Executive summary
Every cleared medical device has already passed one evidence test. The problem is that revenue depends on a second one. Regulatory review asks whether a device is safe and effective enough to sell. Procurement review asks whether adopting it is defensible enough to buy — against current practice, against the institution's economics, and against the risk tolerance of a committee that answers for the decision. The two tests share vocabulary but not thresholds, and the distance between them is where cleared products stall.
~1 in 4
of FDA-approved AI-enabled medical devices explicitly reported that no clinical performance studies had been conducted at approval
Windecker et al., JAMA Network Open, 2025 [1]
6–18 months
typical hospital capital purchasing cycle for equipment requiring committee approval and budget allocation
Hinrichs-Krapels et al., BMJ Open, 2022 [2]
64%
of U.S. hospitals reported using a value analysis committee as early as 2012 — a practice that has since become standard infrastructure
Hristova-Neeley et al., Value in Health, 2015 [4]
This brief examines the evidence threshold problem: the structural mismatch between the evidence companies build for regulators and the evidence institutions require to purchase. It is not a regulatory problem. It is a sequencing problem — evidence gets scoped to the gate behind the company rather than the gate ahead of it, and the cost surfaces at the exact moment the company believed the hard part was over.
Two gates, two questions
Medtech commercialization is a sequence of external gates, and each gate applies its own test. Regulatory clearance is one of the earlier gates, not the last one. The evaluator changes after clearance — from a regulator assessing the device to an institution assessing the decision to adopt the device — and the evidence question changes with it.
The regulator's question is bounded: does this device perform as claimed, within an acceptable risk profile? The buyer's question is comparative and economic: is this device better enough than what we do today to justify its cost, its workflow disruption, its training burden, and the institutional risk of being wrong? A device can answer the first question completely and the second one not at all.
Post-authorization utilization data makes the gap visible. Wu et al. report more than 500 FDA-approved medical AI devices in the U.S., yet clinical usage remains highly concentrated — only two procedure categories exceeded 10,000 insurance claims in their dataset.[3] The implication is not specific to AI. It is that clearance, by itself, does not produce adoption. Clearance is a floor. The buyer runs a different test.
The evidence that clears a device is scoped to a question no buyer is asking.
What the buyer's evidence test actually is
The institutional purchase decision in U.S. hospitals runs through value analysis. As early as 2012, 64% of hospitals reported using some form of value analysis committee, and the accompanying research found that facility-specific value propositions and data are frequently required to secure product approval.[4] In the years since, the VAC has become standard purchasing infrastructure rather than an exception — a committee of clinical, supply chain, nursing, financial, and administrative stakeholders through which new products must pass before the institution is permitted to buy them.
Committee structure changes what evidence means. The clinical champion who loves the device is one voice among several, and the other voices are evaluating cost, comparative outcomes, and operational burden. A systematic review of high-cost device purchasing found that hospital processes involve multiple stakeholders across clinical, engineering, procurement, and finance functions, with documented risk of delay and conflict when those stakeholders are not aligned early — and that value analysis review adds multi-month evaluation cycles requiring economic justification, workflow validation, and cross-department alignment even after clinical validation is complete.[2]
The evidence hierarchy inside that review is also not the regulator's hierarchy. An analysis of hospital purchasing evidence describes value analysis committees as working from three sources — peer-reviewed published literature, the testimony of their own staff physicians, and materials supplied by the manufacturer — and treats manufacturer-supplied evidence as the weakest tier, on the reasonable grounds that it is likely to be favorable and drawn from controlled conditions the institution's real-world population may not reproduce.[5]
Operators who have carried products through these committees describe the bar in plain terms. In published interviews with medical device portfolio companies, executives report committees increasingly expecting Level I–II clinical evidence — randomized, controlled trials — alongside health economic data published in peer-reviewed journals, and note that committees are examining both the quantity and the quality of clinical data more rigorously than in the past.[6] These are practitioner accounts rather than peer-reviewed findings, and they should be read as such. But they describe the same direction of travel the peer-reviewed literature documents: the purchasing threshold is rising, and it is rising along dimensions regulatory review was never designed to test.
Where the mismatch shows up in the data
The clearest documentation of the evidence gap comes from studying what cleared products actually carry with them at approval.
A 2025 cross-sectional study of 903 FDA-approved AI-enabled medical devices found that clinical performance studies were reported for approximately half of the devices analyzed, while one-quarter explicitly stated that no such studies had been conducted at approval.[1] These devices are legally marketable. What they are not, in many cases, is equipped for the comparative and economic questions a purchasing committee will ask — an evidence-clarity gap that directly weakens procurement and partnership confidence.
The pattern extends beyond devices. A cross-sectional analysis of digital health companies published in the Journal of Medical Internet Research found clinical robustness — measured by clinical trials and regulatory filings — to be low across a substantial share of the companies analyzed, and found no meaningful relationship between clinical robustness and total funding raised.[7] Capital, in other words, is not a proxy for evidence. A company can be well funded, cleared, and still under-evidenced for the decision that produces revenue.
Reimbursement research documents the same threshold problem one gate further downstream. A 2025 systematic review found that fragmented reimbursement pathways and insufficient cost-benefit evidence can materially slow or prevent the adoption of healthcare innovation, particularly for disruptive products — the products least able to borrow an existing evidence template.[8]
The problem is not confined to the U.S. Under the European Medical Device Regulation, clinical evaluation requirements have increased substantially, and in a quantitative study of medical device manufacturers, the single most-reported challenge was determining how much data is needed to constitute sufficient clinical evidence — compounded by inconsistent clinical data requirements across notified bodies.[9] The threshold, on both sides of the Atlantic, is real, rising, and unevenly specified. Companies are being asked to clear a bar whose height they are left to estimate.
Why companies aim at the wrong threshold
The evidence threshold problem is structural, not a failure of effort or intent. Four mechanisms produce it repeatedly.
01
Evidence is scoped to the nearest gate
Studies get designed around clearance endpoints because clearance is the gate directly ahead. Procurement endpoints — comparative performance, institutional economics, workflow burden — are deferred as post-launch questions. By the time they are asked, the study that could have answered them has already been run, on the wrong endpoints.
02
The purchasing threshold is invisible from outside
Committee requirements are institution-specific, rarely published, and inconsistently applied.[2][4] Most companies discover the actual bar one rejection at a time, converting what could have been a research question into an expensive field experiment.
03
Capital timing punishes the fix
Runway pressure pushes teams to launch on clearance-grade evidence rather than fund the studies purchasing committees expect. Meanwhile, the funding market has moved the other direction: investors are concentrating earlier capital on companies that can demonstrate clinical validation and revenue traction,[10] which makes an under-evidenced launch harder to recover from than it used to be. The companies most in need of an evidence bridge are the least able to pause and build one.
04
Proxies substitute for outcomes
Bench performance, usability data, and pilot observations get presented where committees expect outcome data and institution-relevant economics. This is a category error, not a quantity problem — more of the wrong kind of evidence does not cross the threshold, and committees that discount manufacturer-supplied materials in the first place[5] discount proxy-grade materials fastest.
These are not unpredictable barriers.
The purchasing threshold can be characterized before it is encountered.
What purchase-ready evidence looks like
The practical response to the evidence threshold problem is not to run every possible study. It is to scope evidence to the gate ahead — to characterize the purchasing threshold early enough that study design, claims, and commercial sequencing can be built against it. For a leadership team, that reduces to a small set of questions:
- What evidence tier does the target committee treat as decision-grade — and is that documented from real purchasing processes, or assumed?
- Which claims in the company's current materials would a committee ask to substantiate, and what actually exists behind each one?
- What does the economic case look like in the institution's numbers — its costs, its patient population, its workflow — rather than the company's model?
- Which evidence gap gates the next commercial step, and which merely gates the step after it? Sequencing evidence investment is as important as making it.
- What would the first three committee rejections teach that could be learned before the first submission?
Answering these questions before launch does not guarantee adoption. It prevents the most common version of stall: a cleared product carrying clearance-grade evidence into a purchase-grade review, and discovering the difference with revenue on the line.
About RŌG Health
RŌG Health works with medtech and medical device teams at commercialization inflection points — fundraising, pivotal pilots, strategic partnerships, and early revenue conversion — when external stakeholders will pressure-test the company's commercial logic whether the company is ready or not.
Endnotes
- [1] Windecker D, et al., "Generalizability of FDA-Approved AI-Enabled Medical Devices for Clinical Use," JAMA Network Open (2025). https://jamanetwork.com/journals/jamanetworkopen/fullarticle/2833324
- [2] Hinrichs-Krapels S, et al., "Purchasing High-Cost Medical Devices and Equipment in Hospitals: A Systematic Review," BMJ Open (2022). https://pmc.ncbi.nlm.nih.gov/articles/PMC9438058/
- [3] Wu K, Wu E, Theodorou B, et al., "Characterizing the Clinical Adoption of Medical AI Devices through U.S. Insurance Claims," NEJM AI (Jan 2024). https://assets.ctfassets.net/otzakoj1abuh/61nlSm2dw6ajoODa0gDITb/7fefc7d9953d70ccdb35d3d2b2126455/NEJM-AI_Issue1_Jan2024.pdf
- [4] Hristova-Neeley D, et al., "The Rise of the Value Analysis Committee at US Hospitals, Better or Worse for Medical Device Companies?" Value in Health (2015). https://www.valueinhealthjournal.com/article/S1098-3015(15)00349-6/fulltext
- [5] Avalon Health Economics, "Medical Device Economics and the Health System Purchaser in the Era of Novel Payment Mechanisms," Issue Brief. https://avalonecon.com/medical-device-economics-and-the-health-system-purchaser-in-the-era-of-novel-payment-mechanisms/
- [6] River Cities Capital Funds, "Navigating the Value Analysis Committee" (portfolio company operator interviews; industry commentary, not peer-reviewed). https://rccf.com/navigating-the-value-analysis-committee/
- [7] Day S, et al., "Assessing the Clinical Robustness of Digital Health Startups: Cross-Sectional Observational Analysis," Journal of Medical Internet Research (2022). https://www.jmir.org/2022/6/e37677/
- [8] Allers S, Eijkenaar F, van Raaij EM, Schut FT, "Patterns in the Influence of Funding and Reimbursement on the Development and Implementation of Healthcare Innovation: A Systematic Review," Journal of Open Innovation (2025). https://www.sciencedirect.com/science/article/pii/S2199853125000253
- [9] Kearney B, McDermott O, "The Challenges for Manufacturers of the Increased Clinical Evaluation in the European Medical Device Regulations: A Quantitative Study," Therapeutic Innovation & Regulatory Science (2023). https://pmc.ncbi.nlm.nih.gov/articles/PMC10276779/
- [10] Silicon Valley Bank, "Healthcare Investments and Exits Report," 17th Edition (January 2026). https://www.svb.com/news/company-news/ai-investment-accounted-for-nearly-half-of-healthcare-investment-in-2025-silicon-valley-bank-releases-17th-healthcare-investments-and-exits-report/
© 2026 RŌG Health. All rights reserved.
This document is provided for informational and analytical purposes only. While every effort has been made to ensure the accuracy of the data and sources referenced, RŌG Health makes no representations or warranties, express or implied, regarding completeness, accuracy, or applicability to any specific situation.
The findings and interpretations presented reflect a synthesis of publicly available data and RŌG Health's independent analysis of commercialization dynamics in medtech and medical devices. They should not be construed as financial, legal, regulatory, or investment advice. All third-party sources are cited where applicable. Any errors or omissions are unintentional.
No part of this publication may be reproduced, distributed, or transmitted in any form without prior written permission from RŌG Health, except for brief quotations with appropriate attribution.
Begin with clarity.
Answer a few questions about where you stand. Leave with a clear, defensible next step.