Considerations for conducting systematic reviews: evaluating the performance of different methods for de-duplicating references
Bibliographic record
Abstract
BACKGROUND: Systematic reviews involve searching multiple bibliographic databases to identify eligible studies. As this type of evidence synthesis is increasingly pursued, the use of various electronic platforms can help researchers improve the efficiency and quality of their research. We examined the accuracy and efficiency of commonly used electronic methods for flagging and removing duplicate references during this process. METHODS: A heterogeneous sample of references was obtained by conducting a similar topical search in MEDLINE, Embase, Cochrane Central Register of Controlled Trials, and PsycINFO databases. References were de-duplicated via manual abstraction to create a benchmark set. The default settings were then used in Ovid multifile search, EndNote desktop, Mendeley, Zotero, Covidence, and Rayyan to de-duplicate the sample of references independently. Using the benchmark set as reference, the number of false-negative and false-positive duplicate references for each method was identified, and accuracy, sensitivity, and specificity were determined. RESULTS: We found that the most accurate methods for identifying duplicate references were Ovid, Covidence, and Rayyan. Ovid and Covidence possessed the highest specificity for identifying duplicate references, while Rayyan demonstrated the highest sensitivity. CONCLUSION: This study reveals the strengths and weaknesses of commonly used de-duplication methods and provides strategies for improving their performance to avoid unintentionally removing eligible studies and introducing bias into systematic reviews. Along with availability, ease-of-use, functionality, and capability, these findings are important to consider when researchers are selecting database platforms and supporting software programs for conducting systematic reviews.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Direct model labels (unvalidated)
Per-model category and study-design labels from the labeling rounds. They are machine output, unvalidated, and the disagreement between models ships as data. No study design here is MEDLINE-validated yet.
| Model arm | Categories | Study design | Confidence |
|---|---|---|---|
| gemma | Metaresearch Domain: Methods · Genre: Empirical About the Canadian research system: no · About a Canadian topic: no | Simulation or modeling | low |
| gpt | Metaresearch Domain: Methods · Genre: Empirical About the Canadian research system: no · About a Canadian topic: no | Other design | high |
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.869 | 0.959 |
| Meta-epidemiology (narrow) | 0.006 | 0.007 |
| Meta-epidemiology (broad) | 0.014 | 0.018 |
| Bibliometrics | 0.041 | 0.039 |
| Science and technology studies | 0.008 | 0.014 |
| Scholarly communication | 0.022 | 0.024 |
| Open science | 0.012 | 0.010 |
| Research integrity | 0.011 | 0.007 |
| Insufficient payload (model declined to judge) | 0.004 | 0.002 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedLabeled directly by 2 models reading the full record.
The models disagree on parts of this classification; every voice is preserved in the section at the end of the page.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".