NPI Process for Medical Wearables: EVT → DVT → PVT → MP Milestone Guide

Most ODM projects do not fail because of technology. They fail because someone skipped a gate review at DVT, and nobody noticed until the pilot build came back with a 62% yield.

This guide walks through the complete New Product Introduction (NPI) process for medical wearable devices — from the moment a product requirements document is frozen to the day the first container ships. It covers what each phase validates, what documentation you should expect at each gate, where 70% of projects encounter their first serious delay, and how to read a milestone review checklist without being a hardware engineer.

If you are a brand owner evaluating an ODM partner, or a project manager responsible for an upcoming wearable launch, this is the framework you need to assess whether a timeline is realistic — and whether a partner is being transparent.

Why 70% of ODM Wearable Projects Miss Their First Ship Date

The pattern is remarkably consistent. A project kicks off with an aggressive 6-month timeline. EVT completes on schedule. The team feels good. Then DVT starts, and within two weeks, three things happen simultaneously: a tooling issue delays the housing, reliability testing reveals a connector problem, and the regulatory team realizes the battery certification won’t be ready in time for the submission window.

None of these are surprises to an experienced NPI team. All of them are preventable with the right gate review discipline.

The root cause is almost always the same: DVT was treated as a continuation of EVT rather than a fundamentally different phase. EVT asks “does it work?” DVT asks “will it work every time, under stress, across the tolerance stack, in the hands of real users?” The jump in rigor is significant, and projects that underestimate it pay with 8-12 weeks of unplanned delay.

The NPI Timeline: 12 Stages from PRD Freeze to Mass Production

Every medical wearable NPI follows a predictable sequence. The exact names vary by company, but the logic is universal. Here is the full chain with typical cycle times for a smartwatch-class device:

Stage	Phase	Typical Duration	Key Output
1. PRD Freeze	Planning	Week 0	Signed product requirements document
2. Industrial Design Lock	Planning	Week 2-4	ID renders, CMF document, mechanical constraints
3. Schematic Design	EVT Prep	Week 4-6	Block diagram, BOM v0.1, key component selection
4. PCB Layout + Stacking	EVT Prep	Week 6-10	Gerber files, stack-up drawing, DFM review
5. EVT Build	EVT	Week 10-14	~50-100 units, functional validation
6. EVT Gate Review	EVT	Week 14	EVT test report, issue list, DVT readiness
7. Tooling + DVT Sample Build	DVT	Week 14-18	T0/T1 samples, tooling verification
8. DVT Reliability Testing	DVT	Week 18-22	Reliability test matrix results
9. DVT Gate Review	DVT	Week 22	DVT report, compliance pre-scan, BOM freeze
10. PVT Pilot Build	PVT	Week 22-26	~200-500 units, production line validation
11. PVT Gate Review	PVT	Week 26	Yield data, CPK report, FAI report
12. MP Ramp	MP	Week 26+	Mass production release, ongoing yield monitoring

Realistic total: 6-8 months for a new smartwatch or smart ring from PRD freeze to MP. Projects that claim 4 months are either reusing an existing platform or skipping steps — and the skipped steps will resurface as field failures.

EVT — Engineering Validation Test: Does the Design Work?

EVT is the first physical build. The goal is not to produce a shippable device. It is to prove that the schematic translates into a functional board, that the mechanical design fits together, and that the core sensors produce data.

What EVT validates

• Power-on and basic function: All power rails come up within spec. The MCU boots. The display lights up. The sensors initialize.
• RF performance: Bluetooth and Wi-Fi connectivity at close range. Antenna tuning is measured but not yet optimized.
• Sensor signal acquisition: PPG, ECG, accelerometer, and temperature sensors produce raw data. Accuracy is measured but not required to meet final spec at this stage.
• Mechanical fit: The PCB sits in the housing. The buttons align. The band attaches. Tolerances are checked but not finalized.
• Charging function: The charging pins make contact and the battery charges. Charge current and cutoff are verified.

EVT build size and expectations

Typically 50-100 units. Yields of 60-80% are normal. EVT boards often have rework wires, and enclosures may be 3D-printed or soft-tooled. This is expected and should not alarm anyone.

EVT deliverables

• EVT test report with pass/fail for each functional block
• Issue list with severity classification (critical / major / minor)
• Updated BOM with identified alternate sources for long-lead components
• Preliminary power consumption measurement
• Antenna performance baseline (VSWR, efficiency)

EVT Gate Review checklist

1. All critical functions verified (power, connectivity, sensors, display, charging)
2. All critical issues have a documented root cause and corrective action plan
3. Mechanical stack-up confirmed with actual parts
4. BOM cost within 15% of target
5. Long-lead components identified with confirmed delivery dates
6. DVT test plan approved by both ODM and brand owner
7. Tooling Kick-off authorized

Common EVT pitfall: Rushing to DVT before all critical EVT issues are closed. A connector that fails once in EVT will fail ten times more often in DVT reliability testing. Close every critical issue before signing the gate.

DVT — Design Validation Test: Will It Work Every Time?

DVT is the hardest phase in the NPI process, and it is where most delays originate. The goal shifts from “does it work?” to “does it work across the full tolerance stack, under environmental stress, for the intended service life?”

DVT uses production-intent tooling, production-intent components, and production-intent assembly processes. The units tested should be indistinguishable from what will ship — because if they are not, the test results are not predictive.

What DVT validates

• Sensor accuracy against reference: Final calibration applied. Accuracy measured against a clinical-grade reference device under controlled conditions, across the stated measurement range, on a representative population.
• Reliability testing: A structured matrix of environmental and mechanical stress tests designed to expose weak points before they reach the field.
• Regulatory pre-compliance: EMC pre-scan, safety pre-scan, and SAR pre-scan (if applicable). These are not formal submissions but are close enough to identify issues before the expensive certification round.
• Water and dust ingress: IP rating verification at a certified lab or in-house with calibrated equipment.
• Battery safety: UN 38.3 pre-screening, overcharge/over-discharge protection verification, thermal runaway containment.
• User experience: Haptic feedback, button feel, display readability in sunlight, band comfort over extended wear.

DVT reliability test matrix

Test	Standard Reference	Conditions	Sample Size
Thermal cycling	IEC 60068-2-14	-20°C to +70°C, 100 cycles, 2h dwell	10
High-temperature storage	IEC 60068-2-2	70°C, 96 hours	5
Low-temperature storage	IEC 60068-2-1	-30°C, 96 hours	5
Damp heat cyclic	IEC 60068-2-30	25°C-55°C, 95% RH, 6 cycles	5
Drop test	IEC 60068-2-31	1.2m onto concrete, 6 faces, 3 drops each	5
Vibration	IEC 60068-2-6	10-500 Hz, 1 octave/min, 3 axes	5
Salt spray	IEC 60068-2-11	5% NaCl, 35°C, 48 hours (for corrosion resistance)	3
Sweat resistance	Internal protocol	Artificial sweat (pH 4.7), 35°C, 48h contact	5
Button cycle life	Internal protocol	100,000 cycles minimum	5
Band attachment strength	Internal protocol	Tensile test to failure, minimum 10kg	5
Charging cycle life	Internal protocol	500 full cycles, capacity retention ≥ 80%	5

Why sweat resistance matters: Medical wearables are worn during exercise, sleep, and in some cases 24/7. A device that corrodes after three months of daily wear will generate a wave of returns that no warranty budget can absorb. This is one of the most under-tested failure modes in the industry.

DVT deliverables

• DVT test report with pass/fail for each test item in the reliability matrix
• Sensor accuracy validation report against reference device
• EMC pre-scan report with identified issues and corrective actions
• IP rating test report
• Updated BOM with all alternate sources qualified
• Final ID sign-off with production-intent surface finish and color
• Pilot production process flow document

DVT Gate Review checklist

1. All reliability tests passed, or documented deviations with risk assessment accepted by brand owner
2. Sensor accuracy meets specification across full measurement range
3. Regulatory pre-scans passed or issues with corrective action plan
4. BOM frozen — no further component changes without formal ECN
5. All tooling finalized and qualified
6. Production test fixtures designed and validated
7. PVT build plan approved
8. Packaging design frozen

Common DVT pitfall: Accepting a test failure with a “fix in PVT” note. DVT is the last phase where a design change is relatively inexpensive. A change made during PVT risks invalidating the entire DVT test corpus and triggering a re-test cycle that costs 4-6 weeks. Fix it at DVT or accept the risk in writing.

PVT — Production Validation Test: Can We Build It at Scale?

PVT is not about the design anymore. The design is frozen. PVT is about the production system: the assembly line, the test stations, the operator training, the incoming inspection process, and the supply chain’s ability to deliver consistent quality at volume.

What PVT validates

• Production line yield: First-pass yield (FPY) target is typically ≥ 90% for a mature product. PVT yields of 80-90% are acceptable with a documented improvement plan.
• Process capability (CPK): Critical-to-quality parameters (antenna performance, sensor calibration, water resistance) must demonstrate CPK ≥ 1.33.
• Test station repeatability: Gauge R&R study on each test station. The measurement system must contribute less than 10% of total observed variation.
• Operator training effectiveness: All operators certified on their stations. Training records documented.
• Supply chain consistency: Incoming quality control (IQC) data on at least 3 production batches of each critical component.

PVT build size

Typically 200-500 units, built on the actual production line by trained operators, using production tooling, production test fixtures, and production components from the frozen BOM. The build should be split across at least two shifts to detect shift-dependent variation.

PVT deliverables

• Production yield report with Pareto analysis of top failure modes
• CPK report for all critical-to-quality parameters
• Gauge R&R report for all test stations
• First Article Inspection (FAI) report per ISO 9001 or AS9102 format
• Production process control plan
• In-process inspection checklist
• Outgoing quality control (OQC) sampling plan with AQL levels
• Production bill of materials (PBOM) with work instructions

PVT Gate Review checklist

1. First-pass yield ≥ 85% with documented improvement trajectory to ≥ 90%
2. All critical CPK values ≥ 1.33
3. All test stations qualified (Gauge R&R ≤ 10%)
4. FAI completed and approved
5. All operators certified
6. IQC process established for all critical components
7. OQC sampling plan approved
8. Packaging line validated
9. Mass Production Release authorized

Common PVT pitfall: Releasing to MP with a yield below 85%. The gap between PVT yield and MP yield does not close on its own — it requires dedicated engineering resources. If those resources are not allocated before the MP release, the line will run at low yield indefinitely, burning margin and delaying shipments.

MP — Mass Production: The Steady State

MP is not a gate. It is a continuous monitoring state. The goal is to maintain the quality level validated at PVT while scaling volume.

MP ongoing responsibilities

• Daily yield monitoring: FPY tracked per shift. Any drop below control limits triggers an immediate line stop and root cause analysis.
• Ongoing reliability testing (ORT): A sample pulled from every production batch and subjected to a subset of the DVT reliability matrix. This catches process drift before it reaches the field.
• Supplier quality management: Incoming inspection data trended. Suppliers with declining quality scores placed on controlled shipping until corrective action is verified.
• ECN control: Any change to the frozen BOM or production process requires a formal Engineering Change Notice, with re-validation scope determined by the change impact assessment.
• Field failure analysis: Every returned unit analyzed. Root cause determined. Corrective action tracked to closure. This loop is the single most valuable source of product improvement data.

ODM vs Brand Owner: The RACI Matrix

One of the most common sources of friction in NPI is unclear responsibility boundaries. The ODM owns the execution. The brand owner owns the decisions. Here is the standard division:

Activity	ODM	Brand Owner
Product Requirements Document	Consulted	Accountable
Industrial Design	Responsible	Accountable (sign-off)
Schematic + PCB Design	Accountable	Informed
Component Sourcing	Accountable	Informed (cost approval)
EVT/DVT/PVT Build	Accountable	Informed
Test Plan Approval	Responsible (draft)	Accountable (approval)
Reliability Testing	Accountable	Informed
Gate Review Sign-off	Responsible	Accountable
Regulatory Submission	Responsible (data)	Accountable (filing)
Packaging Design	Responsible	Accountable (approval)
Production Yield	Accountable	Informed
ECN Approval	Responsible (proposal)	Accountable (approval)

R = Responsible (does the work) · A = Accountable (signs off) · C = Consulted (provides input) · I = Informed (receives updates)

The pattern is clear: the ODM does the work; the brand owner makes the decisions. Problems arise when the brand owner tries to do the work (micromanaging the factory) or when the ODM tries to make the decisions (shipping without approval). A well-run NPI respects this boundary.

How to Compress the Timeline — and What You Should Never Skip

Legitimate compression strategies

• Platform reuse: Starting from an existing qualified platform (same MCU, same sensor suite, same connectivity module) can cut 8-12 weeks from the schedule. This is the single biggest lever.
• Parallel tooling and EVT: If the ID is locked and the mechanical design is low-risk, tooling can start during EVT rather than waiting for EVT sign-off. This requires confidence in the mechanical design and acceptance of tooling modification risk.
• Concurrent certification: Regulatory submissions can begin during DVT if the design is stable and pre-scans are clean. Do not submit before DVT gate review — a post-submission design change can invalidate the filing.
• Overlapping PVT and MP preparation: Packaging, manuals, and production SOPs can be prepared during PVT rather than after.

What you should never skip or compress

• DVT reliability testing: There is no shortcut. Accelerated life testing takes time because physics takes time. Skipping a 96-hour humidity test because “we’re behind schedule” is how you ship devices that corrode in the field.
• Gate reviews: A gate review is not a meeting. It is a documented checkpoint with acceptance criteria. Skipping it does not save time — it defers cost to a later phase where it is 10x more expensive.
• PVT yield ramp: Do not ship from the PVT build. The PVT build exists to find production issues. Ship from MP, after the yield curve has stabilized.
• Field failure analysis loop: The first 1,000 units in the field will teach you more than all of DVT combined. Set up the feedback loop before MP starts, not after the first returns arrive.

Frequently Asked Questions

What do EVT, DVT, and PVT stand for in medical wearable development?

EVT (Engineering Validation Test) validates that the design works functionally — the first physical build proves the schematic, mechanical fit, and sensor acquisition. DVT (Design Validation Test) validates that the design works reliably under stress — production-intent units undergo a structured reliability test matrix including thermal cycling, drop, vibration, sweat resistance, and sensor accuracy validation. PVT (Production Validation Test) validates that the production system can build the device at scale with consistent quality — pilot builds on the actual production line measure yield, process capability (CPK), and test station repeatability.

How long does it take to develop a medical wearable from concept to mass production?

A realistic timeline from PRD freeze to mass production is 6-8 months for a new smartwatch or smart ring. Projects that start from an existing qualified platform can reduce this to 4-5 months. Claims of 3-4 months for a fully new design typically involve skipping DVT reliability testing or PVT yield validation — risks that will surface as field failures or production quality issues.

What reliability tests are required for medical wearable devices?

The core reliability test matrix includes: thermal cycling (-20°C to +70°C), high and low temperature storage, damp heat cyclic (95% RH), drop testing (1.2m onto concrete), vibration testing (10-500 Hz), salt spray corrosion testing, artificial sweat resistance testing, button cycle life (100,000+ cycles), band attachment strength (≥10kg tensile), and charging cycle life (500 cycles with ≥80% capacity retention). For medical-grade devices, sensor accuracy validation against a clinical reference device is also required during DVT.

Why do most ODM wearable projects experience delays at the DVT stage?

DVT is where the gap between “functional prototype” and “production-ready product” becomes visible. The three most common triggers are: tooling issues discovered only when production-intent housings arrive, reliability test failures that require design changes (which then require re-testing), and regulatory pre-compliance scans that reveal EMC or safety issues. These are all normal parts of DVT — the delay occurs when they were not planned for in the schedule. A realistic DVT plan includes 2-4 weeks of buffer for issue resolution and re-testing.

What is the difference between an ODM’s and a brand owner’s responsibilities during NPI?

The ODM is accountable for execution: PCB design, component sourcing, build management, test execution, and production yield. The brand owner is accountable for decisions: requirements sign-off, industrial design approval, test plan approval, gate review sign-off, and regulatory filing. The ODM proposes; the brand owner disposes. Problems arise when the brand owner tries to manage the factory floor, or when the ODM makes unilateral decisions that should require brand owner approval — such as substituting a component without an ECN.

---

Planning a medical wearable project and need a realistic NPI timeline? Our engineering team can provide a phase-by-phase schedule based on your product requirements, including reliability test plans and regulatory submission milestones. Contact us to discuss your project.