Aceso: PICO-guided Evidence Summarization on Medical Literature

Boundary-Aware Dual Biaffine Model for Sequential Sentence Classification in Biomedical Documents

Assigning appropriate rhetorical roles, such as "background," "intervention," and "outcome," to sentences in biomedical documents can streamline the process for physicians to locate evidence and resources for medical treatment and decision-making. While sequence labeling and span-based methods are frequently employed for this task, the former disregards a document's semantic structure, resulting in a lack of semantic coherence across continuous sentences. Span-based approaches, on the other hand, either necessitate the enumeration of all potential spans, which can be time-consuming, or may lead to the misclassification of sentences over extended spans. Consequently, an approach is required that models the semantic structure of documents explicitly and captures boundary information to achieve precise and effective sentence labeling in biomedical documents. To address these challenges, we propose a new approach, the boundary-aware dual biaffine model, which explicitly models the semantic structure of documents and incorporates boundary information via a dual biaffine layer. We introduce a dynamic programming algorithm to minimize missing labels and overlapping predictions, and achieve globally optimal decoding results. We evaluate our approach on three benchmark datasets, namely PubMed 20 k RCT, PubMed-PICO and NICTA-PIBOSO. The experimental results demonstrate that our approach outperforms strong baselines and achieves state-of-the-art performance on PubMed 20 k RCT and PubMed-PICO. Additionally, our method also achieves competitive results on NICTA-PIBOSO.

The association between childhood adverse experiences and internet addiction: A meta-analysis

Many studies have explored the association between adverse childhood experiences (ACEs) and Internet addiction (IA), yet the research findings on the association between them are inconclusive. We conducted a systematic search on 7 databases to identify the relevant studies published until January 2023, and analyzed the findings from 37 studies across 12 countries involving 45,364 participants aged 8 to 67 years (51 % women). Results indicated a positive correlation (r = 0.21) was found between ACE and IA around the world, which differed among continents. It was found that all ACE subtypes were significantly associated with IA (range r = 0.16 to 0.25). Meta-regression showed a stronger association among younger individuals without moderating effects of gender or publication year. In conclusion, this study sheds light on the significant association between ACEs and IA, emphasizing the need for targeted interventions and preventive measures. Future research could delve into specific interventions aimed at mitigating the impact of ACEs on IA, such as cognitive-behavioral therapies or metacognitive therapy interventions. Additionally, investigating cultural factors that may influence this association could provide valuable insights into tailored approaches for different populations. Understanding these dynamics is crucial for developing effective strategies to address IA and its underlying factors.

Leveraging Summary Guidance on Medical Report Summarization

This study presents three deidentified large medical text datasets, named DISCHARGE, ECHO and RADIOLOGY, which contain 50 K, 16 K and 378 K pairs of report and summary that are derived from MIMIC-III, respectively. We implement convincing baselines of automated abstractive summarization on the created datasets with pre-trained encoder-decoder language models, including BERT2BERT, BERTShare, RoBERTaShare, Pegasus, ProphetNet, T5-large, BART and GSUM. Further, based on the BART model, we leverage the sampled summaries from the training set as prior knowledge guidance, for encoding additional contextual representations of the guidance with the encoder and enhancing the decoding representations in the decoder. The experimental results confirm the improvement of ROUGE scores and BERTScore made by the proposed method.

The use of artificial intelligence for automating or semi-automating biomedical literature analyses: a scoping review

OBJECTIVE

Evidence-based medicine (EBM) is a decision-making process based on the conscious and judicious use of the best available scientific evidence. However, the exponential increase in the amount of information currently available likely exceeds the capacity of human-only analysis. In this context, artificial intelligence (AI) and its branches such as machine learning (ML) can be used to facilitate human efforts in analyzing the literature to foster EBM. The present scoping review aimed to examine the use of AI in the automation of biomedical literature survey and analysis with a view to establishing the state-of-the-art and identifying knowledge gaps.

MATERIALS AND METHODS

Comprehensive searches of the main databases were performed for articles published up to June 2022 and studies were selected according to inclusion and exclusion criteria. Data were extracted from the included articles and the findings categorized.

RESULTS

The total number of records retrieved from the databases was 12,145, of which 273 were included in the review. Classification of the studies according to the use of AI in evaluating the biomedical literature revealed three main application groups, namely assembly of scientific evidence (n=127; 47%), mining the biomedical literature (n=112; 41%) and quality analysis (n=34; 12%). Most studies addressed the preparation of systematic reviews, while articles focusing on the development of guidelines and evidence synthesis were the least frequent. The biggest knowledge gap was identified within the quality analysis group, particularly regarding methods and tools that assess the strength of recommendation and consistency of evidence.

CONCLUSION

Our review shows that, despite significant progress in the automation of biomedical literature surveys and analyses in recent years, intense research is needed to fill knowledge gaps on more difficult aspects of ML, deep learning and natural language processing, and to consolidate the use of automation by end-users (biomedical researchers and healthcare professionals).

Safety and efficacy associated with single-fraction high-dose-rate brachytherapy in localized prostate cancer: a systematic review and meta-analysis

OBJECTIVE

Although single-fraction high-dose-rate brachytherapy (SFHDR) for localized prostate cancer has been tried in clinical trials, relevant medical evidence is currently lacking. It is necessary to systematically analyze the safety and efficacy of SFHDR.

METHODS

Comprehensive and systematic searches for eligible studies were performed in PubMed, Embase, and the Cochrane Library databases. The primary endpoints included safety and efficacy, represented by toxic effects and biochemical recurrence-free survival (bRFS), respectively. The proportion rates were used as the effect measure for each study and were presented with corresponding 95% confidence intervals (CI) and related 95% prediction interval (PI). Restricted maximum-likelihood estimator (REML) and the Hartung-Knapp method were used in the meta-analysis.

RESULTS

Twenty-five studies met the inclusion criteria for quantitative analysis, including 1440 patients. The median age of patients was 66.9 years old (62-73 years old) and the median follow-up was 47.5 months (12-75 months). The estimates of cumulative occurrence for severe gastrointestinal (GI) and genitourinary (GU) toxic effects were 0.1% (95% CI 0-0.2%) and 0.4% (95% CI 0-1.2%), and for grade 2 toxic effects were 1.6% (95% CI 0.1-4.7%) and 17.1% (95% CI 5.4-33.5%), respectively. The estimate of 3‑year bRFS was 87.5% (95% CI 84.4-90.3%) and 71.0% (95% CI 63.0-78.3%) for 5‑year bRFS. The pooled bRFS rates for low-risk patients were 99.0% (95% CI 85.2-100.0%) at 3 years and 80.9% (95% CI 75.4-85.9%) at 5 years, and the risk group was found to be statistically correlated with bRFS (3-year bRFS, P < 0.01; 5‑year bRFS, P = 0.04).

CONCLUSION

SFHDR is associated with favorable tolerability and suboptimal clinical benefit in patients with localized prostate cancer. Ongoing and planned high-quality prospective studies are necessary to verify its safety and efficacy.

The scientific publication score - a new tool for summarizing evidence and data quality criteria of biomedical publications

The number of biomedical research articles increases by over 2.5 million publications each year, making it difficult to stay up to date. In this study, we introduce a standardized search and evaluation tool to combat this issue. Employing crowdsourcing, a large database of publications is gathered. Using a standardized data entry format, coined the "scientific publication score" (SPS), specific publication results can be easily aggregated, thereby allowing fast and accurate comparisons for clinical questions. The SPS combines two quality dimensions. The first captures the quality of evidence of the study using the evidence criteria defined by the Centre for Evidence-Based Medicine, Oxford, UK. The second is more fine-grained and considers the magnitude of statistical analyses on individual and specific results. From 2014 to 2019, experts of the European Bone and Joint Infection Society (EBJIS) were asked to enter data of relevant publications about prosthetic joint infection. Data and evidence levels of specific results were averaged, summarized and ranked. A total of 366 publications were divided into two groups: (I) risk factors (e.g., host-related factors, pre- and postoperative issues) with 243 publications and (II) diagnostic methods (e.g., laboratory tests, imaging methods) with 123 publications. After ranking, the highest score for risk factors of prosthetic joint infection were calculated by the SPS for anemia (mean 3.50 ±  SD 0.91), malignancy (mean 3.17 ±  SD 0.29) and previous alloarthroplasty (mean 3.00 ±  SD 0.35). A comparison of the full SPS ranking with the ranking determined at the 2018 International Consensus Meeting (ICM) on Musculoskeletal Infection resulted in a Spearman rank correlation coefficient of 0.48 and a p  value of 0.0382. The diagnostic methods ranked highest by the SPS were aspirate leucocyte count (mean 3.15 ±  SD 1.21), interleukin 6 (mean 3.14 ±  SD 1.07) and aspirate (neutrophils over 80 %) (mean 3.12 ±  SD 0.63). The comparison to the ICM ranking yielded a Spearman rank correlation coefficient of 0.91 and a p  value of 0.0015. Our pilot study evaluated a new tool for the quality assessment of specific results based on the quality of the source publication. The SPS is suitable for a ranking of specific results by evidence and data quality criteria important for systematic reviews.

Mouthwashes and the Effect on the Viral Load of SARS-CoV-2 in Saliva: A Literature Review

BACKGROUND: At present, several active ingredients have been investigated in mouthwashes having certain virucidal properties, which could reduce the viral load of SARS-CoV-2 to avoid contamination in medical or dental practice. AIM: The objective of this review is to analyze the available evidence regarding mouthwashes and their effect on the salivary viral load of SARS-CoV-2. METHODS: Records were retrieved from databases such as PubMed, Scopus, Web of Science, and Virtual Health Library up to June 21, 2022. Randomized or non-randomized clinical trials were included where saliva samples and laboratory or in vitro studies were used in the presence of saliva. RESULTS: After a systematic selection process, 11 clinical studies that evaluated at least one mouthwash within clinical protocols and three laboratory studies that evaluated the virucidal efficacy against SARS-CoV-2 in the presence of saliva were finally included. CONCLUSION: There are oral disinfectants with virucidal action in saliva samples, under clinical and laboratory conditions, capable of reducing the viral load of SARS-CoV-2. Cetylpyridinium chloride, chlorhexidine, and povidone-iodine present the best results so far. However, it was also possible to find active principles of recent appearance that, based on favorable exploratory results, needs further investigation on their efficacy and possible adverse events.

Multiple Embeddings Enhanced Multi-Graph Neural Networks for Chinese Healthcare Named Entity Recognition

Named Entity Recognition (NER) is a natural language processing task for recognizing named entities in a given sentence. Chinese NER is difficult due to the lack of delimited spaces and conventional features for determining named entity boundaries and categories. This study proposes the ME-MGNN (Multiple Embeddings enhanced Multi-Graph Neural Networks) model for Chinese NER in the healthcare domain. We integrate multiple embeddings at different granularities from the radical, character to word levels for an extended character representation, and this is fed into multiple gated graph sequence neural networks to identify named entities and classify their types. The experimental datasets were collected from health-related news, digital health magazines and medical question/answer forums. Manual annotation was conducted for a total of 68,460 named entities across 10 entity types (body, symptom, instrument, examination, chemical, disease, drug, supplement, treatment and time) in 30,692 sentences. Experimental results indicated our ME-MGNN model achieved an F1-score result of 75.69, outperforming previous methods. In practice, a series of model analysis implied that our method is effective and efficient for Chinese healthcare NER.

Data extraction methods for systematic review (semi)automation: Update of a living systematic review

Background: The reliable and usable (semi)automation of data extraction can support the field of systematic review by reducing the workload required to gather information about the conduct and results of the included studies. This living systematic review examines published approaches for data extraction from reports of clinical studies. Methods: We systematically and continually search PubMed, ACL Anthology, arXiv, OpenAlex via EPPI-Reviewer, and the  dblp computer science bibliography. Full text screening and data extraction are conducted within an open-source living systematic review application created for the purpose of this review. This living review update includes publications up to December 2022 and OpenAlex content up to March 2023. Results: 76 publications are included in this review. Of these, 64 (84%) of the publications addressed extraction of data from abstracts, while 19 (25%) used full texts. A total of 71 (93%) publications developed classifiers for randomised controlled trials. Over 30 entities were extracted, with PICOs (population, intervention, comparator, outcome) being the most frequently extracted. Data are available from 25 (33%), and code from 30 (39%) publications. Six (8%) implemented publicly available tools Conclusions: This living systematic review presents an overview of (semi)automated data-extraction literature of interest to different types of literature review. We identified a broad evidence base of publications describing data extraction for interventional reviews and a small number of publications extracting epidemiological or diagnostic accuracy data. Between review updates, trends for sharing data and code increased strongly: in the base-review, data and code were available for 13 and 19% respectively, these numbers increased to 78 and 87% within the 23 new publications. Compared with the base-review, we observed another research trend, away from straightforward data extraction and towards additionally extracting relations between entities or automatic text summarisation. With this living review we aim to review the literature continually.