Mobile App Development in Zurich, Switzerland

MedTech / Healthcare

A Zurich-based MedTech startup had developed a wearable device for continuous respiratory monitoring — worn as a chest patch, it measures breathing rate, breathing pattern, blood oxygen estimation, and cough frequency. The device was designed for: post-operative respiratory monitoring (patients discharged after thoracic surgery who need 7-14 days of respiratory monitoring), COPD management (chronic patients tracking respiratory function daily), and clinical trials (continuous respiratory data collection for pharmaceutical studies). The startup had CE marking for the device hardware but needed a companion mobile app that would: receive data from the wearable via Bluetooth Low Energy, display respiratory metrics in patient-friendly visualizations, alert patients and clinicians when metrics exceed thresholds (e.g., breathing rate above 25/minute, SpO2 estimate below 92%), and transmit data to the clinical dashboard used by physicians. The regulatory challenge: the companion app processes medical data and generates alerts that influence clinical decisions — classifying it as a Class IIa medical device under EU MDR. The app itself needed CE marking, separate from the device hardware. The startup had attempted to build the app with a generalist app development agency — after 6 months and CHF 180,000 spent, the agency delivered an app that worked technically but had no documentation suitable for regulatory submission (no design history file, no risk management file, no IEC 62304 lifecycle documentation). The startup needed to restart app development with a team that understood medical device software regulation.

We rebuilt the companion app over 16 weeks, with integrated regulatory documentation: (1) Regulatory architecture: before writing code, we established: intended use statement (precisely defining what the app does and does not do — critical for regulatory classification), risk management plan per ISO 14971 (identifying hazards: what happens if the app fails to display an alert? What happens if Bluetooth connectivity drops during a critical episode?), software safety classification per IEC 62304 (the app was classified as Class B software — software that could contribute to a hazardous situation but not directly cause serious injury), and design input requirements (traceable from user needs to design requirements to implementation to verification). Every feature has a traceable path from clinical need through risk analysis to implementation to testing. (2) Patient app (iOS and Android): real-time respiratory dashboard showing: breathing rate (breaths/minute with trend graph), breathing pattern (visualized as a waveform), estimated SpO2 (with color-coded normal/warning/critical zones), and cough count (daily count with week-over-week trend). Alert system: configurable thresholds (set by the prescribing physician through the clinical dashboard) trigger push notifications and audible alarms. Alerts are classified: yellow (attention — one metric borderline), orange (warning — one metric exceeded), red (urgent — multiple metrics exceeded or critical threshold breached). Each alert includes a clear instruction: "Your breathing rate has been elevated for 30 minutes. If you feel short of breath, contact your care team. Tap here to call." (3) Bluetooth Low Energy data synchronization: the wearable transmits compressed respiratory data every 30 seconds during active monitoring. The app handles: BLE connection management (automatic reconnection after drops — common in hospital environments with BLE congestion), data integrity verification (checksum validation — medical data must be bit-accurate), offline data buffering (if BLE disconnects, the wearable stores data locally and the app retrieves it on reconnection — no data gaps), and background data reception (the app continues receiving data when backgrounded or when the phone is locked — critical for overnight monitoring). (4) Clinical data transmission: the app transmits anonymized respiratory data to the cloud clinical dashboard every 5 minutes (or immediately for critical alerts). Data transmission uses end-to-end encryption with certificate pinning. The clinical dashboard (a separate web application for physicians) displays: patient respiratory trends, alert history, and comparison to post-operative recovery benchmarks. (5) Quadrilingual support: the app supports German, French, Italian, and English — covering all Swiss language regions plus international clinical trial sites. Medical terminology is validated per language by clinical advisors (medical translation is not standard translation — the German medical term for "breathing rate" is "Atemfrequenz," not a literal translation). (6) Regulatory documentation: delivered alongside the app code: design history file (complete traceability from user needs to verification results), risk management file (ISO 14971 — hazard analysis, risk evaluation, risk control, residual risk assessment), software development lifecycle documentation (IEC 62304 — software development plan, architecture description, detailed design, unit and integration test plans, and test results), cybersecurity documentation (MDCG 2019-16 — threat model, security architecture, vulnerability management plan), and clinical evaluation contribution (the app's performance data — alert accuracy, data transmission reliability — contributed to the device's overall clinical evaluation per MDR Article 61).

Regulatory submission: the app achieved CE marking as a Class IIa medical device 8 weeks after development completion. The notified body reviewer specifically noted that the software documentation was among the most complete they had reviewed for a companion app — contrasting with the typical experience of medical device companies submitting apps with incomplete software lifecycle documentation. Clinical deployment: the app was deployed in 3 Swiss hospitals (Universitätsspital Zürich, Inselspital Bern, CHUV Lausanne) for post-operative respiratory monitoring. 180 patients used the app during the first 6 months. Alert accuracy: 94.2% of alerts were clinically relevant (true positive rate — the patient genuinely needed attention). 5.8% were false positives (mostly caused by movement artifacts during physical therapy). Zero false negatives (no critical respiratory events were missed). Clinical outcome: physicians reported that the app-based monitoring detected 12 respiratory deterioration episodes that would have been missed with standard 4-hourly nursing observations — potentially preventing 3 ICU readmissions (estimated cost avoidance: CHF 45,000-75,000 per avoided ICU readmission). Data reliability: 99.7% data transmission completeness (0.3% gaps due to extended BLE disconnections in areas with heavy RF interference). Bluetooth reconnection time: average 8 seconds after disconnect. Clinical trial adoption: 2 pharmaceutical companies licensed the device + app combination for respiratory endpoint measurement in clinical trials (combined licensing revenue: CHF 1.2 million in the first year). Development cost: CHF 320,000 (vs. CHF 180,000 spent on the failed first attempt that produced no usable regulatory documentation). The regulatory-integrated approach cost 78% more but delivered a CE-marked product — the cheaper approach produced nothing usable.