Exact Matching Research Articles

Effectiveness of R-CHOP versus R-CHOEP for treatment of young patients with high-risk diffuse large B-cell lymphoma: A Danish observational population-based study.

Etoposide to standard R-CHOP is used for high-risk diffuse large B-cell lymphoma (DLBCL) in some countries. Due to the lack of randomized trials, a real-world data study using matching methods was used to test the potential effectiveness of R-CHOEP over R-CHOP. This study included patients from the Danish Lymphoma Register diagnosed between 2006 and 2020 at the age of 18-60 years with de novo DLBCL and age-adjusted IPI ≥2. R-CHOEP treated patients were matched 1:1 without replacement to R-CHOP treated patients using a hybrid exact and genetic matching technique. Primary endpoints were progression-free survival (PFS) and overall survival (OS). In total, 396 patients were included; 213 received R-CHOEP and 183 received R-CHOP. Unadjusted 5-year PFS and OS for R-CHOEP were 69% (95% Confidence intervals [CI]; 63%-76%) and 79% (CI;73%-85%) versus 62% (CI;55%-70%) and 76% (CI;69%-82%) for R-CHOP (log-rank test, PFS p = .25 and OS p = .31). A total of 127 patients treated with R-CHOEP were matched to 127 patients treated with R-CHOP. Matching-adjusted 5-year PFS and OS were 65% (CI; 57%-74%) and 79% (CI; 72%-84%) for R-CHOEP versus 63% (CI; 55%-73%) and 79% (CI;72%-87%) for R-CHOP (log-rank test, PFS p = .90 and OS p = .63). The present study did not confirm superiority of R-CHOEP over R-CHOP for young patients with high-risk DLBCL.

Recognition of target domain Japanese speech using language model replacement

End-to-end (E2E) automatic speech recognition (ASR) models, which consist of deep learning models, are able to perform ASR tasks using a single neural network. These models should be trained using a large amount of data; however, collecting speech data which matches the targeted speech domain can be difficult, so speech data is often used that is not an exact match to the target domain, resulting in lower performance. In comparison to speech data, in-domain text data is much easier to obtain. Thus, traditional ASR systems use separately trained language models and HMM-based acoustic models. However, it is difficult to separate language information from an E2E ASR model because the model learns both acoustic and language information in an integrated manner, making it very difficult to create E2E ASR models for specialized target domain which are able to achieve sufficient recognition performance at a reasonable cost. In this paper, we propose a method of replacing the language information within pre-trained E2E ASR models in order to achieve adaptation to a target domain. This is achieved by deleting the “implicit” language information contained within the ASR model by subtracting the source-domain language model trained with a transcription of the ASR’s training data in a logarithmic domain. We then integrate a target domain language model through addition in the logarithmic domain. This subtraction and addition to replace of the language model is based on Bayes’ theorem. In our experiment, we first used two datasets of the Corpus of Spontaneous Japanese (CSJ) to evaluate the effectiveness of our method. We then we evaluated our method using the Japanese Newspaper Article Speech (JNAS) and CSJ corpora, which contain audio data from the read speech and spontaneous speech domain, respectively, to test the effectiveness of our proposed method at bridging the gap between these two language domains. Our results show that our proposed language model replacement method achieved better ASR performance than both non-adapted (baseline) ASR models and ASR models adapted using the conventional Shallow Fusion method.

Question Answering System Based on Bideirectional Long-Short-Term Memory (Bilstm)

In the modern world, Q&A systems are essential for promoting better communication between people and technology. These systems play an important role in collecting information quickly and efficiently, and this leads to great progress in learning, teaching and development in many areas of life. Using deep learning techniques, this research addresses the problem of excellent prediction of the questions that need to be answered. We created a question-and-answer system using "Bidirectional Long Short-Term Memory (BiLSTM)", a modern neural network known for its accuracy and results in text analysis and natural language understanding. This technique is more effective in understanding questions and producing very accurate answers because of its special ability to pay attention to preceding and following information in a sentence. Preprocessing was used to remove unnecessary, unimportant and time-consuming data. The "Stanford Question Answering Dataset version 2 (SQuAD 2.0)"was used, which is considered one of the important datasets used in the field of machine learning and natural language processing. The following evaluation metrics were used to evaluate the model’s performance: “Mean Average Precision(MAP), Mean Reciprocal Rank(MRR), Recall, Precision, Loss, F1 Score, and Exact Match (EM). The results, based on 150 epochs (EPOGs) and 128 batch sizes with a cleaned dataset split into 70% training and 30% test/validation (15% each), are as follows:"Precision (0.966), Loss (0.591), F1-score (0.966), Recall (0.967), EM (0.967), MRR(0.918), MAP (0.776), and accuracy(0.966)". Interestingly, the highest performance was observed when using the accuracy measure.

The effect of chronic kidney disease on short-term single-level lumbar fusion outcomes

IntroductionChronic kidney disease (CKD) has an increasing global prevalence and has previously been associated with increased complications and morbidity after spine surgery. Understanding the isolated effect of CKD on short-term patient outcomes is critical for optimizing perioperative risk management and healthcare utilization. ObjectiveThe aim of this study is to utilize coarsened exact matching (CEM) to analyze the isolated effect of CKD on short-term patient outcomes in single-level posterior lumbar fusion surgery. MethodsA retrospective analysis of 4680 consecutive patients undergoing single-level, posterior-only lumbar fusion was performed. Univariate logistic regression comparing the odds of outcomes in patients with CKD (n=40) to patients without medical comorbidities (n=2329) was performed. CEM was then employed to match patients with CKD to those without any comorbidities 1:1 on ten patient characteristics known to affect neurosurgical outcomes. Primary outcomes included intraoperative complications, length of stay, discharge disposition, and 30-day Emergency Department (ED) visits, readmissions, reoperations, and mortality. ResultsIn a univariate logistic regression, CKD was associated with increased risk of 30-day ED visits (OR=3.53, p=0.003) but not complication, discharge disposition, or 30-day readmissions or reoperations. Between otherwise exactly matched patients (n=72), CKD similarly remained associated with an increased risk of 30-day ED visits (OR=7.00, p=0.034) and not with other outcomes. ConclusionBetween otherwise exactly matched patients undergoing single-level posterior lumbar fusion, CKD was related to increased risk of 30-day ED utilization but not other markers indicative of inferior surgical outcomes. Further study must investigate the reasons for increased ED visitation and implement risk-mitigation strategies for these patients.

Fine-Grained Recognition of Mixed Signals with Geometry Coordinate Attention.

With the advancement of technology, signal modulation types are becoming increasingly diverse and complex. The phenomenon of signal time-frequency overlap during transmission poses significant challenges for the classification and recognition of mixed signals, including poor recognition capabilities and low generality. This paper presents a recognition model for the fine-grained analysis of mixed signal characteristics, proposing a Geometry Coordinate Attention mechanism and introducing a low-rank bilinear pooling module to more effectively extract signal features for classification. The model employs a residual neural network as its backbone architecture and utilizes the Geometry Coordinate Attention mechanism for time-frequency weighted analysis based on information geometry theory. This analysis targets multiple-scale features within the architecture, producing time-frequency weighted features of the signal. These weighted features are further analyzed through a low-rank bilinear pooling module, combined with the backbone features, to achieve fine-grained feature fusion. This results in a fused feature vector for mixed signal classification. Experiments were conducted on a simulated dataset comprising 39,600 mixed-signal time-frequency plots. The model was benchmarked against a baseline using a residual neural network. The experimental outcomes demonstrated an improvement of 9% in the exact match ratio and 5% in the Hamming score. These results indicate that the proposed model significantly enhances the recognition capability and generalizability of mixed signal classification.

Rapid assessment of heavy metal accumulation capability of Sedum alfredii using hyperspectral imaging and deep learning

Hyperaccumulators are the material basis and key to the phytoremediation of heavy metal contaminated soils. Conventional methods for screening hyperaccumulators are highly dependent on the time- and labor-consuming sampling and chemical analysis. In this study, a novel spectral approach assisted with multi-task deep learning was proposed to streamline accumulating ecotype screening, heavy metal stress discrimination, and heavy metals quantification in plants. The significant Cd/Zn co-hyperaccumulator Sedum alfredii and its non-accumulating ecotype were stressed by Cd, Zn, and Pb. Spectral images of leaves were rapidly acquired by hyperspectral imaging. The self-designed deep learning architecture was composed of a shallow network (ENet) for accumulating ecotype identification, and a multi-task network (HMNet) for heavy metal stress type and accumulation prediction simultaneously. To further assess the robustness of the networks, they were compared with conventional machine learning models (i.e., partial least squares (PLS) and support vector machine (SVM)) on a series of evaluation metrics of classification, multi-label classification, and regression. S. alfredii with heavy metals accumulation capability was identified by ENet with 100 % accuracy. HMNet reduced overfitting and outperformed machine learning models with the average exact match ratio (EMR) of heavy metal stress discrimination increased by 7.46 %, and residual prediction deviations (RPD) of heavy metal concentrations prediction increased by 53.59 %. The method succeeded in rapidly and accurately discriminating heavy metal stress with EMRs over 91 % and accuracies over 96 %, and in predicting heavy metals accumulation with an average RPD of 3.29 for Zn, 2.57 for Cd, and 2.53 for Pb, indicating the satisfactory practicability and potential for sensing heavy metals accumulation. This study provides a relatively novel spectral method to facilitate hyperaccumulator screening and heavy metals accumulation prediction in the phytoremediation process.

Rapid credit expansion and firm behavior: A case study from Türkiye

This study examines the impact of a sudden and remarkable credit expansion experienced in Türkiye during the first half of 2022 on firm behavior by utilizing a novel dataset containing the universe of all incorporated firms in Türkiye. The results of the combination of coarsened exact matching and difference-in-differences methodology show that, amid the credit expansion coupled with the deterioration in expectations of inflation and depreciation in local currency, the firms with higher credit usage tended to engage in alternative channels rather than undertaking real economic activities in short term, such as pulling-forward imports and input demand and taking position against local currency to gain financial profit more than those with less credit usage during the first half of 2022. While SMEs with higher credit usage increased their imports, domestic input purchases, foreign currency purchases, and foreign currency-denominated debt settlement before maturity, large firms increased their foreign currency purchases relative to their counterparts with less credit usage.

Sudden weather disasters as triggers for ethnic protest in autocracies?

Under what conditions do disastrous storms and floods provoke anti-government protests in autocracies? While grievances, as they often occur after disastrous weather events, are necessary preconditions for protests, they are by no means sufficient. This is particularly true in undemocratic states, where protesters typically fear repressive sanctions. Therefore, this essay hypothesizes that disastrous storms and floods primarily result in anti-government mobilization in autocracies when they occur in the homelands of marginalized ethnic groups. In such scenarios, the immediate hardships caused by the disaster come on top of underlying more structural grievances. Moreover, groups with shared ethnic identities are more likely to mobilize for collective action. This argument is tested using quantitative cross-national protest-event data from the Mass Mobilization in Autocracies Database (MMAD). The results of numerous regression analyses on a sample pre-processed with Coarsened Exact Matching to address potential concerns of endogeneity yield robust support for the outlined hypothesis.

Directors tenure diversity and corporate sustainability performance: The non-linear evidence from Indonesia public listed companies

The primary objective of this study is to thoroughly investigate the association between director tenure diversity and corporate sustainability performance. This study utilizes a sample comprising 578 firm-year observations from non-financial companies listed on the Indonesia Stock Exchange. To test the hypothesis, the study employs the Ordinary Least Squares method, complemented by a series of endogeneity tests. This study reveals that the sustainability performance of corporations in Indonesia falls significantly short of satisfactory levels. Furthermore, the study indicates that there is a negative association between tenure diversity and sustainability performance, demonstrating a U-shaped curve pattern. To ensure the robustness of our findings, we performed additional analysis using coarsened exact matching and Heckman (1979) two-stage least square methodologies, confirming that the results remained consistent with those of the initial test. Intriguingly, our supplementary analyses also revealed an inverse association between tenure diversity in the boardroom and sustainability performance within companies. This study makes a significant contribution to the corporate governance literature by elucidating the inverse association between director tenure diversity and sustainability performance. In doing so, it enhances the originality and novelty of existing studies, particularly within the context of developing countries, such as Indonesia. This study exhibits novelty by embracing a quantitative approach to measure sustainability performance, revealing an intriguing inverse association between sustainability performance and ESG initiatives within the companies.

The Impact of Preoperative Spinal Injection Timing on the Postoperative Complications of Lumbar Fusion

ObjectiveTo determine the impact of ESIs on postoperative surgical complications. MethodsThis retrospective all-payer database analysis identified 202,181 adult patients undergoing one- to three-level transforaminal lumbar interbody fusion (TLIF) from 2010 to 2020. 1:1 exact matching on comorbidities and demographics was performed, creating two cohorts: (1) patients who received an ESI within 90 days of surgery and (2) patients who did not receive an ESI. The primary outcome was surgical complication rates between groups at 30 days postoperatively. For the secondary outcome, patients were stratified based on injection time before surgery: 1-30, 31-45, 46-60, 61-75, and 76-90 days. Logistic regression was performed between groups to identify temporal associations of complication rates. The p-value was set to 0.05 for the primary analysis, and the Bonferroni correction was utilized for the secondary outcome. ResultsExact matching produced 12,491 pairs for analysis. Groups were well-matched on demographics, comorbidities, and fusion levels. 30-day postoperative rates of surgical complications, hematomas, wound disruptions, or surgical site infections did not differ between groups (p>0.05). The rate of CSF leak was increased in the ESI group (0.19% vs. 0.09%, p=0.042). When temporally stratified, patients receiving an ESI within 30 days had significantly higher odds of CSF leak (OR: 4.24, 95% CI: 1.97-9.14). ConclusionsPatients who receive an ESI within 30 days of TLIF are at an increased risk for CSF leak. While the incidence of CSF leak remains small, it may be advisable to avoid ESIs at least 30 days before surgery for certain patients.

Automated description generation for software patches

Software patches are pivotal in refining and evolving codebases, addressing bugs, vulnerabilities, and optimizations. Patch descriptions provide detailed accounts of changes, aiding comprehension and collaboration among developers. However, manual description creation poses challenges in terms of time consumption and variations in quality and detail. In this paper, we propose PatchExplainer, an approach that addresses these challenges by framing patch description generation as a machine translation task. In PatchExplainer, we leverage explicit representations of critical elements, historical context, and syntactic conventions. Moreover, the translation model in PatchExplainer is designed with an awareness of description similarity. Particularly, the model is explicitly trained to recognize and incorporate similarities present in patch descriptions clustered into groups, improving its ability to generate accurate and consistent descriptions across similar patches. The dual objectives maximize similarity and accurately predict affiliating groups. Our experimental results on a large dataset of real-world software patches show that PatchExplainer consistently outperforms existing methods, with improvements up to 189% in BLEU, 5.7X in Exact Match rate, and 154% in Semantic Similarity, affirming its effectiveness in generating software patch descriptions.

Taxonomic classification with maximal exact matches in KATKA kernels and minimizer digests.

For taxonomic classification, we are asked to index the genomes in a phylogenetic tree such that later, given a DNA read, we can quickly choose a small subtree likely to contain the genome from which that read was drawn. Although popular classifiers such as Kraken use -mers, recent research indicates that using maximal exact matches (MEMs) can lead to better classifications. For example, we can ■ build an augmented FM-index over the the genomes in the tree concatenated in left-to-right order; ■ for each MEM in a read, find the interval in the suffix array containing the starting positions of that MEM's occurrences in those genomes; ■ find the minimum and maximum values stored in that interval; ■ take the lowest common ancestor (LCA) of the genomes containing the characters at those positions. This solution is practical, however, only when the total size of the genomes in the tree is fairly small. In this paper we consider applying the same solution to three lossily compressed representations of the genomes' concatenation: ■ a KATKA kernel, which discards characters that are not in the first or last occurrence of any -tuple, for a parameter ; a minimizer digest; ■ a KATKA kernel of a minimizer digest. With a test dataset and these three representations of it, simulated reads and various parameter settings, we checked how many reads' longest MEMs occurred only in the sequences from which those reads were generated ("true positive" reads). For some parameter settings we achieved significant compression while only slightly decreasing the true-positive rate.

An Online Assessment : Survey Panel

Abstract : The Electronic Test System (ETS) is crafted to streamline the assessment process, allowing examiners to effortlessly create and administer exams. Examiners can incorporate a variety of question types, including multiple-choice and open-ended text responses. The system utilizes advanced techniques to automate the evaluation of assessments, drawing from a database of questions curated by administrators. The ETS boasts intelligent capabilities for evaluating written responses, recognizing not only exact matches but also synonymous terms within the answers. Additionally, the system offers a manual override feature, enabling examiners to manually review and adjust exam results when necessary. With a user-friendly interface, the ETS accommodates examiners, candidates, and super administrators. Users can assume roles as either examiners or candidates. Upon logging in, candidates can view and complete their scheduled assessments from any location with internet access, removing the need for traditional tools like pen and paper. The ETS aims to provide a comprehensive and efficient assessment solution. This fully automated system saves time, delivers quick and accurate results, and allows candidates the flexibility to complete exams conveniently from any location. Keywords: Assessment, Evaluator, Participant, Administrator, Manual Review, Automated Evaluation System

CEO overseas experience and sustainability report disclosure: Evidence from Indonesia

AbstractThis study examines the relationship between chief executive officers' (CEOs) overseas experience and social responsibility (SR) disclosure, as well as how CEOs from reputable universities can moderate this relationship. Using 488 firm‐year observations from firms listed on the Bursa Efek Indonesia (Indonesian Stock Exchange) and data from the Global Reporting Initiative databases from 2010 to 2021, this study finds a significant negative association between foreign‐educated CEOs (CEO overseas experience) and SR disclosure. This study also finds that CEOs from reputable universities can weaken this relationship. The results are robust to endogeneity tests using the coarsened exact match (CEM) and Heckman's two‐stage regressions. Furthermore, overseas experience CEOs have a significantly negative impact on all dimensions of SR disclosure, except for the environmental and product dimensions. Investors may consider CEOs' foreign education as an additional factor in their investment decisions. This study suggests that firms with CEOs with international exposure tend to exhibit lower levels of SR disclosures. This insight can aid investors in evaluating sustainability and ethical performance by aligning their investment portfolios with CSR objectives and preferences.

The intersection between traditional roles and a fragmented labor market: a propensity score matching analysis of gender wage gap in Ecuador

PurposeThis study aims at estimating the gender wage gap in Ecuador, and its evolution over the last decade and a half, exploring its heterogeneity through different working conditions (formal/informal, full employment/underemployment, short term/long term and tenure/no tenure) and workers personal characteristics (education level; age and children).Design/methodology/approachPropensity score matching (PSM) and coarsened exact matching (CEM) are used to examine the gender pay inequality of wage earners in Ecuador, using the National Employment, Unemployment and Underemployment Survey (ENEMDU) data set from 2007 to 2022.FindingsResults show a persistent gender pay gap, evidencing a significant heterogeneity through the different dimensions taken into account, in terms of working conditions and workers personal characteristics. The evolution of the pay gap during the years analyzed hardly shows any reduction of differences in earnings between men and women; on the contrary, women exposure to precarious and unregulated jobs seems to be increasing wage inequality.Practical implicationsThe results make the case for active policies oriented not only at containing the negative effects of the traditional division of labor within the family but also at improving labor law enforcement, mitigating informality and workers rapid turnover.Originality/valueThis study is one of the few that use matching techniques to study the gender wage gap and the first in Ecuador; the time span taken into account is larger than previous studies, allowing a medium-long run perspective across different economic phases.

The Relationship between Participation in Extracurricular Arts and Sports Activities and Adolescents' Social and Emotional Skills: An Empirical Analysis Based on the OECD Social and Emotional Skills Survey.

In light of the 'double reduction' policy, which affords adolescents increased time for extracurricular pursuits, the strategic organization of these activities' form and content is imperative. Prior research has established a robust correlation between adolescent participation in extracurricular arts and sports and the enhancement in their social and emotional skills. Nevertheless, the relationship between extracurricular arts and sports activities and the various dimensions of social and emotional skills, as well as the connection between participation in different types of these activities and the enhancement in social and emotional skills, requires further investigation. Utilizing the Theory of Multiple Intelligence and data from the OECD-SSES2019 Suzhou (China) student survey, this study employs Ordinary Least Squares (OLS) and coarsened exact matching (CEM) methodologies to address these gaps. In China, participation in extracurricular arts and sports activities is significantly positively associated with various dimensions of social and emotional skills, with a synergistic effect observed between these activities in enhancing these skills. Additionally, this study finds age-related heterogeneity in the relationship between participation in extracurricular arts and sports activities and the improvement in social and emotional skills.

MedT5SQL: a transformers-based large language model for text-to-SQL conversion in the healthcare domain.

In response to the increasing prevalence of electronic medical records (EMRs) stored in databases, healthcare staff are encountering difficulties retrieving these records due to their limited technical expertise in database operations. As these records are crucial for delivering appropriate medical care, there is a need for an accessible method for healthcare staff to access EMRs. To address this, natural language processing (NLP) for Text-to-SQL has emerged as a solution, enabling non-technical users to generate SQL queries using natural language text. This research assesses existing work on Text-to-SQL conversion and proposes the MedT5SQL model specifically designed for EMR retrieval. The proposed model utilizes the Text-to-Text Transfer Transformer (T5) model, a Large Language Model (LLM) commonly used in various text-based NLP tasks. The model is fine-tuned on the MIMICSQL dataset, the first Text-to-SQL dataset for the healthcare domain. Performance evaluation involves benchmarking the MedT5SQL model on two optimizers, varying numbers of training epochs, and using two datasets, MIMICSQL and WikiSQL. For MIMICSQL dataset, the model demonstrates considerable effectiveness in generating question-SQL pairs achieving accuracy of 80.63%, 98.937%, and 90% for exact match accuracy matrix, approximate string-matching, and manual evaluation, respectively. When testing the performance of the model on WikiSQL dataset, the model demonstrates efficiency in generating SQL queries, with an accuracy of 44.2% on WikiSQL and 94.26% for approximate string-matching. Results indicate improved performance with increased training epochs. This work highlights the potential of fine-tuned T5 model to convert medical-related questions written in natural language to Structured Query Language (SQL) in healthcare domain, providing a foundation for future research in this area.

Injecting the score of the first-stage retriever as text improves BERT-based re-rankers

In this paper we propose a novel approach for combining first-stage lexical retrieval models and Transformer-based re-rankers: we inject the relevance score of the lexical model as a token into the input of the cross-encoder re-ranker. It was shown in prior work that interpolation between the relevance score of lexical and Bidirectional Encoder Representations from Transformers (BERT) based re-rankers may not consistently result in higher effectiveness. Our idea is motivated by the finding that BERT models can capture numeric information. We compare several representations of the Best Match 25 (BM25) and Dense Passage Retrieval (DPR) scores and inject them as text in the input of four different cross-encoders. Since knowledge distillation, i.e., teacher-student training, proved to be highly effective for cross-encoder re-rankers, we additionally analyze the effect of injecting the relevance score into the student model while training the model by three larger teacher models. Evaluation on the MSMARCO Passage collection and the TREC DL collections shows that the proposed method significantly improves over all cross-encoder re-rankers as well as the common interpolation methods. We show that the improvement is consistent for all query types. We also find an improvement in exact matching capabilities over both the first-stage rankers and the cross-encoders. Our findings indicate that cross-encoder re-rankers can efficiently be improved without additional computational burden or extra steps in the pipeline by adding the output of the first-stage ranker to the model input. This effect is robust for different models and query types.

Guardian of Image: Evaluating Corporate Reputation as a Shield Against COVID-19 Fallout on Firm Performance

The COVID-19 pandemic has wreaked havoc on the global economy and posed a significant challenge for companies to survive. This research provides empirical evidence on how the pandemic can adversely affect company performance. Furthermore, the study also unravels how a company's reputation can act as a shield amid the onslaught of the pandemic, thereby moderating the negative relationship between COVID-19 and company performance. The research utilizes 2,847 firm-year observations from non-financial companies listed on the Indonesia Stock Exchange from 2017 to 2022. The analysis employed in this study uses moderated regression analysis (MRA) with STATA software. The results indicate that a company's reputation can moderate, weakening the negative relationship between COVID-19 and company performance. These findings are reinforced by robustness testing using Coarsened Exact Matching (CEM) and alternative performance measurements. This information is highly valuable for companies and stakeholders, indicating that the risks arising from the COVID-19 pandemic can be mitigated by strengthening the company's reputation.

Structured peer review: pilot results from 23 Elsevier journals.

Reviewers rarely comment on the same aspects of a manuscript, making it difficult to properly assess manuscripts' quality and the quality of the peer review process. The goal of this pilot study was to evaluate structured peer review implementation by: 1) exploring whether and how reviewers answered structured peer review questions, 2) analysing reviewer agreement, 3) comparing that agreement to agreement before implementation of structured peer review, and 4) further enhancing the piloted set of structured peer review questions. Structured peer review consisting of nine questions was piloted in August 2022 in 220 Elsevier journals. We randomly selected 10% of these journals across all fields and IF quartiles and included manuscripts that received two review reports in the first 2 months of the pilot, leaving us with 107 manuscripts belonging to 23 journals. Eight questions had open-ended fields, while the ninth question (on language editing) had only a yes/no option. The reviews could also leave Comments-to-Author and Comments-to-Editor. Answers were independently analysed by two raters, using qualitative methods. Almost all the reviewers (n = 196, 92%) provided answers to all questions even though these questions were not mandatory in the system. The longest answer (Md 27 words, IQR 11 to 68) was for reporting methods with sufficient details for replicability or reproducibility. The reviewers had the highest (partial) agreement (of 72%) for assessing the flow and structure of the manuscript, and the lowest (of 53%) for assessing whether interpretation of the results was supported by data, and for assessing whether the statistical analyses were appropriate and reported in sufficient detail (52%). Two thirds of the reviewers (n = 145, 68%) filled out the Comments-to-Author section, of which 105 (49%) resembled traditional peer review reports. These reports contained a Md of 4 (IQR 3 to 5) topics covered by the structured questions. Absolute agreement regarding final recommendations (exact match of recommendation choice) was 41%, which was higher than what those journals had in the period from 2019 to 2021 (31% agreement, P = 0.0275). Our preliminary results indicate that reviewers successfully adapted to the new review format, and that they covered more topics than in their traditional reports. Individual question analysis indicated the greatest disagreement regarding the interpretation of the results and the conducting and the reporting of statistical analyses. While structured peer review did lead to improvement in reviewer final recommendation agreements, this was not a randomized trial, and further studies should be performed to corroborate this. Further research is also needed to determine whether structured peer review leads to greater knowledge transfer or better improvement of manuscripts.