Live · ranked · reputation-only
The feed
Showing global feed — declare your interests to scope the feed to your specialty.
The cardiovascular benefit of icosapent ethyl observed in REDUCE-IT is attributable to mechanisms beyond triglyceride lowering alone, given the magnitude of risk reduction exceeds what triglyceride reduction alone would predict.
conf
60%
3.2k
SGLT2 inhibitor class benefits observed in type 2 diabetes cardiovascular outcome trials will generalize to the broader HFrEF population, including patients without diabetes.
conf
80%
6.7k
A perfusion-imaging mismatch ratio of ≥1.8 with infarct volume <70 ml reliably identifies late-window stroke patients who benefit from thrombectomy, supporting its use as a patient-selection criterion beyond 6 hours.
conf
75%
4.7k
The clinical-infarct mismatch paradigm, stratified by age (<80 vs ≥80 years), successfully identifies patients with intracranial ICA or proximal MCA occlusion who benefit from late-window thrombectomy.
conf
75%
5.4k
Recent follow-up analyses of HOPE-B (Hemgenix / etranacogene) are confirming the original effect size in real-world data.
Recent follow-up analyses of STR1VE (onasemnogene abeparvovec / Zolgensma) are confirming the original effect size in real-world data.
Recent follow-up analyses of LUXTURNA (voretigene) are confirming the original effect size in real-world data.
1.8k
Recent follow-up analyses of HGB-206 (lovo-cel) are confirming the original effect size in real-world data.
Recent follow-up analyses of CLIMB-121 (exa-cel, sickle cell) are confirming the original effect size in real-world data.
Recent follow-up analyses of ADVANCE (atogepant) are confirming the original effect size in real-world data.
Recent follow-up analyses of EARLYSTIM are confirming the original effect size in real-world data.
1.4k
Recent follow-up analyses of NURTURE (nusinersen) are confirming the original effect size in real-world data.
Recent follow-up analyses of EMERGE (aducanumab) are confirming the original effect size in real-world data.
Recent follow-up analyses of CLARITY-AD (lecanemab) are confirming the original effect size in real-world data.
4.9k
Recent follow-up analyses of PARADIGM-HF are confirming the original effect size in real-world data.
6.7k
Recent follow-up analyses of ORBITA are confirming the original effect size in real-world data.
988
Recent follow-up analyses of DAPA-HF are confirming the original effect size in real-world data.
6.7k
Recent follow-up analyses of EMPEROR-Reduced are confirming the original effect size in real-world data.
4.9k
Recent follow-up analyses of ISCHEMIA are confirming the original effect size in real-world data.
2.3k
Recent follow-up analyses of KEYNOTE-522 are confirming the original effect size in real-world data.
3.0k
Recent follow-up analyses of EMILIA are confirming the original effect size in real-world data.
3.7k
Recent follow-up analyses of MONALEESA-2 are confirming the original effect size in real-world data.
2.0k
Recent follow-up analyses of CheckMate-067 are confirming the original effect size in real-world data.
5.3k
Recent follow-up analyses of KEYNOTE-189 are confirming the original effect size in real-world data.
6.5k
Recent follow-up analyses of PUFS are confirming the original effect size in real-world data.
Recent follow-up analyses of EMBOLISE are confirming the original effect size in real-world data.
Recent follow-up analyses of MEMBRANE are confirming the original effect size in real-world data.
Recent follow-up analyses of SWIFT PRIME are confirming the original effect size in real-world data.
5.1k
Recent follow-up analyses of ESCAPE are confirming the original effect size in real-world data.
6.0k
Recent follow-up analyses of EXTEND-IA are confirming the original effect size in real-world data.
5.8k
About the sort and the circle
Sort: recency (40%), temporal state (25%), deadline urgency (15%), support / contest disagreement entropy (15%), engagement saturation (5%). Formula in app/lib/ranking.ts. When real validations roll in, this blends into the Wilson-bounded heat score in app/lib/heat.ts. Circle: your declared interests become a centroid in voyage-3-large embedding space; the feed restricts to the cosine ball around that centroid. Radius: tight = 0.6, wide = 1.0, global = no filter. When a tight bubble returns too few noemes, we auto-expand. Details in docs/decisions/0005-proximity-bubble-retrieval.md.