CER literature reviews an auditor can follow, every cycle

Anja leads clinical evidence for an implant portfolio, owning the literature reviews behind Clinical Evaluation Reports and recurring post-market surveillance under EU MDR.

MDR and MEDDEV 2.7/1 Rev 4 require a pre-specified protocol, documented search, criteria fixed before searching, a documented reason for every exclusion, and both favorable and unfavorable data appraised from full text. Across a database, a spreadsheet, and a Word document, each review was slow and the audit trail was rebuilt by hand.

Paperguide runs the systematic method in one workspace and makes the documentation a byproduct, so every decision is logged with its criterion and evidence and the audit trail is complete without a separate write-up.

Every clinical evaluation cycle, Anja’s team faces the same question about a portfolio implant, most recently a drug-eluting coronary stent: what does the real-world literature say about its safety and performance, and can she prove, paper by paper, why each study was in or out? Under MDR, that proof is not optional. The regulation expects a documented reason for every exclusion and demands that unfavorable findings be weighed alongside favorable ones.

She sets up the review’s protocol once, and the screening criteria encode exactly what MDR asks for: does the study report clinical safety or performance data, is the device or a valid equivalent in scope, and, critically, does it capture adverse outcomes and failure modes rather than only positive results? As papers are screened against those criteria, any study that fails carries the specific criterion it failed, in the order she set, so the exclusion log she used to assemble by hand at the end now writes itself as the review runs. For the studies that make it through, she extracts the fields her CER needs, device or equivalent, study design, sample size, follow-up, endpoints, adverse events, effect estimates, into a structured table where every value links back to the exact sentence in the source PDF. When a Notified Body reviewer later asks why a particular study was excluded, or where a specific adverse-event figure came from, the answer is already in the review: the criterion, the evidence, and the person who confirmed it. Papers behind paywalls aren’t a dead end either: open-access PDFs are fetched automatically, and for paywalled papers the team sets an institutional proxy URL to retrieve full text, so screening picks them back up.

The payoff shows up when the review is audited. Every exclusion carries its criterion and the evidence behind it, and every decision is logged with the person who confirmed it, so when an auditor asks why a study was left out, the answer is already on record rather than reconstructed after the fact, which is exactly the defensibility a Notified Body wants to see.