Evidax TRACE™

A purpose-built workflow to generate comparative evidence

AI-augmented, human-in-the-loop, built for drug development.

TRACE is the next frontier in evidence synthesis, producing decision-ready evidence on accelerated timelines.

We built TRACE in-house to solve a bottleneck in evidence generation. Conventional evidence synthesis is slow, expensive, and error-prone.

This is how evidence synthesis usually goes

Two or more reviewers manually screen thousands of records and extract data by hand. This has consequences when the synthesis informs decisions in drug development.

The time

A single systematic review and meta-analysis takes 1,139 working hours on average[1]. It runs 67 weeks from start to publication with an average team of five reviewers[2].

The cost

The estimated cost of a review done this way is $141,000[3].

The accuracy problem

One reviewer screening abstracts misses 13% of relevant studies[4]. Data extraction errors appear in up to half of all reviews[5]. Up to 85% of widely cited meta-analyses contain statistical errors, and 45% mistakenly use standard error in place of standard deviation[6].

Read those stats again.

It does not have to be this way.

Evidax TRACE™ is built differently

TRACE does not replace the expert. It supports the expert.

A senior methodologist leads every review. They set the protocol, decide which evidence is relevant, and are accountable for every decision.

TRACE provides the methodologist with AI assistance inside a purpose-built workflow, accelerating timelines and improving accuracy.

Regulators and HTA bodies support this approach under clear conditions: human oversight, validation, and transparent reporting. The FDA and EMA set these principles for AI across the medicines lifecycle[7,8]. NICE says AI should augment rather than replace human involvement[9].

This is exactly what TRACE was built for. Regulatory-grade evidence synthesis on an accelerated timeline.

Every review is aligned with the rigour of PRISMA and Cochrane guidance, guaranteed. Every decision has a complete audit trail, and you see all of it on the Evidax Workspace. And we are transparently and publicly validating the workflow.

The TRACE workflow

The Evidax TRACE workflow A senior methodologist sets the protocol and is accountable across four stages: records, screening, extraction and synthesis. Each stage has a human checkpoint above it and AI assistance below it, and every decision is written to an audit trail on the Evidax Workspace. Senior methodologist: sets the protocol, decides what is relevant, accountable throughout Records Search and de-duplicate Screening Title, abstract, full text Extraction Outcomes and study data Synthesis Pooled analysis and report AI assistance inside the workflow, with a complete audit trail on the Evidax Workspace

We are publicly validating TRACE™

Many consultancy firms claim speed, rigour, and accuracy.

We would rather prove it than claim it.

Evidax TRACE is now under validation for its performance in study selection and data extraction, compared with dual independent human review as the reference standard.

And we are doing this transparently:

  • pre-defined evaluation criteria
  • publicly accessible protocols
  • evaluation aligned with best practice guidance (Responsible use of AI in evidence SynthEsis) [10]

The validation is embedded within a real-world benchmarking meta-analysis of first-line abemaciclib plus aromatase inhibitor in HR-positive, HER2-negative metastatic breast cancer.

References

  1. Allen & Olkin 1999, JAMA. https://doi.org/10.1001/jama.282.7.634
  2. Borah et al. 2017, BMJ Open. https://doi.org/10.1136/bmjopen-2016-012545
  3. Michelson & Reuter 2019, Contemp Clin Trials Commun. https://doi.org/10.1016/j.conctc.2019.100443
  4. Gartlehner et al. 2020, J Clin Epidemiol. https://doi.org/10.1016/j.jclinepi.2020.01.005
  5. Mathes et al. 2017, BMC Med Res Methodol. https://doi.org/10.1186/s12874-017-0431-4
  6. Kadlec et al. 2023, Sports Med. https://doi.org/10.1007/s40279-022-01766-0
  7. US Food and Drug Administration 2025. https://www.fda.gov/regulatory-information/search-fda-guidance-documents/considerations-use-artificial-intelligence-support-regulatory-decision-making-drug-and-biological
  8. European Medicines Agency 2024. https://www.ema.europa.eu/en/documents/scientific-guideline/reflection-paper-use-artificial-intelligence-ai-medicinal-product-lifecycle_en.pdf
  9. National Institute for Health and Care Excellence 2024. https://www.nice.org.uk/corporate/ecd11
  10. Thomas et al. 2026. Open Science Framework. DOI 10.17605/OSF.IO/FWAUD. https://osf.io/cqa82

Regulatory-grade evidence synthesis, on an accelerated timeline.

Get in touch