PubMed, read for you.

How a review runs here

One real example, start to finish: exercise therapy for knee osteoarthritis pain.

1. 1

   Ask a question

   Type your research question. We draft the PubMed search and the inclusion criteria; you edit both.

   Does exercise therapy reduce knee osteoarthritis pain?

   "Osteoarthritis, Knee"[mh] AND ("Exercise Therapy"[mh] OR exercise*[tiab] …) AND randomi*ed[tiab]

   1,052 records found

2. 2

   Screen abstracts

   Every title and abstract is scored 0–100 against your criteria, with a one-line reason. It pauses at a checkpoint so you can check before the rest.

   90

   Kept — RCT in knee OA adults; aerobic exercise vs control; reports VAS/WOMAC pain.

   5

   Set aside — Narrative review, not a primary trial — fails the design criterion.

   51 kept of the first 150

3. 3

   Check full texts

   Open-access full texts are read criterion by criterion; every verdict quotes the paper.

   Adults with knee osteoarthritis

   “…a diagnosis of unilateral clinical knee OA (per the American College of Rheumatology guidelines)…”

   8 included of 19 read

4. 4

   Ask across them

   Ask questions across what you kept. Every claim cites a paper, and its quote is checked word for word.

   Which programmes reduced knee pain?

   A strengthening-based home programme reduced pain vs medication (MD −1.95 on NPRS) 42376359

   4 of 4 quotes found

   16 of 17 quotes found in 3 answers

5. 5

   Find what was missed

   PubMed's similar-article and citation links from your included studies surface papers the search missed; screen them with the same criteria.

   Related to your included studies

   • Similar to 42216405
   • Cites 42016045
   • Cited by 41985188

   6.4× the inclusion rate of the search, top 20

Latest literature