Methodology & Evidence Standards
From product claim to decision-grade economic view.
Imaging Benchmark routes every estimate through baseline, constraint, evidence, monetization, and realization logic. This page documents the pipeline and the discipline that bound it.
The five-step pipeline.
- 01
Evidence Sourcing
Published studies, validated clinical data, field-verified metrics, product user reviews, and aggregate perception data.
Each source is tagged with a published / vendor-disclosed / inferred / interview evidence class and a confidence tier (high / medium / low). Vendor case studies and regulatory documents carry their own weights; they are never silently aggregated with peer-reviewed clinical studies.
- 02
KPI Verification
Each indicator traced to source methodology, sample characteristics, and reproducibility context.
Scan-time, read-time, and throughput KPIs reference the underlying study population and inclusion criteria. If a benchmark cannot be traced back to a specific sample, it does not appear in the model.
- 03
Environment Modeling
Estimate ranges incorporate your staffing, volume, payer mix, and workflow constraints.
The FastFit Profile and Advanced Operational Profile narrow generic ranges into facility-calibrated ones. Initial Opportunity Range uses FastFit inputs alone; Calibrated Value Range incorporates the AOP's staffing, demand growth, and bottleneck signals.
- 04
Integration
Compatibility with your existing systems has not been assessed. Estimates assume the integrations required for the modeled benefits are available.
- 05
Output Synthesis
Product pages bring the evidence, the estimate ranges and the calculation behind them together for committee review and procurement evaluation.
Every output records the assumption set and calculation version behind it. Two ranges produced six months apart under different methodology are clearly labeled as such; they never silently overwrite each other.
Model discipline
What we will not assume.
No automatic monetization of time savings.
Time saved only matters financially when there is a capture mechanism.
No unweighted vendor claims.
A multi-site study, vendor case study, regulatory document, and buyer comment do not carry the same weight.
Inferences stay inferences.
Inferred incompatibility and inferred compatibility move no dollar figure, stay visible, are labelled as inferences, and are never presented as confirmed in the filters; a limitation affecting one function stays identified as such.
No paid influence on scoring.
Vendor participation, if allowed, must be labeled and cannot influence benchmarks, scoring, modeling, or editorial judgment.
Every estimate range records the exact assumption set and calculation version used to produce it. Methodology updates land as new assumption rows; prior estimates keep the versions they were computed under and remain reproducible.