Additional file 9. Details and results for applying our model on a dialysis-related dataset. Table S1. Number of relevant articles retrieved with and without machine learning algorithm using the demonstration dataset (n=882 records). The research objective was to review CRTs in the hemodialysis setting to report key methodological and ethical issues.

Additional file 4: Description of the gamma and c parameters. Fig. S2. This figure shows the decision boundaries and support vectors for different settings of the regularization parameter (C) and Kernel coefficient (gamma). We used the mglearn package to create this figure [27].

Supplemental material, sj-docx-2-psg-10.1177_22925503231161066 for Randomized Controlled Trial Comparing the Clinical Effectiveness of Collagenase Injection (Xiaflex®) and Palmar Fasciectomy in the Management of Dupuytren's Contracture by Achilles Thoma, Jessica Murphy and Lucas Gallo, Bimpe Ayeni, Lehana Thabane in Plastic Surgery

Abstract Background Cluster randomized trials (CRTs) are becoming an increasingly important design. However, authors of CRTs do not always adhere to requirements to explicitly identify the design as cluster randomized in titles and abstracts, making retrieval from bibliographic databases difficult. Machine learning algorithms may improve their identification and retrieval. Therefore, we aimed to develop machine learning algorithms that accurately determine whether a bibliographic citation is a CRT report. Methods We trained, internally validated, and externally...

Additional file 1. Checklist for Reporting Results of Internet E-Surveys (CHERRIES)

Additional file 2. Additional details for the external dataset.

Additional file 4: Description of the gamma and c parameters. Fig. S2. This figure shows the decision boundaries and support vectors for different settings of the regularization parameter (C) and Kernel coefficient (gamma). We used the mglearn package to create this figure [27].

Additional file 1: Table S1. Search strategy for EMBASE and MEDLINE. Table S2. The journals searched for dental speciality.

Additional file 1. Checklist for Reporting Results of Internet E-Surveys (CHERRIES)

Additional file 2. Sepsis Survey Development and Format

Additional file 3. Sepsis Knowledge Question Scoring Key

Additional file 4. Awareness of Sepsis Additional Results. Table S1. Distribution of responses for items related to awareness of sepsis. Figure S1. Regional differences in awareness of sepsis and interest in learning about sepsis

Additional file 5. Knowledge of Sepsis Additional Results. Figure S1. Sepsis awareness, perceived knowledge, and evaluated knowledge by respondent characteristics. Table S1. Distribution of responses for items related to knowledge of sepsis. Figure S2. Distribution of responses seeking medical support when ill with symptoms of sepsis

Additional file 10. Examples of titles and abstracts that were classified incorrectly by our algortihms.

Additional file 1. Justification for using a CRT search filter.

Additional file 5. Continuous skip-gram architecture.

Additional file 9. Details and results for applying our model on a dialysis-related dataset. Table S1. Number of relevant articles retrieved with and without machine learning algorithm using the demonstration dataset (n=882 records). The research objective was to review CRTs in the hemodialysis setting to report key methodological and ethical issues.

Supplemental material, sj-docx-3-psg-10.1177_22925503231161066 for Randomized Controlled Trial Comparing the Clinical Effectiveness of Collagenase Injection (Xiaflex®) and Palmar Fasciectomy in the Management of Dupuytren's Contracture by Achilles Thoma, Jessica Murphy and Lucas Gallo, Bimpe Ayeni, Lehana Thabane in Plastic Surgery

Additional file 3. Sepsis Knowledge Question Scoring Key

Additional file 6. Distribution of responses for items related to sepsis information access

Additional file 1. Justification for using a CRT search filter.

Additional file 5. Continuous skip-gram architecture.

Additional file 10. Examples of titles and abstracts that were classified incorrectly by our algortihms.

Additional file 3: Fig. S1. A general architecture for a convolutional neural work used for text classification.

Additional file 7: Fig. S3. Probability plot for the CRTs in the first external data classified as a CRT (Figure A, 665 CRTs) and non-CRTs classified as a CRT (Figure B, 1251 non-CRTs). The x-axis depicts the stacked ensemble model's prediction of the article being classified as a CRT. The y-axis represents the proportion of all documents that had the corresponding probability.