Search

Background: Misdiagnosis, arbitrary charges, annoying queues, and clinic waiting times among others are long-standing phenomena in the medical industry across the world. These factors can contribute to patient anxiety about misdiagnosis by clinicians. However, with the increasing growth in use of big data in biomedical and health care communities, the performance of artificial intelligence (Al) techniques of diagnosis is improving and can help avoid medical practice errors, including under the current circumstance of COVID-19., Objective: This study aims to visualize and measure patients' heterogeneous preferences from various angles of AI diagnosis versus clinicians in the context of the COVID-19 epidemic in China. We also aim to illustrate the different decision-making factors of the latent class of a discrete choice experiment (DCE) and prospects for the application of AI techniques in judgment and management during the pandemic of SARS-CoV-2 and in the future., Methods: A DCE approach was the main analysis method applied in this paper. Attributes from different dimensions were hypothesized: diagnostic method, outpatient waiting time, diagnosis time, accuracy, follow-up after diagnosis, and diagnostic expense. After that, a questionnaire is formed. With collected data from the DCE questionnaire, we apply Sawtooth software to construct a generalized multinomial logit (GMNL) model, mixed logit model, and latent class model with the data sets. Moreover, we calculate the variables' coefficients, standard error, P value, and odds ratio (OR) and form a utility report to present the importance and weighted percentage of attributes., Results: A total of 55.8% of the respondents (428 out of 767) opted for AI diagnosis regardless of the description of the clinicians. In the GMNL model, we found that people prefer the 100% accuracy level the most (OR 4.548, 95% CI 4.048-5.110, P<.001). For the latent class model, the most acceptable model consists of 3 latent classes of respondents. The attributes with the most substantial effects and highest percentage weights are the accuracy (39.29% in general) and expense of diagnosis (21.69% in general), especially the preferences for the diagnosis "accuracy" attribute, which is constant across classes. For class 1 and class 3, people prefer the AI + clinicians method (class 1: OR 1.247, 95% CI 1.036-1.463, P<.001; class 3: OR 1.958, 95% CI 1.769-2.167, P<.001). For class 2, people prefer the AI method (OR 1.546, 95% CI 0.883-2.707, P=.37). The OR of levels of attributes increases with the increase of accuracy across all classes., Conclusions: Latent class analysis was prominent and useful in quantifying preferences for attributes of diagnosis choice. People's preferences for the "accuracy" and "diagnostic expenses" attributes are palpable. AI will have a potential market. However, accuracy and diagnosis expenses need to be taken into consideration., (©Taoran Liu, Winghei Tsang, Fengqiu Huang, Oi Ying Lau, Yanhui Chen, Jie Sheng, Yiwei Guo, Babatunde Akinwunmi, Casper JP Zhang, Wai-Kit Ming. Originally published in the Journal of Medical Internet Research (http://www.jmir.org), 23.02.2021.)

Early, ideally pre-symptomatic, recognition of common diseases (e.g., heart disease, cancer, diabetes, Alzheimer's disease) facilitates early treatment or lifestyle modifications, such as diet and exercise. Sensitive, specific identification of diseases using blood samples would facilitate early recognition. We explored the potential of disease identification in high dimensional blood microRNA (miRNA) datasets using a powerful data reduction method: principal component analysis (PCA). Using Qlucore Omics Explorer (QOE), a dynamic, interactive visualization-guided bioinformatics program with a built-in statistical platform, we analyzed publicly available blood miRNA datasets from the Gene Expression Omnibus (GEO) maintained at the National Center for Biotechnology Information at the National Institutes of Health (NIH). The miRNA expression profiles were generated from real time PCR arrays, microarrays or next generation sequencing of biologic materials (e.g., blood, serum or blood components such as platelets). PCA identified the top three principal components that distinguished cohorts of patients with specific diseases (e.g., heart disease, stroke, hypertension, sepsis, diabetes, specific types of cancer, HIV, hemophilia, subtypes of meningitis, multiple sclerosis, amyotrophic lateral sclerosis, Alzheimer's disease, mild cognitive impairment, aging, and autism), from healthy subjects. Literature searches verified the functional relevance of the discriminating miRNAs. Our goal is to assemble PCA and heatmap analyses of existing and future blood miRNA datasets into a clinical reference database to facilitate the diagnosis of diseases using routine blood draws., Competing Interests: No competing interests.

As whole exome sequencing (WES) becomes more widely used in the clinical realm, a wealth of unanalyzed information will be routinely generated. Using WES read depth data to predict copy number variation (CNV) could extend the diagnostic utility of this previously underutilized data by providing clinically important information such as previously unsuspected deletions or duplications. We evaluated ExomeDepth, a free R package, in addition to an aneuploidy prediction method, to detect CNVs in WES data. First, in a blinded pilot study, five out of five genomic alterations were correctly identified from clinical samples with previously defined chromosomal gains or losses, including submicroscopic deletions, duplications, and chromosomal trisomy. We then examined CNV calls among 53 patients participating in the NCGENES research study and undergoing WES, who had existing clinical chromosomal microarray (CMA) data that could be used for validation. For unique CNVs that overlap well with WES coverage regions, sensitivity was 89% for deletions and 65% for duplications. While specificity of the algorithm calls remains a concern, this is less of an issue at high threshold filtering levels. When applied to all 672 patients from the exome sequencing study, ExomeDepth identified eleven diagnostically relevant CNVs ranging in size from a two exon deletion to whole chromosome duplications, as well as numerous other CNVs with varying clinical significance. This opportunistic analysis of WES data yields an additional 1.6% of patients in this study with pathogenic or likely pathogenic CNVs that are clinically relevant to their phenotype as well as clinically relevant secondary findings. Finally, we demonstrate the potential value of copy number analysis in cases where a single heterozygous likely or known pathogenic single nucleotide alteration is identified in a gene associated with an autosomal recessive condition., Competing Interests: The authors have declared that no competing interests exist.

We review the recent development of mobile detection instruments used for medical diagnostics, and consider the relative advantages of approaches that utilize the internal sensing capabilities of commercially available mobile communication devices (such as smartphones and tablet computers) compared to those that utilize a custom external sensor module. In this review, we focus specifically upon mobile medical diagnostic platforms that are being developed to serve the need in global health, personalized medicine, and point-of-care diagnostics., (Copyright © 2016 Elsevier Inc. All rights reserved.)

Purpose: To validate a scoring system that evaluates the ability of Internet search engines to correctly predict diagnoses when symptoms are used as search terms., Methods: We developed a five point scoring system to evaluate the diagnostic accuracy of Internet search engines. We identified twenty diagnoses common to a primary care setting to validate the scoring system. One investigator entered the symptoms for each diagnosis into three Internet search engines (Google, Bing, and Ask) and saved the first five webpages from each search. Other investigators reviewed the webpages and assigned a diagnostic accuracy score. They rescored a random sample of webpages two weeks later. To validate the five point scoring system, we calculated convergent validity and test-retest reliability using Kendall's W and Spearman's rho, respectively. We used the Kruskal-Wallis test to look for differences in accuracy scores for the three Internet search engines., Results: A total of 600 webpages were reviewed. Kendall's W for the raters was 0.71 (p<0.0001). Spearman's rho for test-retest reliability was 0.72 (p<0.0001). There was no difference in scores based on Internet search engine. We found a significant difference in scores based on the webpage's order on the Internet search engine webpage (p=0.007). Pairwise comparisons revealed higher scores in the first webpages vs. the fourth (corr p=0.009) and fifth (corr p=0.017). However, this significance was lost when creating composite scores., Conclusions: The five point scoring system to assess diagnostic accuracy of Internet search engines is a valid and reliable instrument. The scoring system may be used in future Internet research., (Copyright © 2013 Elsevier Ireland Ltd. All rights reserved.)

Background: Searching for online information to interpret symptoms is an increasingly prevalent activity among patients, even among older adults. As older adults typically have complex health care needs, their risk of misinterpreting symptoms via online self-diagnosis may be greater. However, limited research has been conducted with older adults in the areas of symptom interpretation and human-computer interaction., Objective: The intent of the study was to describe the processes that a sample of older adults may use to diagnose symptoms online as well as the processes that predict accurate diagnosis., Methods: We conducted a series of "think-aloud" protocols with 79 adults aged 50 years or older. Participants received one of two vignettes that depicted symptoms of illness. Participants talked out loud about their thoughts and actions while attempting to diagnose the symptoms with and without the help of common Internet tools (Google and WebMD's Symptom Checker). Think-aloud content was categorized using an adapted Q-sort and general inductive approach. We then compared the think-aloud content of participants who were accurate in their diagnosis with those who were not., Results: Nineteen descriptive codes were identified from the think-aloud content. The codes touched upon Web navigation, attempts to organize and evaluate online health information, and strategies to diagnose symptoms. Participants most frequently relied on a strategy where they reviewed and then rejected the online diagnoses if they contained additional symptoms than those that were depicted in the vignette. Finally, participants who were inaccurate in their diagnosis reported being confused by the diagnosis task, lacking confidence in their diagnosis, and using their past experiences with illness to guide diagnosis more frequently than those participants who accurately diagnosed the symptoms., Conclusions: Older adult participants tended to rely on matching strategies to interpret symptoms, but many still utilized existing medical knowledge and previous illness experiences as a guide for diagnosis. Many participants also had difficulty navigating the Internet tools, which suggests an increased need for navigation aids in Web design. Furthermore, participants who were inaccurate in their diagnosis had more difficulty with the Internet tools and confusion with the task than those who were accurate. Future work in this area may want to utilize additional study design such as eye-tracking to further understand the coordination between Web navigation, online symptom information processing, and diagnostic strategies.

Cell heterogeneity and the inherent complexity due to the interplay of multiple molecular processes within the cell pose difficult challenges for current single-cell biology. We introduce an approach that identifies a disease phenotype from multiparameter single-cell measurements, which is based on the concept of "supercell statistics", a single-cell-based averaging procedure followed by a machine learning classification scheme. We are able to assess the optimal tradeoff between the number of single cells averaged and the number of measurements needed to capture phenotypic differences between healthy and diseased patients, as well as between different diseases that are difficult to diagnose otherwise. We apply our approach to two kinds of single-cell datasets, addressing the diagnosis of a premature aging disorder using images of cell nuclei, as well as the phenotypes of two non-infectious uveitides (the ocular manifestations of Behçet's disease and sarcoidosis) based on multicolor flow cytometry. In the former case, one nuclear shape measurement taken over a group of 30 cells is sufficient to classify samples as healthy or diseased, in agreement with usual laboratory practice. In the latter, our method is able to identify a minimal set of 5 markers that accurately predict Behçet's disease and sarcoidosis. This is the first time that a quantitative phenotypic distinction between these two diseases has been achieved. To obtain this clear phenotypic signature, about one hundred CD8(+) T cells need to be measured. Although the molecular markers identified have been reported to be important players in autoimmune disorders, this is the first report pointing out that CD8(+) T cells can be used to distinguish two systemic inflammatory diseases. Beyond these specific cases, the approach proposed here is applicable to datasets generated by other kinds of state-of-the-art and forthcoming single-cell technologies, such as multidimensional mass cytometry, single-cell gene expression, and single-cell full genome sequencing techniques.

Background: Few studies have examined the impact of novel indices of comorbidity, frailty, and disability on outcomes after transcatheter aortic valve replacement (TAVR).Objectives: This study analyzed patients from the Medtronic CoreValve U.S. Pivotal Trial program to develop a simple scoring system that incorporates standard and novel predictor variables.Methods: A multidisciplinary heart team used objective criteria, such as The Society of Thoracic Surgeons Predicted Risk of Mortality (STS PROM), and subjective criteria to assess patients' eligibility for TAVR. The analysis included 3,687 patients randomly divided (2:1) into a derivation cohort (n = 2,482) and a validation cohort (n = 1,205). The study evaluated predictors of all-cause death, which were used to calculate a risk score for each patient.Results: The overall mortality rate was 5.8% at 30 days and 22.8% at 1 year. Home oxygen use, assisted living, albumin levels <3.3 g/dl, and age >85 years predicted death at 30 days. Home oxygen use, albumin levels <3.3 g/dl, falls in the past 6 months, STS PROM score >7%, and severe (≥5) Charlson comorbidity score predicted death at 1 year. A simple scoring system created on the basis of these multivariable predictors effectively stratified risk at 30 days and 1 year into low-risk, moderate-risk, and high-risk subsets. This score showed a 3-fold difference in mortality rates for the low-risk and high-risk subsets at 30 days (3.6% and 10.9%, respectively) and 1 year (12.3% and 36.6%, respectively). The 1-year mortality model was more stable than the 30-day model (C-statistics: 0.79 vs. 0.75).Conclusions: A simple score dominated by novel predictors of outcome effectively stratified early and late mortality rates in extreme-risk and high-risk patients and may assist in selecting appropriate candidates for TAVR. (Safety and Efficacy Study of the Medtronic CoreValve System in the Treatment of Symptomatic Severe Aortic Stenosis in High Risk and Very High Risk Subjects Who Need Aortic Valve Replacement; NCT01240902). [ABSTRACT FROM AUTHOR]