A hybrid solution for extracting structured medical information from unstructured data in medical records via a double-reading/entry system

Using machine learning to mine mental health diagnostic groups from emergency department presentations before and during the COVID-19 pandemic

PURPOSE

The COVID-19 pandemic had a profound negative effect on mental health worldwide. The hospital emergency department plays a pivotal role in responding to mental health crises. Understanding data trends relating to hospital emergency department usage is beneficial for service planning, particularly around preparing for future pandemics. Machine learning has been used to mine large volumes of unstructured data to extract meaningful data in relation to mental health presentations. This study aims to analyse trends in five mental health-related presentations to an emergency department before and during, the COVID-19 pandemic.

METHODS

Data from 690,514 presentations to two Australian, public hospital emergency departments between April 2019 to February 2022 were assessed. A machine learning-based framework, Mining Emergency Department Records, Evolutionary Algorithm Data Search (MEDREADS), was used to identify suicidality, psychosis, mania, eating disorder, and substance use.

RESULTS

While the mental health-related presentations to the emergency department increased during the COVID-19 pandemic compared to pre-pandemic levels, the proportion of mental health presentations relative to the total emergency department presentations decreased. Several troughs in presentation frequency were identified across the pandemic period, which occurred consistently during the public health lockdown and restriction periods.

CONCLUSION

This study implemented novel machine learning techniques to analyse mental health presentations to an emergency department during the COVID-19 pandemic. Results inform understanding of the use of emergency mental health services during the pandemic, and highlight opportunities to further investigate patterns in presentation.

A Scoping Review of Adopted Information Extraction Methods for RCTs

Background

Randomized controlled trials (RCTs) provide the strongest evidence for therapeutic interventions and their effects on groups of subjects. However, the large amount of unstructured information in these trials makes it challenging and time-consuming to make decisions and identify important concepts and valid evidence. This study aims to explore methods for automating or semi-automating information extraction from reports of RCT studies.

Methods

We conducted a systematic search of PubMed, ACM Digital Library, and Web of Science to identify relevant articles published between January 1, 2010, and 2022. We focused on published Natural Language Processing (NLP), machine learning, and deep learning methods that automate or semi-automate key elements of information extraction in the context of RCTs.

Results

A total of 26 publications were included, which discussed the automatic extraction of key characteristics of RCTs using various PICO frameworks (PIBOSO and PECODR). Among these publications, 14 (53.8%) extracted key characteristics based on PICO, PIBOSO, and PECODR, while 12 (46.1%) discussed information extraction methods in RCT studies. Common approaches mentioned included word/phrase matching, machine learning algorithms such as binary classification using the Naïve Bayes algorithm and powerful BERT network for feature extraction, support vector machine for data classification, conditional random field, non-machine-dependent automation, and machine learning or deep learning approaches.

Conclusion

The lack of publicly available software and limited access to existing software makes it difficult to determine the most powerful information extraction system. However, deep learning models like Transformers and BERT language models have shown better performance in natural language processing.

Validation of a Natural Language Processing Algorithm for the Extraction of the Sleep Parameters from the Polysomnography Reports

Background: There is a need to better understand the association between sleep and chronic diseases. In this study we developed a natural language processing (NLP) algorithm to mine polysomnography (PSG) free-text notes from electronic medical records (EMR) and evaluated the performance. Methods: Using the Veterans Health Administration EMR, we identified 46,093 PSG studies using CPT code 95,810 from 1 October 2000−30 September 2019. We randomly selected 200 notes to compare the accuracy of the NLP algorithm in mining sleep parameters including total sleep time (TST), sleep efficiency (SE) and sleep onset latency (SOL), wake after sleep onset (WASO), and apnea-hypopnea index (AHI) compared to visual inspection by raters masked to the NLP output. Results: The NLP performance on the training phase was >0.90 for precision, recall, and F-1 score for TST, SOL, SE, WASO, and AHI. The NLP performance on the test phase was >0.90 for precision, recall, and F-1 score for TST, SOL, SE, WASO, and AHI. Conclusions: This study showed that NLP is an accurate technique to extract sleep parameters from PSG reports in the EMR. Thus, NLP can serve as an effective tool in large health care systems to evaluate and improve patient care.

Digital Disruption and Big Data in Healthcare - Opportunities and Challenges

Background

As the amount of medical data in the electronic medical records system (EMR) is increasing tremendously, the required time to read it by health providers is growing by the exact proportionality. This means that physicians must increase the time spared for each patient again by the precise proportionality. This may lead to exposing the accuracy and quality of the course of action to be taken for the patients. Increasing the physician’s required time for one patient means that the physician can see fewer patients. This will create an issue with the medical management authority as more physicians are needed, and higher expenses will be required.

Purpose

The two questions that arise here are 1. Identify the potential opportunities and challenges for extensive data analysis in the healthcare sector. 2. Evaluate different ways in which big medical data can be analyzed?

Methods

The authors identified the four concerned parties representing the four potential solutions dimensions to answer these two questions. These parties are 1. physicians, 2. health information systems management (HISM) departments, mainly the EMR system, and 3. Health management departments 4. Relevant Health Information Systems (HIS) parties. A literature review and 25 interviews were conducted. The interviews covered 1: Two global organizations: John Hopkins and Joint Commission International (JCI), 2: Three United Arab Emirates-based health organizations: Department of health in Abu Dhabi, SEHA in Abu Dhabi, Dubai health Authority (DHA) in Dubai, 3: 10 Physicians from different specialties, 4: Five EMR managers and 5: Five IT (Information Technology) professionals representing the HIS parties. Qualitative analysis is used as the approach for data analysis.

Results

Identifying the managerial and the technical recommendations to be utilized mainly based on digital disruption technologies, tools, and processes.

Conclusion

Healthcare has been slow in embracing digital disruption and transformation. In most areas, it is still in the initial stages. Recommendations are based on the UAE cases, highlighting the specific technologies and their features.

Big Data Reality Check (BDRC) for public health: to what extent the environmental health and health services research did meet the 'V' criteria for big data? A study protocol

INTRODUCTION

Big data technologies have been talked up in the fields of science and medicine. The V-criteria (volume, variety, velocity and veracity, etc) for defining big data have been well-known and even quoted in most research articles; however, big data research into public health is often misrepresented due to certain common misconceptions. Such misrepresentations and misconceptions would mislead study designs, research findings and healthcare decision-making. This study aims to identify the V-eligibility of big data studies and their technologies applied to environmental health and health services research that explicitly claim to be big data studies.

METHODS AND ANALYSIS

Our protocol follows Preferred Reporting Items for Systematic Review and Meta-Analysis Protocols (PRISMA-P). Scoping review and/or systematic review will be conducted. The results will be reported using PRISMA for Scoping Reviews (PRISMA-ScR), or PRISMA 2020 and Synthesis Without Meta-analysis guideline. Web of Science, PubMed, Medline and ProQuest Central will be searched for the articles from the database inception to 2021. Two reviewers will independently select eligible studies and extract specified data. The numeric data will be analysed with R statistical software. The text data will be analysed with NVivo wherever applicable.

ETHICS AND DISSEMINATION

This study will review the literature of big data research related to both environmental health and health services. Ethics approval is not required as all data are publicly available and involves confidential personal data. We will disseminate our findings in a peer-reviewed journal.

PROSPERO REGISTRATION NUMBER

CRD42021202306.

Video-assisted thoracoscopic surgery lobectomy might be a feasible alternative for surgically resectable pathological N2 non-small cell lung cancer patients

Background

The majority of previous studies of the clinical outcome of video‐assisted thoracoscopic surgery (VATS) versus open lobectomy for pathological N2 non‐small cell lung cancer (pN2 NSCLC) have been single‐center experiences with small patient numbers. The aim of this study was therefore to investigate these procedures but in a large cohort of Chinese patients with pathological N2 NSCLC in real‐world conditions.

Methods

Patients who underwent lobectomy for pN2 NSCLC by either VATS or thoracotomy were retrospectively reviewed from 10 tertiary hospitals between January 2014 and September 2017. Perioperative outcomes and overall survival of the patients were analyzed. Cox regression analysis was performed to identify potential prognostic factors. Propensity‐score analysis was performed to reduce cofounding biases and compare the clinical outcomes between both groups.

Results

Among 2144 pN2 NSCLC, 1244 patients were managed by VATS and 900 by open procedure. A total of 305 (24.5%) and 344 patients died during VATS and the thoracotomy group during a median follow‐up of 16.7 and 15.6 months, respectively. VATS lobectomy patients had better overall survival when compared with those undergoing the open procedure (P < 0.0001). Multivariate COX regression analysis showed VATS lobectomy independently favored overall survival (HR = 0.75, 95% CI: 0.621–0.896, P = 0.0017). Better perioperative outcomes, including less blood loss, shorter drainage time and hospital stay, were also observed in patients undergoing VATS lobectomy (P < 0.05). After propensity‐score matching, 169 patients in each group were analyzed, and no survival difference were found between the two groups. Less blood loss was observed in the VATS group, but there was a longer operation time.

Conclusions

VATS lobectomy might be a feasible alternative to conventional open surgery for resectable pN2 NSCLC.

Key points

Significant findings of the study: VATS lobectomy has comparative OS in pN2 NSCLC versus open procedure in resectable patients.

What this study adds: VATS lobectomy might be feasible for pN2 NSCLC.

Graph-Representation of Patient Data: a Systematic Literature Review

Graph theory is a well-established theory with many methods used in mathematics to study graph structures. In the field of medicine, electronic health records (EHR) are commonly used to store and analyze patient data. Consequently, it seems straight-forward to perform research on modeling EHR data as graphs. This systematic literature review aims to investigate the frontiers of the current research in the field of graphs representing and processing patient data. We want to show, which areas of research in this context need further investigation. The databases MEDLINE, Web of Science, IEEE Xplore and ACM digital library were queried by using the search terms health record, graph and related terms. Based on the "Preferred Reporting Items for Systematic Reviews and Meta-Analysis" (PRISMA) statement guidelines the articles were screened and evaluated using full-text analysis. Eleven out of 383 articles found in systematic literature review were finally included for analysis in this literature review. Most of them use graphs to represent temporal relations, often representing the connection among laboratory data points. Only two papers report that the graph data were further processed by comparing the patient graphs using similarity measurements. Graphs representing individual patients are hardly used in research context, only eleven papers considered such kind of graphs in their investigations. The potential of graph theoretical algorithms, which are already well established, could help increasing this research field, but currently there are too few papers to estimate how this area of research will develop. Altogether, the use of such patient graphs could be a promising technique to develop decision support systems for diagnosis, medication or therapy of patients using similarity measurements or different kinds of analysis.

Tracking Knowledge Evolution in Cloud Health Care Research: Knowledge Map and Common Word Analysis

Background

With the continuous development of the internet and the explosive growth in data, big data technology has emerged. With its ongoing development and application, cloud computing technology provides better data storage and analysis. The development of cloud health care provides a more convenient and effective solution for health. Studying the evolution of knowledge and research hotspots in the field of cloud health care is increasingly important for medical informatics. Scholars in the medical informatics community need to understand the extent of the evolution of and possible trends in cloud health care research to inform their future research.

Objective

Drawing on the cloud health care literature, this study aimed to describe the development and evolution of research themes in cloud health care through a knowledge map and common word analysis.

Methods

A total of 2878 articles about cloud health care was retrieved from the Web of Science database. We used cybermetrics to analyze and visualize the keywords in these articles. We created a knowledge map to show the evolution of cloud health care research. We used co-word analysis to identify the hotspots and their evolution in cloud health care research.

Results

The evolution and development of cloud health care services are described. In 2007-2009 (Phase I), most scholars used cloud computing in the medical field mainly to reduce costs, and grid computing and cloud computing were the primary technologies. In 2010-2012 (Phase II), the security of cloud systems became of interest to scholars. In 2013-2015 (Phase III), medical informatization enabled big data for health services. In 2016-2017 (Phase IV), machine learning and mobile technologies were introduced to the medical field.

Conclusions

Cloud health care research has been rapidly developing worldwide, and technologies used in cloud health research are simultaneously diverging and becoming smarter. Cloud–based mobile health, cloud–based smart health, and the security of cloud health data and systems are three possible trends in the future development of the cloud health care field.

Optimizing drug development in oncology by clinical trial simulation: Why and how?

In therapeutic research, the safety and efficacy of pharmaceutical products are necessarily tested on humans via clinical trials after an extensive and expensive preclinical development period. Methodologies such as computer modeling and clinical trial simulation (CTS) might represent a valuable option to reduce animal and human assays. The relevance of these methods is well recognized in pharmacokinetics and pharmacodynamics from the preclinical phase to postmarketing. However, they are barely used and are poorly regarded for drug approval, despite Food and Drug Administration and European Medicines Agency recommendations. The generalization of CTS could be greatly facilitated by the availability of software for modeling biological systems, by clinical trial studies and hospital databases. Data sharing and data merging raise legal, policy and technical issues that will need to be addressed. Development of future molecules will have to use CTS for faster development and thus enable better patient management. Drug activity modeling coupled with disease modeling, optimal use of medical data and increased computing speed should allow this leap forward. The realization of CTS requires not only bioinformatics tools to allow interconnection and global integration of all clinical data but also a universal legal framework to protect the privacy of every patient. While recognizing that CTS can never replace 'real-life' trials, they should be implemented in future drug development schemes to provide quantitative support for decision-making. This in silico medicine opens the way to the P4 medicine: predictive, preventive, personalized and participatory.

Rationale and design of Patient-centered Retrospective Observation of Guideline-Recommended Execution for Stroke Sufferers in China: China PROGRESS

Background

In 2009, China launched ambitious healthcare reform plans to provide affordable and equitable basic healthcare for all patients, including the substantial number of patients who had a stroke. However, little is known about the pattern of evidence-based stroke care and outcomes across hospitals, regions and time during the last decade.

Aims

The Patient-centered Retrospective Observation of Guideline-Recommended Execution for Stroke Sufferers in China (China PROGRESS) Study aims to use findings from a representative sample of Chinese hospitals over the last decade to improve future stroke care for patients hospitalised with ischaemic stroke (IS) or transient ischaemic attack (TIA).

Design

The China PROGRESS Study will use a two-stage cluster sampling method to identify over 32000 patient records from 208 hospitals across the Eastern, Central and Western geographical regions in China. To assess the temporal trends in patient characteristics, treatment and outcomes, study investigators will select records from 2005, 2010 and 2015. A double data reading/entry system will be developed to conduct this assessment. A central coordinating centre will monitor case ascertainment, data abstraction and data management. Analyses will examine patient characteristics, testing patterns, in-hospital treatment and outcomes, and variations across regions and across time.

Conclusions

The China PROGRESS Study is the first nationally representative study that aims to better understand care quality and outcomes for patients with IS or TIA before and after the national healthcare reform in China. This initiative will translate findings into clinical practices that improve care quality for patients who had a stroke and policy recommendations that allow these changes to be implemented widely.

Ethics approval

This study has also been approved by the central institutional review board (IRB) at Beijing Tiantan Hospital.

Merging Data Diversity of Clinical Medical Records to Improve Effectiveness

Medicine is a knowledge area continuously experiencing changes. Every day, discoveries and procedures are tested with the goal of providing improved service and quality of life to patients. With the evolution of computer science, multiple areas experienced an increase in productivity with the implementation of new technical solutions. Medicine is no exception. Providing healthcare services in the future will involve the storage and manipulation of large volumes of data (big data) from medical records, requiring the integration of different data sources, for a multitude of purposes, such as prediction, prevention, personalization, participation, and becoming digital. Data integration and data sharing will be essential to achieve these goals. Our work focuses on the development of a framework process for the integration of data from different sources to increase its usability potential. We integrated data from an internal hospital database, external data, and also structured data resulting from natural language processing (NPL) applied to electronic medical records. An extract-transform and load (ETL) process was used to merge different data sources into a single one, allowing more effective use of these data and, eventually, contributing to more efficient use of the available resources.

Priorities to Overcome Barriers Impacting Data Science Application in Emergency Care Research

For a variety of reasons including cheap computing, widespread adoption of electronic medical records, digitalization of imaging and biosignals, and rapid development of novel technologies, the amount of health care data being collected, recorded, and stored is increasing at an exponential rate. Yet despite these advances, methods for the valid, efficient, and ethical utilization of these data remain underdeveloped. Emergency care research, in particular, poses several unique challenges in this rapidly evolving field. A group of content experts was recently convened to identify research priorities related to barriers to the application of data science to emergency care research. These recommendations included: 1) developing methods for cross-platform identification and linkage of patients; 2) creating central, deidentified, open-access databases; 3) improving methodologies for visualization and analysis of intensively sampled data; 4) developing methods to identify and standardize electronic medical record data quality; 5) improving and utilizing natural language processing; 6) developing and utilizing syndrome or complaint-based based taxonomies of disease; 7) developing practical and ethical framework to leverage electronic systems for controlled trials; 8) exploring technologies to help enable clinical trials in the emergency setting; and 9) training emergency care clinicians in data science and data scientists in emergency care medicine. The background, rationale, and conclusions of these recommendations are included in the present article.

Circulating anti-p16a IgG autoantibodies as a potential prognostic biomarker for non-small cell lung cancer

It has been reported that p16 protein is overexpressed in many types of solid cancer and its aberrant expression may trigger the immune response, leading to the secretion of anti‐p16 antibodies. Here, we developed an in‐house ELISA with three p16‐derived linear peptide antigens to examine plasma anti‐p16 antibody levels in patients with non‐small cell lung cancer (NSCLC). Blood samples were taken from 200 control subjects and 211 patients with NSCLC prior to anticancer therapy. A Mann–Whitney U test demonstrated that plasma anti‐p16a IgG levels were significantly higher in NSCLC patients than in control subjects (Z = −11.14, P < 0.001). However, neither plasma anti‐p16b nor plasma anti‐p16c IgG levels showed significant differences in patients with NSCLC as compared to control subjects. Moreover, further analysis indicated that anti‐p16a IgG levels increased with tumor stages, and patients with late stage NSCLC, namely group IV, had the highest IgG levels among four subgroups. Receiver operating characteristic analysis revealed that the anti‐p16a IgG assay had a sensitivity of 32.7% against a specificity of 95.0% in group IV, while Kaplan–Meier survival analysis revealed no significant difference in overall survival between patients with high anti‐p16a IgG levels and those with low anti‐p16a IgG levels (χ² = 0.24, P = 0.63). In conclusion, anti‐p16a IgG may be suitable for use as a prognostic biomarker for NSCLC.

Alteration of circulating natural autoantibodies to CD25-derived peptide antigens and FOXP3 in non-small cell lung cancer

Natural autoantibody is a key component for immune surveillance function. Regulatory T (Treg) cells play indispensable roles in promoting tumorigenesis via immune escape mechanisms. Both CD25 and FOXP3 are specific markers for Treg cells and their natural autoantibodies may be involved in anticancer activities. This work was designed to develop an in-house enzyme-linked immunosorbent assay (ELISA) to examine plasma natural IgG against CD25 and FOXP3 in non-small cell lung cancer (NSCLC). Compared with control subjects, NSCLC patients had significantly higher levels of plasma IgG for CD25a (Z = -8.05, P < 0.001) and FOXP3 (Z = -4.17, P < 0.001), lower levels for CD25b (Z = -3.58, P < 0.001), and a trend toward lower levels for CD25c (Z = -1.70, P = 0.09). Interestingly, the anti-CD25b IgG assay had a sensitivity of 25.0% against a specificity of 95.0% in an early stage patients (T1N0M0) who showed the lowest anti-CD25b IgG levels among 4 subgroups classified based on staging information. Kaplan-Meier survival analysis showed that patients with high anti-FOXP3 IgG levels had shorter survival than those with low anti-FOXP3 IgG levels (χ2 = 3.75, P = 0.05). In conclusion, anti-CD25b IgG may be a promising biomarker in terms of screening individuals at high risk of lung cancer.

Plasma anti-BIRC5 IgG may be a useful marker for evaluating the prognosis of nonsmall cell lung cancer

A recent study demonstrated that circulating levels of IgG antibodies against linear peptide antigens derived from baculoviral IAP repeat-containing protein 5 isoform 2 (BIRC5) and myc proto-oncogene protein (MYC) were significantly increased in nonsmall cell lung cancer (NSCLC). This study was undertaken to replicate this initial work in an independent sample. An enzyme-linked immunosorbent assay (ELISA) was developed in-house to examine plasma IgG antibodies for three linear peptide antigens derived from BIRC5a, BIRC5b, and MYC in 211 patients with NSCLC and 200 control subjects. A Mann-Whitney U-test demonstrated that plasma anti-BIRC5a IgG levels, but not anti-BIRC5b or anti-MYC IgG levels, were significantly higher in NSCLC patients than control subjects, especially in male patients. Both squamous cell cancer and adenocarcinoma showed increased anti-BIRC5a IgG levels, but the IgG levels were not found to be changed significantly in the early stage of NSCLC. Kaplan-Meier survival analysis showed that NSCLC patients with high anti-BIRC5b IgG levels had better prognosis and longer overall survival (OS) than patients with low anti-BIRC5b IgG levels, although this significant difference failed to survive the adjustment for age, gender, NSCLC stages, and types. Plasma anti-BIRC5a and MYC IgG levels did not show significant associations with OS. In conclusion, Plasma anti-BIRC5 IgG may be a useful marker for assessment of prognosis of NSCLC but not for early diagnosis of this malignancy.

Feasibility study of structured diagnosis methods for functional dyspepsia in Korean medicine clinics

Background

Functional dyspepsia (FD) is the seventh most common disease encountered in Korean medicine (KM) clinics. Despite the large number of FD patients visiting KM clinics, the accumulated medical records have no utility in evidence development, due to being unstructured. This study aimed to construct a standard operating procedure (SOP) with appropriate structured diagnostic methods for FD, and assess the feasibility for use in KM clinics.

Methods

Two rounds of professional surveys were conducted by 10 Korean internal medicine professors to select the representative diagnostic methods. A feasibility study was conducted to evaluate compliance and time required for using the structured diagnostic methods by three specialists in two hospitals.

Results

As per the results of the professional survey, five questionnaires and one basic diagnostic method were selected. An SOP was constructed based on the survey results, and a feasibility study showed that the SOP compliance score (out of 5) was 3.45 among the subjects, and 3.25 among the practitioners. The SOP was acceptable and was not deemed difficult to execute. The total execution time was 136.5 minutes, out of which the gastric emptying test time was 129 minutes.

Conclusion

This feasibility study of the SOP with structured diagnostic methods for FD confirmed it was adequate for use in KM clinics. It is expected that these study findings will be helpful to clinicians who wish to conduct observational studies as well as to generate quantitative medical records to facilitate Big Data research.

Clinical Research Informatics: Contributions from 2016

Objectives: To summarize key contributions to current research in the field of Clinical Research Informatics (CRI) and to select the best papers published in 2016. Methods: A bibliographic search using a combination of MeSH and free terms on CRI was performed using PubMed, followed by a double-blind review in order to select a list of candidate best papers to be then peer-reviewed by external reviewers. A consensus meeting between the two section editors and the editorial team was organized to finally conclude on the selection of best papers. Results: Among the 452 papers published in 2016 in the various areas of CRI and returned by the query, the full review process selected four best papers. The authors of the first paper utilized a comprehensive representation of the patient medical record and semi-automatically labeled training sets to create phenotype models via a machine learning process. The second selected paper describes an open source tool chain securely connecting ResearchKit compatible applications (Apps) to the widely-used clinical research infrastructure Informatics for Integrating Biology and the Bedside (i2b2). The third selected paper describes the FAIR Guiding Principles for scientific data management and stewardship. The fourth selected paper focuses on the evaluation of the risk of privacy breaches in releasing genomics datasets. Conclusions: A major trend in the 2016 publications is the variety of research on "real-world data" - healthcare-generated data, person health data, and patient-reported outcomes -highlighting the opportunities provided by new machine learning techniques as well as new potential risks of privacy breaches.

A Framework for Interfacing Unstructured Data Into Business Process From Enterprise Social Networks

The increased number of Enterprise Social Networks (ESN) business applications has had a major impact on organizations' business processes improvements by allowing the involvement of human interactions to these process. However, these applications generate unstructured data which create barriers and challenges to offering the data in the form of web services in a SOA environment, which again impacts negatively the business process. In this context, the authors propose a framework to interface ESN unstructured data into BP using text mining techniques. The Term frequency-inverse document frequency is used as a weighting schema in this framework. After that, the cosine similarity and k-mean are utilized to find similar values from different documents and cluster documents into groups respectively. The result of the evaluation of the framework shows promising results for retrieving social unstructured data. These results can be published into the SOA enterprise service bus using the RESTful web services.