The Rigor and Transparency Index Quality Metric for Assessing Biological and Medical Science Methods

Automatic categorization of self-acknowledged limitations in randomized controlled trial publications

OBJECTIVE

Acknowledging study limitations in a scientific publication is a crucial element in scientific transparency and progress. However, limitation reporting is often inadequate. Natural language processing (NLP) methods could support automated reporting checks, improving research transparency. In this study, our objective was to develop a dataset and NLP methods to detect and categorize self-acknowledged limitations (e.g., sample size, blinding) reported in randomized controlled trial (RCT) publications.

METHODS

We created a data model of limitation types in RCT studies and annotated a corpus of 200 full-text RCT publications using this data model. We fine-tuned BERT-based sentence classification models to recognize the limitation sentences and their types. To address the small size of the annotated corpus, we experimented with data augmentation approaches, including Easy Data Augmentation (EDA) and Prompt-Based Data Augmentation (PromDA). We applied the best-performing model to a set of about 12K RCT publications to characterize self-acknowledged limitations at larger scale.

RESULTS

Our data model consists of 15 categories and 24 sub-categories (e.g., Population and its sub-category DiagnosticCriteria). We annotated 1090 instances of limitation types in 952 sentences (4.8 limitation sentences and 5.5 limitation types per article). A fine-tuned PubMedBERT model for limitation sentence classification improved upon our earlier model by about 1.5 absolute percentage points in F1 score (0.821 vs. 0.8) with statistical significance (p<.001). Our best-performing limitation type classification model, PubMedBERT fine-tuning with PromDA (Output View), achieved an F1 score of 0.7, improving upon the vanilla PubMedBERT model by 2.7 percentage points, with statistical significance (p<.001).

CONCLUSION

The model could support automated screening tools which can be used by journals to draw the authors' attention to reporting issues. Automatic extraction of limitations from RCT publications could benefit peer review and evidence synthesis, and support advanced methods to search and aggregate the evidence from the clinical trial literature.

Advocating for Generalizability: Accepting Inherent Variability in Translation of Animal Research Outcomes

Advancing scientific discovery requires investigators to embrace research practices that increase transparency and disclosure about materials, methods, and outcomes. Several research advocacy and funding organizations have produced guidelines and recommended practices to enhance reproducibility through detailed and rigorous research approaches; however, confusion around vocabulary terms and a lack of adoption of suggested practices have stymied successful implementation. Although reproducibility of research findings cannot be guaranteed due to extensive inherent variables in attempts at experimental repetition, the scientific community can advocate for generalizability in the application of data outcomes to ensure a broad and effective impact on the comparison of animals to translation within human research. This report reviews suggestions, based upon work with National Institutes of Health advisory groups, for improving rigor and transparency in animal research through aspects of experimental design, statistical assessment, and reporting factors to advocate for generalizability in the application of comparative outcomes between animals and humans.

Research rigor and reproducibility in research education: A CTSA institutional survey

We assessed the rigor and reproducibility (R&R) activities of institutions funded by the National Center for Advancing Translational Sciences (NCTSA) through a survey and website search (N = 61). Of 50 institutional responses, 84% reported incorporating some form of R&R training, 68% reported devoted R&R training, 30% monitored R&R practices, and 10% incentivized them. Website searches revealed 9 (15%) freely available training curricula, and 7 (11%) institutional programs specifically created to enhance R&R. NCATS should formally integrate R&R principles into its translational science models and institutional requirements.

Do organisms need an impact factor? Citations of key biological resources including model organisms reveal usage patterns and impact

Research resources like transgenic animals and antibodies are the workhorses of biomedicine, enabling investigators to relatively easily study specific disease conditions. As key biological resources, transgenic animals and antibodies are often validated, maintained, and distributed from university based stock centers. As these centers heavily rely largely on grant funding, it is critical that they are cited by investigators so that usage can be tracked. However, unlike systems for tracking the impact of papers, the conventions and systems for tracking key resource usage and impact lag behind. Previous studies have shown that about 50% of the resources are not findable, making the studies they are supporting irreproducible, but also makes tracking resources difficult. The RRID project is filling this gap by working with journals and resource providers to improve citation practices and to track the usage of these key resources. Here, we reviewed 10 years of citation practices for five university based stock centers, characterizing each reference into two broad categories: findable (authors could use the RRID, stock number, or full name) and not findable (authors could use a nickname or a common name that is not unique to the resource). The data revealed that when stock centers asked their communities to cite resources by RRID, in addition to helping stock centers more easily track resource usage by increasing the number of RRID papers, authors shifted from citing resources predominantly by nickname (~50% of the time) to citing them by one of the findable categories (~85%) in a matter of several years. In the case of one stock center, the MMRRC, the improvement in findability is also associated with improvements in the adherence to NIH rigor criteria, as determined by a significant increase in the Rigor and Transparency Index for studies using MMRRC mice. From this data, it was not possible to determine whether outreach to authors or changes to stock center websites drove better citation practices, but findability of research resources and rigor adherence was improved.

A survey of researchers' methods sharing practices and priorities

Missing or inaccessible information about the methods used in scientific research slows the pace of discovery and hampers reproducibility. Yet little is known about how, why, and under what conditions researchers share detailed methods information, or about how such practices vary across social categories like career stage, field, and region. In this exploratory study, we surveyed 997 active researchers about their attitudes and behaviors with respect to methods sharing. The most common approach reported by respondents was private sharing upon request, but a substantial minority (33%) had publicly shared detailed methods information independently of their research findings. The most widely used channels for public sharing were connected to peer-reviewed publications, while the most significant barriers to public sharing were found to be lack of time and lack of awareness about how or where to share. Insofar as respondents were moderately satisfied with their ability to accomplish various goals associated with methods sharing, we conclude that efforts to increase public sharing may wish to focus on enhancing and building awareness of existing solutions-even as future research should seek to understand the needs of methods users and the extent to which they align with prevailing practices of sharing.

Evaluation of Transparency and Openness Guidelines in Physical Therapist Journals

OBJECTIVE

The goals of this study were to evaluate the extent that physical therapist journals support open science research practices by adhering to the Transparency and Openness Promotion (TOP) guidelines and to assess the relationship between journal scores and their respective journal impact factor (JIF).

METHODS

Scimago, mapping studies, the National Library of Medicine, and journal author guidelines were searched to identify physical therapist journals for inclusion. Journals were graded on 10 standards (29 available total points) related to transparency with data, code, research materials, study design and analysis, preregistration of studies and statistical analyses, replication, and open science badges. The relationship between journal transparency and openness scores and their JIF was determined.

RESULTS

Thirty-five journals' author guidelines were assigned transparency and openness factor scores. The median score (interquartile range) across journals was 3.00 out of 29 (3.00) points (for all journals the scores ranged from 0 to 8). The 2 standards with the highest degree of implementation were design and analysis transparency (reporting guidelines) and study preregistration. No journals reported on code transparency, materials transparency, replication, and open science badges. TOP factor scores were a significant predictor of JIF scores.

CONCLUSION

There is low implementation of the TOP standards by physical therapist journals. TOP factor scores demonstrated predictive abilities for JIF scores. Policies from journals must improve to make open science practices the standard in research. Journals are in an influential position to guide practices that can improve the rigor of publication which, ultimately, enhances the evidence-based information used by physical therapists.

IMPACT

Transparent, open, and reproducible research will move the profession forward by improving the quality of research and increasing the confidence in results for implementation in clinical care.

Open Science 2.0: Towards a truly collaborative research ecosystem

Conversations about open science have reached the mainstream, yet many open science practices such as data sharing remain uncommon. Our efforts towards openness therefore need to increase in scale and aim for a more ambitious target. We need an ecosystem not only where research outputs are openly shared but also in which transparency permeates the research process from the start and lends itself to more rigorous and collaborative research. To support this vision, this Essay provides an overview of a selection of open science initiatives from the past 2 decades, focusing on methods transparency, scholarly communication, team science, and research culture, and speculates about what the future of open science could look like. It then draws on these examples to provide recommendations for how funders, institutions, journals, regulators, and other stakeholders can create an environment that is ripe for improvement.

Methodology reporting improved over time in 176,469 randomized controlled trials

OBJECTIVES

To describe randomized controlled trial (RCT) methodology reporting over time.

STUDY DESIGN AND SETTING

We used a deep learning-based sentence classification model based on the Consolidated Standards of Reporting Trials (CONSORT) statement, considered minimum requirements for reporting RCTs. We included 176,469 RCT reports published between 1966 and 2018. We analyzed the reporting trends over 5-year time periods, grouping trials from 1966 to 1990 in a single stratum. We also explored the effect of journal impact factor (JIF) and medical discipline.

RESULTS

Population, Intervention, Comparator, Outcome (PICO) items were commonly reported during each period, and reporting increased over time (e.g., interventions: 79.1% during 1966-1990 to 87.5% during 2010-2018). Reporting of some methods information has increased, although there is room for improvement (e.g., sequence generation: 10.8-41.8%). Some items are reported infrequently (e.g., allocation concealment: 5.1-19.3%). The number of items reported and JIF are weakly correlated (Pearson's r (162,702) = 0.16, P < 0.001). The differences in the proportion of items reported between disciplines are small (<10%).

CONCLUSION

Our analysis provides large-scale quantitative support for the hypothesis that RCT methodology reporting has improved over time. Extending these models to all CONSORT items could facilitate compliance checking during manuscript authoring and peer review, and support metaresearch.

Auto-STEED: A data mining tool for automated extraction of experimental parameters and risk of bias items from in vivo publications

Background

Systematic reviews, i.e., research summaries that address focused questions in a structured and reproducible manner, are a cornerstone of evidence-based medicine and research. However, certain systematic review steps such as data extraction are labour-intensive which hampers their applicability, not least with the rapidly expanding body of biomedical literature.

Objective

To bridge this gap, we aimed at developing a data mining tool in the R programming environment to automate data extraction from neuroscience in vivo publications. The function was trained on a literature corpus (n=45 publications) of animal motor neuron disease studies and tested in two validation corpora (motor neuron diseases, n=31 publications; multiple sclerosis, n=244 publications).

Results

Our data mining tool Auto-STEED (Automated and STructured Extraction of Experimental Data) was able to extract key experimental parameters such as animal models and species as well as risk of bias items such as randomization or blinding from in vivo studies. Sensitivity and specificity were over 85 and 80%, respectively, for most items in both validation corpora. Accuracy and F-scores were above 90% and 0.9 for most items in the validation corpora. Time savings were above 99%.

Conclusions

Our developed text mining tool Auto-STEED is able to extract key experimental parameters and risk of bias items from the neuroscience in vivo literature. With this, the tool can be deployed to probe a field in a research improvement context or to replace one human reader during data extraction resulting in substantial time-savings and contribute towards automation of systematic reviews. The function is available on Github.

Evaluating LC-HRMS metabolomics data processing software using FAIR principles for research software

BACKGROUND

Liquid chromatography-high resolution mass spectrometry (LC-HRMS) is a popular approach for metabolomics data acquisition and requires many data processing software tools. The FAIR Principles - Findability, Accessibility, Interoperability, and Reusability - were proposed to promote open science and reusable data management, and to maximize the benefit obtained from contemporary and formal scholarly digital publishing. More recently, the FAIR principles were extended to include Research Software (FAIR4RS).

AIM OF REVIEW

This study facilitates open science in metabolomics by providing an implementation solution for adopting FAIR4RS in the LC-HRMS metabolomics data processing software. We believe our evaluation guidelines and results can help improve the FAIRness of research software.

KEY SCIENTIFIC CONCEPTS OF REVIEW

We evaluated 124 LC-HRMS metabolomics data processing software obtained from a systematic review and selected 61 software for detailed evaluation using FAIR4RS-related criteria, which were extracted from the literature along with internal discussions. We assigned each criterion one or more FAIR4RS categories through discussion. The minimum, median, and maximum percentages of criteria fulfillment of software were 21.6%, 47.7%, and 71.8%. Statistical analysis revealed no significant improvement in FAIRness over time. We identified four criteria covering multiple FAIR4RS categories but had a low %fulfillment: (1) No software had semantic annotation of key information; (2) only 6.3% of evaluated software were registered to Zenodo and received DOIs; (3) only 14.5% of selected software had official software containerization or virtual machine; (4) only 16.7% of evaluated software had a fully documented functions in code. According to the results, we discussed improvement strategies and future directions.

Reproducibility and rigor in rheumatology research

The pillars of scientific progress in rheumatology are experimentation and observation, followed by the publication of reliable and credible results. These data must then be independently verified, validated, and replicated. Peer and journal-specific technical and statistical reviews are paramount to improving rigor and reproducibility. In addition, research integrity, ethics, and responsible conduct training can help to reduce research misconduct and improve scientific evidence. As the number of published articles in rheumatology grows, the field has become critical for determining reproducibility. Prospective, longitudinal, randomized controlled clinical trials are the gold standard for evaluating clinical intervention efficacy and safety in this space. However, their applicability to larger, more representative patient populations with rheumatological disorders worldwide could be limited due to time, technical, and cost constraints involved with large-scale clinical trials. Accordingly, analysis of real-world, patient-centered clinical data retrieved from established healthcare inventories, such as electronic health records, medical billing reports, and disease registries, are increasingly used to report patient outcomes. Unfortunately, it is unknown whether this clinical research paradigm in rheumatology could be deployed in medically underserved regions.

A qualitative study of the barriers to using blinding in in vivo experiments and suggestions for improvement

In animal experiments, blinding (also known as masking) is a methodological strategy to reduce the risk that scientists, animal care staff, or other staff involved in the research may consciously or subconsciously influence the outcome. Lack of masking has been shown to correlate with an overestimation of treatment efficacy and false positive findings. We conducted exploratory interviews across academic and a commercial setting to discuss the implementation of masking at four stages of the experiment: during allocation and intervention, during the conduct of the experiment, during the outcome assessment, and during the data analysis. The objective was to explore the awareness, engagement, perceptions, and the barriers to implementing masking in animal experiments. We conducted multiple interviews, to explore 30 different experiments, and found examples of excellent practice but also areas where masking was rarely implemented. Significant barriers arose from the operational and informatic systems implemented. These systems have prioritised the management of welfare without considering how to allow researchers to use masking in their experiments. For some experiments, there was a conflict between the management of welfare for an individual animal versus delivering a robust experiment where all animals are treated in the same manner. We identified other challenges related to the level of knowledge on the purpose of masking or the implementation and the work culture. The exploration of these issues provides insight into how we, as a community, can identify the most significant barriers in a given research environment. Here, we offer practical solutions to enable researchers to implement masking as standard. To move forward, we need both the individual scientists to embrace the use of masking and the facility managers and institutes to engage and provide a framework that supports the scientists.

Rigor, Reproducibility, and Transparency in Shared Research Resources: Follow-Up Survey and Recommendations for Improvements

Rigor, reproducibility, and transparency (RR&T) are essential components of all scientific pursuits. Shared research resources, also known as core facilities, are on the frontlines of ensuring robust RR&T practices. The Association of Biomolecular Resource Facilities Committee on Core Rigor and Reproducibility conducted a follow-up survey 4 years after the initial 2017 survey to determine if core facilities have seen a positive impact of new RR&T initiatives (including guidance from the National Institutes of Health, new scientific journal requirements on transparency and data provenance, and educational tools from professional organizations). While there were fewer participants in the most recent survey, the respondents' opinions on the role of core facilities and level of best practices adoption remained the same. Overall, the respondents agreed that procedures should be implemented by core facilities to ensure scientific RR&T. They also indicated that there is a strong correlation between institutions that emphasize RR&T and core customers using this expertise in grant applications and publications. The survey also assessed the impact of the COVID-19 pandemic on core operations and RR&T. The answers to these pandemic-related questions revealed that many of the strategies aimed at increasing efficiencies are also best practices related to RR&T, including the development of standard operating procedures, supply chain management, and cross training. Given the consistent and compelling awareness of the importance of RR&T expressed by core directors in 2017 and 2021 contrasted with the lack of apparent improvements over this time period, the authors recommend an adoption of RR&T statements by all core laboratories. Adhering to the RR&T guidelines will result in more efficient training, better compliance, and improved experimental approaches empowering cores to become "rigor champions."

Planning experiments: Updated guidance on experimental design and analysis and their reporting III

Scientists who plan to publish in British Journal of Pharmacology (BJP) must read this article before undertaking a study. This editorial provides guidance for the design of experiments. We have published previously two guidance documents on experimental design and analysis (Curtis et al., 2015; Curtis et al., 2018). This update clarifies and simplifies the requirements on design and analysis for BJP manuscripts. This editorial also details updated requirements following an audit and discussion on best practice by the BJP editorial board. Explanations for the requirements are provided in the previous articles. Here, we address new issues that have arisen in the course of handling manuscripts and emphasise three aspects of design that continue to present the greatest challenge to authors: randomisation, blinded analysis and balance of group sizes.

Establishing Institutional Scores With the Rigor and Transparency Index: Large-scale Analysis of Scientific Reporting Quality

BACKGROUND

Improving rigor and transparency measures should lead to improvements in reproducibility across the scientific literature; however, the assessment of measures of transparency tends to be very difficult if performed manually.

OBJECTIVE

This study addresses the enhancement of the Rigor and Transparency Index (RTI, version 2.0), which attempts to automatically assess the rigor and transparency of journals, institutions, and countries using manuscripts scored on criteria found in reproducibility guidelines (eg, Materials Design, Analysis, and Reporting checklist criteria).

METHODS

The RTI tracks 27 entity types using natural language processing techniques such as Bidirectional Long Short-term Memory Conditional Random Field-based models and regular expressions; this allowed us to assess over 2 million papers accessed through PubMed Central.

RESULTS

Between 1997 and 2020 (where data were readily available in our data set), rigor and transparency measures showed general improvement (RTI 2.29 to 4.13), suggesting that authors are taking the need for improved reporting seriously. The top-scoring journals in 2020 were the Journal of Neurochemistry (6.23), British Journal of Pharmacology (6.07), and Nature Neuroscience (5.93). We extracted the institution and country of origin from the author affiliations to expand our analysis beyond journals. Among institutions publishing >1000 papers in 2020 (in the PubMed Central open access set), Capital Medical University (4.75), Yonsei University (4.58), and University of Copenhagen (4.53) were the top performers in terms of RTI. In country-level performance, we found that Ethiopia and Norway consistently topped the RTI charts of countries with 100 or more papers per year. In addition, we tested our assumption that the RTI may serve as a reliable proxy for scientific replicability (ie, a high RTI represents papers containing sufficient information for replication efforts). Using work by the Reproducibility Project: Cancer Biology, we determined that replication papers (RTI 7.61, SD 0.78) scored significantly higher (P<.001) than the original papers (RTI 3.39, SD 1.12), which according to the project required additional information from authors to begin replication efforts.

CONCLUSIONS

These results align with our view that RTI may serve as a reliable proxy for scientific replicability. Unfortunately, RTI measures for journals, institutions, and countries fall short of the replicated paper average. If we consider the RTI of these replication studies as a target for future manuscripts, more work will be needed to ensure that the average manuscript contains sufficient information for replication attempts.

Is the future of peer review automated?

The rising rate of preprints and publications, combined with persistent inadequate reporting practices and problems with study design and execution, have strained the traditional peer review system. Automated screening tools could potentially enhance peer review by helping authors, journal editors, and reviewers to identify beneficial practices and common problems in preprints or submitted manuscripts. Tools can screen many papers quickly, and may be particularly helpful in assessing compliance with journal policies and with straightforward items in reporting guidelines. However, existing tools cannot understand or interpret the paper in the context of the scientific literature. Tools cannot yet determine whether the methods used are suitable to answer the research question, or whether the data support the authors' conclusions. Editors and peer reviewers are essential for assessing journal fit and the overall quality of a paper, including the experimental design, the soundness of the study's conclusions, potential impact and innovation. Automated screening tools cannot replace peer review, but may aid authors, reviewers, and editors in improving scientific papers. Strategies for responsible use of automated tools in peer review may include setting performance criteria for tools, transparently reporting tool performance and use, and training users to interpret reports.

Meta-analysis on reporting practices as a source of heterogeneity in in vitro cancer research

Objectives

Heterogeneity of results of exact same research experiments oppose a significant socioeconomic burden. Insufficient methodological reporting is likely to be one of the contributors to results heterogeneity; however, little knowledge on reporting habits of in vitro cancer research and their effects on results reproducibility is available. Exemplified by a commonly performed in vitro assay, we aim to fill this knowledge gap and to derive recommendations necessary for reproducible, robust and translational preclinical science.

Methods

Here, we use systematic review to describe reporting practices in in vitro glioblastoma research using the Uppsala-87 Malignant Glioma (U-87 MG) cell line and perform multilevel random-effects meta-analysis followed by meta-regression to explore sources of heterogeneity within that literature, and any associations between reporting characteristics and reported findings. Literature that includes experiments measuring the effect of temozolomide on the viability of U-87 MG cells is searched on three databases (Embase, PubMed and Web of Science).

Results

In 137 identified articles, the methodological reporting is incomplete, for example, medium glucose level and cell density are reported in only 21.2% and 16.8% of the articles. After adjustments for different drug concentrations and treatment durations, the results heterogeneity across the studies (I²=68.5%) is concerningly large. Differences in culture medium glucose level are a driver of this heterogeneity. However, infrequent reporting of most experimental parameters limits the analysis of reproducibility moderating parameters.

Conclusions

Our results further support the ongoing efforts of establishing consensus reporting practices to elevate durability of results. By doing so, this work can raise awareness of how stricter reporting may help to improve the frequency of successful translation of preclinical results into human application. The authors received no specific funding for this work. A preregistered protocol is available at the Open Science Framework (https://osf.io/9k3dq).

Reproducibility, relevance and reliability as barriers to efficient and credible biomedical technology translation

Biomedical research accuracy and relevance for improving healthcare are increasingly identified as costly problems. Basic research data quality, reporting and methodology, and reproducibility are common factors implicated in this challenge. Preclinical models of disease and therapy, largely conducted in rodents, have known deficiencies in replicating most human conditions. Their translation to human results is acknowledged to be poor for decades. Clinical data quality and quantity is also recognized as deficient; gold standard randomized clinical trials are expensive. Few solid conclusions from clinical studies are replicable and many remain unpublished. The translational pathway from fundamental biomedical research through to innovative solutions handed to clinical practitioners is therefore highly inefficient and costly in terms of wasted resources, early claims from fundamental discoveries never witnessed in humans, and few new, improved solutions available clinically for myriad diseases. Improving this biomedical research strategy and resourcing for reliability, translational relevance, reproducibility and clinical impact requires careful analysis and consistent enforcement at both funding and peer review levels.

Misidentified biomedical resources: Journal guidelines are not a quick fix

Biomedical researchers routinely use a variety of biological models and resources, such as cultured cell lines, antibodies and laboratory animals. Unfortunately, these resources are not flawless: cell lines can be misidentified; for antibodies, problems with specificity, lot‐to‐lot consistency and sensitivity are common; and the reliability of animal models is questioned due to poor translation of animal studies to human clinical trials. In some cases, these problems can render the results of a study meaningless. As a response, some journals have implemented guidelines regarding the use and reporting of cell lines, antibodies and laboratory animals. In our study we use a portfolio of existing and newly created datasets to investigate identification and authentication information of cell lines, antibodies and organisms before and after guideline introduction, compared to journals without guidelines. We observed a general improvement of reporting quality over time, which the implementation of guidelines accelerated only in some cases. We therefore conclude that the effectiveness of journal guidelines is likely to be context dependent, affected by factors such as implementation conditions, research community support and monitoring and resource availability. Hence, journal reporting guidelines in themselves are not a quick fix to repair shortcomings in biomedical resource documentation, even though they can be part of the solution.

Risk of bias assessment in preclinical literature using natural language processing

We sought to apply natural language processing to the task of automatic risk of bias assessment in preclinical literature, which could speed the process of systematic review, provide information to guide research improvement activity, and support translation from preclinical to clinical research. We use 7840 full‐text publications describing animal experiments with yes/no annotations for five risk of bias items. We implement a series of models including baselines (support vector machine, logistic regression, random forest), neural models (convolutional neural network, recurrent neural network with attention, hierarchical neural network) and models using BERT with two strategies (document chunk pooling and sentence extraction). We tune hyperparameters to obtain the highest F1 scores for each risk of bias item on the validation set and compare evaluation results on the test set to our previous regular expression approach. The F1 scores of best models on test set are 82.0% for random allocation, 81.6% for blinded assessment of outcome, 82.6% for conflict of interests, 91.4% for compliance with animal welfare regulations and 46.6% for reporting animals excluded from analysis. Our models significantly outperform regular expressions for four risk of bias items. For random allocation, blinded assessment of outcome, conflict of interests and animal exclusions, neural models achieve good performance; for animal welfare regulations, BERT model with a sentence extraction strategy works better. Convolutional neural networks are the overall best models. The tool is publicly available which may contribute to the future monitoring of risk of bias reporting for research improvement activities.

Converting Access Microbiology to an open research platform: focus group and AI review tool research results

The Microbiology Society will be launching an open research platform in October 2021. Developed using funding from the Wellcome Trust and the Howard Hughes Medical Institute (HHMI), the platform will combine our current sound-science journal, Access Microbiology, with artificial intelligence (AI) review tools and many of the elements of a preprint server. In an effort to improve the rigour, reproducibility and transparency of the academic record, the Access Microbiology platform will host both preprints of articles and their Version of Record (VOR) publications, as well as the reviewer reports, Editor's decision, authors' response to reviewers and the AI review reports. To ensure the platform meets the needs of our community, in February 2020 we conducted focus group meetings with various stakeholders. Using articles previously submitted to Access Microbiology, we undertook testing of a range of potential AI review tools and investigated the technical feasibility and utility of including these tools as part of the platform. In keeping with the open and transparent ethos of the platform, we present here a summary of the focus group feedback and AI review tool testing.

Toward assessing clinical trial publications for reporting transparency

OBJECTIVE

To annotate a corpus of randomized controlled trial (RCT) publications with the checklist items of CONSORT reporting guidelines and using the corpus to develop text mining methods for RCT appraisal.

METHODS

We annotated a corpus of 50 RCT articles at the sentence level using 37 fine-grained CONSORT checklist items. A subset (31 articles) was double-annotated and adjudicated, while 19 were annotated by a single annotator and reconciled by another. We calculated inter-annotator agreement at the article and section level using MASI (Measuring Agreement on Set-Valued Items) and at the CONSORT item level using Krippendorff's α. We experimented with two rule-based methods (phrase-based and section header-based) and two supervised learning approaches (support vector machine and BioBERT-based neural network classifiers), for recognizing 17 methodology-related items in the RCT Methods sections.

RESULTS

We created CONSORT-TM consisting of 10,709 sentences, 4,845 (45%) of which were annotated with 5,246 labels. A median of 28 CONSORT items (out of possible 37) were annotated per article. Agreement was moderate at the article and section levels (average MASI: 0.60 and 0.64, respectively). Agreement varied considerably among individual checklist items (Krippendorff's α= 0.06-0.96). The model based on BioBERT performed best overall for recognizing methodology-related items (micro-precision: 0.82, micro-recall: 0.63, micro-F1: 0.71). Combining models using majority vote and label aggregation further improved precision and recall, respectively.

CONCLUSION

Our annotated corpus, CONSORT-TM, contains more fine-grained information than earlier RCT corpora. Low frequency of some CONSORT items made it difficult to train effective text mining models to recognize them. For the items commonly reported, CONSORT-TM can serve as a testbed for text mining methods that assess RCT transparency, rigor, and reliability, and support methods for peer review and authoring assistance. Minor modifications to the annotation scheme and a larger corpus could facilitate improved text mining models. CONSORT-TM is publicly available at https://github.com/kilicogluh/CONSORT-TM.

CODECHECK: an Open Science initiative for the independent execution of computations underlying research articles during peer review to improve reproducibility

The traditional scientific paper falls short of effectively communicating computational research.  To help improve this situation, we propose a system by which the computational workflows underlying research articles are checked. The CODECHECK system uses open infrastructure and tools and can be integrated into review and publication processes in multiple ways. We describe these integrations along multiple dimensions (importance, who, openness, when). In collaboration with academic publishers and conferences, we demonstrate CODECHECK with 25 reproductions of diverse scientific publications. These CODECHECKs show that asking for reproducible workflows during a collaborative review can effectively improve executability. While CODECHECK has clear limitations, it may represent a building block in Open Science and publishing ecosystems for improving the reproducibility, appreciation, and, potentially, the quality of non-textual research artefacts. The CODECHECK website can be accessed here: https://codecheck.org.uk/.

CODECHECK: an Open Science initiative for the independent execution of computations underlying research articles during peer review to improve reproducibility

The traditional scientific paper falls short of effectively communicating computational research.  To help improve this situation, we propose a system by which the computational workflows underlying research articles are checked. The CODECHECK system uses open infrastructure and tools and can be integrated into review and publication processes in multiple ways. We describe these integrations along multiple dimensions (importance, who, openness, when). In collaboration with academic publishers and conferences, we demonstrate CODECHECK with 25 reproductions of diverse scientific publications. These CODECHECKs show that asking for reproducible workflows during a collaborative review can effectively improve executability. While CODECHECK has clear limitations, it may represent a building block in Open Science and publishing ecosystems for improving the reproducibility, appreciation, and, potentially, the quality of non-textual research artefacts. The CODECHECK website can be accessed here: https://codecheck.org.uk/.

Assessment of transparency indicators across the biomedical literature: How open is open?

Recent concerns about the reproducibility of science have led to several calls for more open and transparent research practices and for the monitoring of potential improvements over time. However, with tens of thousands of new biomedical articles published per week, manually mapping and monitoring changes in transparency is unrealistic. We present an open-source, automated approach to identify 5 indicators of transparency (data sharing, code sharing, conflicts of interest disclosures, funding disclosures, and protocol registration) and apply it across the entire open access biomedical literature of 2.75 million articles on PubMed Central (PMC). Our results indicate remarkable improvements in some (e.g., conflict of interest [COI] disclosures and funding disclosures), but not other (e.g., protocol registration and code sharing) areas of transparency over time, and map transparency across fields of science, countries, journals, and publishers. This work has enabled the creation of a large, integrated, and openly available database to expedite further efforts to monitor, understand, and promote transparency and reproducibility in science.