ROI Case File No.278｜'Medivida's Treatment App Development Labyrinth'

Chapter 1: The Incoming Request—Development Too Long, Endpoint Invisible

The week after resolving the QuantumVista Cloud SWOT analysis case, a consultation arrived from North America regarding a medical tech company's development project stagnation. Case File 278 of Volume 22 "The Pursuit of Reproducibility" is a story about moving forward a plan-collapsed project with small improvement cycles.

"Detective, we've been developing a treatment support app for diabetes patients since two years ago. We planned everything from requirements definition, design, development, testing, to maintenance operations. Yet two years later, we still haven't finished requirements definition."

Rachel Cohen, Development Director of Medivida from Boston, visited 221B Baker Street with a completely exhausted expression. In her hands were 300-page requirements definition documents contrasted sharply with zero development progress reports.

"We're a Massachusetts company developing medical devices and apps. We're trying to create an FDA (U.S. Food and Drug Administration)-approved treatment app. But requirements are so massive, we can't start development."

Medivida's Development Stagnation: - Founded: 2018 (medical tech startup) - Project start: January 2023 - Elapsed period: 24 months - Requirements document: 300 pages (incomplete) - Development progress: 0% - Invested hours: Cumulative 18,000 hours - Development cost: 320 million yen (requirements definition only)

Rachel's expression held deep anxiety.

"The problem is we're trying to create a 'perfect plan.' FDA approval criteria, physician requests, patient needs, technical constraints, security requirements, privacy regulations... Trying to perfectly define everything upfront, requirements definition never ends."

Plan Bloat: - Initial requirement items: 120 items - Current requirement items: 1,847 items (15x inflation) - Monthly meeting additions: Average 80 items - Stakeholders: Physicians, nurses, patients, regulators, insurers, IT department - Result of incorporating everyone's requests: Plan nobody can execute

Meeting Records: - Physician: "Want to record not just blood sugar but diet, exercise, sleep" - Nurse: "Need medication management feature" - Patient representative: "Make operation simple" - Regulatory: "Need all data traceability for FDA approval" - Security: "Medical information requires strict encryption and access control" - Insurer: "Make treatment effectiveness evidence recordable"

"Trying to meet all requests, the plan became massive and we're immobilized."

Chapter 2: Plan-Do-Check-Act Breathing—The Art of Small Cycles

"Rachel, how is your current project management conducted?"

To my question, Rachel answered.

"Waterfall-type development process. Perfect requirements definition before moving to design, perfect design before moving to development. The policy is completing each phase reliably before proceeding to the next."

Current Process (Waterfall Type): - Phase 1: Requirements definition (planned 12 months → actual 24 months, incomplete) - Phase 2: Basic design (planned 6 months → not started) - Phase 3: Detailed design (planned 6 months → not started) - Phase 4: Development (planned 12 months → not started) - Phase 5: Testing (planned 6 months → not started) - Phase 6: FDA application (planned 12 months → not started) - Total: Planned 54 months, 24 months elapsed with 0% progress

I explained the importance of continuous improvement.

"With complex projects like medical apps, you can't create perfect plans from the start. PDCA—Plan, Do, Check, Act. By repeating this small cycle, you advance reliably."

⬜️ ChatGPT｜Catalyst of Concepts

"Don't wait for perfect plans. Plan small, execute, learn, improve"

🟧 Claude｜Alchemist of Narratives

"Plans are maps, execution is the journey. No matter how perfect the map, the journey won't begin"

🟦 Gemini｜Compass of Reason

"PDCA is the technology of advancement. A thousand small steps lead farther than one big step"

The three members began analysis. Gemini deployed the "Medical App-Specialized PDCA" framework on the whiteboard.

PDCA Cycle's 4 Stages: 1. Plan - Set small goals and hypotheses 2. Do - Execute the plan 3. Check - Measure results, validate hypotheses 4. Act - Reflect learnings in next plan

"Rachel, let's advance Medivida's development with small PDCA cycles."

Chapter 3: Setting Phase 0—The Courage to Start with Minimum Features

Phase 1: Project Redefinition (1 week)

First, we organized the 1,847 requirement items by "priority."

Question: "What is this app's most important value?"

Stakeholder Interviews: - Physician: "Blood sugar recording and treatment suggestions based on it" - Patient: "Easily record daily blood sugar" - FDA regulatory: "Proof of safety and effectiveness as medical device"

Redefined Core Value: "Diabetes patients can record blood sugar, physicians can utilize it for treatment decisions"

Phase 2: Minimum Feature Set (MVP) Decision (3 days)

We classified the 1,847 items into "Must (essential)," "Should (desirable)," "Could (if possible)."

Must (Essential for FDA approval and treatment): Only 8 items 1. Manual blood sugar input 2. Save input data 3. Time-series graph display 4. Data sharing with physician 5. Data encryption 6. Access log recording (regulatory compliance) 7. Patient consent acquisition 8. Abnormal value alerts

Should (Nice to have): 32 items - Meal logging, exercise logging, medication management, automatic input (linked devices) etc.

Could (Future): 1,807 items - AI diagnostic support, family notifications, insurer linkage etc.

Decision: Start development with Phase 0, only 8 items

Rachel asked anxiously.

"Can we get FDA approval with just 8 features?"

"Phase 0's purpose isn't FDA approval. It's validating the hypothesis that 'blood sugar recording helps treatment.' Once validated, we add features in the next cycle."

Phase 3: First PDCA (4-week cycle)

Plan: - Goal: Develop 8-feature prototype in 4 weeks - Hypothesis: "Patients record blood sugar daily, physicians can utilize it for treatment" - Measurement indicators: Recording continuation rate, physician utilization rate

Do: - Week 1: Basic design and database construction - Week 2: Blood sugar input screen and graph display - Week 3: Physician sharing feature and security implementation - Week 4: Testing and improvement

Check: Test participants: 10 diabetes patients, 3 physicians (4-week usage)

Results: - Recording continuation rate: 70% (7 of 10 recorded daily) - Physician utilization rate: 100% (all 3 used for treatment decisions) - Patient voices: "Simple and sustainable" (8 people) - Physician voices: "Trends clear in graphs" (3 people)

Discovered Issues: - Want automatic input from blood sugar meters (6 people) - Want to record meal content too (4 people) - Alert timing too early (2 people)

Act: Features to prioritize in next cycle clarified 1. Blood sugar meter linkage (automatic input) 2. Simple meal logging 3. Alert setting customization

Chapter 4: Cycles 2-5 Acceleration—Evolution Through Learning

Phase 4: 2nd Cycle (4 weeks)

Plan: - Additional features: Meter linkage, simple meal logging, alert settings - Hypothesis: "Automatic input reduces recording burden, improving continuation rate"

Do: Implemented linkage API with 2 meter manufacturers

Check: Test participants: 25 (previous 10 + new 15) - Recording continuation rate: 70% → 88% (automatic input effect) - Recording accuracy improved (reduced manual input errors) - Meal logging usage: 52% (photo + simple memo format popular)

Act: Standardized meter linkage, strengthened meal logging

Phase 5: 3rd Cycle (4 weeks)

Plan: - Additional features: Medication management, exercise logging, physician comment feature - Hypothesis: "Communication with physician improves patient motivation"

Check: Test participants: 50 - Physician comment feature usage: Physicians 100%, patients 92% - Patient satisfaction: 4.6/5 - Blood sugar improvement: Average HbA1c 7.8% → 7.2% (0.6 point improvement)

Phase 6: 4th Cycle - FDA Application Preparation (8 weeks)

Plan: - Goal: FDA Class II medical device approval application - Preparation: Clinical data collection, safety evaluation, application document creation

Do: - 3-month clinical trial (200 patients, 15 physicians) - Security audit, data integrity verification - FDA application documents (510k) creation

Check: - Clinical trial results: Average 0.7 point HbA1c improvement - Serious adverse events/incidents: Zero - Data integrity: 100% (all records traceable)

Act: Official FDA application

Phase 7: 5th Cycle - Market Launch Preparation (12 weeks)

While awaiting FDA review, we prepared for market launch.

Plan: - Partnership negotiations with insurers - Trial deployment at medical institutions - Support structure construction

Do: - Partnership contracts with 2 major insurers - Pilot operation at 20 hospitals/clinics

Check: - Medical institution adoption rate: 85% - Patient treatment continuation rate improvement: +32% vs conventional - Medical cost reduction effect: 120,000 yen per patient annually

Results after 18 months (5 PDCA cycles completed):

Development Metrics: - Development period: Planned 54 months → Actual 18 months (67% shorter) - Development cost: Planned 1.5 billion yen → Actual 520 million yen (65% reduction) - Implemented features: Initial 1,847 items → Actual 42 items (focused on necessary features)

FDA Approval: - Application to approval: 8 months - Approved as Class II medical device - Approval rate: Single-attempt approval (no additional materials requested)

Business Results: - Partner medical institutions: 120 - Registered patients: 8,500 - Monthly revenue: 42 million yen (insurance reimbursement model) - Patient satisfaction: 4.7/5

Clinical Effectiveness: - HbA1c improvement: Average 0.8 points - Hypoglycemic events: 42% reduction - Treatment continuation rate: +38% improvement vs conventional

Chapter 5: The Detective's PDCA Diagnosis—The Art of Winning Long Battles

Holmes compiled the comprehensive analysis.

"Rachel, the essence of PDCA is 'continuous advancement.' Rather than spending 2 years creating perfect plans, build imperfectly in 4 weeks and repeat improvements. Long-term projects like medicine especially—small improvement accumulation generates maximum results."

Final Report after 24 months:

Medivida became a leading company in the U.S. diabetes treatment digital health market.

Final Achievements: - Registered patients: 8,500 → 42,000 - Partner medical institutions: 120 → 580 - Annual revenue: 500 million yen → 2.8 billion yen - Next-generation app (hypertension) also development started (expected completion in 6 months with PDCA)

Rachel's letter contained deep gratitude:

"Through PDCA, we transformed from 'an organization bound by plans' to 'an organization continuously learning.' Most important was the courage to start small without waiting for perfection. Repeating 4-week cycles 5 times, we completed 2-year plans in 18 months. Now all projects proceed with PDCA. Plans don't need perfection—just keep improving, we understood."

The Detective's Perspective—Improvement is the Journey's Compass

That night, I contemplated project management's essence.

The true value of PDCA lies in humility. You can't create perfect plans from the start. Then plan small, execute, learn, improve. Spinning this cycle fast generates reliable advancement.

Plans are maps, but execution is the journey. No matter how perfect the map, the journey won't begin. Even with imperfect maps, once you start walking, the path becomes visible.

"The key to escaping labyrinths isn't perfect maps. Take one step, hit walls, change direction, walk again. That repetition."

The next case will also depict a moment when continuous improvement opens a company's future.

"Perfect plans are beautiful illusions. But small steps are reliable reality"—From the detective's notes

--

Related files