Application of AI in Sepsis: Citation Network Analysis and Evidence Synthesis

BACKGROUND

Artificial intelligence (AI) has garnered considerable attention in the context of sepsis research, particularly in personalized diagnosis and treatment. Conducting a bibliometric analysis of existing publications can offer a broad overview of the field and identify current research trends and future research directions.

OBJECTIVE

The objective of this study is to leverage bibliometric data to provide a comprehensive overview of the application of AI in sepsis.

METHODS

We conducted a search in the Web of Science Core Collection database to identify relevant articles published in English until August 31, 2023. A predefined search strategy was used, evaluating titles, abstracts, and full texts as needed. We used the Bibliometrix and VOSviewer tools to visualize networks showcasing the co-occurrence of authors, research institutions, countries, citations, and keywords.

RESULTS

A total of 259 relevant articles published between 2014 and 2023 (until August) were identified. Over the past decade, the annual publication count has consistently risen. Leading journals in this domain include Critical Care Medicine (17/259, 6.6%), Frontiers in Medicine (17/259, 6.6%), and Scientific Reports (11/259, 4.2%). The United States (103/259, 39.8%), China (83/259, 32%), United Kingdom (14/259, 5.4%), and Taiwan (12/259, 4.6%) emerged as the most prolific countries in terms of publications. Notable institutions in this field include the University of California System, Emory University, and Harvard University. The key researchers working in this area include Ritankar Das, Chris Barton, and Rishikesan Kamaleswaran. Although the initial period witnessed a relatively low number of articles focused on AI applications for sepsis, there has been a significant surge in research within this area in recent years (2014-2023).

CONCLUSIONS

This comprehensive analysis provides valuable insights into AI-related research conducted in the field of sepsis, aiding health care policy makers and researchers in understanding the potential of AI and formulating effective research plans. Such analysis serves as a valuable resource for determining the advantages, sustainability, scope, and potential impact of AI models in sepsis.

Artificial Intelligence in Kidney Disease: A Comprehensive Study and Directions for Future Research

Artificial intelligence (AI) has emerged as a promising tool in the field of healthcare, with an increasing number of research articles evaluating its applications in the domain of kidney disease. To comprehend the evolving landscape of AI research in kidney disease, a bibliometric analysis is essential. The purposes of this study are to systematically analyze and quantify the scientific output, research trends, and collaborative networks in the application of AI to kidney disease. This study collected AI-related articles published between 2012 and 20 November 2023 from the Web of Science. Descriptive analyses of research trends in the application of AI in kidney disease were used to determine the growth rate of publications by authors, journals, institutions, and countries. Visualization network maps of country collaborations and author-provided keyword co-occurrences were generated to show the hotspots and research trends in AI research on kidney disease. The initial search yielded 673 articles, of which 631 were included in the analyses. Our findings reveal a noteworthy exponential growth trend in the annual publications of AI applications in kidney disease. Nephrology Dialysis Transplantation emerged as the leading publisher, accounting for 4.12% (26 out of 631 papers), followed by the American Journal of Transplantation at 3.01% (19/631) and Scientific Reports at 2.69% (17/631). The primary contributors were predominantly from the United States (n = 164, 25.99%), followed by China (n = 156, 24.72%) and India (n = 62, 9.83%). In terms of institutions, Mayo Clinic led with 27 contributions (4.27%), while Harvard University (n = 19, 3.01%) and Sun Yat-Sen University (n = 16, 2.53%) secured the second and third positions, respectively. This study summarized AI research trends in the field of kidney disease through statistical analysis and network visualization. The findings show that the field of AI in kidney disease is dynamic and rapidly progressing and provides valuable information for recognizing emerging patterns, technological shifts, and interdisciplinary collaborations that contribute to the advancement of knowledge in this critical domain.

Benzodiazepine Use and the Risk of Dementia in the Elderly Population: An Umbrella Review of Meta-Analyses

The prevalence of dementia among the elderly is high, and it is the leading cause of death globally. However, the relationship between benzodiazepine use and dementia risk has produced inconsistent results, necessitating an updated review of the evidence. To address this, we conducted an umbrella review of meta-analyses to summarize the available evidence on the association between benzodiazepine use and dementia risk and evaluate its credibility. We systematically evaluated the meta-analyses of observational studies that examined the connection between benzodiazepine use and dementia risk. For each meta-analysis, we collected the overall effect size, heterogeneity, risk of bias, and year of the most recent article and graded the evidence based on pre-specified criteria. We also used AMSTAR, a measurement tool to evaluate systematic reviews, to assess the methodological quality of each study. Our review included five meta-analyses encompassing 30 studies, and the effect size of the association between benzodiazepine use and dementia risk ranged from 1.38 to 1.78. Nonetheless, the evidence supporting this relationship was weak, and the methodological quality of the studies included was low. In conclusion, our findings revealed limited evidence of a link between benzodiazepine use and dementia risk, and more research is required to determine a causal connection. Physicians should only prescribe benzodiazepine for appropriate indications.

Artificial intelligence in diabetic retinopathy: Bibliometric analysis

BACKGROUND

The use of artificial intelligence in diabetic retinopathy has become a popular research focus in the past decade. However, no scientometric report has provided a systematic overview of this scientific area.

AIMS

We utilized a bibliometric approach to identify and analyse the academic literature on artificial intelligence in diabetic retinopathy and explore emerging research trends, key authors, co-authorship networks, institutions, countries, and journals. We further captured the diabetic retinopathy conditions and technology commonly used within this area.

METHODS

Web of Science was used to collect relevant articles on artificial intelligence use in diabetic retinopathy published between January 1, 2012, and December 31, 2022 . All the retrieved titles were screened for eligibility, with one criterion that they must be in English. All the bibliographic information was extracted and used to perform a descriptive analysis. Bibliometrix (R tool) and VOSviewer (Leiden University) were used to construct and visualize the annual numbers of publications, journals, authors, countries, institutions, collaboration networks, keywords, and references.

RESULTS

In total, 931 articles that met the criteria were collected. The number of annual publications showed an increasing trend over the last ten years. Investigative Ophthalmology & Visual Science (58/931), IEEE Access (54/931), and Computers in Biology and Medicine (23/931) were the most journals with most publications. China (211/931), India (143/931, USA (133/931), and South Korea (44/931) were the most productive countries of origin. The National University of Singapore (40/931), Singapore Eye Research Institute (35/931), and Johns Hopkins University (34/931) were the most productive institutions. Ting D. (34/931), Wong T. (28/931), and Tan G. (17/931) were the most productive researchers.

CONCLUSION

This study summarizes the recent advances in artificial intelligence technology on diabetic retinopathy research and sheds light on the emerging trends, sources, leading institutions, and hot topics through bibliometric analysis and network visualization. Although this field has already shown great potential in health care, our findings will provide valuable clues relevant to future research directions and clinical practice.

A Meta-Analysis of Proton Pump Inhibitor Use and the Risk of Acute Kidney Injury: Geographical Differences and Associated Factors

Proton pump inhibitors (PPIs) are widely prescribed in medical practice for the treatment of several gastrointestinal disorders. Previous epidemiology studies have reported the association between PPI use and the risk of AKI, although the magnitude of the association between PPIs and the risk of acute kidney injury (AKI) remains uncertain. Therefore, we conducted a meta-analysis to determine the relationship between PPI therapy and the risk of AKI. We systematically searched for relevant articles published before January 2023 on PubMed, Scopus, and Web of Science. In addition, we conducted a manual search of the bibliographies of potential articles. Two independent reviewers examined the appropriateness of all studies for inclusion. We pooled studies that compared the risk of AKI with PPI against their control using a random effect model. The search criteria based on PRISMA guidelines yielded 568 articles. Twelve observational studies included 2,492,125 individuals. The pooled adjusted RR demonstrated a significant positive association between PPI therapy and the risk of AKI (adjusted RR 1.75, 95% CI: 1.40-2.19, p < 0.001), and it was consistent across subgroups. A visual presentation of the funnel plot and Egger's regression test showed no evidence of publication bias. Our meta-analysis indicated that persons using PPIs exhibited an increased risk of AKI. North American individuals had a higher risk of AKI compared to Asian and European individuals. However, the pooled effect from observational studies cannot clarify whether the observed association is a causal effect or the result of some unmeasured confounding factors. Hence, the biological mechanisms underlying this association are still unclear and require further research.

Increased risk of atrial fibrillation in patients with psoriasis: A meta-analysis of observational studies

BACKGROUND

Several epidemiological studies have shown that psoriasis increases the risk of developing atrial fibrillation but evidence of this is still scarce.

AIMS

Our objective was to systematically review, synthesise and critique the epidemiological studies that provided information about the relationship between psoriasis and atrial fibrillation risk.

METHODS

We searched through PubMed, EMBASE and the bibliographies for articles published between 1 January 2000, and 1 November 2017, that reported on the association between psoriasis and atrial fibrillation. All abstracts, full-text articles and sources were reviewed with duplicate data excluded. Summary relative risks (RRs) with 95% CI were pooled using a random effects model.

RESULTS

We identified 252 articles, of these eight unique abstracts underwent full-text review. We finally selected six out of these eight studies comprising 11,187 atrial fibrillation patients. The overall pooled relative risk (RR) of atrial fibrillation was 1.39 (95% CI: 1.257-1.523, P < 0.0001) with significant heterogeneity (I2 = 80.316, Q = 45.723, τ2 = 0.017, P < 0.0001) for the random effects model. In subgroup analysis, the greater risk was found in studies from North America, RR 1.482 (95% CI: 1.119-1.964, P < 0.05), whereas a moderate risk was observed in studies from Europe RR 1.43 (95% CI: 1.269-1.628, P < 0.0001).

LIMITATIONS

We were only able to include six studies with 11,178 atrial fibrillation patients, because only a few such studies have been published.

CONCLUSION

Our results showed that psoriasis is significantly associated with an increased risk of developing atrial fibrillation. Therefore, physicians should monitor patient's physical condition on a timely basis.

Deep Learning for the Diagnosis of Esophageal Cancer in Endoscopic Images: A Systematic Review and Meta-Analysis

Esophageal cancer, one of the most common cancers with a poor prognosis, is the sixth leading cause of cancer-related mortality worldwide. Early and accurate diagnosis of esophageal cancer, thus, plays a vital role in choosing the appropriate treatment plan for patients and increasing their survival rate. However, an accurate diagnosis of esophageal cancer requires substantial expertise and experience. Nowadays, the deep learning (DL) model for the diagnosis of esophageal cancer has shown promising performance. Therefore, we conducted an updated meta-analysis to determine the diagnostic accuracy of the DL model for the diagnosis of esophageal cancer. A search of PubMed, EMBASE, Scopus, and Web of Science, between 1 January 2012 and 1 August 2022, was conducted to identify potential studies evaluating the diagnostic performance of the DL model for esophageal cancer using endoscopic images. The study was performed in accordance with PRISMA guidelines. Two reviewers independently assessed potential studies for inclusion and extracted data from retrieved studies. Methodological quality was assessed by using the QUADAS-2 guidelines. The pooled accuracy, sensitivity, specificity, positive and negative predictive value, and the area under the receiver operating curve (AUROC) were calculated using a random effect model. A total of 28 potential studies involving a total of 703,006 images were included. The pooled accuracy, sensitivity, specificity, and positive and negative predictive value of DL for the diagnosis of esophageal cancer were 92.90%, 93.80%, 91.73%, 93.62%, and 91.97%, respectively. The pooled AUROC of DL for the diagnosis of esophageal cancer was 0.96. Furthermore, there was no publication bias among the studies. The findings of our study show that the DL model has great potential to accurately and quickly diagnose esophageal cancer. However, most studies developed their model using endoscopic data from the Asian population. Therefore, we recommend further validation through studies of other populations as well.

Proton Pump Inhibitors Use and the Risk of Pancreatic Cancer: Evidence from Eleven Epidemiological Studies, Comprising 1.5 Million Individuals

Previous epidemiological studies have shown that proton pump inhibitor (PPI) may modify the risk of pancreatic cancer. We conducted an updated systematic review and meta-analysis of observational studies assessing the effect of PPI on pancreatic cancer. PubMed, Embase, Scopus, and Web of Science were searched for studies published between 1 January 2000, and 1 May 2022. We only included studies that assessed exposure to PPI, reported pancreatic cancer outcomes, and provided effect sizes (hazard ratio or odds ratio) with 95% confidence intervals (CIs). We calculated an adjusted pooled risk ratio (RR) with 95%CIs using the random-effects model. Eleven studies (eight case-control and three cohorts) that reported 51,629 cases of pancreatic cancer were included. PPI was significantly associated with a 63% increased risk of pancreatic cancer (RRadj. 1.63, 95%CI: 1.19-2.22, p = 0.002). Subgroup analysis showed that the pooled RR for rabeprazole and lansoprazole was 4.08 (95%CI: 0.61-26.92) and 2.25 (95%CI: 0.83-6.07), respectively. Moreover, the risk of pancreatic cancer was established for both the Asian (RRadj. 1.37, 95%CI: 0.98-1.81) and Western populations (RRadj.2.76, 95%CI: 0.79-9.56). The findings of this updated meta-analysis demonstrate that the use of PPI was associated with an increased risk of pancreatic cancer. Future studies are needed to improve the quality of evidence through better verification of PPI status (e.g., patient selection, duration, and dosages), adjusting for possible confounders, and ensuring long-term follow-up.

Early Diabetes Prediction: A Comparative Study Using Machine Learning Techniques

Most screening tests for Diabetes Mellitus (DM) in use today were developed using electronically collected data from Electronic Health Record (EHR). However, developing and under-developing countries are still struggling to build EHR in their hospitals. Due to the lack of HER data, early screening tools are not available for those countries. This study develops a prediction model for early DM by direct questionnaires for a tertiary hospital in Bangladesh. Information gain technique was used to reduce irreverent features. Using selected variables, we developed logistic regression, support vector machine, K-nearest neighbor, Naïve Bayes, random forest (RF), and neural network models to predict diabetes at an early stage. RF outperformed other machine learning algorithms achieved 100% accuracy. These findings suggest that a combination of simple questionnaires and a machine learning algorithm can be a powerful tool to identify undiagnosed DM patients.

Clinical Usefulness of Drug-Disease Interaction Alerts from a Clinical Decision Support System, MedGuard, for Patient Safety: A Single Center Study

Clinical decision support systems have been widely used in healthcare, yet few studies have concurrently measured the clinical effectiveness of CDSSs, and the appropriateness of alerts with physicians' response to alerts. We conducted a retrospective analysis of prescriptions caused disease-medication related alerts. Medication orders for outpatients' prescriptions, all aged group were included in this study. All the prescriptions were reviewed, and medication orders compared with a widely used medication reference (UpToDate) and other standard guidelines. We reviewed 1,409 CDS alerts (2.67% alert rate) on 52,654 prescriptions ordered during the study period. 545 (38.70%) of alerts were overridden. Override appropriateness was 2.20% overall. However, the rate of alert acceptance was higher, ranging from 11.11 to 92.86%. The MedGuard system had a lower overridden rate than other systems reported in previous studies. The acceptance rate of alerts by physicians was high. Moreover, false-positive rate was low. The MedGuard system has the potential to reduce alert fatigue and to minimize the risk of patient harm.

Predicting Hepatocellular Carcinoma With Minimal Features From Electronic Health Records: Development of a Deep Learning Model

BACKGROUND

Hepatocellular carcinoma (HCC), usually known as hepatoma, is the third leading cause of cancer mortality globally. Early detection of HCC helps in its treatment and increases survival rates.

OBJECTIVE

The aim of this study is to develop a deep learning model, using the trend and severity of each medical event from the electronic health record to accurately predict the patients who will be diagnosed with HCC in 1 year.

METHODS

Patients with HCC were screened out from the National Health Insurance Research Database of Taiwan between 1999 and 2013. To be included, the patients with HCC had to register as patients with cancer in the catastrophic illness file and had to be diagnosed as a patient with HCC in an inpatient admission. The control cases (non-HCC patients) were randomly sampled from the same database. We used age, gender, diagnosis code, drug code, and time information as the input variables of a convolution neural network model to predict those patients with HCC. We also inspected the highly weighted variables in the model and compared them to their odds ratio at HCC to understand how the predictive model works.

RESULTS

We included 47,945 individuals, 9553 of whom were patients with HCC. The area under the receiver operating curve (AUROC) of the model for predicting HCC risk 1 year in advance was 0.94 (95% CI 0.937-0.943), with a sensitivity of 0.869 and a specificity 0.865. The AUROC for predicting HCC patients 7 days, 6 months, 1 year, 2 years, and 3 years early were 0.96, 0.94, 0.94, 0.91, and 0.91, respectively.

CONCLUSIONS

The findings of this study show that the convolutional neural network model has immense potential to predict the risk of HCC 1 year in advance with minimal features available in the electronic health records.

Artificial Intelligence in Gastric Cancer: Identifying Gastric Cancer Using Endoscopic Images with Convolutional Neural Network

Simple Summary

Gastric cancer (GC) is one of the most newly diagnosed cancers and the fifth leading cause of death globally. Previous studies reported that the detection rate of gastric cancer (EGC) at an earlier stage is low, and the overall false-negative rate with esophagogastroduodenoscopy (EGD) is up to 25.8%, which often leads to inappropriate treatment. Accurate diagnosis of EGC can reduce unnecessary interventions and benefits treatment planning. Convolutional neural network (CNN) models have recently shown promising performance in analyzing medical images, including endoscopy. This study shows that an automated tool based on the CNN model could improve EGC diagnosis and treatment decision.

Abstract

Gastric cancer (GC) is one of the most newly diagnosed cancers and the fifth leading cause of death globally. Identification of early gastric cancer (EGC) can ensure quick treatment and reduce significant mortality. Therefore, we aimed to conduct a systematic review with a meta-analysis of current literature to evaluate the performance of the CNN model in detecting EGC. We conducted a systematic search in the online databases (e.g., PubMed, Embase, and Web of Science) for all relevant original studies on the subject of CNN in EGC published between 1 January 2010, and 26 March 2021. The Quality Assessment of Diagnostic Accuracy Studies-2 was used to assess the risk of bias. Pooled sensitivity, specificity, positive likelihood ratio, negative likelihood ratio, and diagnostic odds ratio were calculated. Moreover, a summary receiver operating characteristic curve (SROC) was plotted. Of the 171 studies retrieved, 15 studies met inclusion criteria. The application of the CNN model in the diagnosis of EGC achieved a SROC of 0.95, with corresponding sensitivity of 0.89 (0.88–0.89), and specificity of 0.89 (0.89–0.90). Pooled sensitivity and specificity for experts endoscopists were 0.77 (0.76–0.78), and 0.92 (0.91–0.93), respectively. However, the overall SROC for the CNN model and expert endoscopists was 0.95 and 0.90. The findings of this comprehensive study show that CNN model exhibited comparable performance to endoscopists in the diagnosis of EGC using digital endoscopy images. Given its scalability, the CNN model could enhance the performance of endoscopists to correctly stratify EGC patients and reduce work load.

A simulated measurement for COVID-19 pandemic using the effective reproductive number on an empirical portion of population: epidemiological models

COVID-19 as a global pandemic has had an unprecedented impact on the entire world. Projecting the future spread of the virus in relation to its characteristics for a specific suite of countries against a temporal trend can provide public health guidance to governments and organizations. Therefore, this paper presented an epidemiological comparison of the traditional SEIR model with an extended and modified version of the same model by splitting the infected compartment into asymptomatic mild and symptomatic severe. We then exposed our derived layered model into two distinct case studies with variations in mitigation strategies and non-pharmaceutical interventions (NPIs) as a matter of benchmarking and comparison. We focused on exploring the United Arab Emirates (a small yet urban centre (where clear sequential stages NPIs were implemented). Further, we concentrated on extending the models by utilizing the effective reproductive number (R t) estimated against time, a more realistic than the static R 0, to assess the potential impact of NPIs within each case study. Compared to the traditional SEIR model, the results supported the modified model as being more sensitive in terms of peaks of simulated cases and flattening determinations.

Logical Observation Identifiers Names and Codes (LOINC®) Applied to Microbiology: A National Laboratory Mapping Experience in Taiwan

Background and Objective: Logical Observation Identifiers Names and Codes (LOINC) is a universal standard for identifying laboratory tests and clinical observations. It facilitates a smooth information exchange between hospitals, locally and internationally. Although it offers immense benefits for patient care, LOINC coding is complex, resource-intensive, and requires substantial domain expertise. Our objective was to provide training and evaluate the performance of LOINC mapping of 20 pathogens from 53 hospitals participating in the National Notifiable Disease Surveillance System (NNDSS). Methods: Complete mapping codes for 20 pathogens (nine bacteria and 11 viruses) were requested from all participating hospitals to review between January 2014 and December 2016. Participating hospitals mapped those pathogens to LOINC terminology, utilizing the Regenstrief LOINC mapping assistant (RELMA) and reported to the NNDSS, beginning in January 2014. The mapping problems were identified by expert panels that classified frequently asked questionnaires (FAQs) into seven LOINC categories. Finally, proper and meaningful suggestions were provided based on the error pattern in the FAQs. A general meeting was organized if the error pattern proved to be difficult to resolve. If the experts did not conclude the local issue’s error pattern, a request was sent to the LOINC committee for resolution. Results: A total of 53 hospitals participated in our study. Of these, 26 (49.05%) used homegrown and 27 (50.95%) used outsourced LOINC mapping. Hospitals who participated in 2015 had a greater improvement in LOINC mapping than those of 2016 (26.5% vs. 3.9%). Most FAQs were related to notification principles (47%), LOINC system (42%), and LOINC property (26%) in 2014, 2015, and 2016, respectively. Conclusions: The findings of our study show that multiple stage approaches improved LOINC mapping by up to 26.5%.

Improved diagnosis-medication association mining to reduce pseudo-associations

BACKGROUND AND OBJECTIVE

Association rule mining has been adopted to medical fields to discover prescribing patterns or relationships among diseases and/or medications; however, it has generated unreasonable associations among these entities. This study aims to identify the real-world profile of disease-medication (DM) associations using the modified mining algorithm and assess its performance in reducing DM pseudo-associations.

METHODS

We retrieved data from outpatient records between January 2011 and December 2015 in claims databases maintained by the Health and Welfare Data Science Center, Ministry of Health and Welfare, Taiwan. The association rule mining's lift (Q-value) was adopted to quantify DM associations, referred to as Q₁ for the original algorithm and as Q₂ for the modified algorithm. One thousand DM pairs with positive Q₁-values (Q₁⁺) and negative or no Q₂-values (Q₂^- or Q₂^∅) were selected as the validation dataset, in which two pharmacists assessed the DM associations.

RESULTS

A total of 3,120,449 unique DM pairs were identified, of which there were 333,347 Q₁⁺Q₂^- pairs and 429,931 Q₁⁺Q₂^∅ pairs. Q₁⁺Q₂^- rates were relatively high in ATC classes C (29.91%) and R (30.24%). Classes L (69.91%) and V (52.52%) demonstrated remarkably high Q₁⁺Q₂^∅ rates. For the 1000 pairs in the validation, 93.7% of the Q₁⁺Q₂^- or Q₁⁺Q₂^∅ DM pairs were assessed as pseudo-associations. However, classes M (5.3%), H (4.5%), and B (4.1%) showed the highest rates of plausible associations falsely given Q₂^- or Q₂^∅ by the modified algorithm.

CONCLUSIONS

The modified algorithm demonstrated high accuracy to identify pseudo-associations regarded as positive associations by the original algorithm and would potentially be applied to improve secondary databases to facilitate research on real-world prescribing patterns and further enhance drug safety.

Risk of cancer in long-term levothyroxine users: Retrospective population-based study

Levothyroxine is a widely prescribed medication for the treatment of an underactive thyroid. The relationship between levothyroxine use and cancer risk is largely underdetermined. To investigate the magnitude of the possible association between levothyroxine use and cancer risk, this retrospective case‐control study was conducted using Taiwan’s Health and Welfare Data Science Center database. Cases were defined as all patients who were aged ≥20 years and had a first‐time diagnosis for cancer at any site for the period between 2001 and 2011. Multivariable conditional logistic regression models were used to calculate an adjusted odds ratio (AOR) to reduce potential confounding factors. A total of 601 733 cases and 2 406 932 controls were included in the current study. Levothyroxine users showed a 50% higher risk of cancer at any site (AOR: 1.50, 95% CI: 1.46‐1.54; P < .0001) compared with non–users. Significant increased risks were also observed for brain cancer (AOR: 1.90, 95% CI: 1.48‐2.44; P < .0001), skin cancer (AOR: 1.42, 95% CI: 1.17‐1.72; P < .0001), pancreatic cancer (AOR: 1.27, 95% CI: 1.01‐1.60; P = .03), and female breast cancer (AOR: 1.24, 95% CI: 1.15‐1.33; P < .0001). Our study results showed that levothyroxine use was significantly associated with an increased risk of cancer, particularly brain, skin, pancreatic, and female breast cancers. Levothyroxine remains a highly effective therapy for hypothyroidism; therefore, physicians should carefully consider levothyroxine therapy and monitor patients’ condition to avoid negative outcomes. Additional studies are needed to confirm these findings and to evaluate the potential biological mechanisms.

A State-of-the-Art Survey on Artificial Intelligence to Fight COVID-19

Artificial intelligence (AI) has shown immense potential to fight COVID-19 in many ways. This paper focuses primarily on AI's role in managing COVID-19 using digital images, clinical and laboratory data analysis, and a summary of the most recent articles published last year. We surveyed the use of AI for COVID-19 detection, screening, diagnosis, the progression of severity, mortality, drug repurposing, and other tasks. We started with the technical overview of all models used to fight the COVID-19 pandemic and ended with a brief statement of the current state-of-the-art, limitations, and challenges.

Application of Artificial Intelligence for Screening COVID-19 Patients Using Digital Images: Meta-analysis

BACKGROUND

The COVID-19 outbreak has spread rapidly and hospitals are overwhelmed with COVID-19 patients. While analysis of nasal and throat swabs from patients is the main way to detect COVID-19, analyzing chest images could offer an alternative method to hospitals, where health care personnel and testing kits are scarce. Deep learning (DL), in particular, has shown impressive levels of performance when analyzing medical images, including those related to COVID-19 pneumonia.

OBJECTIVE

The goal of this study was to perform a systematic review with a meta-analysis of relevant studies to quantify the performance of DL algorithms in the automatic stratification of COVID-19 patients using chest images.

METHODS

A search strategy for use in PubMed, Scopus, Google Scholar, and Web of Science was developed, where we searched for articles published between January 1 and April 25, 2020. We used the key terms "COVID-19," or "coronavirus," or "SARS-CoV-2," or "novel corona," or "2019-ncov," and "deep learning," or "artificial intelligence," or "automatic detection." Two authors independently extracted data on study characteristics, methods, risk of bias, and outcomes. Any disagreement between them was resolved by consensus.

RESULTS

A total of 16 studies were included in the meta-analysis, which included 5896 chest images from COVID-19 patients. The pooled sensitivity and specificity of the DL models in detecting COVID-19 were 0.95 (95% CI 0.94-0.95) and 0.96 (95% CI 0.96-0.97), respectively, with an area under the receiver operating characteristic curve of 0.98. The positive likelihood, negative likelihood, and diagnostic odds ratio were 19.02 (95% CI 12.83-28.19), 0.06 (95% CI 0.04-0.10), and 368.07 (95% CI 162.30-834.75), respectively. The pooled sensitivity and specificity for distinguishing other types of pneumonia from COVID-19 were 0.93 (95% CI 0.92-0.94) and 0.95 (95% CI 0.94-0.95), respectively. The performance of radiologists in detecting COVID-19 was lower than that of the DL models; however, the performance of junior radiologists was improved when they used DL-based prediction tools.

CONCLUSIONS

Our study findings show that DL models have immense potential in accurately stratifying COVID-19 patients and in correctly differentiating them from patients with other types of pneumonia and normal patients. Implementation of DL-based tools can assist radiologists in correctly and quickly detecting COVID-19 and, consequently, in combating the COVID-19 pandemic.

Application of Artificial Intelligence in COVID-19 Pandemic: Bibliometric Analysis

The application of artificial intelligence (AI) to health has increased, including to COVID-19. This study aimed to provide a clear overview of COVID-19-related AI publication trends using longitudinal bibliometric analysis. A systematic literature search was conducted on the Web of Science for English language peer-reviewed articles related to AI application to COVID-19. A search strategy was developed to collect relevant articles and extracted bibliographic information (e.g., country, research area, sources, and author). VOSviewer (Leiden University) and Bibliometrix (R package) were used to visualize the co-occurrence networks of authors, sources, countries, institutions, global collaborations, citations, co-citations, and keywords. We included 729 research articles on the application of AI to COVID-19 published between 2020 and 2021. PLOS One (33/729, 4.52%), Chaos Solution Fractals (29/729, 3.97%), and Journal of Medical Internet Research (29/729, 3.97%) were the most common journals publishing these articles. The Republic of China (190/729, 26.06%), the USA (173/729, 23.73%), and India (92/729, 12.62%) were the most prolific countries of origin. The Huazhong University of Science and Technology, Wuhan University, and the Chinese Academy of Sciences were the most productive institutions. This is the first study to show a comprehensive picture of the global efforts to address COVID-19 using AI. The findings of this study also provide insights and research directions for academic researchers, policymakers, and healthcare practitioners who wish to collaborate in these domains in the future.

Development of a Web-Based System for Exploring Cancer Risk With Long-term Use of Drugs: Logistic Regression Approach

Background

Existing epidemiological evidence regarding the association between the long-term use of drugs and cancer risk remains controversial.

Objective

We aimed to have a comprehensive view of the cancer risk of the long-term use of drugs.

Methods

A nationwide population-based, nested, case-control study was conducted within the National Health Insurance Research Database sample cohort of 1999 to 2013 in Taiwan. We identified cases in adults aged 20 years and older who were receiving treatment for at least two months before the index date. We randomly selected control patients from the patients without a cancer diagnosis during the 15 years (1999-2013) of the study period. Case and control patients were matched 1:4 based on age, sex, and visit date. Conditional logistic regression was used to estimate the association between drug exposure and cancer risk by adjusting potential confounders such as drugs and comorbidities.

Results

There were 79,245 cancer cases and 316,980 matched controls included in this study. Of the 45,368 associations, there were 2419, 1302, 662, and 366 associations found statistically significant at a level of P<.05, P<.01, P<.001, and P<.0001, respectively. Benzodiazepine derivatives were associated with an increased risk of brain cancer (adjusted odds ratio [AOR] 1.379, 95% CI 1.138-1.670; P=.001). Statins were associated with a reduced risk of liver cancer (AOR 0.470, 95% CI 0.426-0.517; P<.0001) and gastric cancer (AOR 0.781, 95% CI 0.678-0.900; P<.001). Our web-based system, which collected comprehensive data of associations, contained 2 domains: (1) the drug and cancer association page and (2) the overview page.

Conclusions

Our web-based system provides an overview of comprehensive quantified data of drug-cancer associations. With all the quantified data visualized, the system is expected to facilitate further research on cancer risk and prevention, potentially serving as a stepping-stone to consulting and exploring associations between the long-term use of drugs and cancer risk.

Obesity and Mortality Among Patients Diagnosed With COVID-19: A Systematic Review and Meta-Analysis

Coronavirus disease 2019 (COVID-19) has already raised serious concern globally as the number of confirmed or suspected cases have increased rapidly. Epidemiological studies reported that obesity is associated with a higher rate of mortality in patients with COVID-19. Yet, to our knowledge, there is no comprehensive systematic review and meta-analysis to assess the effects of obesity and mortality among patients with COVID-19. We, therefore, aimed to evaluate the effect of obesity, associated comorbidities, and other factors on the risk of death due to COVID-19. We did a systematic search on PubMed, EMBASE, Google Scholar, Web of Science, and Scopus between January 1, 2020, and August 30, 2020. We followed Cochrane Guidelines to find relevant articles, and two reviewers extracted data from retrieved articles. Disagreement during those stages was resolved by discussion with the main investigator. The random-effects model was used to calculate effect sizes. We included 17 articles with a total of 543,399 patients. Obesity was significantly associated with an increased risk of mortality among patients with COVID-19 (RRadjust: 1.42 (95%CI: 1.24-1.63, p < 0.001). The pooled risk ratio for class I, class II, and class III obesity were 1.27 (95%CI: 1.05-1.54, p = 0.01), 1.56 (95%CI: 1.11-2.19, p < 0.01), and 1.92 (95%CI: 1.50-2.47, p < 0.001), respectively). In subgroup analysis, the pooled risk ratio for the patients with stroke, CPOD, CKD, and diabetes were 1.80 (95%CI: 0.89-3.64, p = 0.10), 1.57 (95%CI: 1.57-1.91, p < 0.001), 1.34 (95%CI: 1.18-1.52, p < 0.001), and 1.19 (1.07-1.32, p = 0.001), respectively. However, patients with obesity who were more than 65 years had a higher risk of mortality (RR: 2.54; 95%CI: 1.62-3.67, p < 0.001). Our study showed that obesity was associated with an increased risk of death from COVID-19, particularly in patients aged more than 65 years. Physicians should aware of these risk factors when dealing with patients with COVID-19 and take early treatment intervention to reduce the mortality of COVID-19 patients.

Association between Anemia and Risk of Parkinson Disease

METHODS

We systematically searched articles on electronic databases such as PubMed, Embase, Scopus, and Google Scholar between January 1, 2000 and July 30, 2020. Articles were independently evaluated by two authors. We included observational studies (case-control and cohort) and calculated the risk ratios (RRs) for associated with anemia and PD. Heterogeneity among the studies was assessed using the Q and I ² statistic. We utilized the random-effect model to calculate the overall RR with 95% CI.

RESULTS

A total of 342 articles were identified in the initial searches, and 7 full-text articles were evaluated for eligibility. Three articles were further excluded for prespecified reasons including insufficient data and duplications, and 4 articles were included in our systematic review and meta-analysis. A random effect model meta-analysis of all 4 studies showed no increased risk of PD in patients with anemia (N = 4, RR_adjusted = 1.17 (95% CI: 0.94-1.45, p = 0.15). However, heterogeneity among the studies was significant (I ² = 92.60, p = <0.0001). The pooled relative risk of PD in female patients with anemia was higher (N = 3, RR_adjusted = 1.14 (95% CI: 0.83-1.57, p = 0.40) as compared to male patients with anemia (N = 3, RR_adjusted = 1.09 (95% CI: 0.83-1.42, p = 0.51).

CONCLUSION

This is the first meta-analysis that shows that anemia is associated with higher risk of PD when compared with patients without anemia. However, more studies are warranted to evaluate the risk of PD among patients with anemia.

Clinical Characteristics and Neonatal Outcomes of Pregnant Patients With COVID-19: A Systematic Review

Background and Objective: Coronavirus disease 2019 (COVID-19) characterized by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has created serious concerns about its potential adverse effects. There are limited data on clinical, radiological, and neonatal outcomes of pregnant women with COVID-19 pneumonia. This study aimed to assess clinical manifestations and neonatal outcomes of pregnant women with COVID-19.

Methods: We conducted a systematic article search of PubMed, EMBASE, Scopus, Google Scholar, and Web of Science for studies that discussed pregnant patients with confirmed COVID-19 between January 1, 2020, and April 20, 2020, with no restriction on language. Articles were independently evaluated by two expert authors. We included all retrospective studies that reported the clinical features and outcomes of pregnant patients with COVID-19.

Results: Forty-seven articles were assessed for eligibility; 13 articles met the inclusion criteria for the systematic review. Data is reported for 235 pregnant women with COVID-19. The age range of patients was 25–40 years, and the gestational age ranged from 8 to 40 weeks plus 6 days. Clinical characteristics were fever [138/235 (58.72%)], cough [111/235 (47.23%)], and sore throat [21/235 (8.93%)]. One hundred fifty six out of 235 (66.38%) pregnant women had cesarean section, and 79 (33.62%) had a vaginal delivery. All the patients showed lung abnormalities in CT scan images, and none of the patients died. Neutrophil cell count, C-reactive protein (CRP) concentration, ALT, and AST were increased but lymphocyte count and albumin levels were decreased. Amniotic fluid, neonatal throat swab, and breastmilk samples were taken to test for SARS-CoV-2 but all found negativ results. Recent published evidence showed the possibility of vertical transmission up to 30%, and neonatal death up to 2.5%. Pre-eclampsia, fetal distress, PROM, pre-mature delivery were the major complications of pregnant women with COVID-19.

Conclusions: Our study findings show that the clinical, laboratory and radiological characteristics of pregnant women with COVID-19 were similar to those of the general populations. The possibility of vertical transmission cannot be ignored but C-section should not be routinely recommended anymore according to latest evidences and, in any case, decisions should be taken after proper discussion with the family. Future studies are needed to confirm or refute these findings with a larger number of sample sizes and a long-term follow-up period.

Machine Learning Approach to Reduce Alert Fatigue Using a Disease Medication-Related Clinical Decision Support System: Model Development and Validation

Background

Computerized physician order entry (CPOE) systems are incorporated into clinical decision support systems (CDSSs) to reduce medication errors and improve patient safety. Automatic alerts generated from CDSSs can directly assist physicians in making useful clinical decisions and can help shape prescribing behavior. Multiple studies reported that approximately 90%-96% of alerts are overridden by physicians, which raises questions about the effectiveness of CDSSs. There is intense interest in developing sophisticated methods to combat alert fatigue, but there is no consensus on the optimal approaches so far.

Objective

Our objective was to develop machine learning prediction models to predict physicians’ responses in order to reduce alert fatigue from disease medication–related CDSSs.

Methods

We collected data from a disease medication–related CDSS from a university teaching hospital in Taiwan. We considered prescriptions that triggered alerts in the CDSS between August 2018 and May 2019. Machine learning models, such as artificial neural network (ANN), random forest (RF), naïve Bayes (NB), gradient boosting (GB), and support vector machine (SVM), were used to develop prediction models. The data were randomly split into training (80%) and testing (20%) datasets.

Results

A total of 6453 prescriptions were used in our model. The ANN machine learning prediction model demonstrated excellent discrimination (area under the receiver operating characteristic curve [AUROC] 0.94; accuracy 0.85), whereas the RF, NB, GB, and SVM models had AUROCs of 0.93, 0.91, 0.91, and 0.80, respectively. The sensitivity and specificity of the ANN model were 0.87 and 0.83, respectively.

Conclusions

In this study, ANN showed substantially better performance in predicting individual physician responses to an alert from a disease medication–related CDSS, as compared to the other models. To our knowledge, this is the first study to use machine learning models to predict physician responses to alerts; furthermore, it can help to develop sophisticated CDSSs in real-world clinical settings.

Development of an Artificial Intelligence-Based Automated Recommendation System for Clinical Laboratory Tests: Retrospective Analysis of the National Health Insurance Database

BACKGROUND

Laboratory tests are considered an essential part of patient safety as patients' screening, diagnosis, and follow-up are solely based on laboratory tests. Diagnosis of patients could be wrong, missed, or delayed if laboratory tests are performed erroneously. However, recognizing the value of correct laboratory test ordering remains underestimated by policymakers and clinicians. Nowadays, artificial intelligence methods such as machine learning and deep learning (DL) have been extensively used as powerful tools for pattern recognition in large data sets. Therefore, developing an automated laboratory test recommendation tool using available data from electronic health records (EHRs) could support current clinical practice.

OBJECTIVE

The objective of this study was to develop an artificial intelligence-based automated model that can provide laboratory tests recommendation based on simple variables available in EHRs.

METHODS

A retrospective analysis of the National Health Insurance database between January 1, 2013, and December 31, 2013, was performed. We reviewed the record of all patients who visited the cardiology department at least once and were prescribed laboratory tests. The data set was split into training and testing sets (80:20) to develop the DL model. In the internal validation, 25% of data were randomly selected from the training set to evaluate the performance of this model.

RESULTS

We used the area under the receiver operating characteristic curve, precision, recall, and hamming loss as comparative measures. A total of 129,938 prescriptions were used in our model. The DL-based automated recommendation system for laboratory tests achieved a significantly higher area under the receiver operating characteristic curve (AUROCmacro and AUROCmicro of 0.76 and 0.87, respectively). Using a low cutoff, the model identified appropriate laboratory tests with 99% sensitivity.

CONCLUSIONS

The developed artificial intelligence model based on DL exhibited good discriminative capability for predicting laboratory tests using routinely collected EHR data. Utilization of DL approaches can facilitate optimal laboratory test selection for patients, which may in turn improve patient safety. However, future study is recommended to assess the cost-effectiveness for implementing this model in real-world clinical settings.

Deep Learning for Accurate Diagnosis of Glaucomatous Optic Neuropathy Using Digital Fundus Image: A Meta-Analysis

We conducted a study to evaluate the algorithms based on deep learning to automatically diagnosis of GON from digital fundus images. A systematic articles search was conducted in PubMed, EMBASE, Google Scholar for the study that investigated the performance of deep learning algorithms for the detection of GON. A total of eight studies were included in this study, of which 5 studies were used to conduct our meta-analysis. The pooled AUROC for detecting GON was 0.98. However, the sensitivity and specificity of deep learning to detect GON were 0.90 (95% CI: 0.90-0.91), and 0.94 (95%CI: 0.93-0.94), respectively.

Use of Mobile Phone App Interventions to Promote Weight Loss: Meta-Analysis

BACKGROUND

Obesity and lack of physical activity are major health risk factors for many life-threatening diseases, such as cardiovascular diseases, type 2 diabetes, and cancer. The use of mobile app interventions to promote weight loss and boost physical activity among children and adults is fascinating owing to the demand for cutting-edge and more efficient interventions. Previously published studies have examined different types of technology-based interventions and their impact on weight loss and increase in physical activity, but evidence regarding the impact of only a mobile phone app on weight loss and increase in physical activity is still lacking.

OBJECTIVE

The main objective of this study was to assess the efficacy of a mobile phone app intervention for reducing body weight and increasing physical activity among children and adults.

METHODS

PubMed, Google Scholar, Scopus, EMBASE, and the Web of Science electronic databases were searched for studies published between January 1, 2000, and April 30, 2019, without language restrictions. Two experts independently screened all the titles and abstracts to find the most appropriate studies. To be included, studies had to be either a randomized controlled trial or a case-control study that assessed a mobile phone app intervention with body weight loss and physical activity outcomes. The Cochrane Collaboration Risk of Bias tool was used to examine the risk of publication bias.

RESULTS

A total of 12 studies involving a mobile phone app intervention were included in this meta-analysis. Compared with the control group, the use of a mobile phone app was associated with significant changes in body weight (-1.07 kg, 95% CI -1.92 to -0.21, P=.01) and body mass index (-0.45 kg/m2, 95% CI -0.78 to -0.12, P=.008). Moreover, a nonsignificant increase in physical activity was observed (0.17, 95% CI -2.21 to 2.55, P=.88).

CONCLUSIONS

The findings of this study demonstrate the promising and emerging efficacy of using mobile phone app interventions for weight loss. Future studies are needed to explore the long-term efficacy of mobile app interventions in larger samples.

Use of Mobile Phone App Interventions to Promote Weight Loss: Meta-Analysis

Background

Obesity and lack of physical activity are major health risk factors for many life-threatening diseases, such as cardiovascular diseases, type 2 diabetes, and cancer. The use of mobile app interventions to promote weight loss and boost physical activity among children and adults is fascinating owing to the demand for cutting-edge and more efficient interventions. Previously published studies have examined different types of technology-based interventions and their impact on weight loss and increase in physical activity, but evidence regarding the impact of only a mobile phone app on weight loss and increase in physical activity is still lacking.

Objective

The main objective of this study was to assess the efficacy of a mobile phone app intervention for reducing body weight and increasing physical activity among children and adults.

Methods

PubMed, Google Scholar, Scopus, EMBASE, and the Web of Science electronic databases were searched for studies published between January 1, 2000, and April 30, 2019, without language restrictions. Two experts independently screened all the titles and abstracts to find the most appropriate studies. To be included, studies had to be either a randomized controlled trial or a case-control study that assessed a mobile phone app intervention with body weight loss and physical activity outcomes. The Cochrane Collaboration Risk of Bias tool was used to examine the risk of publication bias.

Results

A total of 12 studies involving a mobile phone app intervention were included in this meta-analysis. Compared with the control group, the use of a mobile phone app was associated with significant changes in body weight (−1.07 kg, 95% CI −1.92 to −0.21, P=.01) and body mass index (−0.45 kg/m2, 95% CI −0.78 to −0.12, P=.008). Moreover, a nonsignificant increase in physical activity was observed (0.17, 95% CI −2.21 to 2.55, P=.88).

Conclusions

The findings of this study demonstrate the promising and emerging efficacy of using mobile phone app interventions for weight loss. Future studies are needed to explore the long-term efficacy of mobile app interventions in larger samples.

Appropriateness of Overridden Alerts in Computerized Physician Order Entry: Systematic Review

BACKGROUND

The clinical decision support system (CDSS) has become an indispensable tool for reducing medication errors and adverse drug events. However, numerous studies have reported that CDSS alerts are often overridden. The increase in override rates has raised questions about the appropriateness of CDSS application along with concerns about patient safety and quality of care.

OBJECTIVE

The aim of this study was to conduct a systematic review to examine the override rate, the reasons for the alert override at the time of prescribing, and evaluate the appropriateness of overrides.

METHODS

We searched electronic databases, including Google Scholar, PubMed, Embase, Scopus, and Web of Science, without language restrictions between January 1, 2000 and March 31, 2019. Two authors independently extracted data and crosschecked the extraction to avoid errors. The quality of the included studies was examined following Cochrane guidelines.

RESULTS

We included 23 articles in our systematic review. The range of average override alerts was 46.2%-96.2%. An average of 29.4%-100% of the overrides alerts were classified as appropriate, and the rate of appropriateness varied according to the alert type (drug-allergy interaction 63.4%-100%, drug-drug interaction 0%-95%, dose 43.9%-88.8%, geriatric 14.3%-57%, renal 27%-87.5%). The interrater reliability for the assessment of override alerts appropriateness was excellent (kappa=0.79-0.97). The most common reasons given for the override were "will monitor" and "patients have tolerated before."

CONCLUSIONS

The findings of our study show that alert override rates are high, and certain categories of overrides such as drug-drug interaction, renal, and geriatric were classified as inappropriate. Nevertheless, large proportions of drug duplication, drug-allergy, and formulary alerts were appropriate, suggesting that these groups of alerts can be primary targets to revise and update the system for reducing alert fatigue. Future efforts should also focus on optimizing alert types, providing clear information, and explaining the rationale of the alert so that essential alerts are not inappropriately overridden.

Deep learning algorithms for detection of diabetic retinopathy in retinal fundus photographs: A systematic review and meta-analysis

BACKGROUND

Diabetic retinopathy (DR) is one of the leading causes of blindness globally. Earlier detection and timely treatment of DR are desirable to reduce the incidence and progression of vision loss. Currently, deep learning (DL) approaches have offered better performance in detecting DR from retinal fundus images. We, therefore, performed a systematic review with a meta-analysis of relevant studies to quantify the performance of DL algorithms for detecting DR.

METHODS

A systematic literature search on EMBASE, PubMed, Google Scholar, Scopus was performed between January 1, 2000, and March 31, 2019. The search strategy was based on the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) reporting guidelines, and DL-based study design was mandatory for articles inclusion. Two independent authors screened abstracts and titles against inclusion and exclusion criteria. Data were extracted by two authors independently using a standard form and the Quality Assessment of Diagnostic Accuracy Studies (QUADAS-2) tool was used for the risk of bias and applicability assessment.

RESULTS

Twenty-three studies were included in the systematic review; 20 studies met inclusion criteria for the meta-analysis. The pooled area under the receiving operating curve (AUROC) of DR was 0.97 (95%CI: 0.95-0.98), sensitivity was 0.83 (95%CI: 0.83-0.83), and specificity was 0.92 (95%CI: 0.92-0.92). The positive- and negative-likelihood ratio were 14.11 (95%CI: 9.91-20.07), and 0.10 (95%CI: 0.07-0.16), respectively. Moreover, the diagnostic odds ratio for DL models was 136.83 (95%CI: 79.03-236.93). All the studies provided a DR-grading scale, a human grader (e.g. trained caregivers, ophthalmologists) as a reference standard.

CONCLUSION

The findings of our study showed that DL algorithms had high sensitivity and specificity for detecting referable DR from retinal fundus photographs. Applying a DL-based automated tool of assessing DR from color fundus images could provide an alternative solution to reduce misdiagnosis and improve workflow. A DL-based automated tool offers substantial benefits to reduce screening costs, accessibility to healthcare and ameliorate earlier treatments.

Deep Learning Approach for the Development of a Novel Predictive Model for Prostate Cancer

We developed a deep learning approach for accurate prediction of PCA patients one year earlier with minimal features from electronic health records. The area under the receiver operating curve for prediction of PCA was 0.94. Moreover, the sensitivity and specificity of CNN were 0.87 and 0.88, respectively.

Artificial Intelligence in Ophthalmology: A Meta-Analysis of Deep Learning Models for Retinal Vessels Segmentation

BACKGROUND AND OBJECTIVE

Accurate retinal vessel segmentation is often considered to be a reliable biomarker of diagnosis and screening of various diseases, including cardiovascular diseases, diabetic, and ophthalmologic diseases. Recently, deep learning (DL) algorithms have demonstrated high performance in segmenting retinal images that may enable fast and lifesaving diagnoses. To our knowledge, there is no systematic review of the current work in this research area. Therefore, we performed a systematic review with a meta-analysis of relevant studies to quantify the performance of the DL algorithms in retinal vessel segmentation.

METHODS

A systematic search on EMBASE, PubMed, Google Scholar, Scopus, and Web of Science was conducted for studies that were published between 1 January 2000 and 15 January 2020. We followed the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) procedure. The DL-based study design was mandatory for a study's inclusion. Two authors independently screened all titles and abstracts against predefined inclusion and exclusion criteria. We used the Quality Assessment of Diagnostic Accuracy Studies (QUADAS-2) tool for assessing the risk of bias and applicability.

RESULTS

Thirty-one studies were included in the systematic review; however, only 23 studies met the inclusion criteria for the meta-analysis. DL showed high performance for four publicly available databases, achieving an average area under the ROC of 0.96, 0.97, 0.96, and 0.94 on the DRIVE, STARE, CHASE_DB1, and HRF databases, respectively. The pooled sensitivity for the DRIVE, STARE, CHASE_DB1, and HRF databases was 0.77, 0.79, 0.78, and 0.81, respectively. Moreover, the pooled specificity of the DRIVE, STARE, CHASE_DB1, and HRF databases was 0.97, 0.97, 0.97, and 0.92, respectively.

CONCLUSION

The findings of our study showed the DL algorithms had high sensitivity and specificity for segmenting the retinal vessels from digital fundus images. The future role of DL algorithms in retinal vessel segmentation is promising, especially for those countries with limited access to healthcare. More compressive studies and global efforts are mandatory for evaluating the cost-effectiveness of DL-based tools for retinal disease screening worldwide.

Association Between Atrial Fibrillation and Dementia: A Meta-Analysis

Background: A potential evidence from previous epidemiological studies remains conflicting findings regarding the association between atrial fibrillation (AF) and dementia risk. We, therefore, carried out a meta-analysis of relevant studies to investigate the magnitude of the association between AF and dementia risk.

Methods: We performed a systematic literature search of PubMed, EMBASE, and Google Scholar for potential studies between January 1, 1990, and December 31, 2018, with no restriction on the publication language. All potential studies were independently assessed by two reviewers. We only included observational studies that calculated the odds ratio (OR)/hazards ratio (HR) for dementia associated with atrial fibrillation. We first assessed the heterogeneity among study-specific HRs using the Q statistic and I² statistic. We then used the random-effects model to obtain the overall HR and its 95% CI for all studies. We also tested and corrected for publication bias by funnel plot–based methods. The quality of each study was assessed with the Newcastle Ottawa Scale.

Results: A total of 16 studies with 2,415,356 individuals, and approximately 200,653 cases of incidence dementia were included in this study. Patients with AF had a greater risk of incidence dementia than those without AF (random-effect hazard ratio HR: 1.36, 95% CI: 1.23–1.51, p < 0.0001; I² = 83.58). Funnel plot and Egger test did not reveal significant publication bias. However, limitations of the study included high heterogeneity and varying degrees of confounder adjustment across individual studies.

Conclusion: This study serves as added evidence supporting the hypothesis that AF is associated with an increased risk of dementia. More studies are needed to establish whether optimal treatment of AF can reduce or mitigate the risk of dementia.

Artificial Intelligence in Diabetic Retinopathy: Insights from a Meta-Analysis of Deep Learning

The demand for AI to improve patients outcome has been increased; we, therefore, aim to establish the diagnostic values of AI in diabetic retinopathy by pooling the published studies of deep learning on this subject. A total of eight studies included which evaluated deep learning in a total of 706,922 retinal images. The overall pooled area under receiver operating curve (AUROC) was 98.93% (95%CI:98.37%-99.49%). However, the overall pooled sensitivity and specificity for detecting referable diabetic retinopathy (RDR) was 74% (95% CI: 73%-74%), and 95% (95% CI: 95%-95%). The findings of this study show that deep learning had high sensitivity and specificity for identifying diabetic retinopathy.

An artificial intelligence approach to early predict non-ST-elevation myocardial infarction patients with chest pain

BACKGROUND AND AIMS

Hospital admission rate for the patients with chest pain has already been increased worldwide but no existing risk score has been designed to stratify non-ST-elevation myocardial infarction (NSTEMI) from non-cardiogenic chest pain. Clinical diagnosis of chest pain in the emergency department is always highly subjective and variable. We, therefore, aimed to develop an artificial intelligence approach to predict stable NSTEMI that would give valuable insight to reduce misdiagnosis in the real clinical setting.

METHODS

A standard protocol was developed to collect data from chest pain patients who had visited the emergency department between December 2016 and February 2017. All the chest pain patients with aged <20 years were primarily included in this study. However, STEMI, previous history of ACS, and out-of-hospital cardiac arrest were excluded from our study. An artificial neural network (ANN) model was then developed to predict NSTEMI patients. The accuracy, sensitivity, specificity, and receiver operating characteristic curve was used to measure the performance of this model.

RESULTS

A total of 268 chest pain patients were included in this study; of those, 47 (17.5%) was stable NSTEMI, and 221 (82.5%) was unstable angina patients. Serval risk factors such as cardiac risk factor, systolic blood pressure, hemoglobin, corrected QT interval (QTc), PR interval, glutamic-oxaloacetic transaminase, glutamic pyruvic transaminase and troponin were independently associated with stable NSTEMI. The area under the receiver operating characteristic (AUROC) and accuracy of ANN were 98.4, and 92.86. Additionally, the sensitivity, specificity, positive predictive value, and negative predictive value of the ANN model was 90.91, 93.33, 76.92, and 97.67 respectively.

CONCLUSION

Our prediction model showed a higher accuracy to predict NSTEMI patients. This model has a potential application in disease detection, monitoring, and prognosis of chest pain at risk of AMI.

Prediction of fatty liver disease using machine learning algorithms

BACKGROUND AND OBJECTIVE

Fatty liver disease (FLD) is a common clinical complication; it is associated with high morbidity and mortality. However, an early prediction of FLD patients provides an opportunity to make an appropriate strategy for prevention, early diagnosis and treatment. We aimed to develop a machine learning model to predict FLD that could assist physicians in classifying high-risk patients and make a novel diagnosis, prevent and manage FLD.

METHODS

We included all patients who had an initial fatty liver screening at the New Taipei City Hospital between 1st and 31st December 2009. Classification models such as random forest (RF), Naïve Bayes (NB), artificial neural networks (ANN), and logistic regression (LR) were developed to predict FLD. The area under the receiver operating characteristic curve (ROC) was used to evaluate performances among the four models.

RESULTS

A total of 577 patients were included in this study; of those 377 patients had fatty liver. The area under the receiver operating characteristic (AUROC) of RF, NB, ANN, and LR with 10 fold-cross validation was 0.925, 0.888, 0.895, and 0.854 respectively. Additionally, The accuracy of RF, NB, ANN, and LR 87.48, 82.65, 81.85, and 76.96%.

CONCLUSION

In this study, we developed and compared the four classification models to predict fatty liver disease accurately. However, the random forest model showed higher performance than other classification models. Implementation of a random forest model in the clinical setting could help physicians to stratify fatty liver patients for primary prevention, surveillance, early treatment, and management.

Increase Risk of Multiple Sclerosis in Patients with Psoriasis Disease: An Evidence of Observational Studies

BACKGROUND

Psoriasis, a common chronic inflammatory disease, increases the risk of developing multiple sclerosis (MS), but evidence for this outcome is still unclear. However, we performed a meta-analysis of relevant studies to quantify the magnitude of the association between psoriasis and MS. It will help to assess the current state of knowledge, fill the gaps in our existing concern, and make a recommendation for future research.

METHODS

PubMed, EMBASE, and the bibliographies of articles were searched for studies published between January 1, 1990, and November 1, 2017, which reported on the association between psoriasis and MS. Articles were included if they (1) were published in English, (2) reported patients with psoriasis, and the outcome of interest was MS, (3) provided OR/RR/HR with 95% CI or sufficient information to calculate the 95% CI, and (4) if ≥50 patients. All abstracts, full-text articles, and sources were reviewed, with duplicate data excluded. Summary relative risk (ORs) with 95% CI was pooled using a random-effects model. Subgroup and sensitivity analyses were also conducted.

RESULTS

We selected 11 articles out of 785 unique abstracts for full-text review using our predetermined selection criteria, and 9 out of these 11 studies met all of our inclusion criteria. The overall pooled increased of developing MS in patients with psoriasis was RR 1.607 (95% CI 1.322-1.953, p < 0.0001) with low heterogeneity (I2 = 37.41%, Q = 12.782, τ2 = 0.027) for the random effect model. In the subgroup analysis, the MS risk in the patient with psoriasis was also significantly higher in the 6 studies from Europe RR 1.57 (95% CI 1.26-1.94, p < 0.001) with moderate heterogeneity (I2 = 50.66%, Q = 10.13, τ2 = 0.03) for the random effect model.

CONCLUSION

Our results showed that psoriasis is significantly associated with an increased risk of developing MS. Physicians should carefully be observed symptoms and empower their patients to improve existing knowledge and quality of life. Further studies are warranted to establish the mechanisms underlying this relationship.

Proton pump inhibitors and risk of hip fracture: a meta-analysis of observational studies

We performed a meta-analysis of relevant studies to quantify the magnitude of the association between proton pump inhibitors (PPIs) and risk of hip fracture. Patients with PPIs had a greater risk of hip fracture than those without PPI therapy (RR 1.20, 95% CI 1.14-1.28, p < 0.0001). These results could be taken into consideration with caution, and patients should also be concerned about the inappropriate use of PPIs.

INTRODUCTION

Proton pump inhibitors (PPIs) are generally considered as first-line medicine with great safety profile, commonly prescribed for gastroesophageal reflux disease (GERD) and peptic ulcer disease. However, several epidemiological studies documented that long-term use of PPIs may be associated with an increased risk of hip fracture. Although, the optimal magnitude of the hip fracture risk is still undetermined. We, therefore, performed a meta-analysis of relevant studies to quantify the magnitude of the association between PPIs and risk of hip fracture.

METHODS

We collected relevant articles using MEDLINE, EMBASE, Google Scholar, and Web of Science from January 1, 1990, to March 31, 2018. We included only the large (n ≥ 500) observational studies with a follow-up duration of at least one year in which the hip fracture patients were identified by a standard procedure. Two of the authors extracted data from each included study independently according to a standardized protocol.

RESULTS

A total of 24 observational studies with 2,103,800 participants (319,568 hip fracture patients) met all the eligibility criteria. Patients with PPIs had a greater risk of hip fracture than those without PPI therapy (RR 1.20, 95% CI 1.14-1.28, p < 0.0001). An increased association was also observed in both low and medium doses of PPI taken and hip fracture risk (RR 1.17, 95% CI 1.05-1.29, p = 0.002; RR 1.28, 95% CI 1.14-1.44, p < 0.0001), but it appeared to be even greater among the patients with higher dose (RR 1.30, 95% CI 1.20-1.40, p < 0.0001). Moreover, the overall pooled risk ratios were 1.20 (95% CI 1.15-1.25, p < 0.0001) and 1.24 (95% CI 1.10-1.40, p < 0.0001) for the patients with short- and long-term PPI therapy, respectively, compared with PPI non-users.

CONCLUSION

Our results suggest that PPI use is significantly associated with an increased risk of hip fracture development, which is not observed in H2RA exposure. Physicians should, therefore, exercise caution when considering a long-term PPI treatment to their patients who already have an elevated risk of hip fracture. In addition, patients should be concerned about the inappropriate use of PPIs; if necessary, then, they should continue to receive it with a clear indication.

Non-steroidal anti-inflammatory drugs and risk of Parkinson's disease in the elderly population: a meta-analysis

PURPOSE

Several studies have explored the impact of non-steroidal anti-inflammatory drugs (NSAIDs) and the risk of Parkinson disease (PD). However, the extent to which NSAIDs may increase or decrease the risk of PD remains unresolved. We, therefore, performed a meta-analysis of relevant studies to quantify the magnitude of the association between NSAID use and PD risk in the elderly population.

METHODS

The electronic databases such as PubMed, EMBASE, Scopus, Google Scholar, and Web of Science were used to search the relevant articles published between January 1990 and December 2017. Large (n ≥ 1000) observational design studies with a follow-up at least 1 year were considered. Two authors independently extracted information from the included studies. Random effect model was used to calculate risk ratios (RRs) with 95% confidence interval (Cl).

RESULTS

A total of 17 studies with 2,498,258 participants and nearly 14,713 PD patients were included in the final analysis. The overall pooled RR of PD was 0.95 (95%CI 0.860-1.048) with significant heterogeneity (I² = 63.093, Q = 43.352, p < 0.0001). In the subgroup analysis, the overall pooled RR of PD was 0.90 (95%CI 0.738-1.109), 0.96 (95%CI 0.882-1.055), and 0.99 (95%CI 0.841-0.982) from the studies of North America, Europe, and Asia. Additionally, long-term use, study design, individual NSAID use, and risk of PD were also evaluated.

CONCLUSION

Despite the neuroprotective potential of NSAIDs demonstrated in some experimental studies, our findings suggest that there is no association between NSAIDs and the risk of Parkinson disease at the population level. Until further evidence is established, clinicians need to be vigilant ensuring that the use of NSAIDs remains restricted to their approved anti-inflammatory and analgesic effect.

Recent Advancement of Clinical Information Systems: Opportunities and Challenges

Objectives:  Clinical information systems (CISs) have generated opportunities for meaningful improvements both in patient care and workflow but there is still a long way to perfection. Healthcare providers are still facing challenges of data exchange, management, and integration due to lack of functionality among these systems. Our objective here is to systematically review, synthesize, and summarize the literature that describes the current stage of clinical information systems, so as to assess the current state of knowledge, and identify benefits and challenges.

Methods:  PubMed, EMBASE, and the bibliographies of articles were searched for studies published until September 1, 2017, which reported on significant advancement of clinical information systems, as well as problems and opportunities in this field. Studies providing the most detailed information were included and the others were kept only as references.

Results:  We selected 23 papers out of 1,026 unique abstracts for full-text review using our selection criteria, and 20 out of these 23 studies met all of our inclusion criteria. We focused on three major areas: 1) Ambulatory and inpatients clinical information systems; 2) Specialty information systems; and 3) Ancillary information systems. As CIS can support evidence-based practices that, in turn, improve patient's safety, quality and efficacy of care, advancement, acceptability, and adaptability of CIS have increased worldwide. Although, the demand for CIS functionality is rising fast, current CISs still have data integration challenges and lack of functionality to exchange patient information from all or some parts of the healthcare system. These limitations can be attributed to technical, human, and organizational factors

Conclusion:  Clinical information systems provide tremendous opportunities to reduce clinical errors such as medication errors and diagnostic errors and to support healthcare professionals by offering up-to-date patient information. They promise to improve workflow and efficiency of care, thus boosting the overall quality of healthcare.

Risk of Hemorrhagic Stroke in Patients Exposed to Nonsteroidal Anti-Inflammatory Drugs: A Meta-Analysis of Observational Studies

BACKGROUND AND AIM

Nonsteroidal anti-inflammatory drugs (NSAIDs) are one of the most common pain relief medications, but the risk of hemorrhagic stroke in patients taking these medications is unclear. In this study, our aim was to systematically review, synthesize, and critique the epidemiological studies that evaluate the association between NSAIDs and hemorrhagic stroke risk. We therefore assessed the current state of knowledge, filling the gaps in our existing concern, and make a recommendation for future research.

METHODS

We searched for articles in PubMed, EMBASE, Scopus, and Web of Science between January 1, 1990, and July 30, 2017, which reported on the association between the use of NSAIDs and hemorrhagic stroke. The search was limited to studies published in English. The quality of the included studies was assessed in accordance with the Cochrane guidelines and the Newcastle-Ottawa criteria. Summary risk ratios (RRs) with 95% CI were pooled using a random-effects model. Subgroup and sensitivity analyses were also conducted.

RESULTS

We selected 15 out of the 785 unique abstracts for full-text review using our selection criteria, and 13 out of these 15 studies met all of our inclusion criteria. The overall pooled RR of hemorrhagic stroke was 1.332 (95% CI 1.105-1.605, p = 0.003) for the random effect model. In the subgroup analysis, a significant risk was observed among meloxicam, diclofenac, and indomethacin users (RR 1.48; 95% CI 1.149-1.912, RR 1.392; 95% CI 1.107-1.751, and RR 1.363; 95% CI 1.088-1.706). In addition, a greater risk was found in studies from Asia (RR 1.490, 95% CI 1.226-1.811) followed by Europe (RR 1.393, 95% CI 1.104-1.757) and Australia (RR 1.361, 95% CI 0.755-2.452).

CONCLUSION

Our results indicated that the use of NSAIDs is significantly associated with a higher risk of developing hemorrhagic stroke. These results should be interpreted with caution because they may be confounded owing to the observational design of the individual studies. Nevertheless, we recommend that NSAIDs should be used judiciously, and their efficacy and safety should be monitored proactively.

Applications of Machine Learning in Fatty Live Disease Prediction

: Fatty liver disease (FLD) is considered the most prevalent form of chronic liver disease worldwide. The prediction of fatty liver disease is an important factor for effective treatment and reduce serious health consequences. We, therefore construct a prediction model based on machine learning algorithms. A dataset was developed with ten attributes that included 994 liver patients in which 533 patients were females and others were male. Random Forest (RF), Support Vector Machine (SVM), Artificial Neural Network (ANN), and Logistic Regression (RF) data mining technique with 10-fold cross-validation was used in the proposed model for the prediction of fatty liver disease. The performances were evaluated with accuracy, sensitivity, specificity, positive predictive value, and negative predictive value. In this proposed model, logistic regression technique provides a better result (Accuracy 76.30%, sensitivity 74.10%, and specificity 64.90%) among all other techniques. This study demonstrates that machine learning models particularly logistic regression model provides a higher accurate prediction for fatty liver diseases based on medical data from electronic medical. This model can be used as a valuable tool for clinical decision making.

Exploring association between statin use and breast cancer risk: an updated meta-analysis

PURPOSE

The benefits of statin treatment for preventing cardiac disease are well established. However, preclinical studies suggested that statins may influence mammary cancer growth, but the clinical evidence is still inconsistent. We, therefore, performed an updated meta-analysis to provide a precise estimate of the risk of breast cancer in individuals undergoing statin therapy.

METHODS

For this meta-analysis, we searched PubMed, the Cochrane Library, Web of Science, Embase, and CINAHL for published studies up to January 31, 2017. Articles were included if they (1) were published in English; (2) had an observational study design with individual-level exposure and outcome data, examined the effect of statin therapy, and reported the incidence of breast cancer; and (3) reported estimates of either the relative risk, odds ratios, or hazard ratios with 95% confidence intervals (CIs). We used random-effect models to pool the estimates.

RESULTS

Of 2754 unique abstracts, 39 were selected for full-text review, and 36 studies reporting on 121,399 patients met all inclusion criteria. The overall pooled risks of breast cancer in patients using statins were 0.94 (95% CI 0.86-1.03) in random-effect models with significant heterogeneity between estimates (I ² = 83.79%, p = 0.0001). However, we also stratified by region, the duration of statin therapy, methodological design, statin properties, and individual stain use.

CONCLUSIONS

Our results suggest that there is no association between statin use and breast cancer risk. However, observational studies cannot clarify whether the observed epidemiologic association is a causal effect or the result of some unmeasured confounding variable. Therefore, more research is needed.

E-Health Literacy and Health Information Seeking Behavior Among University Students in Bangladesh

Web 2.0 has become a leading health communication platform and will continue to attract young users; therefore, the objective of this study was to understand the impact of Web 2.0 on health information seeking behavior among university students in Bangladesh. A random sample of adults (n = 199, mean 23.75 years, SD 2.87) participated in a cross-sectional, a survey that included the eHealth literacy scale (eHEALS) assessed use of Web 2.0 for health information. Collected data were analyzed using a descriptive statistical method and t-tests. Finally logistic regression analyses were conducted to determine associations between sociodemographic, social determinants, and use of Web 2.0 for seeking and sharing health information. Almost 74% of older Web 2.0 users (147/199, 73.9%) reported using popular Web 2.0 websites, such as Facebook and Twitter, to find and share health information. Current study support that current Web-based health information seeking and sharing behaviors influence health-related decision making.