Natural language processing in urology: Automated extraction of clinical information from histopathology reports of uro-oncology procedures.

Objectives: We aimed to automate routine extraction of clinically relevant unstructured information from uro-oncological histopathology reports by applying rule-based and machine learning (ML)/deep learning (DL) methods to develop an oncology focused natural language processing (NLP) algorithm. Methods: Our algorithm employs a combination of a rule-based approach and support vector machines/neural networks (BioBert/Clinical BERT), and is optimised for accuracy. We randomly extracted 5772 uro-oncological histology reports from 2008 to 2018 from electronic health records (EHRs) and split the data into training and validation datasets in an 80:20 ratio. The training dataset was annotated by medical professionals and reviewed by cancer registrars. The validation dataset was annotated by cancer registrars and defined as the gold standard with which the algorithm outcomes were compared. The accuracy of NLP-parsed data was matched against these human annotation results. We defined an accuracy rate of >95% as "acceptable" by professional human extraction, as per our cancer registry definition. Results: There were 11 extraction variables in 268 free-text reports. We achieved an accuracy rate of between 61.2% and 99.0% using our algorithm. Of the 11 data fields, a total of 8 data fields met the acceptable accuracy standard, while another 3 data fields had an accuracy rate between 61.2% and 89.7%. Noticeably, the rule-based approach was shown to be more effective and robust in extracting variables of interest. On the other hand, ML/DL models had poorer predictive performances due to highly imbalanced data distribution and variable writing styles between different reports and data used for domain-specific pre-trained models. Conclusion: We designed an NLP algorithm that can automate clinical information extraction accurately from histopathology reports with an overall average micro accuracy of 93.3%.

Development of a real-world database for asthma and COPD: The SingHealth-Duke-NUS-GSK COPD and Asthma Real-World Evidence (SDG-CARE) collaboration.

PURPOSE: The SingHealth-Duke-GlaxoSmithKline COPD and Asthma Real-world Evidence (SDG-CARE) collaboration was formed to accelerate the use of Singaporean real-world evidence in research and clinical care. A centerpiece of the collaboration was to develop a near real-time database from clinical and operational data sources to inform healthcare decision making and research studies on asthma and chronic obstructive pulmonary disease (COPD). METHODS: Our multidisciplinary team, including clinicians, epidemiologists, data scientists, medical informaticians and IT engineers, adopted the hybrid waterfall-agile project management methodology to develop the SingHealth COPD and Asthma Data Mart (SCDM). The SCDM was developed within the organizational data warehouse. It pulls and maps data from various information systems using extract, transform and load (ETL) pipelines. Robust user testing and data verification was also performed to ensure that the business requirements were met and that the ETL pipelines were valid. RESULTS: The SCDM includes 199 data elements relevant to asthma and COPD. Data verification was performed and found the SCDM to be reliable. As of December 31, 2019, the SCDM contained 36,407 unique patients with asthma and COPD across the spectrum from primary to tertiary care in our healthcare system. The database updates weekly to add new data of existing patients and to include new patients who fulfil the inclusion criteria. CONCLUSIONS: The SCDM was systematically developed and tested to support the use RWD for clinical and health services research in asthma and COPD. This can serve as a platform to provide research and operational insights to improve the care delivered to our patients.

Blockchain applications in health care for COVID-19 and beyond: a systematic review.

The COVID-19 pandemic has had a substantial and global impact on health care, and has greatly accelerated the adoption of digital technology. One of these emerging digital technologies, blockchain, has unique characteristics (eg, immutability, decentralisation, and transparency) that can be useful in multiple domains (eg, management of electronic medical records and access rights, and mobile health). We conducted a systematic review of COVID-19-related and non-COVID-19-related applications of blockchain in health care. We identified relevant reports published in MEDLINE, SpringerLink, Institute of Electrical and Electronics Engineers Xplore, ScienceDirect, arXiv, and Google Scholar up to July 29, 2021. Articles that included both clinical and technical designs, with or without prototype development, were included. A total of 85 375 articles were evaluated, with 415 full length reports (37 related to COVID-19 and 378 not related to COVID-19) eventually included in the final analysis. The main COVID-19-related applications reported were pandemic control and surveillance, immunity or vaccine passport monitoring, and contact tracing. The top three non-COVID-19-related applications were management of electronic medical records, internet of things (eg, remote monitoring or mobile health), and supply chain monitoring. Most reports detailed technical performance of the blockchain prototype platforms (277 [66·7%] of 415), whereas nine (2·2%) studies showed real-world clinical application and adoption. The remaining studies (129 [31·1%] of 415) were themselves of a technical design only. The most common platforms used were Ethereum and Hyperledger. Blockchain technology has numerous potential COVID-19-related and non-COVID-19-related applications in health care. However, much of the current research remains at the technical stage, with few providing actual clinical applications, highlighting the need to translate foundational blockchain technology into clinical use.

Early warning score validation methodologies and performance metrics: a systematic review.

BACKGROUND: Early warning scores (EWS) have been developed as clinical prognostication tools to identify acutely deteriorating patients. In the past few years, there has been a proliferation of studies that describe the development and validation of novel machine learning-based EWS. Systematic reviews of published studies which focus on evaluating performance of both well-established and novel EWS have shown conflicting conclusions. A possible reason is the heterogeneity in validation methods applied. In this review, we aim to examine the methodologies and metrics used in studies which perform EWS validation. METHODS: A systematic review of all eligible studies from the MEDLINE database and other sources, was performed. Studies were eligible if they performed validation on at least one EWS and reported associations between EWS scores and inpatient mortality, intensive care unit (ICU) transfers, or cardiac arrest (CA) of adults. Two reviewers independently did a full-text review and performed data abstraction by using standardized data-worksheet based on the TRIPOD (Transparent reporting of a multivariable prediction model for individual prognosis or diagnosis) checklist. Meta-analysis was not performed due to heterogeneity. RESULTS: The key differences in validation methodologies identified were (1) validation dataset used, (2) outcomes of interest, (3) case definition, time of EWS use and aggregation methods, and (4) handling of missing values. In terms of case definition, among the 48 eligible studies, 34 used the patient episode case definition while 12 used the observation set case definition, and 2 did the validation using both case definitions. Of those that used the patient episode case definition, 18 studies validated the EWS at a single point of time, mostly using the first recorded observation. The review also found more than 10 different performance metrics reported among the studies. CONCLUSIONS: Methodologies and performance metrics used in studies performing validation on EWS were heterogeneous hence making it difficult to interpret and compare EWS performance. Standardizing EWS validation methodology and reporting can potentially address this issue.

Factors affecting patient enablement in an Asian setting: a mixed methods study.

INTRODUCTION: Patient-centred medical care has been rising in importance since the turn of the century. It entails treating patients in relation to their biopsychosocial outlook so as to support the management of their conditions. The extent to which a patient is enabled to acquire skills and knowledge can be measured with the Patient Enablement Instrument (PEI) proposed by Howie and colleagues, and it has been noted to be more reflective of a good consultation compared to patient satisfaction scores. This study aimed to determine the level of patient enablement in the Singaporean context and the factors facilitating it. METHODS: We conducted an embedded mixed method study with primary care patients in two phases: (a) a PEI questionnaire was completed by 150 patients; and (b) a qualitative approach using focused group discussions and individual interviews was used to explore factors associated with high enablement. RESULTS: The mean PEI score was 4.5 ± 4.4, with significantly higher scores among patients attending specialised primary care clinics. Important physician factors were doctors' advice, attitude and relationship with the patient. Critical system factors included good continuity of care, workload and financial support, while patient factors included their beliefs, preparedness, inquisitiveness and trust, with considerable impact from the influence of community. CONCLUSION: The PEI score in the Singaporean context is similar to that of other Asian contexts, but slightly higher than that reported in Western studies. Good doctor-patient relationships, efficient systems facilitating continuity of care, and motivated and informed patients all contribute to increased enablement.