Sources: health-report-2026-07-22-to-2026-07-23.json (Apple Watch sleep stages, exported 08:10 on 07-23) + OSCAR_Greg_by_Day_2026-07-22.csv (ResMed AirSense/AirFlow 11 via OSCAR). Matched by calendar date of sleep-session start (22:27 on 07-22).
Leak avg crossed the 1.0 L/min threshold on the second consecutive elevated night (07-21: 0.85 → 07-22: 1.08 L/min), with the 90th-percentile leak doubling (1.2 → 2.4 L/min) and a new max of 31.2 L/min. The previous report set 1.0 L/min on consecutive nights as the watch threshold for mask seal attention. Tonight crossed it. Leak is rising session-over-session, not spiking randomly. If 07-23 is again ≥ 0.8 L/min avg, a mask cushion check or fit adjustment is recommended.
Despite the elevated leak, sleep quality was excellent: awake time only 10.0 min (lowest in the 7-night dataset), REM sleep hit 31.6% (125.9 min) — the highest percentage recorded across all matched nights — and sleep efficiency reached 97.6%. The leak is not currently translating into sleep disruption.
Deep sleep returned to 12.0% (47.7 min) from 07-21's exceptional 28.4%. This is the expected normalization after a slow-wave recovery night, not a decline. The 7-night mean is 15.2%; tonight is modestly below it, consistent with a reversion to baseline after an outlier.
AHI week-over-week is still declining (trailing 7d: 1.62 vs prior 7d: 1.82), extending the downward trend from the 3.24 early-July peak. This is positive despite the modest uptick from 0.88 to 1.59 tonight.
| Metric | Value |
|---|---|
| Bedtime (session start) | 22:27 |
| Time in bed | 6.81 h (408.4 min) |
| Total sleep duration | 6.64 h (398.4 min) |
| Sleep efficiency (TST/TIB) | 97.6% |
| Awake time | 10.0 min — lowest in 7-night dataset |
| Core sleep | 3.75 h (224.8 min) — 56.4% |
| Deep sleep | 0.79 h (47.7 min) — 12.0% |
| REM sleep | 2.10 h (125.9 min) — 31.6% — highest in dataset |
| CPAP usage | 6.9 h |
| AHI | 1.59 |
| RDI | 1.74 |
| OA / H / CA / UA / RERA | 7 / 4 / 0 / 0 / 1 |
| Pressure avg | 7.60 cmH₂O |
| Pressure 90th %ile | 10.48 cmH₂O |
| Pressure max | 12.0 cmH₂O |
| Leak avg | 1.08 L/min ↑ (2nd consecutive elevated night) |
| Leak max | 31.2 L/min |
| Leak 90th %ile | 2.4 L/min (doubled from 1.2 prior night) |
| Date | AHI | RERA | Leak avg | Leak 90th | Deep % | Deep (min) | REM % | Awake (min) | Efficiency |
|---|---|---|---|---|---|---|---|---|---|
| 07-16 | 1.50 | 3 | 0.13 | 0 | 16.6% | ~70 | 20.6% | 12.5 | 96.3% |
| 07-17 | 2.78 | 0 | 0.27 | 1.2 | 16.4% | ~69 | 30.0% | 7.0 | 98.2% |
| 07-18 | 2.35 | 0 | 0.54 | 0 | 14.7% | ~63 | 23.1% | 36.6 | 90.3% |
| 07-19 | 1.29 | 1 | 0.92 | 0 | 10.1% | ~43 | 17.0% | 14.0 | 96.2% |
| 07-20 | 0.94 | 0 | 0.03 | 0 | 8.3% | 29.6 | 21.0% | 21.1 | 94.4% |
| 07-21 | 0.88 | 1 | 0.85 | 1.2 | 28.4% | 106.9 | 24.8% | 27.1 | 93.3% |
| 07-22 | 1.59 | 1 | 1.08 | 2.4 | 12.0% | 47.7 | 31.6% | 10.0 | 97.6% |
The leak profile has shifted across two nights: - Leak avg: 0.85 → 1.08 (+0.23 L/min) - Leak 90th: 1.2 → 2.4 L/min (doubled) - Leak max: 22.8 → 31.2 L/min
The rising 90th-percentile is the more informative signal — it indicates the leak is sustained over more of the night, not just a brief positional spike. Prior high-leak episodes (07-18, 07-19) had 90th percentile of 0 (i.e., no sustained component); the last two nights both show real 90th-pct values.
Split at leak ≤ 0.30 vs. > 0.30 L/min:
the user r (leak avg vs. deep%): +0.19 — directionally flat; the dataset remains too small for meaningful inference. the user r (leak avg vs. awake min): +0.02 — effectively zero across 7 nights. The 07-21 recovery night and tonight's excellent sleep continuity have decoupled leak from awake time in this sample.
Working interpretation: The mask seal issue is a hardware problem worth fixing on its own terms, independent of sleep quality impact. The fact that tonight showed excellent sleep (10 min awake, 31.6% REM) despite 1.08 avg leak may indicate tolerance, positional luck, or that the leaks are occurring during REM-dense periods where the machine compensates. But tolerance to moderate leak doesn't rule out worsening: if avg leak reaches 2+ L/min, therapy efficacy risk rises.
| Metric | Tonight (07-22) | 7-day avg (07-16–07-22) | Prior 7-day avg (07-15–07-21) | 30-day avg (06-23–07-22) |
|---|---|---|---|---|
| AHI | 1.59 | 1.62 | 1.82 | 2.70 |
| RDI | 1.74 | 1.76 | 2.02 | — |
| RERA | 1 | 0.86 | 1.29 | ~1.14 |
| Pressure avg (cmH₂O) | 7.60 | 8.10 | 8.27 | 8.49 |
| Leak avg (L/min) | 1.08 | 0.55 | 0.43 | 1.05† |
| Usage hours | 6.9 | 6.45 | 6.42 | 6.53 |
| Deep sleep % | 12.0% | 15.2%‡ | 15.8%‡ | — |
| REM sleep % | 31.6% | 24.0%‡ | 22.8%‡ | — |
| Awake time (min) | 10.0 | 18.3‡ | 19.7‡ | — |
†30-day leak avg includes Jul 01 (6.22 L/min outlier from mask failure); without that night: (31.63 - 6.22) / 29 = 0.88 L/min.
‡Sleep-stage averages cover only the 7 matched nights (Jul 16–22) with both Apple Watch and CPAP data.
| Week | AHI avg | RERA avg | Leak avg | n |
|---|---|---|---|---|
| 06-22 to 06-28 | 2.84 | ~0.86 | 1.98 | 7 |
| 06-29 to 07-05 | 3.24 | ~1.29 | 1.41 | 7 |
| 07-06 to 07-12 | 3.00 | ~1.14 | 0.43 | 7 |
| 07-13 to 07-19 | 2.22 | 2.00 | 0.42 | 7 |
| 07-16 to 07-22 (trailing 7d) | 1.62 | 0.86 | 0.55 | 7 |
AHI continues declining week-over-week (1.62 is a new trailing-7-day low). RERA has also dropped back from the 2.00 spike in the 07-13–19 window. Leak avg has ticked up slightly due to the last two nights but is still well below the mid-June 1.98 level. Pressure average continues to drift lower (8.49 → 8.27 → 8.10 → 7.60 tonight), consistent with the machine dialing back as respiratory load decreases — a positive efficiency signal.
The primary watch item remains mask seal. Two consecutive nights of rising leak (avg 0.85 → 1.08, 90th pct 1.2 → 2.4) with an increasing max is a clear pattern. If tomorrow (07-23) shows leak avg ≥ 0.8 or 90th pct ≥ 2.0, a mask cushion inspection is warranted — look for wear, deformation, or pillow compression that would shift fit during the night. The otherwise strong sleep metrics tonight (97.6% efficiency, 10 min awake, 31.6% REM) are genuinely good and suggest the leak hasn't yet impacted sleep architecture, but the trend direction is worth addressing before it does.
Deep sleep at 12% is a normal baseline value after the 07-21 recovery spike; expect this to stabilize in the 12–18% range over the coming week. REM at 31.6% is the highest recorded value and is a positive signal — no action needed.