Clinical Evidence Assessment
“What does the literature actually support?”
Systematic literature reviews and meta-analyses against a pre-registered PICO framework — pooled estimates, heterogeneity, and certainty of evidence, documented to survive a notified body's scrutiny.
Typical Deliverables
The Problem
A narrative summary of “the literature supports our device” does not survive scrutiny. Notified bodies and reviewers want a reproducible search strategy, explicit inclusion criteria, a quantified pooled effect where studies allow it, and an honest account of heterogeneity and bias. Most literature sections in clinical evaluation reports are built the other way around — evidence selected to fit a pre-decided conclusion, rather than a conclusion derived from a pre-registered, systematic search.
The Approach
Every evidence assessment starts with a PICO question, registered before the search begins: Population, Intervention, Comparator, Outcome. The search strategy, screening criteria, and quality-appraisal method follow directly from that question — not the other way around.
Where studies are comparable enough to combine, a meta-analysis produces a pooled effect estimate with a quantified confidence interval, heterogeneity statistics (I², τ², prediction intervals), and sensitivity analyses that test whether the conclusion holds under reasonable alternative assumptions. Where they are not, the review says so explicitly — a defensible narrative synthesis beats a misleading pooled number.
Service Catalog
Systematic Literature Reviews
Comprehensive, protocol-driven literature searches conducted according to PRISMA guidelines, with documented search strings across the relevant databases, transparent screening at title/abstract and full-text stages, and structured quality assessment. Serves as the evidentiary foundation for clinical evaluation reports, study design justification, and state-of-the-art claims.
PICO Framework & Protocol Registration
The review question is operationalized into Population, Intervention, Comparator, and Outcome before a single database is searched, together with pre-specified eligibility criteria and analysis plan. Registered where appropriate (e.g. PROSPERO) so the review cannot be quietly reshaped by whatever the search happens to find.
Meta-Analysis & Pooled Effect Estimates
Fixed-effect, common-effect, and random-effects meta-analysis of extracted study data — risk ratios, odds ratios, mean differences, standardized mean differences, or hazard ratios, as appropriate. All analyses run in R (metafor), fully reproducible, with forest plots, funnel plots, and publication-bias diagnostics (Egger's test, trim-and-fill) included where relevant.
Heterogeneity & Sensitivity Analysis
Quantified between-study heterogeneity (I², τ², H², Cochran's Q) with prediction intervals that show the plausible range for a new study — not just the average effect. Pre-specified subgroup and meta-regression analyses probe likely sources of heterogeneity, and sensitivity analyses (leave-one-out, alternative models, influence diagnostics) test whether the conclusion is robust.
GRADE Certainty-of-Evidence Assessment
Structured grading of the certainty of evidence for each outcome — risk of bias, inconsistency, indirectness, imprecision, and publication bias — presented in a GRADE summary-of-findings table. Gives reviewers and internal stakeholders a transparent, standardized read on how much weight the evidence can actually bear.
State-of-the-Art Reviews for CERs
A state-of-the-art review positioned specifically for MEDDEV 2.7/1 Rev. 4 and MDCG 2020-6 requirements: alternative treatment options, current clinical practice, benchmark safety and performance data, and how the device compares against that benchmark — feeding directly into the benefit-risk determination of the CER.
Evidence assessment rarely stands alone. A systematic review often surfaces the evidence gap that a pivotal study needs to close, or the state-of-the-art baseline a post-market clinical roadmap needs to benchmark against. The same evidence base carries through every downstream document.
Regulatory Framework
All evidence assessment work is conducted in compliance with:
- —PRISMA 2020 — Preferred Reporting Items for Systematic Reviews and Meta-Analyses
- —MEDDEV 2.7/1 Rev. 4 — Clinical evaluation guidance
- —MDCG 2020-6 — Clinical evidence for legacy devices / equivalence
- —GRADE — Grading of Recommendations Assessment, Development and Evaluation
- —EU Medical Device Regulation (EU) 2017/745 (MDR)
- —Cochrane Handbook for Systematic Reviews of Interventions
Need a defensible evidence base for your device?
Contact