Morning evidence scan
Impact Signals
Daily PubMed evidence triage for people who need the point fast: what changed, who it could help, how strong the evidence looks, and which benefit–cost assumptions need verification.
Loaded studies0Recent feed in memory
Visible now0After your filters
Rank logicMCDAPlus on-device clicks
Research triage, not medical advice
Impact Signals summarizes and ranks studies so you can decide what deserves deeper review. Do not use summaries, scores, or benefit–cost estimates to diagnose, treat, prescribe, or change care without reading the full study and consulting qualified professionals.
What to verify
- Full-text methods, setting, eligibility, and outcomes.
- Risk of bias, harms, conflicts, funding, and replication.
- Whether benefit–cost metrics are source-derived or assumptions-based.
Research triage only — not medical advice. Open a study to review source metadata, certainty, bias, harms, equity, benefit–cost assumptions, source links, comments, and a print-ready PDF.
Loading latest studies… cached results will appear first when available.
About Impact Signals
Impact Signals scans new PubMed studies, ranks findings for triage, and explains the practical signal in plain English — who may benefit, how strong the evidence looks, what harms or equity issues to check, and which benefit–cost assumptions need verification.
How ranking worksSignal score and MCDA fields are triage aids that weigh relevance, effect, certainty, rigor, equity/harms, implementation, reproducibility, and timeliness when available.
What certainty meansCertainty, study design, sample size, risk of bias, funding, conflicts, registration, and replication should be checked before drawing conclusions.
Benefit–cost lensBCR, net-benefit, horizon, and discount-rate fields are assumptions-based unless explicitly source-derived. Treat them as prompts for deeper economic review.
This board is not medical advice, clinical guidance, diagnosis, treatment recommendation, or policy approval. Verify every finding in PubMed or the full text before applying it.