An investigator has proposed a study
The scientific question is theirs. The decision about what to support, and on what terms, is yours, and it is usually made before anyone has assessed whether the data would be usable.
Home / Neuromodulation CRO / Investigator-initiated studies
Neuromodulation · Investigator-led evidence · Regulation (EU) 2017/745Turn an emerging clinical signal into structured evidence and a clear development decision.
An investigator proposes a study. A physician reports something interesting in a population outside the device's current intended purpose. Eclevar helps you decide whether the signal is worth pursuing, choose the evidence route that fits it, and design the study so the data is still usable when the decision matters.
Opportunities in neuromodulation rarely arrive through the clinical development plan. They arrive from outside it: an implanting physician using the system in a population your protocol never described, an academic group proposing a new anatomical target, a commercial team seeing demand the clinical evidence does not yet support. Most sponsors are in one of three positions.
The scientific question is theirs. The decision about what to support, and on what terms, is yours, and it is usually made before anyone has assessed whether the data would be usable.
The signal looks real but the evidence gap is undefined, and you need a route before you can budget.
The results are encouraging. The question is whether what you hold supports anything beyond a publication.
The mistake in all three is the same: treating clinical interest as though it were clinical evidence. Enthusiasm from an experienced implanter is a reason to look carefully, not a reason to fund a study, and not evidence that a new indication can be supported.
Before agreeing to support an investigator-led study, a manufacturer should be able to answer seven questions. Most can answer three.
An investigator may originate the hypothesis, the concept and the scientific leadership. None of that reduces what useful evidence requires: clear responsibilities, a credible protocol, defined endpoints, monitoring proportionate to the risk, data-quality standards, a statistical plan, governance, publication planning, and transparency about what the manufacturer provided.
The level of manufacturer involvement is not fixed, and should not be. An investigator-initiated study does not require the operational model of a sponsor-led pivotal investigation, but it does require that somebody has decided, deliberately and in writing, which parts of that model apply.
That a study is described as investigator-initiated tells you who originated it. It does not by itself determine which clinical-investigation route under Regulation (EU) 2017/745 applies.
That depends on the purpose of the investigation, the CE-marking status of the device, whether the use sits within or outside the existing intended purpose, the study design, who holds the sponsor role, the jurisdiction and applicable national requirements. Article 62 concerns investigations conducted for the conformity-assessment purposes described in Article 62(1). Article 74 addresses investigations involving CE-marked devices, including post-market clinical follow-up within the intended purpose, under the conditions set out there; where a CE-marked device is investigated outside its intended purpose, Article 74(2) points back to Articles 62 to 81. Article 82 concerns other investigations not performed for the purposes listed in Article 62(1), and Member States may impose additional requirements.
Eclevar assesses and documents the applicable route with the sponsor and against relevant local requirements. The responsibilities and pathway that apply depend on study purpose, device status, intended purpose, jurisdiction and regulatory context. Eclevar does not make the final legal determination.
And what that obliges the investigator or the manufacturer to do.
Assessed against the factors above rather than assumed from the study's origin.
Agreed in writing before enrollment, including whether the manufacturer may use the dataset in a clinical evaluation.
Decided at the start so the dataset stays structured, traceable and potentially reusable in the broader evidence program where appropriate, rather than reconstructed afterwards.
None of the four can be retrofitted.
An investigator-initiated study is one of five routes. Choosing between them is the decision this page exists to support.
| Route | Best used when | Core question | Key risk |
|---|---|---|---|
| Investigator-initiated study | An investigator holds a credible hypothesis | Does this signal merit further investigation? | Protocol and data quality make the evidence unusable later |
| Sponsor-supported academic study | The manufacturer supports investigator evidence while keeping governance boundaries clear | How much support is appropriate, and where does it stop? | Unclear responsibilities and inconsistent data standards across centers |
| Sponsor-led exploratory study | The manufacturer needs control of protocol, endpoints and data | Is the evidence important enough to justify a sponsor-led program? | Cost committed before the hypothesis is confirmed |
| Registry or real-world evidence | Longitudinal use or outcomes need observation at scale | Can real-world data answer this question? | Structure and completeness decide everything, and both are set at design |
| Sponsor-led clinical investigation for potential indication expansion | Existing evidence justifies a controlled program | What evidence package would the expanded indication require? | Committing to a design before the evidence gap is defined |
Figure 1 · Where each route sits
Routes are not ranked. A registry can answer a question a controlled investigation cannot, and a small exploratory study can settle one no registry would reach. Fitness depends on the question, not the size of the study.
For a CE-marked device, where the question concerns long-term clinical evidence within its existing intended purpose, PMCF or another appropriate post-market evidence model may be more relevant than an exploratory indication-expansion study: see neuromodulation PMCF and registry programs and, for registry methodology, medical device registries.
| What you have | Evidence maturity | Route usually indicated | Likely next step |
|---|---|---|---|
| A single-center observation, one clinician | Anecdotal | Exploratory investigator-initiated study, or no study yet | Define the hypothesis and the endpoint before committing |
| Consistent observations across several centers | Emerging | Multicenter investigator-initiated study, or a structured registry | Standardize definitions across sites before enrollment |
| Structured preliminary evidence with a defined endpoint | Preliminary | Sponsor-supported or sponsor-led exploratory study | Formal evidence-gap assessment |
| Reproducible effect in a strategically important population | Developing | Sponsor-led feasibility or pivotal design work | Development plan built with regulatory input |
| A CE-marked device in established use, the question being durability at scale | Post-market | PMCF or registry, within the existing intended purpose | Route decided against the existing intended purpose |
If you cannot place your program on this table, that is the first thing worth an hour of discussion.
Seven stages. The output of stage seven is a decision, and stop is a legitimate one.
Clinical rationale, evidence gap, intended population, literature and comparator landscape, endpoint and operational feasibility, site capability, and the implications for your development plan.
Research question, objectives, eligibility criteria, endpoint and responder definition, follow-up duration, safety endpoints, device-specific variables and statistical assumptions. Written so the investigator's scientific ownership survives the process.
Patient volume in the specific indication your protocol defines rather than total implant numbers, research infrastructure, device experience, coordinator capacity, follow-up and data-quality capability, and competing studies already open. Reputation is not a qualification criterion, and the most experienced neuromodulation centers in Europe are usually already committed elsewhere. See feasibility and site selection.
Project coordination, site activation, monitoring proportionate to risk and design, vendor coordination, documentation and governance. An exploratory single-center study and a multicenter sponsor-led program do not receive the same operational model. See clinical trial monitoring.
This is where most investigator-generated neurostimulation evidence is lost. Not through negligence, but through decisions taken early, for good local reasons, that quietly remove any use for the data beyond the paper it was collected for.
Reuse fails because there is no stable analysis population. Prevented by a predefined endpoint architecture, with any change handled as a documented amendment rather than an evolving convention. Eclevar drafts the endpoint rationale against the literature for the indication and fixes it before enrollment.
Reuse fails because the exposure cannot be reconstructed, so the study cannot describe what therapy was delivered. Prevented by structured programming variables: amplitude, frequency, pulse width, active contacts, program identifier, and the date and reason for each change, captured as queryable fields. Eclevar designs that structure from the start and reconciles device configuration exports against the study record where the system supports it.
Reuse fails because longitudinal comparison weakens and time-dependent endpoints stop being interpretable. Prevented by pre-specified visit windows and explicit rules for what counts as within them, enforced in the data capture design rather than left to site convention.
Reuse fails because the reoperation picture, which is what a reviewer asks about, cannot be assembled. Prevented by a prospective event taxonomy agreed before first patient in and reconciled with the manufacturer's vigilance and complaint terminology, then trained into the sites.
Reuse fails because nothing can be adjusted for and subgroup findings cannot be defended. Prevented by a baseline sized to the questions the results will be asked to answer, and an analysis plan fixed before database lock.
Figure 3 · The reuse chain and its gates
Doing this early is not about taking over the study. The manufacturer's ability to use the dataset is decided at design, by people who are usually not thinking about it. Where full delivery is needed, see clinical data management.
An exploratory design that ignores the modality collects the wrong variables.
Common to all: stimulation parameters and programming changes, lead and electrode configuration, anatomical target, implant history, revision, explant and reoperation, device deficiencies, therapy adherence, concomitant medication, patient-reported outcomes, quality of life, responder definition and durability of effect.
Pain response and responder threshold, trial-to-implant conversion, analgesic use as a confounder, lead migration and revision, stimulation settings over time. See spinal cord stimulation clinical trial CRO.
Indication-specific outcome measures, assessor training and rating consistency, programming changes, imaging where the protocol requires it, anatomical targeting. See deep brain stimulation clinical trial CRO.
Titration as a protocol variable, diary and electronic outcome completeness, long baseline periods, and adherence as an exposure measure rather than a compliance statistic. See vagus nerve stimulation studies.
This page owns the decision about the evidence route. Those pages own delivery of the study once the route is chosen. Broader neuromodulation capability sits on the neuromodulation CRO capability page.
There is a moment when an investigator-led study has produced what it can produce and the next question needs a different vehicle. Recognizing it early is the difference between a development program and a publication.
Observed, described, and worth a structured look.
The effect is tested again, or the question is narrowed.
What would have to be shown, and by which route.
Built jointly with regulatory, against the intended purpose.
Designed for the decision it has to support.
Figure 4. Each rung has an exit. Climbing is a decision, not a default.
A framework in which every path leads to a larger project is not a decision framework. Recommending that a program should not progress yet is usually the cheapest advice a sponsor receives.
Six outcomes are legitimate at any point: stop, refine, replicate, continue as an investigator-initiated study, move to sponsor-led development, or design a clinical investigation. What changes when the manufacturer becomes sponsor: the obligations, the documentation standard, the monitoring model, the cost and control of the dataset. Where the next step is first clinical use in a new population, see early feasibility and first-in-human studies.
Most investigator-led programs stall internally not on science but on uncertainty about who is responsible for what.
Holds the scientific hypothesis, clinical leadership and academic ownership where applicable.
Holds defined support, device information, the wider evidence strategy and governance proportionate to the study model.
Provides independent operational and methodological infrastructure: study design support, operational planning, data architecture, site support, monitoring, biostatistics, project governance and medical writing.
The scope of manufacturer support, data ownership and reuse, publication planning and timing, the data capture standard, the monitoring approach, and what is disclosed about manufacturer involvement.
Who holds the sponsor role and its statutory obligations, the applicable clinical-investigation route, submission and notification requirements, insurance, safety reporting duties and national requirements. These depend on study design, jurisdiction and regulatory status, and are established with qualified local advice.
Exploratory evidence is not weak evidence. It is differently framed, and the framing has to be decided rather than inherited.
The decisions that bind the future program are made early: how the hypothesis is expressed, what the sample-size assumptions rest on, whether the analysis is exploratory or confirmatory, the endpoint hierarchy, the responder definition and its published basis, the approach to longitudinal and missing data, and which subgroup analysis is prespecified. See medical device biostatistics.
Intended purpose is the line that decides most of this. If the proposed use sits outside the population, the anatomical target or the clinical claim covered by your existing clinical evaluation, the evidence requirement changes, and so does the route.
A proposed indication expansion should be evaluated jointly from clinical, regulatory and evidence-generation perspectives. Assessed separately, the clinical team designs a study the regulatory strategy cannot use, or the regulatory team specifies requirements the study was never built to meet.
Where results are methodologically and regulatorily appropriate, they may contribute to the broader clinical evidence strategy. That is a judgment made per program, not an automatic consequence of a study being completed. See EU MDR regulatory strategy and clinical evaluation report.
Evidence-gap architecture, endpoint and responder rationale built for the indication, structured stimulation and device-configuration data, revision and explant taxonomies, and post-market follow-up designed against the clinical evaluation it supports.
For Nihon Kohden, Eclevar delivered clinical evaluation strategy, a Clinical Evaluation Report, a systematic literature review, a state-of-the-art assessment and clinical-data appraisal for an electromyography and evoked-potential system. That system is a diagnostic clinical-neurophysiology platform and does not deliver therapeutic neurostimulation. It is included as evidence of neurological clinical-evidence methodology, not of neurostimulation study delivery.
Chief Operating Officer
Clinical evidence strategy informed by first-hand Notified Body review experience, applied to how a Class III evidence dossier is structured and defended.
Chief Data Officer
Data architecture for stimulation settings and device configuration, database governance, statistics and the analysis-ready dataset.
Project Delivery Lead, France and United Kingdom
Program delivery and governance across the clinical, data, writing and quality workstreams.
Head of Medical Writing
Protocols, study reports, clinical evaluation reports and Notified Body responses.
Indication-specific medical oversight is provided by the sponsor's investigators and, where required by the program, by appropriately qualified independent specialists. Former positions are stated for biographical context only. Eclevar MedTech is an independent contract research organization. It is not affiliated with, accredited by or endorsed by any Notified Body.
For sponsors with an emerging clinical opportunity and no decision yet. The smallest useful first step.
Deliverables. Clinical hypothesis assessment, evidence-gap map, proposed evidence route with the reasoning, key endpoint questions, data-reuse risks in the design as proposed, initial regulatory-route considerations, initial feasibility considerations, and a recommendation of proceed, refine, generate preliminary evidence first, or stop.
No regulatory acceptance is promised or implied by this review.
For an identified investigator-led opportunity that is going ahead.
Scope. Protocol and endpoint support, data architecture, statistical planning, study operations, monitoring proportionate to the design, and reporting. Scope is set to the study model rather than to a standard package.
For programs where early evidence supports broader development.
Scope. Evidence review of what you already hold, indication-expansion evidence strategy, feasibility, clinical investigation design and full-service execution.
Yes, at whichever level fits the study model, from a design review to full operational delivery. We do not impose a sponsor-led model on a study that does not need one.
When the hypothesis is credible, the population is defined, the endpoint is measurable, and the terms of support, data ownership and data standard are agreed in writing beforehand. If any of those is unresolved, the decision is premature rather than negative.
It can contribute, where it was structured so the data is usable and the results are methodologically and regulatorily appropriate. It does not by itself extend an intended purpose, and publication alone does not change the device's intended purpose or the regulatory basis on which it is placed on the market. Whether the dataset can be used at all usually depends on decisions taken before enrollment.
Where appropriate methodologically and regulatorily, yes, judged per program against the intended purpose, the design and the data quality. The starting point is an evidence review of what you actually hold.
Fix the endpoint and responder definition before enrollment, capture stimulation settings and programming history as structured fields, pre-specify visit windows, define the revision and explant taxonomy in advance, and finalize the analysis plan before database lock.
That depends on the study, the jurisdiction and the regulatory status of the device. It is a legal question, not a naming convention, and it is settled before the first patient with qualified local advice.
Yes, individually or together. The first three are hard to separate: the endpoint decision determines the data structure, and the data structure determines what the statistics can do. Monitoring should be proportionate to study design, risk, critical data and the responsibilities of the parties; the appropriate model may range from focused central oversight to more extensive remote and on-site monitoring.
Tell us what you are seeing, in which population, at how many centers, and what you are being asked to support. We will identify the evidence route, the endpoint questions, the data-reuse risks and the realistic next step.
Confidentiality and NDA arrangements can be agreed before detailed program materials are reviewed.