A Multisite Randomized Trial of Portable Sleep Studies and Positive Airway Pressure Autotitration Versus Laboratory-Based Polysomnography for the Diagnosis and Treatment of Obstructive Sleep Apnea: The HomePAP Study

The researchers compared two complete clinical pathways for diagnosing and treating obstructive sleep apnea (OSA):

**Home pathway (HOME):** Patients used a portable "Level 3" sleep monitor at home overnight, which recorded airflow, breathing effort, oxygen levels, and heart rate but did not measure brain waves (EEG) or eye movements. If the home test showed moderate-to-severe OSA (AHI ≥ 15), patients then used an automatic CPAP machine (autoPAP) at home for one week. The autoPAP machine adjusted pressure automatically throughout the night to find the minimum pressure needed to keep the airway open. After that week, the 90th percentile pressure from the autoPAP data was used to prescribe a fixed-pressure CPAP machine for ongoing use.

**Laboratory pathway (LAB):** Patients underwent a standard overnight in-laboratory polysomnography (PSG) with full EEG monitoring. If they had moderate-to-severe OSA, they returned for a second night in the lab for a manual CPAP titration, where a technician adjusted pressure levels while monitoring sleep stages and breathing.

**Primary outcome:** CPAP adherence at 3 months, measured as average hours of use per night at effective pressure.

**Secondary outcomes:** CPAP acceptance (whether patients agreed to use CPAP after diagnosis), time from enrollment to treatment initiation, changes in daytime sleepiness (Epworth Sleepiness Scale, ESS), and functional outcomes (Functional Outcomes of Sleep Questionnaire, FOSQ).

**Sample size:** 373 participants enrolled; 186 randomized to HOME, 187 to LAB. After excluding those who did not meet the AHI ≥ 15 threshold on their initial test, 106 in HOME and 107 in LAB remained eligible and continued to CPAP treatment.

**Population:** Consecutive new referrals to seven AASM-accredited sleep centers across the United States.

**Inclusion criteria:** Age 18 years or older, high probability of moderate-to-severe OSA based on a clinical algorithm (including snoring, witnessed apneas, gasping/choking, and obesity), and an Epworth Sleepiness Scale score ≥ 12 (indicating excessive daytime sleepiness).

**Exclusion criteria:** Prior OSA diagnosis or CPAP use, severe comorbid conditions (e.g., heart failure, COPD, neuromuscular disease), or conditions that would make home monitoring unsafe (e.g., oxygen desaturation below 85% during screening).

**Demographics:** Mean age approximately 48 years, mean BMI approximately 37 kg/m² (class II obesity), approximately 60% male, predominantly Caucasian. Mean baseline AHI was approximately 40 events/hour (severe OSA). Mean baseline ESS was approximately 15 (very sleepy).

**Apnea-Hypopnea Index (AHI):** Number of apneas (complete airway collapse) plus hypopneas (partial collapse with ≥ 3% oxygen desaturation) per hour of sleep. Measured via full PSG in the lab and via portable monitor (Type 3 device) at home.

**Epworth Sleepiness Scale (ESS):** Self-report questionnaire (0–24 scale, higher = sleepier). A score ≥ 12 indicates excessive daytime sleepiness.

**Functional Outcomes of Sleep Questionnaire (FOSQ):** Self-report measure of how sleepiness affects daily activities (5 domains: activity level, vigilance, intimacy, general productivity, social outcomes). Lower scores indicate greater impairment.

**CPAP adherence:** Downloaded from the CPAP machine's internal data card. Measured as average hours of use per night at effective pressure (the pressure prescribed, not just time the machine was on) and percentage of nights used for ≥ 4 hours.

**CPAP acceptance:** Whether the patient agreed to take home a CPAP machine after diagnosis and titration.

**Time to treatment:** Days from enrollment to first night of CPAP use.

**Study design:** Randomized, open-label, parallel-group, unblinded, multicenter clinical trial. This is a non-inferiority design — the researchers wanted to show that the home pathway was not worse than the lab pathway by more than a pre-specified margin.

**Randomization:** Participants were randomized 1:1 to HOME or LAB after initial screening but before any diagnostic testing. Randomization was stratified by site and by whether participants had a bed partner (since bed partners can influence CPAP adherence).

**Blinding:** This was an open-label study. Neither participants nor clinicians were blinded to which pathway they received. This is a major limitation because patients who know they are in the "convenient" home arm may be more motivated to use CPAP, and clinicians may communicate different expectations.

**Duration:** The study ran from enrollment through 3 months of CPAP use. The home pathway took approximately 1–2 weeks (home test + 1 week autoPAP). The lab pathway took approximately 2–4 weeks (scheduling two overnight lab visits). Follow-up occurred at 1 month and 3 months after CPAP initiation.

**Statistical approach:** The primary analysis used a non-inferiority test with a margin of −1 hour for CPAP usage (i.e., home was considered non-inferior if the lower bound of the 95% confidence interval for the difference in usage was greater than −1 hour). They also tested for superiority if non-inferiority was established. Analyses were intention-to-treat (all randomized participants who had any data) and per-protocol (those who completed the full pathway).

**What this design can and cannot prove:**

**Can prove:** That a home-based pathway is non-inferior to a lab-based pathway for CPAP adherence and functional outcomes in a specific population (high-probability, moderate-to-severe OSA, very sleepy patients). The randomized design controls for known and unknown confounders between groups.

**Cannot prove:** That the home pathway works for mild OSA (AHI < 15) or for patients without significant daytime sleepiness. Cannot prove that the home pathway is superior for all outcomes (though it was superior for adherence). Cannot prove that the home pathway is safe for patients with comorbidities (excluded from this study). Cannot determine which component of the home pathway (the portable monitor vs. the autoPAP vs. the reduced burden) drove the adherence difference.

**Major methodological weaknesses:**

1. **Open-label design:** Patients and clinicians knew which arm they were in. This could inflate adherence in the HOME arm due to expectation effects or reduced burden.

2. **High dropout:** Of 373 enrolled, only 213 (57%) remained eligible after initial testing. Many were excluded because their AHI was < 15 on the initial test. This limits generalizability to the broader population of suspected OSA patients.

3. **No sham control:** There was no placebo CPAP group, so the study cannot separate the specific effects of CPAP from the non-specific effects of being in a treatment program.

4. **Industry funding:** The study was funded by Philips Respironics (a CPAP manufacturer). While the authors report no conflicts, industry-funded studies tend to show more favorable results for the sponsor's products.

5. **Short follow-up:** Three months is relatively short for a chronic condition. Adherence often declines further after 6–12 months.

**Primary outcome — CPAP adherence at 3 months:**

Average nightly use at effective pressure: **HOME: 4.7 ± 2.1 hours vs. LAB: 3.7 ± 2.4 hours** (difference = 1.0 hour, 95% CI: 0.4 to 1.6 hours, p < 0.001 for non-inferiority, p < 0.01 for superiority).

Percentage of nights used ≥ 4 hours: **HOME: 62.8 ± 29.2% vs. LAB: 49.4 ± 36.1%** (difference = 12.6%, p < 0.01).

The home pathway was statistically superior to the lab pathway for adherence.

**Secondary outcomes:**

**CPAP acceptance:** No significant difference. Approximately 90% of eligible participants in both arms accepted CPAP therapy.

**Time to treatment:** No significant difference. Mean time from enrollment to first CPAP use was approximately 21 days in both arms (though the home arm had fewer visits, scheduling logistics balanced this out).

**Titration pressure:** No significant difference. The 90th percentile pressure from autoPAP in the HOME arm was similar to the manually determined pressure in the LAB arm (approximately 10–11 cm H₂O).

**Effective titration:** No significant difference. Both pathways achieved adequate pressure settings in > 90% of cases.

**Change in Epworth Sleepiness Scale (ESS):** Both groups improved significantly from baseline (mean baseline ESS ≈ 15). At 3 months: HOME decreased by approximately 6 points, LAB decreased by approximately 5 points. The difference between groups was not statistically significant (p = 0.20).

**Change in Functional Outcomes of Sleep Questionnaire (FOSQ):** Both groups improved significantly. No significant difference between groups.

**Adverse events:** No serious adverse events related to the home pathway. Minor events (mask discomfort, skin irritation) were similar in both groups.

**Subgroup analyses (exploratory):**

Patients with higher baseline ESS (sleepier) had greater adherence in both arms.

Patients with bed partners had slightly higher adherence in the HOME arm, possibly because the home test was less disruptive to the partner.

**CPAP adherence:** The home pathway led to patients using CPAP for about **1 hour more per night** on average. This is a clinically meaningful difference. For context, CPAP is typically considered "adherent" if used ≥ 4 hours per night on ≥ 70% of nights. The home arm averaged 4.7 hours (close to the threshold), while the lab arm averaged 3.7 hours (below it). The 12.6% higher rate of nights with ≥ 4 hours use means that for every 8 nights, the home group used CPAP adequately on 5 nights, while the lab group used it adequately on 4 nights.

**Daytime sleepiness:** Both groups improved by about 5–6 points on the ESS, which is a large effect (typical minimal clinically important difference is 2–3 points). This means the average patient went from "very sleepy" (ESS 15) to "normal range" (ESS 9–10). The home pathway was not better for sleepiness reduction, despite better adherence, suggesting that even 3.7 hours per night may be sufficient for some patients.

**Functional outcomes:** Both groups improved by approximately 2 points on the FOSQ (scale 5–20), which is considered a moderate improvement in daily functioning.

**Acknowledged by authors:**

Open-label design may have biased adherence in the HOME arm.

High rate of screen failures (43% of enrolled participants did not meet AHI ≥ 15 on initial test), limiting generalizability to lower-risk populations.

Short follow-up (3 months) does not capture long-term adherence patterns.

The home pathway used a specific portable monitor and autoPAP device; results may not generalize to other devices.

The study was conducted at AASM-accredited centers with experienced staff; results may differ in less specialized settings.

**Additional critical limitations:**

**No blinding of outcome assessment:** CPAP adherence was measured objectively from machine data, which is robust. But ESS and FOSQ are self-report and could be influenced by patient expectations.

**Selection bias:** Participants were already motivated enough to seek care at a sleep center and had high baseline sleepiness (ESS ≥ 12). Results may not apply to less motivated or less sleepy patients.

**No cost-effectiveness analysis:** The study did not formally compare costs, though home testing is generally cheaper. This limits practical decision-making for healthcare systems.

**No objective measure of sleep quality:** The home monitor did not measure sleep stages (REM, NREM), so it cannot distinguish between sleep and wake. This could lead to underestimation of AHI if the patient was awake for part of the night.

**Industry funding:** Philips Respironics provided the devices and funded the study. While the authors report no conflicts, industry-funded trials of CPAP tend to show larger effect sizes than independent trials.

For someone running their own n=1 experiment to optimize sleep apnea diagnosis or CPAP adherence:

**What to test:**

Compare a home-based diagnostic pathway (using a validated Type 3 portable monitor like WatchPAT or ApneaLink) followed by autoPAP for 1 week, versus a traditional in-lab PSG. If you have already been diagnosed, test whether using autoPAP (which adjusts pressure automatically) leads to better adherence than a fixed-pressure prescription from a lab titration.

**Minimum meaningful duration:**

For adherence: At least 3 months. The study showed differences at 3 months, but adherence often declines further. Run your experiment for 3–6 months to see if the effect is sustained.

For symptom improvement: 2–4 weeks. ESS and FOSQ changes were apparent by 1 month in this study.

**What to measure (specific metrics):**

**Primary:** Average hours of CPAP use per night at effective pressure (download from machine data card). Aim for ≥ 4 hours/night on ≥ 70% of nights.

**Secondary:** Epworth Sleepiness Scale (ESS) weekly; Functional Outcomes of Sleep Questionnaire (FOSQ) monthly; subjective sleep quality (0–10 scale each morning); partner reports of snoring/apneas.

**Objective:** If using a home monitor, record AHI, oxygen desaturation index (ODI), and time below 90% oxygen saturation (T90).

**Key confounds to control for:**

**Mask comfort:** A poorly fitting mask is the #1 reason for CPAP non-adherence. Use the same mask type throughout your experiment.

**Humidity and temperature:** CPAP with heated humidification improves adherence. Keep settings constant.

**Bed partner:** Having a partner who encourages CPAP use improves adherence. If your partner changes during the experiment, note it.

**Seasonal allergies or illness:** Nasal congestion reduces CPAP tolerance. Track colds, allergies, and use of decongestants.

**Alcohol and sedatives:** These worsen OSA and may change pressure requirements. Keep intake consistent or record it.

**Sleep position:** Supine (back) sleep worsens OSA. Use a positional device or track sleep position if possible.

**What a positive result would look like:**

**For home vs. lab diagnosis:** Your home-based AHI is within 5 events/hour of your lab-based AHI (the study found good agreement). If the home test shows AHI ≥ 15, you can proceed to treatment without a lab visit.

**For autoPAP vs. fixed CPAP:** Your average nightly use increases by at least 1 hour (the study's effect size) and your percentage of nights with ≥ 4 hours use increases by at least 10%. Your ESS decreases by ≥ 3 points (minimal clinically important difference).

**For adherence improvement overall:** You achieve ≥ 4 hours/night on ≥ 70% of nights, and your ESS drops below 10 (normal range).

**Caveat:** This study was done in patients with high pre-test probability of moderate-to-severe OSA and significant daytime sleepiness (ESS ≥ 12). If you have mild OSA (AHI 5–15) or no daytime sleepiness, the home pathway may still work, but the evidence is weaker. Consider consulting a sleep specialist before self-diagnosing.