictors and their predictive performance yielded a mean area value under the receiver operating characteristic curve of 0.857. The inclusion of metabolite variables in the clinical prediction of future CKD may thus improve the risk prediction in persons with pre- and T2D. The metabolite link with hyperglycemia-related early kidney dysfunction warrants further investigation.

Download Full-text

Prediction of Past SARS-CoV-2 Infections: A Prospective Cohort Study Among Swiss Schoolchildren

Frontiers in Pediatrics ◽

10.3389/fped.2021.710785 ◽

2021 ◽

Vol 9 ◽

Author(s):

Jacob Blankenberger ◽

Sarah R. Haile ◽

Milo A. Puhan ◽

Christoph Berger ◽

Thomas Radtke ◽

...

Keyword(s):

Prediction Models ◽

Characteristic Curve ◽

Community Setting ◽

Predictive Performance ◽

Clinical History ◽

School Level ◽

Sociodemographic Characteristics ◽

Serological Tests ◽

Time Points ◽

June July

Objective: To assess the predictive value of symptoms, sociodemographic characteristics, and SARS-CoV-2 exposure in household, school, and community setting for SARS-CoV-2 seropositivity in Swiss schoolchildren at two time points in 2020.Design: Serological testing of children in primary and secondary schools (aged 6–13 and 12–16 years, respectively) took place in June–July (T1) and October–November (T2) 2020, as part of the longitudinal, school-based study Ciao Corona in the canton of Zurich, Switzerland. Information on sociodemographic characteristics and clinical history was collected with questionnaires to parents; information on school-level SARS-CoV-2 infections was collected with questionnaires to school principals. Community-level cumulative incidence was obtained from official statistics. We used logistic regression to identify individual predictors of seropositivity and assessed the predictive performance of symptom- and exposure-based prediction models.Results: A total of 2,496 children (74 seropositive) at T1 and 2,152 children (109 seropositive) at T2 were included. Except for anosmia (odds ratio 15.4, 95% confidence interval [3.4–70.7]) and headache (2.0 [1.03–3.9]) at T2, none of the individual symptoms were significantly predictive of seropositivity at either time point. Of all the exposure variables, a reported SARS-CoV-2 case in the household was the strongest predictor for seropositivity at T1 (12.4 [5.8–26.7]) and T2 (10.8 [4.5–25.8]). At both time points, area under the receiver operating characteristic curve was greater for exposure-based (T1, 0.69; T2, 0.64) than symptom-based prediction models (T1, 0.59; T2, 0.57).Conclusions: In children, retrospective identification of past SARS-CoV-2 infections based on symptoms is imprecise. SARS-CoV-2 seropositivity is better predicted by factors of SARS-CoV-2 exposure, especially reported SARS-CoV-2 cases in the household. Predicting SARS-CoV-2 seropositivity in children in general is challenging, as few reliable predictors could be identified. For an accurate retrospective identification of SARS-CoV-2 infections in children, serological tests are likely indispensable.Trial registration number: NCT04448717.

Download Full-text

A Bayesian Model for Prediction of Rheumatoid Arthritis from Risk Factors

10.1101/2020.07.09.20150326 ◽

2020 ◽

Author(s):

Leon Lufkin ◽

Marko Budišić ◽

Sumona Mondal ◽

Shantanu Sur

Keyword(s):

Rheumatoid Arthritis ◽

Risk Factors ◽

Predictive Model ◽

Characteristic Curve ◽

Strong Association ◽

Autoimmune Disorder ◽

Mixed Data ◽

Validation Dataset ◽

Third Order ◽

Organ Systems

Rheumatoid arthritis (RA) is a chronic autoimmune disorder that typically manifests as destructive joint inflammation but also affects multiple other organ systems. The pathogenesis of RA is complex where a variety of factors including comorbidities, demographic, and socioeconomic variables are known to influence the incidence and progress of the disease. In this work, we aimed to predict RA from a set of 11 well-known risk factors and their interactions using Bayesian logistic regression. We considered up to third-order interactions between the risk factors and implemented factor analysis of mixed data (FAMD) to account for both the continuous and categorical natures of these variables. The predictive model was further optimized over the area under the receiver operating characteristic curve (AUC) using a genetic algorithm (GA). We use data from the National Health and Nutrition Examination Survey (NHANES). Our optimal predictive model has a smoothed AUC of 0.826 (95% CI: 0.801 −0.850) on a validation dataset and 0.805 (95% CI: 0.781 −0.829) on a holdout test dataset. Our model identified multiple second- and third-order interactions that demonstrate a strong association with RA, implying the potential role of risk factor interactions in the disease mechanism. Interestingly, we find that the inclusion of higher-order interactions in the model only marginally improves overall predictive ability. Our findings on the contribution of RA risk factors and their interaction on disease prediction could be useful in developing strategies for early diagnosis of RA, thus opening potential avenues for improved patient outcomes and reduced healthcare burden to society.

Download Full-text