Delphi survey on the most promising areas and methods to improve systematic reviews’ production and updating
Bibliographic record
Abstract
BACKGROUND: Systematic reviews (SRs) are invaluable evidence syntheses, widely used in biomedicine and other scientific areas. Tremendous resources are being spent on the production and updating of SRs. There is a continuous need to automatize the process and use the workforce and resources to make it faster and more efficient. METHODS: Information gathered by previous EVBRES research was used to construct a questionnaire for round 1 which was partly quantitative, partly qualitative. Fifty five experienced SR authors were invited to participate in a Delphi study (DS) designed to identify the most promising areas and methods to improve the efficient production and updating of SRs. Topic questions focused on which areas of SRs are most time/effort/resource intensive and should be prioritized in further research. Data were analysed using NVivo 12 plus, Microsoft Excel 2013 and SPSS. Thematic analysis findings were used on the topics on which agreement was not reached in round 1 in order to prepare the questionnaire for round 2. RESULTS: Sixty percent (33/55) of the invited participants completed round 1; 44% (24/55) completed round 2. Participants reported average of 13.3 years of experience in conducting SRs (SD 6.8). More than two thirds of the respondents agreed/strongly agreed the following topics should be prioritized: extracting data, literature searching, screening abstracts, obtaining and screening full texts, updating SRs, finding previous SRs, translating non-English studies, synthesizing data, project management, writing the protocol, constructing the search strategy and critically appraising. Participants have not considered following areas as priority: snowballing, GRADE-ing, writing SR, deduplication, formulating SR question, performing meta-analysis. CONCLUSIONS: Data extraction was prioritized by the majority of participants as an area that needs more research/methods development. Quality of available language translating tools has dramatically increased over the years (Google translate, DeepL). The promising new tool for snowballing emerged (Citation Chaser). Automation cannot substitute human judgement where complex decisions are needed (GRADE-ing). TRIAL REGISTRATION: Study protocol was registered at https://osf.io/bp2hu/ .
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
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.740 | 0.790 |
| Meta-epidemiology (narrow) | 0.003 | 0.001 |
| Meta-epidemiology (broad) | 0.064 | 0.005 |
| Bibliometrics | 0.001 | 0.008 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.004 | 0.000 |
| Open science | 0.004 | 0.001 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.000 | 0.010 |
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, unvalidatedMachine predicted; both teacher heads agree on what is shown here.
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".