Respiratory expertise and data science, focused on earlier clinical action
RespiroScreen is a small London team building early warning tools for pulmonology clinics supporting people with COPD and asthma between visits.
Earlier action. Fewer emergency admissions.
More than 100,000 UK patients are admitted to hospital each year following COPD flare-ups. In many cases, deterioration develops over 48 to 96 hours, creating time for clinical action. Until now, clinics have had little visibility into that period between appointments.
RespiroScreen gives clinics a clearer view of changing respiratory health. Each morning, a patient completes a 90-second check-in using their smartphone. Acoustic signal processing and ML models compare the recording with their personal baseline, then send the clinic a graded early warning alert when patterns suggest a developing exacerbation.
Our approach uses acoustic signal processing in a simple home routine. It adds continuity where regular clinical contact is limited. We will keep publishing evidence as the product and study base develop.
Three perspectives, one clinical challenge
Our bootstrapped team combines acoustic signal processing, clinical data infrastructure, and respiratory care experience. We stay close to the problem, with clinical input shaping product decisions from the start.
Brings acoustic signal processing and clinical decision support expertise to the respiratory monitoring problem, connecting technical development with the information clinicians need.
Focuses on the data infrastructure behind continuous monitoring, including distributed systems, medical data flows, and practical delivery for clinic teams.
Brings respiratory clinical perspective to alert design, evidence planning, and the question of what makes a signal useful during real-world patient care.
The standards guiding our work
An early-access programme at one site, involving 47 patients, showed encouraging findings. We state that limitation plainly and are working toward evidence from several sites and publication in peer-reviewed literature. We will not make claims beyond what the data can support.
Alerts should help busy clinical teams make decisions, not create another queue of work. We test each workflow choice against one practical question: does it help a clinician reach the right next step with less effort?
Respiratory recordings require careful handling. UK patient data is held on UK servers and encrypted at rest and in transit. It is not used to train models without patient opt-in. Consent and access should remain understandable for people with different levels of health and digital confidence.
RespiroScreen was founded in 2025 and remains a bootstrapped team. We are not yet CQC-registered or CE-marked as a medical device, and are working toward both milestones. Our evidence is early, so we describe it as such.
Why RespiroScreen exists
RespiroScreen began in 2025 around a straightforward clinical gap: clinics cannot respond to respiratory changes they cannot see. People living with COPD and asthma may have limited contact between visits, while deterioration can progress at home before an emergency admission becomes unavoidable.
The building blocks were available. Smartphone microphones can collect acoustic signals, while ML models can compare breath patterns against a person's baseline. The harder question was clinical: which changes matter, and how should they reach a team without disrupting care?
That question brings technical and respiratory perspectives together. Clinical science informs alert thresholds, dashboard priorities, and patient-facing language. The result is a deliberately focused product, built close to the everyday needs of clinics and the people they support.
Start a free pilot with your clinic
We work with your clinical team during setup. No EHR or IT integration is required. The first 90 days are free.