Living systematic review services that keep your evidence current

A conventional systematic review is a snapshot. The search closes, the synthesis is written, and from that day the answer starts drifting away from the literature, because trials keep publishing whether or not anyone is reading them into the review. Studies of review currency have found that a substantial share of published reviews are out of date within a few years, and in fast-moving fields the conclusions can be overturned within months. For a one-off academic question that may not matter much. For a guideline panel, a policy team, or a clinical service relying on the review to make a live decision, it matters enormously, and it is the problem living mode was designed to solve.

The living approach, formalised through Cochrane’s living evidence programme, replaces the publish-and-forget cycle with active surveillance. The searches keep running, the yield is checked on a declared schedule, and the review is updated when the new evidence warrants it, so the published answer and the state of the literature never drift far apart. The rigour of the underlying review does not change; what changes is that currency becomes a maintained property rather than a decaying one.

1. 1

   Build the base review

   A full systematic review to the usual standard: protocol, comprehensive search, dual screening, extraction, risk-of-bias appraisal, and a narrative or pooled synthesis. Living mode is a maintenance plan, not a shortcut.

2. 2

   Declare the surveillance plan

   We specify the surveillance search frequency, usually monthly or quarterly, the sources monitored, and the update triggers, so every future decision to update or hold is pre-committed rather than improvised.

3. 3

   Run scheduled surveillance

   Saved search strategies are re-run on schedule, new records are deduplicated against the existing yield, and eligible studies are screened in duplicate exactly as in the base review.

4. 4

   Update when triggers fire

   When new evidence meets a declared trigger, we extract, appraise, and fold it into the analysis, re-run the synthesis, and issue a new version with the changes documented.

5. 5

   Report every version, retire deliberately

   Each version carries its search date, updated flow counts, and a change log. When the evidence stabilises or the decision is made, living mode is closed formally, not abandoned.

The discipline of a living review lives in its triggers. Checking the literature monthly does not mean republishing monthly; most surveillance cycles find nothing that changes the answer, and issuing a new version for every minor study would exhaust everyone and dilute the signal. We therefore agree, in the protocol for your review, exactly what forces an update: a new eligible trial above a stated size, evidence that shifts a pooled estimate beyond a stated margin, a change in the certainty of evidence, or the appearance of data on a previously empty outcome. Everything below the threshold is logged and held for the next scheduled update.

Cadence follows the field, not a fixed rule. A question with weekly trial activity may justify monthly surveillance and quarterly synthesis updates; a slower literature may need only quarterly checks. We set the cadence from the observed publication rate in your area and state it in the methods, so readers and editors can see the review’s currency promise and hold it to account.

A living review that cannot show what changed between versions is just a review that gets edited. We keep the record set, the extraction tables, and the analysis code under version control, so every version of the review maps to a frozen snapshot of its data and its meta-analysis outputs. Each update produces a change log stating the new search date, the records added, the studies included and excluded with reasons, any revised risk-of-bias judgements, and the effect of the new data on the pooled estimates and the GRADE certainty ratings. Reviewers, panels, and readers can therefore reconstruct any version and see precisely why the conclusions moved, which is the difference between a living review and a moving target.

Reporting is where many living reviews quietly fall apart, because the flow of records is no longer a single funnel but a base yield plus a stream of surveillance additions. We maintain PRISMA 2020 reporting at every version: cumulative flow counts that reconcile across the base search and each surveillance round, updated exclusion logs, and a methods section that states the current search date and cadence. When an editor or panel member asks what the numbers were at any earlier version, the answer is on file rather than reconstructed from memory. Our flow diagram generator rebuilds the diagram from the updated counts at each version, so the reporting never lags the evidence it describes.

Living mode costs sustained effort, so the honest question is whether your review needs it. A conventional full systematic review with a planned later update serves most questions perfectly well, and if the deadline is the dominant constraint a rapid evidence assessment may fit better still. The table sets out the trade-off, and our guide to how long a systematic review takes covers the baseline timelines behind it. Where living mode is not justified, we say so plainly and quote the simpler option instead.

We quote in two parts, both fixed: a fee for the base review, shaped by the breadth of the question and the expected yield from the comprehensive literature search, and a fee per surveillance and update cycle at the agreed cadence. That structure keeps the ongoing commitment predictable and lets you stop living mode at any planned point without renegotiating the original work. Send us your question, your decision context, and how current the answer needs to stay, and we will return a surveillance plan, a cadence recommendation, and a fixed quote for both parts.