Nyquist AI
Nyquist AI
MCP Prompt Library

15 MCP use cases for regulatory & clinical teams

Real workflows RA, QA, and clinical teams run today with the NyquistAI MCP server — the traditional approach they replaced, and the prompts that replaced it.

Not connected yet? See the MCP setup guide.

Jump to a use case

  1. 01Regulatory Intelligence
  2. 02Risk Analysis (FMEA / Hazard Analysis)
  3. 03Indications for Use (IFU) Definition
  4. 04Clinical & Bench Study Design
  5. 05RA Intelligence / Product Code Analysis
  6. 06Predicate Review
  7. 07Clinical Study Design
  8. 08Pre-Submission Finalization
  9. 09Performance Testing Requirements
  10. 10Submission Architecture & Authoring
  11. 11Signal Detection / Adverse Event Analysis
  12. 12PSUR / PMSR Automation
  13. 13Regulatory Requirements Monitoring
  14. 14CAPA Integration
  15. 15Guidance Gap Analysis
Use Case 01

Regulatory Intelligence

Background

RA teams manually tracked FDA guidance documents, product code changes, predicate clearance history, special controls, and competitive 510(k)/PMA activity. This required daily FDA website monitoring, email alert subscriptions, and expensive regulatory intelligence consultants. Missing a guidance update or new predicate could mean a delayed or rejected submission.

Traditional approach

  1. Subscribe to FDA email alerts (guidance, product codes, recalls)
  2. Manually review FDA website daily/weekly
  3. Maintain internal tracker spreadsheets of relevant guidances
  4. Search 510(k) database for predicates
  5. Download and read full 510(k) summaries for relevant devices
  6. Consult regulatory counsel for interpretation

Timeline: Days to weeks for comprehensive landscape; ongoing daily monitoring required

Example prompts

“Monitor all FDA guidance updates for product code QNP this week and flag anything that affects our 510(k) strategy.”
“Search for all 510(k) clearances for oximeters in the last 3 years. Identify the top 5 predicates for a new submission and extract their regulatory highlights.”
Use Case 02

Risk Analysis (FMEA / Hazard Analysis)

Background

Risk analysis (FMEAs, hazard analyses, benefit-risk assessments) required manual literature review, expert judgment, and regulatory precedent research. Teams started FMEAs from blank templates, relying on engineering memory and consultants at $200-400/hr. Real-world failure mode data from MAUDE and recall databases was rarely incorporated due to the labor involved in mining it.

Traditional approach

  1. Gather historical failure data from internal complaint system
  2. Manually search MAUDE database (complex interface)
  3. Download and read individual adverse event reports
  4. Conduct expert sessions to identify failure modes
  5. Map failure modes to harm categories (ISO 14971)
  6. Assign severity/probability estimates from literature and expert judgment
  7. Draft FMEA in spreadsheet

Timeline: 2-3 weeks for initial FMEA

Example prompts

“Pull all MAUDE reports for hemodialysis catheters (product code MSD) in the last 3 years. Categorize by device problems with frequency counts. Format as an FMEA input table with device problem, frequency, and representative report descriptions.”
Use Case 03

Indications for Use (IFU) Definition

Background

Defining the Indications for Use statement required reading competitor clearances, FDA predicate guidance, and legal review of intended use language. Overly broad IFU = FDA pushback. Too narrow = limited commercial market. RA teams spent weeks calibrating the IFU across internal stakeholders, often with 2-3 rounds of FDA comments before acceptance.

Traditional approach

  1. Research predicate IFU language from 510(k) summaries (manual download + review)
  2. Review applicable FDA guidance for intended use requirements
  3. Draft IFU with RA team
  4. Legal review for scope and liability
  5. Clinical review for accuracy
  6. Multiple internal revision cycles
  7. Submit - often receive FDA comments requiring IFU revision

Timeline: 2-4 weeks; often 2-3 revision cycles

Example prompts

“Compare the Indications for Use language across the top 10 cleared 510(k)s for product code QNP. Identify the common language patterns FDA accepted, and recommend the optimal intended use statement for our gastrointesinal lesion detection system.”
Use Case 04

Clinical & Bench Study Design

Background

Clinical and bench study designs required internal clinical experts, consultants, and extensive literature review to determine test methods, sample sizes, acceptance criteria, and comparators. FDA often requested additional testing during review because the study design didn't match what they had previously accepted - delaying clearance by 6-12 months and adding $100K-$500K in testing costs.

Traditional approach

  1. Review applicable standards (ISO, ASTM, IEC) for test methods
  2. Consult internal clinical/biomedical engineers
  3. Review published literature for accepted methods
  4. Draft study protocol
  5. Internal review and approval
  6. Submit - often receive AI request requiring study changes
  7. Redesign and retest

Timeline: 4-8 weeks for protocol; potential 6-12 month delay

Example prompts

“Analyze all 510(k)s cleared in the last 2 years for product code QJQ. Extract the biocompatibility testing methods, sample sizes, acceptance criteria, and sterilization validation approaches that FDA accepted. Format as a testing matrix grouped by test category.”
Use Case 05

RA Intelligence / Product Code Analysis

Background

Before starting a new device project, RA teams needed to identify the correct regulatory pathway, product code, device classification, special controls, and predicate landscape. No single source had all this. Getting pathway wrong early meant costly pivots 12-18 months into development. For MSD's dialysis system, the team needed rapid product code analysis and time-to-market projections before committing to the 510(k) pathway.

Traditional approach

  1. Search FDA product code database manually
  2. Research device classification regulations (21 CFR)
  3. Review special controls guidance documents
  4. Search 510(k) database for predicate clearances
  5. Consult regulatory counsel for pathway recommendation
  6. Research clearance timelines in FDA's CDRH database
  7. Draft regulatory strategy memo for leadership

Timeline: 2-3 weeks for initial RA strategy

Example prompts

“Find all devices cleared under product code QJQ in the last 5 years, sorted by clearance time. For the fast clearances, extract their predicate strategies, testing approaches, and any patterns associated with faster clearance. Provide a median clearance timeline.”
Use Case 06

Predicate Review

Background

Identifying and reviewing predicates was one of the most time-consuming 510(k) preparation steps. Teams manually searched the database, downloaded PDFs, read hundreds of pages, and determined substantial equivalence comparators. For MSD's dialysis system, there were over 250 cleared devices in the relevant product code - making manual review impossible within a competitive timeline. Traditional approach: 6-12 months. MSD needed it in 1 month.

Traditional approach

  1. Search 510(k) database by product code
  2. Download all relevant 510(k) summaries (often 50-200+ PDFs)
  3. Read and summarize each (30-60 minutes each)
  4. Create comparison matrix in spreadsheet
  5. Identify top predicates based on intended use similarity
  6. Obtain full 510(k) files for top candidates
  7. Document substantial equivalence rationale

Timeline: 3-6 months for thorough predicate review

Example prompts

“From all cleared devices in product code MSD, first narrow to devices with intended use matching hemodialysis access. Then further narrow it to devices with similar catheter tip designs. From the remaining, extract their predicate chains, performance testing summaries, and substantial equivalence arguments. Identify which predicate is most defensible.”
Use Case 07

Clinical Study Design

Background

Clinical evaluation reports and study protocols required clinical experts, reference to predicate clinical data, MEDDEV/MDCG guidances, and FDA recommendations. For novel devices, FDA often required unexpected clinical evidence discovered only after submission via AI requests. For MSD's dialysis system, understanding what clinical data FDA had accepted for predicate dialysis devices was critical to scoping the study appropriately.

Traditional approach

  1. Review MEDDEV 2.7/1 and FDA clinical evidence guidance
  2. Research clinical literature for device type
  3. Consult clinical affairs team and medical advisor
  4. Review predicate submissions for clinical data requirements
  5. Draft clinical evaluation plan and study protocol
  6. Internal review and IRB consultation
  7. Submit - risk of AI request for additional clinical data

Timeline: 4-8 weeks for protocol; potential 6-12 month delay

Example prompts

“Compare recently approved dialysis catheter devices' clinical profiles. For each: (1) summarize the clinical data submitted, (2) identify endpoints and sample sizes FDA accepted, and (3) flag any performance gaps that FDA is likely to require clinical evidence to bridge.”
Use Case 08

Pre-Submission Finalization

Background

Pre-submission meetings require weeks of preparation: drafting specific questions, anticipating FDA responses, formatting the Pre-Sub document per FDA's template, and coordinating inputs from RA, clinical, and engineering. Poorly prepared Pre-Subs yield vague FDA feedback. For MSD, the Pre-Sub was the critical moment to align with FDA on testing strategy before committing resources to performance testing.

Traditional approach

  1. Draft Pre-Sub questions through multiple internal review cycles
  2. Research FDA's likely position on each question
  3. Format per FDA's Q-Sub guidance
  4. Gather device description, intended use, and regulatory history
  5. Coordinate RA, clinical, and engineering inputs
  6. Legal review
  7. Format and submit via eSTAR

Timeline: 3-4 weeks

Example prompts

“Draft our Pre-Submission meeting materials for a 510(k) on a hemodialysis catheter (product code MSD). Based on FDA's feedback patterns in similar 510(k) reviews, identify the 5 highest-risk areas where FDA is likely to request additional information. Format each question per FDA's Pre-Sub guidance template: (1) question background, (2) specific question, (3) our proposed approach, (4) regulatory basis.”
Use Case 09

Performance Testing Requirements

Background

Determining the right performance testing battery (biocompatibility, sterilization, mechanical, electrical, software) required standards research, consultant review, and often iterative FDA questions. The consequence of wrong tests: AI requests requiring additional testing, adding 6-12 months and $100K-$500K. In the MSD case, FDA initially requested $300K in cleaning validation testing.

Traditional approach

  1. Review applicable standards (ISO 10993, IEC 60601, ASTM, etc.)
  2. Consult regulatory affairs and biomedical engineering
  3. Review FDA's 510(k) guidance for device type
  4. Build testing matrix based on internal expertise
  5. Submit - risk of FDA requesting additional tests not initially planned
  6. Conduct additional testing and resubmit

Timeline: 4-8 weeks initial matrix; months of delay if additional tests required

Example prompts

“Pull all MAUDE adverse event reports for hemodialysis catheters (product code MSD). Filter for reports where cleaning or reprocessing failure is listed as the primary cause. Calculate total reports, reports related to cleaning, and provide a summary I can use in my Pre-Sub response to FDA to challenge the cleaning validation requirement.”
Use Case 10

Submission Architecture & Authoring

Background

510(k) and PMA submissions required RA writers to manually compile hundreds of pages across device description, indications, substantial equivalence argument, performance testing summary, labeling, and biocompatibility sections. Each submission started largely from scratch. FDA reviewers frequently commented on formatting issues and missing elements - adding review time even when the science was sound.

Traditional approach

  1. Gather all inputs: test reports, labeling, device description, predicate comparison
  2. Map to FDA's 510(k) format requirements manually
  3. Write each section (device description, IFU, SE argument, testing, labeling)
  4. Internal review cycles (RA, clinical, legal, engineering)
  5. Format per eCopy or eSTAR requirements
  6. Final QC review

Timeline: 4-12 weeks for full submission authoring

Example prompts

“Draft Section 10 (Performance Testing Summary) for our 510(k). Format per FDA's guidance. Compare our results to the predicate's accepted data and highlight substantial equivalence.”
“Draft the Device Description section for our 510(k). Include: intended use, indications, device components, materials, design features, and how it compares to our predicate K202150.”
Use Case 11

Signal Detection / Adverse Event Analysis

Background

MDR/MedWatch signal detection required manual review of MAUDE reports, trending analysis in spreadsheets, and monthly/quarterly safety committee reviews. Teams rarely caught signals proactively - signals were typically identified only after appearing in FDA's recall database or a competitor's Field Safety Notice. For companies with multiple device families, monitoring the entire portfolio was practically impossible manually.

Traditional approach

  1. Monthly/quarterly download of MAUDE reports for product codes
  2. Filter and categorize reports in spreadsheets
  3. Trend analysis against prior period baseline
  4. Present findings at safety committee meeting
  5. Evaluate against internal MDR threshold criteria
  6. File MDRs if threshold met

Timeline: Quarterly cycle; signals often identified 3-6 months after emergence

Example prompts

“Pull all MAUDE reports for product code QNP submitted in the last 180 days. Compare against the prior 180-day period. Identify: (1) new device problems not present in the prior period, (2) device problems showing >20% increase in frequency, and (3) any reports that meet our internal MDR threshold criteria. Output as a signal detection report.”
Use Case 12

PSUR / PMSR Automation

Background

Periodic Safety Update Reports (PSURs) under EU MDR/IVDR and Post-Market Surveillance Reports (PMSRs) required RA teams to manually compile adverse event data from multiple national databases (MAUDE, MHRA, BfArM, TGA), literature updates, complaint rates, and benefit-risk assessments. Each PSUR took 4-8 weeks of dedicated RA/clinical time. For companies with large device portfolios, PSUR requirements were a Notified Body renewal bottleneck.

Traditional approach

  1. Pull adverse event data from each national database (manual downloads)
  2. Compile complaint data from internal quality system
  3. Run updated literature search
  4. Update benefit-risk assessment
  5. Draft PSUR/PMSR following MEDDEV/MDCG guidance structure
  6. Clinical and regulatory review cycles
  7. Submit to Notified Body

Timeline: 4-8 weeks per PSUR; annually or every 2 years per device

Example prompts

“Generate a PSUR adverse event analysis section for our Class II cardiovascular device (product code MHX). Pull MAUDE data for the last 12 months, compare to our prior PSUR baseline (attached), identify any new signals or trend changes, and draft the benefit-risk conclusion section referencing MDCG 2020-6 guidance.”
Use Case 13

Regulatory Requirements Monitoring

Background

Tracking changes to applicable standards (ISO 13485, ISO 14971, IEC 60601, ASTM) and FDA guidances and international regulations (EU MDR, Health Canada, TGA) required dedicated regulatory intelligence staff or expensive subscription services. Companies routinely missed effective dates, discovered compliance gaps during audits, and scrambled to update technical files after standards were already superseded.

Traditional approach

  1. Subscribe to standards bodies (ISO, IEC, ASTM) - expensive
  2. Set up Google Alerts for FDA guidance topics
  3. Subscribe to regulatory intelligence newsletters
  4. Manually track effective dates in spreadsheet
  5. Conduct periodic gap assessments against current technical files
  6. Assign remediation actions to functional teams

Timeline: Ongoing; gaps discovered at audit rather than proactively

Example prompts

“What new FDA guidances have been issued in the last 6 months that apply to our Class II cardiovascular devices? For each applicable guidance: (1) summarize the key requirements, (2) identify whether it is final or draft, (3) flag any requirements that differ from current practice, and (4) provide recommended action.”
Use Case 14

CAPA Integration

Background

CAPA investigations required correlating internal complaint data with MAUDE reports, production records, and risk management files. Root cause analysis was manual and siloed - teams rarely leveraged external data from competitor devices with similar failure modes, missing valuable industry-wide insights. CAPA inadequacy is consistently among the top FDA 483 observations.

Traditional approach

  1. Receive complaint or adverse event triggering CAPA
  2. Review internal complaint history for similar events
  3. Gather production records and nonconformance reports
  4. Conduct root cause analysis (fishbone, 5-why, FTA)
  5. Draft CAPA document per internal SOP
  6. Obtain approvals; implement and verify effectiveness

Timeline: 4-12 weeks per CAPA investigation

Example prompts

“For our CAPA #2024-018 (catheter tip delamination during removal), pull all MAUDE reports for hemodialysis catheters (product code MSD) where tip separation or delamination is the reported problem. For the top 10 most detailed reports: (1) extract the described root cause, (2) identify what corrective actions manufacturers reported, and (3) note whether design change vs. process change was the primary corrective action.”
Use Case 15

Guidance Gap Analysis

Background

When new FDA guidance is finalized, determining what updates are required across all submissions, labeling, SOPs, and technical files required a cross-functional impact assessment lasting 4-8 weeks. Companies with large portfolios could not practically assess impact across all device families and hoped auditors would not find gaps. For cybersecurity, PFAS, and software guidances (all recently updated), the gap analysis burden was particularly acute.

Traditional approach

  1. Read new/updated final FDA guidance
  2. Identify applicable device families
  3. Review current technical files, SOPs, and submissions against new requirements
  4. Meet with cross-functional team to assess gaps
  5. Create remediation plan with owners and dates
  6. Track remediation to closure

Timeline: 4-8 weeks; often incomplete for large portfolios

Example prompts

“Run a gap analysis of our Class II IVD technical files (attached) against the updated FDA guidance on Cybersecurity in Medical Devices. For each requirement: (1) identify the specific obligation, (2) assess whether our documentation addresses it (yes/partial/no), (3) rank the gap by regulatory risk (Critical/High/Medium/Low), and (4) recommend a specific remediation action with estimated effort.”