SHAIPED Survey Form

SHAIPED questionnaire to Health Technology Assessment Body

SHAIPED Stakeholder Survey

About this questionnaire

This questionnaire is part of the SHAIPED project, a European initiative exploring how health data access can support the development, assessment, and deployment of Medical Device Artificial Intelligence (MDAI) in Europe.

Within SHAIPED, this work focuses specifically on the role of Real-World Data (RWD) in Health Technology Assessment (HTA) and reimbursement processes for MDAI, the challenges faced by stakeholders, and the ways in which future Health Data Access Bodies (HDABs) may help address these challenges.

We welcome responses both from experts with direct experience evaluating MDAI and from experts who may not yet have assessed these technologies but can provide valuable insights into how such processes should be shaped in the future.

The questionnaire contains 4 sections. Completing the questionnaire should take approximately 45 minutes.

The results will be made available in a project deliverable published on the SHAIPED webpage, and we will also send the document to the email address used to contact you.

Thank you very much for your time and for sharing your expertise.

For any questions, please contact: info@shaiped.eu

Participant information

Section 0 — Organisation Profile

Your name

Organisation

Please update your organization name, if necessary.

Contact email (optional — for confirmation)

Country

If your organization operates at country level, please select one. Otherwise, leave empty.

Please indicate your current role within your organisation, including your job title and how long you have worked in this role.

Briefly describe your experience relevant to HTA and reimbursement processes, and, where applicable, the assessment of MDAI.

Additional notes on your organisation

Section 1 — General questions about HTA processes for MDAI

For the purposes of this survey, Medical Device Artificial Intelligence (MDAI) refers to AI‑based systems that qualify as medical devices or form part of a medical device, including AI‑based medical devices and AI‑based in vitro diagnostic medical devices. This category includes, for example, computed tomography systems incorporating AI‑driven pattern‑recognition modules.

This definition excludes general‑purpose AI systems; administrative or workflow software without an intended medical purpose; consumer wellness applications; and products that do not have a defined medical purpose.

Within this survey, we distinguish between two types of MDAI:

Locked/Static MDAI: AI‑based medical devices whose algorithms remain fixed once deployed and do not change through use. This includes systems that are later re‑trained, re‑validated, and re‑released as new fixed versions; they remain static between releases.
Unlocked/Adaptive MDAI: AI‑based medical devices designed to adjust or improve their performance after deployment based on new data or real‑world inputs generated during use, dynamically updating AI models and processes. This adaptation may occur periodically (e.g., through batch retraining) or, as methodologies advance, continuously over time.

1.1 Please briefly describe the HTA landscape for MDAI in your country.
Please indicate how MDAI may enter the healthcare system and be used in practice, whether they undergo a formal HTA process for reimbursement, and whether any re-evaluations occur after adoption/deployment.

1.2 Are you aware of any MDAI that has undergone an HTA process in your country?
If yes, please provide examples of MDAI where possible, and include links to relevant sources.

1.3 Based on your experience and professional opinion, what are the main challenges associated with evaluating locked/static MDAI in HTA processes?

Note: here, we consider systems that may be updated and re-evaluated but remain locked between versions. Example: An AI ultrasound tool, which may later be modified, but for which each released version can be evaluated before deployment.

1.4 Based on your experience and professional opinion, how can HTA processes/methods be adapted to accommodate the evolving nature of unlocked/adaptive MDAI?

Note: here, “evolving” means that performance may change after deployment through a defined learning process, either periodically or continuously. Example: an AI-based radiology triage system that is retrained using post-market data, whether through scheduled model updates or continuous learning during deployment.

1.5 Based on your experience and professional opinion, how could the main challenges in HTA for MDAI be addressed during the evaluation phase and at the decision-making phase?
For example, through specific evidence requirements or conditional reimbursement agreements.

Section 2 — Current role of RWD across the MDAI lifecycle

This section explores Real-World Data (RWD) and Real-World Evidence (RWE) in HTA and reimbursement processes for MDAI. We use the EMA definition of RWD: “routinely collected data relating to patient health status or the delivery of health care from a variety of sources other than traditional clinical trials”. This includes RWD used for the training, validation, and testing of MDAI, as well as RWD generated through the deployment and use of MDAI within healthcare systems. Relatedly, RWE refers to the evidence generated through the analysis of such data.

2.1 Please indicate the current role of RWD in HTA processes in your country.
If RWD is not considered at all, select “Not considered” under “Level of use” and leave the rest blank.

Level of use	Type of technology (select one or multiple)	Phase of the technology lifecycle* (select one or multiple)	Type of RWD** (select one or multiple)
Not considered Occasionally used Used routinely	Pharmaceuticals Medical devices Digital Medical Devices MDAI Other	Pre-adoption: Horizon scanning Pre-adoption: Early dialogue / scientific consultation Pre-adoption: Initial reimbursement recommendation or decision Post-adoption: Monitoring adoption Post-adoption: Reassessments (e.g. in coverage with evidence development schemes) Post-adoption: Disinvestment Other	Any Electronic Health Records Administrative data Registries Health surveys Other

* as defined in Trowman et al. (2023)

** as defined in Garrison et al. (2007)

2.2 What are your personal views on the role that RWD and RWE should play in HTA processes for health technologies?

2.3 Based on your experience and professional opinion, what are the most critical limitations affecting the use of RWD and RWE in HTA processes for MDAI?

2.4 Which factors determine whether RWD are considered “fit for purpose”? (select all that apply)

Representativeness

Completeness

Temporal Coverage

Linkability

Standardisation

Provenance Documentation

Validation Against Controlled Data

Regulatory Acceptance Track Record

Other

2.5 Based on your experience and professional opinion, which factors should be considered to determine whether RWD is of sufficient quality in HTA processes for MDAI?

Example: data completeness.

2.6 Is there guidance in your country on the use of RWD in HTA processes? This guidance does not need to be specific to MDAI.
If yes, please specify and, if possible, include a link to the relevant source(s).

2.7 Is there guidance in your country on the use of RWD specifically for MDAI in HTA processes?
If yes, please specify and, if possible, include a link to the relevant source(s).

2.8 Do you perceive misalignment or duplication between regulatory and HTA expectations and guidance regarding the use of RWD?
If yes, please describe where this is most evident.

Privacy Statement

Processing of personal data within SHAIPED is governed by a Joint Controller Agreement (Art. 26 GDPR).

Read the full Privacy Statement (PDF)

I have read and accept the Privacy Statement *