Automatic screening using word embeddings achieved high sensitivity and workload reduction for updating living network meta-analyses

A living critical interpretive synthesis to yield a framework on the production and dissemination of living evidence syntheses for decision-making

BACKGROUND

The COVID-19 pandemic has had an unprecedented impact in the global research production and has also increased research waste. Living evidence syntheses (LESs) seek to regularly update a body of evidence addressing a specific question. During the COVID-19 pandemic, the production and dissemination of LESs emerged as a cornerstone of the evidence infrastructure. This critical interpretive synthesis answers the questions: What constitutes an LES to support decision-making?; when should one be produced, updated, and discontinued?; and how should one be disseminated?

METHODS

Searches included the Cochrane Library, EMBASE (Ovid), Health Systems Evidence, MEDLINE (Ovid), PubMed, and Web of Science up to 23 April 2024 and included articles that provide any insights on addressing the compass questions on LESs. Articles were selected and appraised, and their insights extracted. An interpretive and iterative coding process was used to identify relevant thematic categories and create a conceptual framework.

RESULTS

Among the 16,630 non-duplicate records identified, 208 publications proved eligible. Most were non-empirical articles, followed by actual LESs. Approximately one in three articles were published in response to the COVID-19 pandemic. The conceptual framework addresses six thematic categories: (1) what is an LES; (2) what methodological approaches facilitate LESs production; (3) when to produce an LES; (4) when to update an LES; (5) how to make available the findings of an LES; and (6) when to discontinue LES updates.

CONCLUSION

LESs can play a critical role in reducing research waste and ensuring alignment with advisory and decision-making processes. This critical interpretive synthesis provides relevant insights on how to better organize the global evidence architecture to support their production.

TRIAL REGISTRATION

PROSPERO registration: CRD42021241875.

Evaluating the efficacy of artificial intelligence tools for the automation of systematic reviews in cancer research: A systematic review

To evaluate the performance accuracy and workload savings of artificial intelligence (AI)-based automation tools in comparison with human reviewers in medical literature screening for systematic reviews (SR) of primary studies in cancer research in order to gain insights on improving the efficiency of producing SRs. Medline, Embase, the Cochrane Library, and PROSPERO databases were searched from inception to November 30, 2022. Then, forward and backward literature searches were completed, and the experts in this field including the authors of the articles included were contacted for a thorough grey literature search. This SR was registered on PROSPERO (CRD 42023384772). Among the 3947 studies obtained from search, five studies met the preplanned study selection criteria. These five studies evaluated four AI tools: Abstrackr (four studies), RobotAnalyst (one), EPPI-Reviewer (one), and DistillerSR (one). Without missing final included citations, Abstrackr eliminated 20%-88% of titles and abstracts (time saving of 7-86 hours) and 59% of the full-texts (62 h) from human review across four different cancer-related SRs. In comparison, RobotAnalyst (1% of titles and abstracts, 1 h), EPPI Review (38% of titles and abstracts, 58 h; 59% of full-texts, 62 h), DistillerSR (42% of titles and abstracts, 22 h) also provided similar or lower work savings for single cancer-related SRs. AI-based automation tools exhibited promising but varying levels of accuracy and efficiency during the screening process of medical literature for conducting SRs in the cancer field. Until further progress is made and thorough evaluations are conducted, AI tools should be utilized as supplementary aids rather than complete substitutes for human reviewers.

Methods and guidance on conducting, reporting, publishing, and appraising living systematic reviews: a scoping review

BACKGROUND AND OBJECTIVE

The living systematic review (LSR) approach is based on ongoing surveillance of the literature and continual updating. Most currently available guidance documents address the conduct, reporting, publishing, and appraisal of systematic reviews (SRs), but are not suitable for LSRs per se and miss additional LSR-specific considerations. In this scoping review, we aim to systematically collate methodological guidance literature on how to conduct, report, publish, and appraise the quality of LSRs and identify current gaps in guidance.

METHODS

A standard scoping review methodology was used. We searched MEDLINE (Ovid), EMBASE (Ovid), and The Cochrane Library on August 28, 2021. As for searching gray literature, we looked for existing guidelines and handbooks on LSRs from organizations that conduct evidence syntheses. The screening was conducted by two authors independently in Rayyan, and data extraction was done in duplicate using a pilot-tested data extraction form in Excel. Data was extracted according to four pre-defined categories for (i) conducting, (ii) reporting, (iii) publishing, and (iv) appraising LSRs. We mapped the findings by visualizing overview tables created in Microsoft Word.

RESULTS

Of the 21 included papers, methodological guidance was found in 17 papers for conducting, in six papers for reporting, in 15 papers for publishing, and in two papers for appraising LSRs. Some of the identified key items for (i) conducting LSRs were identifying the rationale, screening tools, or re-revaluating inclusion criteria. Identified items of (ii) the original PRISMA checklist included reporting the registration and protocol, title, or synthesis methods. For (iii) publishing, there was guidance available on publication type and frequency or update trigger, and for (iv) appraising, guidance on the appropriate use of bias assessment or reporting funding of included studies was found. Our search revealed major evidence gaps, particularly for guidance on certain PRISMA items such as reporting results, discussion, support and funding, and availability of data and material of a LSR.

CONCLUSION

Important evidence gaps were identified for guidance on how to report in LSRs and appraise their quality. Our findings were applied to inform and prepare a PRISMA 2020 extension for LSR.

A real-world evaluation of the implementation of NLP technology in abstract screening of a systematic review

The laborious and time-consuming nature of systematic review production hinders the dissemination of up-to-date evidence synthesis. Well-performing natural language processing (NLP) tools for systematic reviews have been developed, showing promise to improve efficiency. However, the feasibility and value of these technologies have not been comprehensively demonstrated in a real-world review. We developed an NLP-assisted abstract screening tool that provides text inclusion recommendations, keyword highlights, and visual context cues. We evaluated this tool in a living systematic review on SARS-CoV-2 seroprevalence, conducting a quality improvement assessment of screening with and without the tool. We evaluated changes to abstract screening speed, screening accuracy, characteristics of included texts, and user satisfaction. The tool improved efficiency, reducing screening time per abstract by 45.9% and decreasing inter-reviewer conflict rates. The tool conserved precision of article inclusion (positive predictive value; 0.92 with tool vs. 0.88 without) and recall (sensitivity; 0.90 vs. 0.81). The summary statistics of included studies were similar with and without the tool. Users were satisfied with the tool (mean satisfaction score of 4.2/5). We evaluated an abstract screening process where one human reviewer was replaced with the tool's votes, finding that this maintained recall (0.92 one-person, one-tool vs. 0.90 two tool-assisted humans) and precision (0.91 vs. 0.92) while reducing screening time by 70%. Implementing an NLP tool in this living systematic review improved efficiency, maintained accuracy, and was well-received by researchers, demonstrating the real-world effectiveness of NLP in expediting evidence synthesis.

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).

Progress and challenges of network meta-analysis

In the past years, network meta-analysis (NMA) has been widely used among clinicians, guideline makers, and health technology assessment agencies and has played an important role in clinical decision-making and guideline development. To inform further development of NMAs, we conducted a bibliometric analysis to assess the current status of published NMA methodological studies, summarized the methodological progress of seven types of NMAs, and discussed the current challenges of NMAs.

Natural language processing was effective in assisting rapid title and abstract screening when updating systematic reviews

BACKGROUND AND OBJECTIVE

To examine whether the use of natural language processing (NLP) technology is effective in assisting rapid title and abstract screening when updating a systematic review.

STUDY DESIGN

Using the searched literature from a published systematic review, we trained and tested an NLP model that enables rapid title and abstract screening when updating a systematic review. The model was a light gradient boosting machine (LightGBM), an ensemble learning classifier which integrates four pretrained Bidirectional Encoder Representations from Transformers (BERT) models. We divided the searched citations into two sets (ie, training and test sets). The model was trained using the training set and assessed for screening performance using the test set. The searched citations, whose eligibility was determined by two independent reviewers, were treated as the reference standard.

RESULTS

The test set included 947 citations; our model included 340 citations, excluded 607 citations, and achieved 96% sensitivity, and 78% specificity. If the classifier assessment in the case study was accepted, reviewers would lose 8 of 180 eligible citations (4%), none of which were ultimately included in the systematic review after full-text consideration, while decreasing the workload by 64.1%.

CONCLUSION

NLP technology using the ensemble learning method may effectively assist in rapid literature screening when updating systematic reviews.

A rule-based approach for automatically extracting data from systematic reviews and their updates to model the risk of conclusion change

Few data-driven approaches are available to estimate the risk of conclusion change in systematic review updates. We developed a rule-based approach to automatically extract information from reviews and updates to be used as features for modelling conclusion change risk. Rules were developed to extract relevant information from published Cochrane reviews and used to construct four features: the number of included trials and participants in the reviews, a measure based on the number of participants, and the time elapsed between the search dates. We compared the performance of random forest, decision tree, and logistic regression to predict the conclusion change risk. The performance was measured by accuracy, precision, recall, F₁ -score, and area under ROC (AU-ROC). One rule was developed to extract the conclusion change information (96% accuracy, 100 reviews), one for the search date (100% accuracy, 100 reviews), one for the number of included clinical trials (100% accuracy, 100 reviews), and 22 for the number of participants (97.3% accuracy, 200 reviews). For unseen reviews, the random forest classifier showed the highest accuracy (80.8%) and AU-ROC (0.80). All classifiers showed relatively similar performance with overlapping 95% confidence interval (CI). The coverage score was shown to be the most useful feature for predicting the conclusion change risk. Features mined from Cochrane reviews and updates can estimate conclusion change risk. If data from more published reviews and updates were made accessible, data-driven methods to predict the conclusion change risk may be a feasible way to support decisions about updating reviews.

Validation of a Semiautomated Natural Language Processing-Based Procedure for Meta-Analysis of Cancer Susceptibility Gene Penetrance

PURPOSE

Quantifying the risk of cancer associated with pathogenic mutations in germline cancer susceptibility genes-that is, penetrance-enables the personalization of preventive management strategies. Conducting a meta-analysis is the best way to obtain robust risk estimates. We have previously developed a natural language processing (NLP) -based abstract classifier which classifies abstracts as relevant to penetrance, prevalence of mutations, both, or neither. In this work, we evaluate the performance of this NLP-based procedure.

MATERIALS AND METHODS

We compared the semiautomated NLP-based procedure, which involves automated abstract classification and text mining, followed by human review of identified studies, with the traditional procedure that requires human review of all studies. Ten high-quality gene-cancer penetrance meta-analyses spanning 16 gene-cancer associations were used as the gold standard by which to evaluate the performance of our procedure. For each meta-analysis, we evaluated the number of abstracts that required human review (workload) and the ability to identify the studies that were included by the authors in their quantitative analysis (coverage).

RESULTS

Compared with the traditional procedure, the semiautomated NLP-based procedure led to a lower workload across all 10 meta-analyses, with an overall 84% reduction (2,774 abstracts v 16,941 abstracts) in the amount of human review required. Overall coverage was 93%-we are able to identify 132 of 142 studies-before reviewing references of identified studies. Reasons for the 10 missed studies included blank and poorly written abstracts. After reviewing references, nine of the previously missed studies were identified and coverage improved to 99% (141 of 142 studies).

CONCLUSION

We demonstrated that an NLP-based procedure can significantly reduce the review workload without compromising the ability to identify relevant studies. NLP algorithms have promising potential for reducing human efforts in the literature review process.

Challenges and opportunities for public health made possible by advances in natural language processing

Natural language processing (NLP) is a subfield of artificial intelligence devoted to understanding and generation of language. The recent advances in NLP technologies are enabling rapid analysis of vast amounts of text, thereby creating opportunities for health research and evidence-informed decision making. The analysis and data extraction from scientific literature, technical reports, health records, social media, surveys, registries and other documents can support core public health functions including the enhancement of existing surveillance systems (e.g. through faster identification of diseases and risk factors/at-risk populations), disease prevention strategies (e.g. through more efficient evaluation of the safety and effectiveness of interventions) and health promotion efforts (e.g. by providing the ability to obtain expert-level answers to any health related question). NLP is emerging as an important tool that can assist public health authorities in decreasing the burden of health inequality/inequity in the population. The purpose of this paper is to provide some notable examples of both the potential applications and challenges of NLP use in public health.

MMiDaS-AE: Multi-modal Missing Data aware Stacked Autoencoder for Biomedical Abstract Screening

Systematic review (SR) is an essential process to identify, evaluate, and summarize the findings of all relevant individual studies concerning health-related questions. However, conducting a SR is labor-intensive, as identifying relevant studies is a daunting process that entails multiple researchers screening thousands of articles for relevance. In this paper, we propose MMiDaS-AE, a Multi-modal Missing Data aware Stacked Autoencoder, for semi-automating screening for SRs. We use a multi-modal view that exploits three representations, of: 1) documents, 2) topics, and 3) citation networks. Documents that contain similar words will be nearby in the document embedding space. Models can also exploit the relationship between documents and the associated SR MeSH terms to capture article relevancy. Finally, related works will likely share the same citations, and thus closely related articles would, intuitively, be trained to be close to each other in the embedding space. However, using all three learned representations as features directly result in an unwieldy number of parameters. Thus, motivated by recent work on multi-modal auto-encoders, we adopt a multi-modal stacked autoencoder that can learn a shared representation encoding all three representations in a compressed space. However, in practice one or more of these modalities may be missing for an article (e.g., if we cannot recover citation information). Therefore, we propose to learn to impute the shared representation even when specific inputs are missing. We find this new model significantly improves performance on a dataset consisting of 15 SRs compared to existing approaches.

Measuring the impact of screening automation on meta-analyses of diagnostic test accuracy

BACKGROUND

The large and increasing number of new studies published each year is making literature identification in systematic reviews ever more time-consuming and costly. Technological assistance has been suggested as an alternative to the conventional, manual study identification to mitigate the cost, but previous literature has mainly evaluated methods in terms of recall (search sensitivity) and workload reduction. There is a need to also evaluate whether screening prioritization methods leads to the same results and conclusions as exhaustive manual screening. In this study, we examined the impact of one screening prioritization method based on active learning on sensitivity and specificity estimates in systematic reviews of diagnostic test accuracy.

METHODS

We simulated the screening process in 48 Cochrane reviews of diagnostic test accuracy and re-run 400 meta-analyses based on a least 3 studies. We compared screening prioritization (with technological assistance) and screening in randomized order (standard practice without technology assistance). We examined if the screening could have been stopped before identifying all relevant studies while still producing reliable summary estimates. For all meta-analyses, we also examined the relationship between the number of relevant studies and the reliability of the final estimates.

RESULTS

The main meta-analysis in each systematic review could have been performed after screening an average of 30% of the candidate articles (range 0.07 to 100%). No systematic review would have required screening more than 2308 studies, whereas manual screening would have required screening up to 43,363 studies. Despite an average 70% recall, the estimation error would have been 1.3% on average, compared to an average 2% estimation error expected when replicating summary estimate calculations.

CONCLUSION

Screening prioritization coupled with stopping criteria in diagnostic test accuracy reviews can reliably detect when the screening process has identified a sufficient number of studies to perform the main meta-analysis with an accuracy within pre-specified tolerance limits. However, many of the systematic reviews did not identify a sufficient number of studies that the meta-analyses were accurate within a 2% limit even with exhaustive manual screening, i.e., using current practice.

Evaluation of an automatic article selection method for timelier updates of the Comet Core Outcome Set database

Curated databases of scientific literature play an important role in helping researchers find relevant literature, but populating such databases is a labour intensive and time-consuming process. One such database is the freely accessible Comet Core Outcome Set database, which was originally populated using manual screening in an annually updated systematic review. In order to reduce the workload and facilitate more timely updates we are evaluating machine learning methods to reduce the number of references needed to screen. In this study we have evaluated a machine learning approach based on logistic regression to automatically rank the candidate articles. Data from the original systematic review and its four first review updates were used to train the model and evaluate performance. We estimated that using automatic screening would yield a workload reduction of at least 75% while keeping the number of missed references around 2%. We judged this to be an acceptable trade-off for this systematic review, and the method is now being used for the next round of the Comet database update.