The central problem of outpatient pulmonary care is not knowledge. Clinicians know what to look for. The problem is time. The physician has limited time with the patient in the exam room, and then months pass between visits. A patient’s respiratory rate climbs. Her rescue inhaler becomes a crutch. Her heart rate spikes overnight and no one is there to see it. The visit is a photograph of a process that unfolds in video. Whatever the patient remembers to report, the clinician documents. The rest disappears.
We built a clinically integrated technology system to close that gap.
A pulmonary patient generates clinical data everywhere: office visits, pulmonary function tests, CT scans, lab panels, sleep studies, hospital records, medication histories, the offices of other physicians. In most practices, that data sits in the chart and nobody looks at it until the next appointment. We pull all of this into one database legible to AI, and surface key information to members of the care team at pivotal moments.
On top of that system of record, we add something no one else has: an FDA-cleared sleep mat that sits under the patient’s mattress. The patient doesn’t touch it, charge it, or think about it. Every night it streams heart rate, respiration rate, sleep stages, and snoring duration. We have thousands deployed, with new patients enrolling every week. For sleep patients, we also pull data directly from CPAP and BiPAP devices: usage hours, mask-leak patterns, AHI trends. For pulmonary rehab patients, a pulse oximeter captures oxygen saturation during exercise.
Every night, our system pulls the data from each patient’s mat and runs it against clinical thresholds. When something crosses a line, an alert fires to one of our nurses.
One night, a mat flagged a patient whose heart rate had hit 91.9 and whose respiratory rate had climbed to 21. The nurse called. The patient’s oxygen was below 80%. She was hospitalized for a COPD exacerbation that no one else had seen coming.
At another clinic, a mat picked up elevated heart rate and respiratory rate in an asthma patient. On her next call, she reported worsening symptoms. A CTA was ordered to rule out a pulmonary embolism, and the scan showed a mosaic-like pattern in her lungs. She was started on medication and is doing much better now. That clinical finding started with a data point nobody would have seen until her next scheduled visit.
In a third case, a nurse reviewing mat data noticed a patient’s heart rate sitting at 67 to 69, below his pacemaker’s set rate of 72. She called him. A week later, he was hospitalized for congestive heart failure. His pacemaker was recalibrated; its own warning system had missed the malfunction. The sleep mat caught what the pacemaker could not.
We presented pilot research at CHEST 2025 showing that overnight sleep mat data, combined with clinical history, carries early warning signals for respiratory hospitalizations.
Every month, our nurses contact thousands of patients across the network. During each call, they work from the full clinical picture: chart notes, diagnostics, the physician’s action plan, and the patient’s nocturnal vitals and sleep architecture. These care coordinators function as an extension of the clinic.
One call caught a woman on 9 liters of oxygen producing frothy brown sputum. The nurse sent her to the ED, where she was diagnosed with a pulmonary embolism. Another call turned up vague complaints: fatigue, poor concentration, trouble speaking. It was a myocardial infarction requiring stent placement. A third patient, whose next appointment was four months away, was using her rescue inhaler five to six times a day. The nurse got her seen within the week.
Not every catch is an emergency. During a routine call, a nurse heard a patient describe chronic exhaustion despite nine hours of sleep. The mat data showed four hours of nightly snoring. A home sleep test confirmed obstructive sleep apnea. After starting therapy, her AHI dropped from 15.9 to 1.3. Nobody had ordered that test. The data made the case before the physician had to.
Nurses also assess inhaler technique on these calls. Nearly one in five patients has a correctable error: wrong priming sequence, no breath hold, skipping the rinse after a steroid inhaler. Small mechanical problems that build up over months and show up as treatment resistance at the next visit. Catching them by phone is unglamorous work. It is also some of the highest-value care in pulmonary medicine.
Behind those calls is an outreach engine that flags patients who need attention: lapsed appointments, overdue orders, declining device engagement. Every contact is tracked and sequenced. An AI scribe captures every visit, drafts the note, and writes it back to the chart so the physician can be present with the patient instead of the screen. We score nurse conversations for clinical accuracy and follow-through, because quality in a growing network does not hold on its own.
The pulmonary specialty will transform in the coming decade, as care algorithms begin to integrate home data, medical records, AI scribe and phone call recordings, and society guidelines into more precise medicine than what exists today. We’re building towards this vision.
