Sources: health-report-2026-07-21-to-2026-07-22.json (Apple Watch sleep stages, exported 07:14 on 07-22) + OSCAR_Greg_by_Day_2026-07-21.csv (ResMed AirSense/AirFlow 11 via OSCAR). Matched by calendar date of sleep-session start (22:27 on 07-21).
Deep sleep rebounded to 28.4% (106.9 min) — the highest value in the entire matched dataset, and a +20.1 pp single-night recovery from the 8.3% record low on 07-20. This is a strong positive signal and directly confirms the cumulative-deficit interpretation from the prior report: after two consecutive excellent CPAP nights (AHI 0.94 → 0.88), slow-wave sleep has begun recovering. The 5-night trailing average of deep sleep was 13.2% (07-16 to 07-20); tonight is +15.2 pp above that.
Leak avg crept back up to 0.85 L/min (from 0.03 the prior night), with a max of 22.8 L/min. The 90th-percentile leak of just 1.2 L/min indicates this was a brief positional event, not a sustained seal failure — the machine ran clean for the vast majority of the night. Not alarming at this level, but worth watching: if leak avg exceeds ~1.0 L/min on consecutive nights, the mask seal may need reseating.
Pressure average fell to 6.82 cmH₂O — the lowest nightly average recorded since the near-zero baseline readings in mid-June — reflecting the machine running optimally at minimal respiratory load. This is consistent with AHI 0.88 (second-lowest on record). Two consecutive nights at or below 0.94 AHI with pressure declining suggests the therapeutic range is well-established.
| Metric | Value |
|---|---|
| Bedtime (session start) | 22:27 |
| Time in bed | 6.72 h (402.9 min) |
| Total sleep duration | 6.26 h (375.8 min) |
| Sleep efficiency (TST/TIB) | 93.3% |
| Awake time | 27.1 min |
| Core sleep | 2.93 h (175.6 min) — 46.7% |
| Deep sleep | 1.78 h (106.9 min) — 28.4% — new record high |
| REM sleep | 1.55 h (93.3 min) — 24.8% |
| CPAP usage | 6.85 h |
| AHI | 0.88 — second-lowest on record |
| RDI | 1.02 |
| OA / H / CA / UA / RERA | 4 / 1 / 1 / 0 / 1 |
| Pressure avg | 6.82 cmH₂O (new dataset low) |
| Pressure 90th %ile | 9.88 cmH₂O |
| Leak avg | 0.85 L/min (up from 0.03 prior night) |
| Leak max | 22.8 L/min (brief episode) |
| Leak 90th %ile | 1.2 L/min |
| Date | AHI | RERA | Leak avg | Deep % | Deep (min) | REM % | Awake (min) | Efficiency |
|---|---|---|---|---|---|---|---|---|
| 07-16 | 1.50 | 3 | 0.13 | 16.6% | ~70 | 20.6% | 12.5 | 96.3% |
| 07-17 | 2.78 | 0 | 0.27 | 16.4% | ~69 | 30.0% | 7.0 | 98.2% |
| 07-18 | 2.35 | 0 | 0.54 | 14.7% | ~63 | 23.1% | 36.6 | 90.3% |
| 07-19 | 1.29 | 1 | 0.92 | 10.1% | ~43 | 17.0% | 14.0 | 96.2% |
| 07-20 | 0.94 | 0 | 0.03 | 8.3% | 29.6 | 21.0% | 21.1 | 94.4% |
| 07-21 | 0.88 | 1 | 0.85 | 28.4% | 106.9 | 24.8% | 27.1 | 93.3% |
Key trend: Deep sleep hit a 5-night low of 8.3% on 07-20 (despite best-ever CPAP metrics), then snapped back to a new high of 28.4% tonight. This is the classic pattern of slow-wave sleep debt recovery: architecture can't rebound in a single night even after the triggering stressor resolves, but rebounds with a day of lag. The CPAP held excellent control on both nights (AHI ≤ 1.0), providing the stable respiratory environment for slow-wave sleep to recover.
Note on tonight's slightly elevated leak: Leak avg 0.85 is modestly above the low-leak range (≤0.30), and deep sleep was still at its highest ever — this doesn't support a simple leak-disrupts-deep-sleep model within a single night. The brief max event (22.8 L/min) likely occurred during an isolated positional shift rather than representing sustained seal failure.
Split at leak ≤ 0.30 vs. > 0.30 L/min:
Important caveat: Tonight has reversed the expected direction of the leak-deep relationship. Previously, low-leak nights were associated with better deep sleep; now, with 07-21 (0.85 leak, 28.4% deep) added to the dataset, the high-leak group shows higher average deep sleep (17.7% vs. 13.8%) and higher awake time (25.9 vs. 13.5 min). The awake-time signal may reflect the brief leak episode; the deep-sleep reversal is best explained by 07-21 being a recovery night where slow-wave pressure drove unusually high deep sleep, independent of the minor leak.
the user r (leak vs. deep%): +0.38 — essentially uncorrelated across only 6 nights; direction reversal is not meaningful at this sample size. the user r (leak vs. awake min): +0.29 — slight positive trend but not interpretable with n=6.
Working interpretation: Deep sleep recovery on 07-21 was driven by slow-wave rebound after cumulative deficit, not by CPAP metrics tonight. The 0.85 leak avg didn't prevent the recovery, but the 22.8 max suggests the mask seal needs monitoring.
| Metric | Tonight (07-21) | 7-day avg (07-15–07-21) | Prior 7-day avg (07-14–07-20) | ~30-day avg |
|---|---|---|---|---|
| AHI | 0.88 | 1.82 | 2.15 | ~2.70 |
| RDI | 1.02 | 2.02 | — | — |
| RERA | 1 | 1.29 | 1.57 | ~1.28 |
| Pressure avg (cmH₂O) | 6.82 | 8.27 | 8.62 | ~8.55† |
| Leak avg (L/min) | 0.85 | 0.43 | 0.35 | ~1.02‡ |
| Usage hours | 6.85 | 6.42 | 6.34 | ~6.45 |
| Deep sleep % | 28.4% | 15.8%§ | ~13.2%§ | — |
| REM sleep % | 24.8% | 22.8%§ | ~22.3%§ | — |
| Awake time (min) | 27.1 | 19.7§ | ~18.2§ | — |
†30-day pressure excludes mid-June zero-pressure baseline readings.
‡30-day leak elevated by Jun 24–27 and Jul 01 high-leak nights.
§Sleep-stage averages cover only nights with Apple Watch data (6 nights: Jul 16–21).
| 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 |
| 07-15 to 07-21 (trailing 7d) | 1.82 | 1.29 | 0.43 | 7 |
The trailing-7-day AHI of 1.82 is the lowest weekly average of the entire tracked period, continuing the downward trend from 3.24 in early July. RERA is declining. Pressure average has dropped from ~8.62 to 8.27, and tonight's 6.82 cmH₂O will pull it lower next week if the trend holds. These are strong positive signals for therapeutic efficacy.
The primary watch item going forward is mask seal. Two nights at AHI < 1.0 with declining pressure is excellent — the goal now is to maintain it. Leak avg 0.85 tonight with a max of 22.8 is worth monitoring: if it recurs across the next 2–3 nights, check mask cushion wear or fit adjustments. Deep sleep has recovered to its highest level yet; if it settles in the 15–22% range over the next week, that would represent a genuine improvement from the 10–12% range of mid-July.