Label-free, low-cost AI ranking for systematic review screening

Abstract

Title and abstract screening is among the most labor-intensive steps of systematic reviews. Jev, a low-cost AI decision model, scores records against eligibility criteria without screening labels. We scored 169,982 records from 128 reviews and, after tuning on development reviews, evaluated 23 held-out reviews once and 31 Cochrane reviews post hoc. Jev outranked the best free ranker in all 23 held-out reviews (mean AUC 0.966 vs 0.831) and came within 0.011 AUC of Claude Opus 5.5 on 1999 records at about 1/700 of its list-price cost (US$0.02/1000 records). Combined with active learning, it raised work saved at 95% recall from 0.659 to 0.806. A fixed score threshold fell below 95% recall in 3 of 28 Cochrane reviews, whereas Jev ranking with statistical stopping (post hoc) reached 95% recall in all 51 reviews, reading 86-88% of records. Label-free ranking with Jev was accurate and inexpensive, but stopping safely required screening decisions.

Off-label use patterns of glucagon-like peptide-1 receptor agonists: a systematic review
UF Ofodile, JC Nwadigo, CC Aghalu
The European Journal of Public Health1 Oct 2026Peer-reviewedCC BY-NC 4.0Full textPharmacology & toxicologyPMCEurope PMC
Factors Influencing Adherence to Mammography Screening Guidelines Among Women From Diverse Cultural Backgrounds: A Systematic Review
Antonia Christoforidou, Marianna Drakopoulou, Alexandra Mantoudi +5
Cureus16 Oct 2026Peer-reviewedCC BYFull textGeneral medicineEurope PMC
Screening for rheumatoid arthritis-associated interstitial lung disease using lung ultrasound: a systematic review and quality appraisal of recent studies
A. D. Copoeru, M. Mocan, A. Farcas +2
medRxiv2 Nov 2026PreprintAll rights reservedFull textGeneral medicinemedRxiv
Artificial intelligence-based tongue image analysis for non-invasive screening of gastric cancer, precancerous lesions, and related gastrointestinal disorders: a systematic review
Rukai Wang, Chunlin Li, Hengyan Zhang +3
Frontiers in medicine16 Oct 2026Peer-reviewedCC BYFull textGeneral medicineEurope PMC
Hepatitis C screening strategies and linkage-to-care approaches in U.S. emergency departments: A systematic review and meta-analysis
Pilar Hernandez-Con, Amanda Joseph, Shreya Dhungana +5
Preventive medicine reports6 Oct 2026Peer-reviewedCC BY-NC-NDFull textGeneral medicineEurope PMC

Cite

BibTeX RIS

BibTeX

@misc{jo2026labelfree,
  title = {{Label-free, low-cost AI ranking for systematic review screening}},
  author = {T. Jo and D. Lee and L. Lee},
  howpublished = {medRxiv},
  year = {2026},
  month = oct,
  doi = {10.64898/2026.10.09.26365173},
  url = {https://www.medrxiv.org/content/10.64898/2026.10.09.26365173v1},
  note = {CC BY},
}

RIS

TY  - UNPB
TI  - Label-free, low-cost AI ranking for systematic review screening
AU  - T. Jo
AU  - D. Lee
AU  - L. Lee
PY  - 2026
DA  - 2026/10/09
PB  - medRxiv
DO  - 10.64898/2026.10.09.26365173
UR  - https://www.medrxiv.org/content/10.64898/2026.10.09.26365173v1
AB  - Title and abstract screening is among the most labor-intensive steps of systematic reviews. Jev, a low-cost AI decision model, scores records against eligibility criteria without screening labels. We scored 169,982 records from 128 reviews and, after tuning on development reviews, evaluated 23 held-out reviews once and 31 Cochrane reviews post hoc. Jev outranked the best free ranker in all 23 held-out reviews (mean AUC 0.966 vs 0.831) and came within 0.011 AUC of Claude Opus 5.5 on 1999 records at about 1/700 of its list-price cost (US$0.02/1000 records). Combined with active learning, it raised work saved at 95% recall from 0.659 to 0.806. A fixed score threshold fell below 95% recall in 3 of 28 Cochrane reviews, whereas Jev ranking with statistical stopping (post hoc) reached 95% recall in all 51 reviews, reading 86-88% of records. Label-free ranking with Jev was accurate and inexpensive, but stopping safely required screening decisions.
ER  -

Abstract

Related

Cite