Your team might get this one wrong. For most device companies, a clinical trial is the biggest cost and the longest delay on the road to market. Some run a study they never needed. Others skip one and then get stuck in regulatory review for months. 

The real question is not “trial or no trial?” It is ‘what level of clinical evidence will the regulator accept for your device, your risk class and your claims?’ 

In this blog, we have provided medical device companies with a five-question framework to answer that questions. It covers three key markets: the FDA, the EU MDR, and CDSCO. By the end of this blog, you will have a clear view of the clinical trial requirement for each market and steps after the trial. 

Why Risk Class Drives the Clinical Evidence Decision 

Regulators follow one simple rule: the higher the potential harm, the more direct human data they want.  

That is why medical device risk classification is the starting point for every evidence decision. 

All three markets use a four-level, risk-based system. The table below shows how the classes roughly line up. 

Risk level	US FDA	EU MDR 2017/745	India MDR 2017 (CDSCO)
Low	Class I	Class I	Class A
Low to moderate	Class II	Class IIa	Class B
Moderate to high	Class II	Class IIb	Class C
High	Class III	Class III	Class D

Two points matter most for decision makers: 

• Your claims set your class, not just your technology: The same hardware can land in a different class depending on what you say it does and who it is for. 

• The classes do not map one to one: A device that is Class II in the US may be Class IIb or Class III in the EU. In India, CDSCO publishes the class for specific device types, so you check the official list rather than decide it yourself. 

The Decision Framework: Five Questions to Ask 

Work through these five questions in order. Each weak answer pushes you closer to a new clinical study. 

1. What is your risk class in each target market? Confirm the classification and applicable pathway with medical device regulatory support before planning your clinical evidence strategy.. 
2. Is there a similar device already approved or accepted by the relevant regulatory authority? In the US, this can be used to show that the new device is substantially similar. In the EU, you need to prove that the devices are equivalent. In India, the similar device must already be approved in the country. 
3. Can non-clinical data answer the safety and performance questions? Think bench testing, biocompatibility, software validation, animal studies and usability testing. 
4. Are you making new clinical claims? A new indication, a new patient group or a claim that you beat the current standard of care usually needs new human data. 
5. Is the existing clinical data enough? This includes published literature, data on the predicate or equivalent device, real-world evidence and post-market data. 

Your answers will point to one of three outcomes: 

Outcome	What it usually looks like
No new clinical data	Bench, lab and literature evidence, plus a clinical evaluation where the market requires one
Targeted clinical study	A small pilot or feasibility study to close one specific gap
Full pivotal study	A larger study designed to prove safety and performance to the regulator

Only a device that clears all five questions avoids new clinical data. In the EU, Class III and implantable devices need a study unless a specific exemption applies, even with a strong predicate.  

Applying the Framework by Risk Class

 

• This is where medical device clinical evidence requirements start to differ sharply between markets. Here is what to expect at each level. 

Low Risk: FDA Class I, EU Class I, CDSCO Class A

 

• A new clinical study is rarely needed at this level. 

• US: Many Class I devices are exempt from 510(k) premarket notification. General controls and good manufacturing practice usually do the work. 

• EU: Even Class I devices need a clinical evaluation under MDR Article 61. For most, this is built from published literature and existing data, not a new study. 

• India: A pivotal clinical investigation in India is required for investigational devices except Class A. From August 2026, Class A non-sterile, non-measuring devices must also show quality management system compliance at registration. 

The common trap here is claim creep. A “wellness” product that starts to promise diagnosis or treatment can quickly move up a class. Also, no study does not mean no evidence work. Plan for post-market surveillance in every market and, in the EU, post-market clinical follow-up (PMCF). 

Moderate Risk: FDA Class II, EU Class IIa and IIb, CDSCO Class B and C

This is the “it depends” zone, and where the framework earns its keep. 

• US: Most Class II devices go through a 510(k). FDA has said it requests clinical data for less than 10% of 510(k) submissions. Clinical data comes in when bench tests cannot settle the question, when your device differs from the predicate in a way that raises new questions, or when the predicate has a newly identified risk. 

• EU: Clinical evaluation is mandatory, and you must justify the level of clinical evidence you rely on. Equivalence is harder to claim under MDR than under the old directives. Note that implantable Class IIb devices fall under the stricter rules for implants described below. 

• India: Importers of Class B devices can usually submit published safety and performance data and a free sale certificate instead of a local study. For Class C, the key question is whether a predicate is already approved in India.

High Risk: FDA Class III, EU Class III and Implantable, CDSCO Class D 

At this level, expect to run a study unless you can prove an exemption. 

• US: A PMA is built on your own clinical data. A study of a significant risk device needs an FDA-approved Investigational Device Exemption (IDE) before the first patient is enrolled (FDA).  

• EU: Article 61(4) requires clinical investigations for Class III and implantable devices, with a few exemptions explained in MDCG 2023-7. These include certain modifications of your own device, equivalence backed by a contract giving full access to the other device’s technical file, and “well-established technologies”.  

• India: A device with no approved predicate is treated as an investigational device. You need CDSCO permission (Form MD-22, granted in Form MD-23) to run the study (CDSCO). 

If none of these routes fit, plan for a full medical device clinical investigation designed to ISO 14155. 

Grey Zones That Change the Answer 

Some situations do not fit neatly into a risk class. These are the ones that most often surprise device teams.

1. Device Changes and Line Extensions

Small changes can still trigger clinical data.  

• In the US, even a new size that goes beyond the current size range may need it.  

• In India, a change in material, design or intended use counts as a major change that needs prior approval. 

2. Software and AI

Software is judged on three things:  

• A valid clinical link between its output and the condition  

• Proof that it processes data correctly  

• Clinical validation in the real target population.  

Clinical validation is expected for every software device, but it does not always need a new prospective study. Well-designed studies on existing patient data can sometimes do the job. 

3. Drug-Device and Biologic-Device Combinations

Part of the evidence may follow drug or biologic rules. Get an early decision on which rules lead before you design any study. 

Significant Risk vs Non-Significant Risk Studies (US) 

According to NIH Clinical Center, a significant risk study needs an approved IDE from FDA. A non-significant risk study follows a shorter set of rules and is overseen by an ethics board (IRB). If you are unsure, you can ask FDA to decide through a Study Risk Determination request. 

New Patient Groups 

Moving into children, older adults or a different disease stage is a new claim. In India, a new intended use, population, material or major design change can make a local study mandatory. 

If You Do Need a Trial: Getting It Right the First Time 

Once sites are activated, clinical trial monitoring helps ensure protocol compliance, participant safety, data quality and timely issue resolution throughout the study. 

A study that the regulator rejects is the most expensive outcome of all. A structured clinical trial feasibility assessment before protocol finalisation can help identify recruitment, site, patient population and operational risks before they become costly study delays. Well-run clinical trials for medical devices tend to share four habits. 

1. Start small, then scale: Device programmes usually begin with a pilot or feasibility study, then move to a pivotal study. A clear protocol is critical at this stage, making clinical trial medical writing an important part of translating the evidence strategy into a regulator-ready study plan. The pilot tests the device, the procedure and your endpoints before you commit to a large sample.
2. Ask the regulator before you enrol: Each market gives you a way to get early feedback:

• US: The FDA Pre-Submission (Pre-Sub) is free, and FDA aims to give written feedback within 70 days. The feedback is not binding, but it is the best way to test your endpoints and sample size. 

• EU: For Class III and certain Class IIb devices, you can consult an expert panel on your clinical plan. Early talks with your Notified Body also help. 

• India: For novel devices, a Subject Expert Committee (SEC) may review your file and decide whether a local study is needed. 

3. Choose endpoints that work for regulators and payers. Choosing appropriate endpoints also requires a defensible statistical strategy, including clinical biostatistics support for sample-size assumptions, analysis planning and interpretation.
4. Design one study for many markets. FDA can accept data from studies run outside the US if they follow Good Clinical Practice and meet the conditions in 21 CFR 812.28. So a well-designed study in India can support filings elsewhere. 

India’s own clinical investigation requirements include CDSCO and ethics committee approval, following the approved study plan and CDSCO GCP, and registering the study in the Clinical Trials Registry of India. 

India also offers practical advantages for device sponsors, including clinical trial services in India supported by large and varied patient groups, Strong hospital networks are valuable, but effective clinical trial site start-up is equally important for getting sites activated efficiently and comparatively lower running costs.:  

• Large and varied patient groups 

• Strong hospital networks  

• Lower running costs.  

Innovate Research supports device studies from protocol and medical writing through site management, monitoring, data management and biostatistics, under ISO 13485 certified quality systems. 

Throughout the study, clinical data management ensures that clinical data are captured, reviewed, cleaned and prepared for reliable statistical analysis and regulatory submission. 

Conclusion 

Risk class sets the starting point, but your claims, your predicate and your data gaps decide the final answer. Low risk devices rarely need a study. Moderate risk devices need careful, market-by-market judgement. High risk devices should plan for one unless a clear exemption applies. 

The companies that move fastest settle their medical device clinical trial requirements before design freeze, not after a regulator’s question. 

Not sure where your device stands? Talk to Innovate Research’s regulatory and clinical teams. We can run a clinical evidence gap assessment across FDA, EU MDR and CDSCO pathways, and help you design the leanest study that will hold up to review.  

Our services include: 

• Feasibility Studies 

• Medical Writing 

• Clinical Site Management 

• Biospecimen Collection 

• Clinical Data Management 

Visit Innovate Research to learn more about our services or Request a proposal. 

FAQs 

Does every medical device need a clinical trial? 

No. Most low risk devices and many moderate risk devices reach the market without a new study. They rely on bench testing, existing literature and data from similar devices. High risk and novel devices are the ones that usually need one. 

Does a 510(k) need clinical data? 

Usually not. FDA has said it requests clinical data for less than 10% of 510(k) submissions. It is more likely when your device has new technology, a new indication or a predicate with a newly found safety risk. 

What is the difference between a clinical evaluation and a clinical investigation? 

A clinical evaluation is a structured review of all existing clinical data on your device, such as literature and post-market data. A clinical investigation is a new study in human patients. In the EU, every device needs a clinical evaluation, but only some need a new investigation. 

Can we use clinical data from another country in our FDA submission? 

Yes. FDA can accept data from studies run outside the US if the study was well designed, followed Good Clinical Practice and meets the conditions in 21 CFR 812.28. Plan this from the start, as fixing gaps later is hard. 

Our device is already CE marked or FDA cleared. Do we still need a trial in India? 

Often not. If your device has been legally sold for at least two years in the US, UK, Australia, Canada, Japan or the EU, CDSCO can waive the local study when it is satisfied with your safety and performance data. Importers of Class A and B devices can usually submit published data and a free sale certificate instead.