Test Accuracy Research Articles

Performance du score de probabilité à 4 niveaux 4PEPS pour le diagnostic de l'embolie pulmonaire dans une population d'Afrique subsaharienne : données du Registre des Embolies Pulmonaires du Centre Hospitalier Universitaire de Bogodogo, Burkina Faso

ObjectiveTo evaluate the performance of the 4PEPS score in the diagnosis of pulmonary embolism at the University Hospital of Bogodogo from January 1, 2021 to July 31, 2023. MethodologyThis was a cross-sectional descriptive and analytical diagnostic study, running from January 1, 2021 to July 31, 2023. It took place in the infectious and tropical diseases departments, including a pneumology unit and a cardiology unit, of the Bogodogo University Hospital. Patients of both sexes with suspected pulmonary embolism who had undergone CT scan were included. The 4PEPS score was calculated and dichotomized into probable and improbable. It thus constituted the diagnostic test. CT scan was the gold standard. The accuracy of the diagnostic test was judged by the area under the ROC curve. An area under the curve between 0.70 and 1 would mean that the score was moderately informative to perfect. ResultsOur study included 472 patients with suspected pulmonary embolism out of a total population of 1228 patients. Hospital prevalence was 21.7%. The mean age of patients was 54.3 years. Females accounted for 52.1% of cases, with a sex ratio of 0.93. The prevalence of pulmonary embolism in the different probability levels of the 4PEPS score was 13.3% for the very low level, 11.7% for the low level, 84.6% for the intermediate level and 93.3% for the high level. Sensitivity and specificity were 92.1% and 86.82% respectively. The positive and negative predictive values were 90.1% and 89.4% respectively. The area under the ROC curve was 0.91. ConclusionIn our study, the 4PEPS score showed good negative and positive predictive values. The use of this score will enable practitioners faced with diagnostic difficulties to make therapeutic decisions, reducing inappropriate prescriptions for thoracic angioscan.

A loop-mediated isothermal amplification test for yaws: a multi-country diagnostic accuracy evaluation

To meet the WHO target of eradicating yaws by 2030, highly sensitive and specific diagnostic tools are needed. A multiplex Treponema pallidum-Haemophilus ducreyi loop-mediated isothermal amplification (TPHD-LAMP) test holds promise as a near-patient diagnostic tool for yaws and H ducreyi. We conducted a prospective evaluation in Cameroon, Côte d'Ivoire, Ghana, and the Republic of the Congo to determine the diagnostic accuracy of the TPHD-LAMP test, as well as to assess its acceptability, feasibility, and cost. Active case searching within schools and communities was used to locate participants with clinically suspicious laws-like lesions. Individuals with serologically confirmed active yaws provided paired lesion swabs between March, 2021, and April, 2023. For each participant, one swab was tested with the TPHD-LAMP at a local district laboratory and the other with reference quantitative PCR (qPCR) tests conducted at national reference laboratories. The primary outcome was TPHD-LAMP test sensitivity and specificity compared with qPCR. Laboratory technicians were interviewed using a multiple-choice survey to gauge acceptability and feasibility of the TPHD-LAMP test. Costs of each test were calculated. Of 3085 individuals with at least one suspected yaws lesion, 531 (17%) were serologically confirmed. We enrolled 493 participants with seropositive yaws and a further 32 with negative serology. The sensitivity of the TPHD-LAMP test for detecting T pallidum was 63% (95% CI 56-70) and the specificity was 66% (95% CI 61-71). Sensitivity and specificity for T pallidum improved to 73% (63-82; p=0·0065) and 75% (68-80; p=0·0003), respectively, in H ducreyi-negative samples. Interviews highlighted challenges in user-friendliness and practicality of the TPHD-LAMP test. The cost of the test per sample was one third of that of qPCR, although the TPHD-LAMP test entailed higher costs to establish the assay. This was the first multi-country diagnostic evaluation of a molecular test for yaws. The TPHD-LAMP testing, in its current form, falls short of the WHO target product profile criteria for yaws diagnostics. These findings highlight the importance of assessing new diagnostics in real-world conditions to ensure their suitability for programmatic use. The EDCTP2 programme supported by the EU.

Adversarial attacks in signature verification: a deep learning approach

Handwritten signature recognition in forensic science is crucial for identity and document authentication. While serving as a legal representation of a person’s agreement or consent to the contents of a document, handwritten signatures de termine the authenticity of a document, identify forgeries, pinpoint the suspects and support other pieces of evidence like ink or document analysis. This work focuses on developing and evaluating a handwritten signature verification sys tem using a convolutional neural network (CNN) and emphasising the model’s efficacy using hand-crafted adversarial attacks. Initially, handwritten signatures have been collected from sixteen volunteers, each contributing ten samples, fol lowed by image normalization and augmentation to boost synthetic data samples and overcome the data scarcity. The proposed model achieved a testing accu racy of 91.35% using an 80:20 train-test split. Additionally, using the five-fold cross-validation, the model achieved a robust validation accuracy of nearly 98%. Finally, the introduction of manually constructed adversarial assaults on the sig nature images undermines the model’s accuracy, bringing the accuracy down to nearly 80%. This highlights the need to consider adversarial resilience while designing deep learning models for classification tasks. Exposing the model to real look-alike fake samples is critical while testing its robustness and refining the model using trial and error methods.

Refining Ovarian Cancer Test accuracy Scores: A multicentre, prospective cohort study investigating diagnostic accuracy among women with symptoms of suspected ovarian cancer (the ROCkeTS study): Results for post-menopausal women

Refining Ovarian Cancer Test accuracy Scores: A multicentre, prospective cohort study investigating diagnostic accuracy among women with symptoms of suspected ovarian cancer (the ROCkeTS study): Results for post-menopausal women

A New Test for Evaluating the Psychological Impact of COVID-19 in Nurses

Introduction During the COVID-19 pandemic, many health care workers were involved and suffered significant physical and psychological strain, and nursing was one of the main players. Clearly and objectively establishing the psychological impact on nurses during the pandemic is necessary for long-term recovery and to be prepared for new outbreaks in the future. Aim The aims of this study were to evaluate the perceived impact on nurses at risk of psychological stress at work during a major health care crisis and determine the psychological impact and its influencing factors. Methods An online questionnaire was designed and distributed between April 9 and April 19, 2020, in Spain. The comprehensive self-conducted questionnaire included demographic information and the Psychological Stress and Adaptation at work Score. For the analysis, descriptive statistics, logistic regressions, and receiver operating characteristic curves were determined. Results A total of 825 nurses (33% working in intensive care units and emergency departments) answered the questionnaire. A logistic regression analysis revealed that nurses with higher psychological stress scores experienced more physical overload (P = .02), higher psychological overload (P < .01), poor teamwork (P = .02), more distress due to new circumstances (P < .01), more personal impact of how society responded to COVID-19 (P < .01), anticipation of negative future consequences (P < .01), and worse balance between family and work (P < .01). The area under the curve for the cutoff point of Psychological Stress and Adaptation at work Score 40 was 0.82 (95% confidence interval, 0.79-0.91), with 70% sensitivity and 81% specificity. The positive predictive value was 72%, and the negative predictive value was 80%. Conclusion Psychological Stress and Adaptation at work Score is a rapid and accurate test for evaluating the psychological well-being of health care staff that can be used when crisis prevents the use of conventional stress diagnosis methods. Higher psychological impact was observed in nurses who worked in the intensive care unit and in the hospital ward.

Brain Tumor Detection and Classification Using MRI Images

Abstract: Brain tumors pose a serious health risk, and prompt and precise identification is essential for successful treatment. This article describes an automated method employing convolutional neural networks (CNNs) to identify and categorize brain cancers in MRI images. The dataset that was used consists of 7,023 MRI scans that were obtained from three different sources: the Br35H dataset, the SARTAJ dataset, and the fig share repository. Four classes are identified from the images: pituitary tumors, gliomas, meningioma’s, and no tumor. With the use of this improved dataset, the CNN model was created and trained, ultimately attaining a 97.5% test accuracy. Moreover, class-wise analysis showed that the precision values for pituitary tumors were 96.1%, 98.8%, 96.1%, and 98.6%, respectively, with equivalent recall rates of 94.7%, 95.4%, 100%, and 99% for gliomas, meningioma’s, and no tumors. With values of 0.966 for gliomas, 0.957 for meningioma’s, 0.994 for no tumor, and 0.975 for pituitary tumors, the F1-scores demonstrate the general resilience of the model. These findings show that the suggested CNNbased method is effective in assisting with brain tumor diagnosis from MRI scans, providing radiologists and other medical practitioners with a useful tool. In order to improve diagnostic accuracy, future study may entail enhancing the model even further and investigating the integration of other imaging modalities

Evaluation of revised UK-NEQAS CSF-xanthochromia method for subarachnoid hemorrhage: outcome data provide evidence for clinical value.

Subarachnoid haemorrhage (SAH) has a high morbidity and mortality and requires prompt diagnosis. In patients with negative findings on computed-tomogram of the brain (CT-Brain) cerebrospinal fluid (CSF)-xanthochromia is considered the test of choice if performed 12 h or more after symptom onset. We audited the accuracy, usefulness and timing of CSF-xanthochromia testing and the interpretation of equivocal CSF-xanthochromia findings. We also investigated mortality outcomes for defined subsets of patients. A retrospective audit of CSF-xanthochromia tests over 8 years was performed. The service uses the revised UK-NEQAS (United Kingdom National External Quality Assessment Service) method. We analysed 543 cases (F=299, median age 44yrs) with 19 cases (3.5 %) having SAH. Sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) of CSF-xanthochromia testing were 100 , 98.1, 65.9, 100 % respectively (equivocal results were counted as positives). 280 cases (F=153, median age 43yrs) had LP performed more than 24 h after the onset of headache (median time to LP=72 h). The sensitivity and specificity of the CSF-xanthochromia were 100 and 97.4 % in this group with NPV 100 % and PPV 66.6 %. 183 (65.4 %) CSF- xanthochromia negative cases in this subgroup had follow up data and survived without SAH occurring in the 12months follow up. In this study, supported by follow up outcome data, we show that CSF-xanthochromia testing using the revised UK-NEQAS method is fit-for-purpose for the use as a second line test to exclude SAH in patients with negative CT-brain including delayed presentation more than 24 h after headache onset.

Selective Identification of Hazardous Gases Using Flexible, Room-Temperature Operable Sensor Array Based on Reduced Graphene Oxide and Metal Oxide Nanoparticles via Machine Learning.

Selective detection and monitoring of hazardous gases with similar properties are highly desirable to ensure human safety. The development of flexible and room-temperature (RT) operable chemiresistive gas sensors provides an excellent opportunity to create wearable devices for detecting hazardous gases surrounding us. However, chemiresistive gas sensors typically suffer from poor selectivity and zero-cross selectivity toward similar types of gases. Herein, a flexible, RT operable chemiresistive gas sensors array is designed, featuring reduced graphene oxide (rGO) and rGO decorated with zinc oxide (ZnO), titanium dioxide (TiO2), and tin dioxide (SnO2) nanoparticles (NPs) on a flexible polyimide (PI) substrate. The sensor array consists of four different sensing layers capable of the selective identification of various hazardous gases such as NO2, NO, and SO2 using machine learning (ML). The gas sensor array exhibits a stable response even when mechanically deformed or exposed to high humidity (up to 60%). Each gas sensor, due to the different metal oxide NPs, shows unique responses in terms of sensitivity, responsiveness, response time, and recovery time to different gases. Consequently, the sensor array generates distinct response patterns that effectively differentiate between the target gases. By leveraging these distinctive recovery patterns and employing a data fusion approach in ML, specific concentrations of target gases can be distinguished. Using ML with fused array sensing data, the training and test accuracies achieved were 98.20 and 97.70%, respectively. This innovative combination of sensor arrays and ML offers significant potential for selective gas detection in environmental monitoring and personal safety applications.

Prognostic and diagnostic implications of impaired rest and exercise-stress left atrial compliance in heart failure with preserved ejection fraction: Insights from the HFpEF Stress Trial

Abstract Background With emerging therapies, early diagnosis of heart failure with preserved ejection fraction (HFpEF) comes to the fore. Whilst the reference standard of exercise-stress right heart catheterisation is well established, the clinical routine struggles between feasibility of exercise-stress and diagnostic accuracy of available tests. Methods The HFpEF Stress Trial (DZHK-17) prospectively enrolled 75 patients with exertional dyspnoea and echocardiographic signs of diastolic dysfunction (E/e’>8) who underwent simultaneous rest and exercise-stress echocardiography and RHC. HFpEF was defined according to pulmonary capillary wedge pressure (HFpEF: PCWP rest: ≥15mmHg stress: ≥25mmHg). Patients were classified as non-cardiac dyspnoea (NCD) in the absence of HFpEF and cardiovascular disease. LA compliance was defined as reservoir strain (Es)/(E/e’). Follow-up was conducted after 4 years to evaluate cardiovascular hospitalisation (CVH). Results The final study population included 68 patients (HFpEF n=34 and NCD n=34) of which 23 reached the clinical endpoint, 1 Patient was lost to follow-up. Patients with HFpEF according to the HFA-PEFF score (≥5 points) had significantly lower LA compliance at rest (p<0.001) compared to patients with a score ≤4. LA compliance at rest outperformed E/e’ (AUC 0.78 vs 0.87, p=0.024) and showed a statistical trend to outperform Es (AUC 0.79 vs 0.87, p=0.090) for the diagnosis of HFpEF. LA compliance at rest predicted CVH (HR 2.83, 95% CI 1.70-4.74, p<0.001) irrespective of concomitant atrial fibrillation. Conclusions LA compliance at rest can be obtained from clinical routine imaging and bears strong diagnostic and prognostic accuracy. Addition of LA compliance can improve the role of echocardiography as the primary test and gatekeeper.

Gene-echocardiography: refining genotype-phenotype correlations in cardiomyopathy patients with restrictive phenotype

Abstract Background The restrictive phenotype, the least frequent and genetically complex form of cardiomyopathy, is the focus of this investigation. The aim is to clarify the association between restrictive phenotype and genetic variants in cardiomyopathy patient. Methods We utilized whole-exome sequencing in a cohort of 568 cardiomyopathy patients and performed echocardiographic evaluations to identify restrictive phenotypes. Comprehensive analysis of cardiac characteristics and functionality was also undertaken. Results Among the cohort, 33 patients exhibited a restrictive phenotype. These patients were distributed as follows: 7 within the restrictive cardiomyopathy group (RCM), 20 in the hypertrophic cardiomyopathy group (HRCM), and 6 in the dilated cardiomyopathy group (HDRCM). A notable prevalence of gene mutations, mainly in the MYBPC3 gene, was found in these individuals. We observed that mutation carriers, characterized by smaller aortic sinus dimension, demonstrated an association with the mutation and diminished myocardial work. Furthermore, MYBPC3 mutation carriers showed a thicker ventricular wall and a higher predisposition for septal hypertrophy. Both HRCM and HDRCM groups displayed consistent hypertrophic manifestations, with the latter demonstrating a higher frequency of interventricular septal hypertrophy. The 5-year survival rate was calculated at 54.5%, with gene mutations not significantly affecting longevity. Conclusion Genetic factors profoundly influence the restrictive phenotype of cardiomyopathy. There is a strong link between restrictive phenotype and genetic variants in cardiomyopathy, providing a foundation for more accurate genetic testing and personalized management of patients. The emerging "gene-echocardiography" concept may refine the accuracy and productivity of genetic counseling and testing in cardiomyopathy.

Federated Learning with tensor networks: a quantum AI framework for healthcare

Abstract The healthcare industries frequently handle sensitive and proprietary data, and due to strict
privacy regulations, they are often reluctant to share data directly. In today’s context, Federated
Learning (FL) stands out as a crucial remedy, facilitating the rapid advancement of distributed machine
learning while effectively managing critical concerns regarding data privacy and governance.
The fusion of federated learning and quantum computing represents a groundbreaking interdisciplinary
approach with immense potential to revolutionize various industries, from healthcare to
finance. In this work, we propose a federated learning framework based on quantum tensor networks
that takes advantage of the principles of many-body quantum physics. Currently, there are
no known classical tensor networks implemented in federated settings. Furthermore, we investigated
the effectiveness and feasibility of the proposed framework by conducting a differential privacy analysis
to ensure the security of sensitive data across healthcare institutions. Experiments on popular
medical image datasets show that the federated quantum tensor network model achieved a mean
receiver-operator characteristic area under the curve (ROC-AUC) of 91-98%, outperforming several
state-of-the-art federated learning methods while using fewer parameters. Experimental results
demonstrate that the quantum federated global model, consisting of highly entangled tensor network
structures, showed better generalization and robustness and achieved higher testing accuracy,
surpassing the performance of locally trained clients under unbalanced data distributions among
healthcare institutions.

The application of Chinese version of SARC-F and SARC-CalF in sarcopenia screening against five definitions: a diagnostic test accuracy study

BackgroundSARC-F questionnaire is a simple and convenient tool for sarcopenia screening, and SARC-CalF is a modified version of it. The developments of their Chinese versions are warranted for the clinical use for Chinese population. This study aimed to culturally adapt the SARC-F questionnaire into Chinese using standardized methods, validate the reliability and diagnostic accuracy of the Chinese version SARC-F and SARC-CalF against five sarcopenia diagnosis criteria, and determine optimal cut-off values for clinical practice in Chinese population.MethodsThe translation and cross-cultural adaptation of SARC-F into Chinese were conducted following the methodological report from European Union Geriatric Medicine Society Sarcopenia Special Interest Group. The Chinese version of SARC-F was validated through a diagnostic test, using diagnostic criteria of sarcopenia recommended by the revised 2019 European Working Group on Sarcopenia in Older People (EWGSOP2) consensus, Asian Working Group for Sarcopenia (AWGS2019) consensus, the International Working Group on Sarcopenia (IWGS), the Foundation for the National Institutes of Health (FNIH) Biomarkers Consortium and the Sarcopenia Definition and Outcomes Consortium (SDOC). Additional analysis was done against the criteria of severe sarcopenia according to the revised EWGSOP2 and AWGS2019.ResultsThe Chinese version of SARC-F was well translated and demonstrated good reliability and acceptability. The diagnostic test included 1859 community-dwelling older individuals from two medical centers. Against five different definitions of sarcopenia, the Chinese version of SARC-F showed reasonable diagnostic accuracy for sarcopenia screening (AUC 0.614–0.821), and was demonstrated low sensitivity (13.7–37.9%) but high specificity (94.8–97.7%) with a cut-off value of ≥ 4. SARC-CalF significantly enhanced the diagnostic accuracy of SARC-F when using definitions of EWGSOP2, AWGS2019 and IWGS (all P ≤ 0.001). A score of ≥ 2 for SARC-F and ≥ 7 for SARC-CalF were established as optimal cut-off points for identifying older individuals as at risk of sarcopenia in Chinese population.ConclusionsThe Chinese version SARC-F is of reasonable reliability and validity for sarcopenia screening. Despite its low sensitivity, it proves to be a useful tool to identify severe cases in community taking advantage of its simplicity. SARC-CalF appears to be a more suitable screening tool for clinical use in detecting sarcopenia.

The New Method for Ground Insulation Monitoring of the Secondary Circuit in Substations

When two-or-more-point grounding faults occur in the secondary circuit of a substation, they will lead to large data measurement errors in the protection, measurement, and control devices, which can easily cause protection malfunction. At present, there is a lack of an online panoramic monitoring system for the secondary circuits of substations, and it is difficult to accurately identify the faults in these secondary circuits. In order to solve this problem, this paper proposes a cloud–edge collaboration framework for the power Internet of Things (IoT), which can accurately identify faults by monitoring the secondary circuit of power grid substations with panoramic online status, reducing maintenance costs and improving overall operational efficiency. Firstly, the collection requirements of current transformer (CT) and potential transformer (PT) fault characteristics in the secondary circuit of the substation are analyzed. The platform architecture of the secondary loop panoramic monitoring system of the cloud-side cooperative substation is proposed, along with the algorithm and data wireless acquisition scheme of the edge computing, cloud computing, and cloud-side system. Furthermore, the edge computing platform and the current and voltage wireless sensors are developed. The maximum error in the accuracy test of the voltage sensor is 1.9180%, while the maximum error in the accuracy test of the current sensor is 1.3406%.

SignifAI: French sign language translation based on deep learning and large language models

In this project, we propose SignifAI which is an AI-powered software that translates French Sign Language into spoken French. The goal of SignifAI is to remove the language barrier that separates the deaf and hearing communities. The translation process proves to be a major challenge for existing methods, as most of them rely on traditional Natural Language Processing (NLP) methods which solve the task in a similar way as spoken languages (e.g., paying more attention to the logics and reasoning). The drastic syntactical difference between sign and spoken languages, however, makes such approaches not as effective and efficient. In SignifAI, we propose to leverage a combination of deep learning models for sign gesture recognition and large language models (LLM) for sentence generation from recognized words. The former allows for precise detection of individual words, while the latter takes the recognized words as inputs and returns a fluently reformulated translated sentence. With extremely limited data available for French sign language, we created our own enhanced dataset through synthetically augmenting existing videos. Utilizing the SlowFast model for this task, we have achieved a maximum test accuracy of 98.9%, and a fluid translation process after implementing ChatGPTs API. The potential applications of SignifAI range from aiding communication between the deaf and the hearing, to helping the deaf place phone calls in numerical services. Ultimately, SignifAI strives to improve all aspects of a deaf persons life, and compensates for their speaking disabilities.

Improved Early Detection of Tube Leaks Faults in Pulverised Coal-fired Boiler Using Deep Feed Forward Neural Network

Boiler tube leaks significantly reduce the operational availability of power units, yet their early detection and prediction have not been fully realised in the industry. This paper introduces a novel approach employing deep feedforward neural networks for early detection of boiler tube leaks in pulverised coal-fired boilers. Early detection enhances repair planning, minimising downtime and production losses. It also improves monitoring and control of boiler tube failures, thereby optimising power plant operations and revenue. Diverse deep neural network models were developed and rigorously tested by leveraging 9 years of operational data (2012–2020). Exhaustive hyper-parameter optimisation, involving seven parameters, substantially improved predictive accuracy. By achieving training and testing accuracies of 82.8% to 99.3%, the study assessed their ability to detect boiler tube leaks over the same 9-year span, providing insights into leak detection capabilities. The models notably predicted all 12 identified tube leak events, averaging a 14-day lead time before boiler shutdown. In addition to leak prediction, a leak detection matrix was devised to analyse residual behaviour and reduce the likelihood of false alarms. However, the models’ predictive performance was observed to be limited to the following year, with satisfactory results for 2021 only. Incorporating the 2021 data into retraining significantly improved the predictions for 2022. The study concludes that while the models demonstrate robust short-term prediction capabilities, they require continuous retraining to maintain accuracy and relevance. This ongoing refinement is essential for keeping the models up-to-date and reliable in predicting future boiler tube leaks.

A Scalable Digital Light Processing 3D Printing Method

The 3D printing method based on digital light processing (DLP) technology can transform liquid resin materials into complex 3D models. However, due to the limitations of digital micromirror device (DMD) specifications, the normal DLP 3D printing method (NDPM) cannot simultaneously process large-size and small-feature parts. Therefore, a scalable DLP 3D printing method (SDPM) was proposed. Different printing resolutions for a part were designed by changing the distance between the projector and the molding liquid level. A scalable DLP printer was built to realize the printing resolution requirements at different sizes. A series of experiments were performed. Firstly, the orthogonal experimental method was used, and the minimum and maximum projection distances were obtained as 20.5 cm and 30.5 cm, respectively. Accordingly, the layer thickness, exposure time, and waiting leveling time were 0.08 mm, 3 s, and 6 s and 0.08 mm, 7 s, and 10 s. Secondly, single-layer column feature printing was finished, which was shown to have two minimum printing resolutions of 101 μm and 157 μm at a projection distance of 20.5 cm and 30.5 cm. Thirdly, a shape accuracy test was conducted by using the SDPM. Compared with the NDPM, the shape accuracy of the small-feature round, diamond, and square parts was improved by 49%, 42%, and 2%, respectively. This study verified that the SDPM can build models with features demonstrating high local shape accuracy.

Deep learning approaches for classification of copper-containing metal scrap in recycling processes

Separating copper from iron scrap is a critical operation in metal recycling and achieving this with low cost sensoric equipment like RGB cameras instead of XRF/XRT is becoming increasingly attractive. In this article, the groundwork for creating an image classification model to separate copper from iron scrap has been performed. Twenty of the most common and most easily available CNN architectures were trained on 2200 metal scrap specimens and evaluated inline on a sensor-based sorting rig for their prediction accuracy and their inference latency to mimic real circumstances in an industrial setting. Out of these evaluated architectures, DenseNet-201 with 98% accuracy in inline tests is recommended if potent hardware is available. Otherwise AlexNet with 92% accuracy or MobileNet-V2 with 90% accuracy are recommended for further investigation and model creation if hardware restrictions apply. Based on the presented results in this article, the initial cumbersome surveyance of the most suitable network architecture can be substantially reduced and the creation of a sorting model can be streamlined. This article thus provides the basis for creating an inline applicable sorting method for scrap metal that uses low cost sensorics equipment and can provide reasonably high accuracy in its prediction.

Acceptability of six-monthly PrEP dispensing supported with interim HIV self-testing to simplify PrEP delivery in Kenya: findings from qualitative research

BackgroundIn Africa, dispensing oral HIV pre-exposure prophylaxis (PrEP) within already strained public health facilities has led to prolonged waiting periods and suboptimal experiences for clients. We sought to explore the acceptability of dispensing PrEP semiannually with interim HIV self-testing (HIVST) versus quarterly PrEP dispensing with clinic-based HIV testing to optimize clinic-delivered PrEP services.MethodsWe conducted a qualitative study within a non-inferiority individual-level randomized controlled trial testing the effect of six-monthly PrEP dispensing with HIVST compared to the standard-of-care three-monthly PrEP dispensing on PrEP clinical outcomes in Kenya (ClinicalTrials.gov: NCT03593629). Eligible participants were ≥ 18 years, refilling PrEP for the first time, and either in an HIV serodifferent relationship (men and women) or singly enrolled (women only). A subset of participants in the intervention group completed serial in-depth interviews (IDIs) at enrollment, six months, and 12 months. We utilized stratified purposive sampling to ensure representation across participant groups. We analyzed our qualitative data thematically using a combination of inductive and deductive approaches, the latter guided by the Theoretical Framework of Acceptability (TFA).ResultsBetween May 2018 and June 2021, we conducted 120 serial IDIs with 55 participants; 64% (35/55) were in a serodifferent relationship, 64% (35/55) were women, and the median age was 32 years (IQR 27–40). Overall, participants found this novel PrEP delivery model highly acceptable; it was well-liked, private (TFA construct: affective attitude), and less burdensome (TFA construct: burden) compared to standard PrEP delivery. Additionally, participants were confident in their ability to participate in the intervention (TFA construct: self-efficacy). Some participants, however, highlighted model disadvantages, including fewer opportunities for in-person counseling and potentially less accurate HIV testing (TFA construct: opportunity costs). Ultimately, most participants reported that the intervention allowed them to achieve their HIV prevention goals (TFA construct: perceived effectiveness) and that their confidence in at-home HIVST and PrEP continuation increased following each semiannual clinic visit.ConclusionsSemiannual PrEP clinic visits supported with six-monthly drug dispensing and interim HIVST was acceptable among PrEP users who experienced the intervention in Kenya. More comprehensive pre-intervention counseling and training on HIVST may help alleviate the client concerns presented, which were often resolved over time with intervention experience.

The Critical Shoulder Angle as a Highly Specific Predictor of a Full-Thickness Rotator Cuff Tear: A Case-Control Study.

The critical shoulder angle (CSA) has become an important topic of study in patients with rotator cuff tears (RCTs). However, there are conflicting data on whether the CSA can differentiate between patients with normal shoulder pathology and full-thickness RCTs on shoulder radiographs. The purpose of this study was to define the relationship between full-thickness RCTs and the CSA. It was hypothesized that patients with full-thickness RCTs would have an increased CSA compared with matched controls. Cross-sectional study; Level of evidence, 3. This retrospective case-control study identified patients with magnetic resonance imaging scans showing full-thickness RCTs between 2009 and 2019. A 1 to 1 propensity score match was performed to identify a control group with normal rotator cuffs while controlling for baseline participant characteristics-including age, sex, body mass index, and tobacco use. A total cohort of 532 was identified, with 266 cases and 266 controls. Two independent observers measured CSAs on true anteroposterior shoulder radiographs. There was no difference in baseline participant characteristics between the RCT and the non-RCT groups (P > .05). The mean CSA for the entire cohort was 33.6°± 4.2°. The CSA did not significantly vary by sex (P = .088) or tobacco usage (P = .16). The mean CSA for the RCT case group, 36.2°± 3.3°, was significantly different from the mean CSA for the control group, 30.9°± 3.3° (P < .0001). The receiver operating characteristic curve analysis produced an area under the curve of 0.88 (P < .0001). At CSAs ≥35°, there was a 67.7% sensitivity and 89.4% specificity for having a full-thickness RCT. Last, each degree of increase in the CSA increased the risk of having an associated RCT by 1.7 times (OR, 1.7 [95% CI, 1.551-1.852]; P < .0001). Patients with RCTs had significantly higher CSAs compared with matched controls. Increased CSA was an independent risk factor for RCTs, with an odds ratio of 1.7 per degree. The CSA is an accurate test (area under the curve, 0.88) with good sensitivity (67.7%) and specificity (89.4%) at values ≥35°. The CSA is a simple, reproducible measurement that can assist in clinical decision-making regarding full-thickness RCTs.

Testing for Equal Average Forecast Accuracy in Possibly Unstable Environments *

We consider the issue of testing the null of equal average forecast accuracy in a model where the forecast error loss differential series has a potentially non-constant mean function over time. We show that when time variation is present in the loss differential mean, the standard Diebold and Mariano (1995) test, which was proposed for evaluating forecasts in a stable environment, has an asymptotic size of zero, and, whilst consistent, can have reduced local power. This arises due to inconsistent estimation of the implicit long run variance estimator, which diverges under a time varying mean. We suggest a modified statistic that replaces the standard long run variance estimator based on full-sample demeaning of the loss differential series with one based on nonparametric local demeaning. The new long run variance estimator is consistent under both the null and alternative when the mean function is time varying or constant, and in both cases, the modified test recovers the asymptotic size and power properties associated with the original test in the constant mean case. The modified test therefore provides a robust method for testing the equal average forecast accuracy null, allowing for instability in the loss differential mean. The benefits of our test are demonstrated via Monte Carlo simulation and two empirical applications.