Sleep medicine generates more data per patient than almost anything else in outpatient pulmonology.

The RealityA study is a report, a set of values and a waveform

⚠️

The Problem

The study arrives as a picture.

A PDF in the document bin is readable once and searchable never. Nothing trends, nothing reports, and every downstream use starts with somebody retyping.

🔧

The Solution

Report as text, key values as fields.

Take the interpretation into the results section and capture the small set of indices that get reused, rather than trying to structure the whole study.

✅

The Resolution

Questions you can answer on a Monday.

Who is due for a follow-up study, whose index moved, who never started therapy — answerable from a report instead of from memory.

Polysomnography: a long study and a short report

An attended in-laboratory study is the most instrumented version of this. NHLBI describes polysomnography as measuring brain waves, heart rate, breathing and blood oxygen level during a full night of sleep, with sensors recording breathing effort and rate, oxygen levels and muscle movements.

Almost none of that belongs in the EHR. The raw recording is enormous and is reviewed in a specialist scoring package, not in a chart. What belongs in the chart is the scored summary, the interpretation, and a pointer back to the study.

Practices that try to push the whole study into the record stall, and practices that push nothing but a PDF end up unable to report. The productive middle is narrower than either: the interpretation as text, a short list of named values, and a link that opens the study in its native viewer with patient context carried across.

Home sleep test data and what the chart receives

Home testing has changed the volume and the shape of what arrives. A home sleep test produces fewer channels and a smaller file, which makes integration easier, and it produces them in higher numbers, which makes structure matter more.

Coverage rules are specific about what qualifies, and the specificity is the reason to capture data rather than images. The Medicare national coverage determination for CPAP in obstructive sleep apnea accepts attended polysomnography or an unattended home test using a Type II, III or IV monitor measuring at least 3 channels, with the treating physician ordering the test and supervising it.

So the record has to show which device type was used and how many channels it captured. If that lives only inside a report image, answering the question later means opening files one at a time.

CPAP compliance data and the 12-week clock

The coverage structure makes adherence data operationally important rather than merely interesting. The same determination limits initial CPAP coverage to a 12-week period to identify beneficiaries who benefit from it, with continued coverage for those who do.

It also sets out the thresholds that qualify a patient in the first place: an AHI or RDI of 15 or more events per hour, or between 5 and 14 events per hour with documented excessive daytime sleepiness, impaired cognition, mood disorders, insomnia, or comorbidities including hypertension, ischaemic heart disease or a history of stroke.

Read that as a data specification. The index, the symptom documentation and the comorbidities all have to be findable, and a twelve-week clock has to start somewhere the practice can see. Where adherence data comes back from a device or a supplier, it is worth treating as a remote monitoring feed rather than as correspondence, because correspondence does not trend.

Scored data versus a PDF of the score

This is the decision that determines everything downstream, and it is usually made by default at go-live.

Pick the small set of values that genuinely get reused — in most practices that is the apnea-hypopnea index, the oxygen desaturation index, the lowest and mean saturation, total sleep time and efficiency, and the device type. Six or seven fields. Capture those discretely and let the rest of the report be searchable narrative.

The failure we see most often is a practice attempting to structure everything, stalling on the mapping, and shipping nothing — then defaulting back to attaching the PDF. Choosing seven values and capturing them properly beats an incomplete attempt at seventy, and it is achievable inside an ordinary workflow rebuild rather than needing a project of its own.

The referral loop and who closes it

Sleep is a referral-heavy service, and the loop is where patients disappear. A primary care physician refers, the study is done, the interpretation is signed, and then one of three things has to happen: therapy starts, a repeat study is scheduled, or the patient is discharged back with a plan.

None of those is automatic, and none of them generates an alert if it fails to occur. A patient who had a study in March and never started therapy will not appear on any list unless somebody built the list.

That report is the highest-value thing in this post and the cheapest to build: patients with a signed interpretation, no therapy recorded, and no follow-up appointment. Give it an owner and run it monthly. Where the follow-up is a conversation about tolerance rather than an examination, a virtual visit closes it faster than a clinic slot.

A note on consumer devices

Patients increasingly arrive with data from a watch. It is worth knowing precisely what that data is and is not.

When FDA cleared a sleep apnea notification feature for over-the-counter use, the clearance came with an explicit boundary: the feature is not intended to diagnose, treat, or aid in the management of sleep apnea, and the absence of a notification is not intended to indicate the absence of sleep apnea.

That is a useful sentence to have to hand. A notification is a reason to evaluate, not a result, and it does not belong in the chart as one. Deciding in advance how consumer data is recorded — as patient-reported context rather than as a finding — avoids a category of confusion later.

• Pick seven values and capture them discretely. Not seventy.
• Record device type and channel count. The coverage rule names them.
• Start the twelve-week clock somewhere visible.
• Build the unclosed-loop report. Signed interpretation, no therapy, no follow-up.

Qventive provides sleep study EHR integration help for medical practices across New Jersey, usually as part of our pulmonology EHR and IT support and a broader pulmonology practice management engagement. Because studies and adherence data move between a testing facility, a supplier and your chart, the design belongs in your HIPAA risk analysis as well as the clinical plan.

Sleep study integration: common questions