By Adnan Usmani, Director of Operations, Discover STM Publishing Ltd. Reviewed by the Discover STM editorial team.
Editors now send about 4.5 review invitations to secure one review, and acceptance has roughly halved since 2018. Small medical journals cannot fix that trend, but they can work with it. Invite in parallel with a clear stop rule, make it easy to say no, widen the pool on purpose, and measure the few numbers that show where reviews get stuck.
In short
- Reviewer acceptance fell from 43% in 2018 to 22% in 2024 in ScholarOne data, while invitations more than doubled.
- Small journals feel this more: narrow fields, the same few experts, no fee, and a less familiar name.
- Send short, personal invitations in small parallel batches, and ask every decliner to suggest a colleague.
- Track four numbers each month, starting with invitations per completed review.
What do the 2026 numbers show?
Silverchair's 2026 Future of Peer Review Report, published in June 2026 (see also the NISO summary), draws on eight years of ScholarOne Manuscripts data, a survey of more than 2,000 reviewers, authors and editors, and expert interviews. The main findings:
- The reviewer pool grew 54% between 2018 and 2025, but invitations grew by more than 100%.
- Acceptance of review invitations fell from 43% in 2018 to 22% in 2024.
- Editors send about 4.5 invitations to secure one review, nearly double the 2018 rate.
- For every 100 invitations, editors wait a combined 407 days on reviewers who decline or never respond.
- Acceptance differs by region. Reviewers in East and Southeast Asia are more likely to accept than those in the US, UK and Western Europe.
- 53% of reviewers say they use AI for peer review tasks, mostly without publisher guidance.
The 407 days figure matters most. Much of the delay comes from waiting on people who were never going to say yes.
We see the same trend in our own editorial offices: reviewer acceptance across the journals we support has dropped a lot in recent years.
Why do small medical journals feel it more?
A large journal can lose half its acceptances and still fill reviews from a deep pool. A small society or university journal usually cannot:
- Narrow fields. A subspecialty manuscript may have only a dozen qualified reviewers in the region.
- The same few experts. Those experts also review for larger journals, sit on committees and run clinics. Every journal asks them first.
- No fee. Reviewing is unpaid everywhere, but larger brands offer more visible prestige in return.
- An unfamiliar journal name. A reviewer who does not know the journal may treat the email as spam or as low priority.
- Volume from elsewhere. Your invitation arrives in an inbox that already holds several others.
What works in practice?
Keep the invitation short and personal. Name the reviewer and say in one line why you chose them. Include the title, the full abstract, the article type and a clear deadline. Put the accept and decline links near the top. A reviewer should be able to decide in under a minute.
Make declining easy and useful. Offer a one-click decline with an optional "suggest a colleague" field. A fast no is worth more to you than silence. Suggested colleagues are often more willing than cold contacts, because a peer named them.
Invite in parallel, with a stop rule. Inviting one reviewer at a time and waiting for each reply adds to those 407 days. In our editorial offices we usually invite five reviewers at a time for a manuscript, and give them seven to nine days to accept or decline before we invite the next group. Decide in advance what happens when enough reviewers agree, for example thanking the others and withdrawing their invitations politely. Write the rule down so every handling editor uses it.
Use author-suggested reviewers, with checks. Author suggestions can be useful in narrow fields, but they need verification. Check that the email address is institutional, not a free webmail account. Match the name to an ORCID record and to recent publications. Look for shared affiliations, recent co-authorship and shared grants with the authors. Never rely on an author-supplied email alone. Fake reviewer identities are a known route to manipulated peer review. Always use at least one reviewer the editor selected independently.
Widen the pool by region and career stage. If acceptance is higher in East and Southeast Asia, it is worth adding qualified reviewers from those regions for suitable topics. Career stage matters too. In November 2026 Nature Portfolio published its first large-scale analysis of its Early Career Researcher Co-Review programme, with an accompanying Nature editorial. In this programme, an invited reviewer brings in a PhD candidate or postdoctoral researcher and keeps responsibility for the final report. The analysis covered 78,661 reviewers and 24,645 research manuscripts at Nature Communications (April 2024 to July 2025) and Nature (February to July 2025). Early career researchers made up around 17% of reviewers at Nature and around 11% at Nature Communications, and more than a quarter of manuscripts included at least one early career reviewer. Women made up around 40% of early career reviewers, against 24% of established reviewers. Among established reviewers who took part, 88% said they are more likely to review if they can co-review with an early career researcher; the report does not give the number of people surveyed. It also found no evidence that co-review slows peer review: turnaround times were comparable. What the report does not yet show is whether co-reviewers go on to accept independent invitations, so treat co-review as a promising route rather than a proven fix.
For a medical journal, a simple version works: invite a senior reviewer and offer the option to co-review with a registrar, fellow or postdoctoral researcher they supervise. Both names go on the record, and both receive credit. The senior reviewer remains responsible for the report.
Build a reliable core group, and protect it. Most small journals depend on a few people who always say yes. That is a risk. Set a limit, for example no more than one review per person in a set period, and give a rest period after each completed review. Keep a short list of reviewers who are ready but under-used, and invite them first.
Be clear about AI. With more than half of reviewers using AI, state your policy in the invitation, including manuscript confidentiality. Do not rely on AI detectors to police it. They are not reliable.
What recognition costs nothing?
- Certificates for each completed review, sent automatically with the thank-you email.
- An annual thank-you list published on the journal website, with the reviewer's consent.
- ORCID review credit, which adds a verified review record without revealing the manuscript.
- Letters for promotion files, signed by the editor-in-chief, on request, for academic promotion or appraisal files.
What should you track each month?
Four numbers are enough to see where reviews get stuck:
- Invitations per completed review. Compare your figure with the 4.5 benchmark, keeping in mind that it comes from a much larger system.
- Days to first agreement. The time from first invitation to the first reviewer saying yes. This shows whether your batches and response windows work.
- Late reviews. The share of agreed reviews that pass the deadline. A rising figure suggests your core group is overloaded.
- Share of new reviewers. The share of completed reviews from people who have not reviewed for you before. If it stays near zero, the pool is not growing.
Change one thing at a time, so you can see what helped.
How can an editorial office help?
A virtual editorial office can take on the routine work: sending invitation batches, chasing replies, applying the stop rule, checking reviewer identities and conflicts, issuing certificates and keeping the monthly figures. Editors keep every editorial decision, including who reviews and what happens to the manuscript. The aim is to give editors more time for judgement and less time for email.
FAQ
How many reviewers should we invite at once?
We usually invite five at a time and allow seven to nine days for a reply. Adjust the number once you know your own invitations per completed review.
Are author-suggested reviewers safe to use?
They can be, if you verify identity through institutional email and ORCID, check for conflicts, and always add at least one reviewer the editor chose independently.
Can early career researchers review for a medical journal?
Yes, as co-reviewers working with an established reviewer who supervises and takes responsibility for the report. Credit both names.
Should we ban reviewers from using AI?
That is the editors' decision. Whatever you choose, state it clearly in the reviewer guidelines and protect manuscript confidentiality. Do not rely on AI detectors to enforce it.
Sources
- NISO. "New Report Highlights Trends and Challenges in Peer Review." NISO I/O, June 2026. https://www.niso.org/niso-io/2026/06/new-report-highlights-trends-and-challenges-peer-review (accessed 19 November 2026)
- Silverchair. "The Numbers Behind the Noise: What Our 2026 Peer Review Report Shows." 2026. https://www.silverchair.com/news/the-numbers-behind-the-noise/ (accessed 19 November 2026)
- "Open up peer review to early-career researchers." Nature 659, 561, editorial, 17 November 2026. https://www.nature.com/articles/d41586-026-03707-7 (accessed 19 November 2026)
- Nature Portfolio. "Widening participation in peer review: How Nature Portfolio journals are supporting early career researchers." 2026. Executive summary, findings and methodology appendix. https://www.nature.com/immersive/ecr-participation-report/index.html (read in full, 19 November 2026)
About Discover STM Publishing Ltd
Discover STM Publishing Ltd, based in Cork, Ireland, has provided publishing services for scholarly journals since 2012, including journal hosting, a virtual editorial office and production. We publish and host nine journals for societies, universities and publishers, and have managed more than 4,500 manuscripts with more than 2,000 reviewers. Editors make all editorial decisions. To talk about your journal's reviewer workflow, write to [email protected].