Sources: health-report-2026-08-17-to-2026-08-18.json (Apple Watch sleep stages, exported 08:27 on 08-18) + OSCAR_Greg_by_Day_2026-08-17.csv (ResMed AirSense/AirFlow 11 via OSCAR). Session start was 23:07 on 08-17 — five minutes later than 08-16's 23:02, still solidly in the healthy 23:00–23:30 window. CPAP hours 6.28 vs. Watch TIB 6.13 (Δ ≈ 9.0 min), matching yesterday's tight alignment.
| Metric | Value |
|---|---|
| Bedtime (session start) | 23:07 on 08-17 (on-schedule) |
| Time in bed | 6.13 h (367.8 min) |
| Total sleep duration | 6.00 h (360.3 min) |
| Sleep efficiency (TST/TIB) | 97.95% — 2nd highest of last 10 nights (behind 08-16's 98.34%); +3.0 pts vs. 7-day (94.9%), +2.9 pts vs. 30-day (95.1%) |
| Awake time | 7.5 min — rank 5/30 (top-17% lowest); −63% vs. 7-day (20.6), −63% vs. 30-day (20.2) |
| Core sleep | 3.63 h (217.8 min) — 60.45% |
| Deep sleep | 0.85 h (51.2 min) — 14.21% — essentially at the 30-day mean (14.48%) |
| REM sleep | 1.52 h (91.3 min) — 25.35% — above 30-day (23.76%); full recovery from 08-16's 19.78% |
| CPAP usage | 6.28 h — near 7-day (6.66), −6.7% vs. 30-day (6.73) |
| AHI | 1.11 — ties 08-10 as lowest since matched history began (rank 5/30, tied with 08-16 at 1.13); −31% vs. 7-day (1.60), −34% vs. 30-day (1.68) |
| RDI | 1.11 (RERA = 0) |
| OA / H / CA / UA / RERA | 5 / 2 / 0 / 0 / 0 — no central events, no RERA, no unclassified apneas (fully clean respiratory profile) |
| OA per hour | 0.80 /h — below trailing 30-day OA-rate (~1.14/h); 4th night out of last 10 with OA/h ≤ 1.0 |
| Pressure avg | 7.57 cmH₂O — lowest since 08-13 (7.01); −0.35 vs. 08-16, cumulative −0.78 from 08-15 peak |
| Pressure 90th %ile | 9.54 cmH₂O — lowest since 08-13; −1.40 vs. 08-16, back below 30-day (10.10) |
| Leak avg | 1.14 L/min — 3rd consecutive uptick (0.31 → 0.91 → 1.14); crosses "very-high" ≥ 1.0 threshold |
| Leak 90th %ile | 1.2 L/min — dropped sharply from 08-16's 4.8; lowest tail-leak reading since 08-13 |
| Leak max | 25.2 L/min (brief; note peak isolated) |
| Compliance flag | 1 (compliant) |
| Date | AHI | OA | H | CA | RERA | Leak avg | Leak 90 | P_avg | P_90 | Hours | Deep % | REM % | Awake (min) | Eff |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 08-05 | 1.49 | 7 | 2 | 1 | 0 | 1.90 | 7.2 | 7.97 | 10.58 | 6.72 | 11.33 | 23.31 | 10.5 | 97.34% |
| 08-06 | 2.83 | 11 | 4 | 2 | 1 | 0.56 | 2.4 | 7.60 | 10.34 | 6.00 | 15.91 | 19.42 | 9.5 | 97.30% |
| 08-08 | 2.34 | 6 | 10 | 0 | 0 | 0.87 | 3.6 | 7.63 | 9.90 | 6.83 | 16.17 | 26.60 | 15.6 | 96.12% |
| 08-09 | 1.34 | 6 | 3 | 0 | 2 | 3.00 | 10.8 | 7.79 | 10.94 | 6.72 | 10.97 | 23.65 | 15.1 | 96.19% |
| 08-10 | 0.58 | 3 | 0 | 0 | 0 | 0.17 | 0.0 | 6.89 | 9.46 | 5.20 | 15.30 | 20.31 | 20.1 | 94.73% |
| 08-11 | 2.57 | 12 | 4 | 0 | 0 | 0.42 | 1.2 | 7.28 | 10.02 | 6.62 | 13.57 | 25.24 | 21.1 | 94.61% |
| 08-12 | 2.68 | 13 | 3 | 2 | 1 | 0.28 | 1.2 | 7.99 | 10.00 | 6.72 | 20.78 | 19.11 | 28.1 | 92.80% |
| 08-13 | 1.29 | 8 | 0 | 2 | 0 | 0.20 | 1.2 | 7.01 | 9.54 | 7.75 | 10.93 | 26.70 | 20.1 | 95.60% |
| 08-14 | 1.63 | 11 | 1 | 0 | 1 | 0.22 | 0.0 | 7.94 | 10.86 | 7.38 | 9.73 | 29.20 | 33.6 | 92.19% |
| 08-16 | 1.13 | 3 | 3 | 1 | 2 | 0.91 | 4.8 | 7.92 | 10.94 | 6.22 | 14.73 | 19.78 | 6.0 | 98.34% |
| 08-17 | 1.11 | 5 | 2 | 0 | 0 | 1.14 | 1.2 | 7.57 | 9.54 | 6.28 | 14.21 | 25.35 | 7.5 | 97.95% |
(08-15 not in table: no matched Watch sleep-stage session that night — CPAP data shows AHI 1.53, OA 8, Leak 0.31, P_avg 8.35, Hours 5.88.)
Two-night composite (08-16 + 08-17): AHI mean 1.12 • OA total 8 across 12.50 h use = 0.64 OA/h • Awake total 13.5 min across 12.13 h TIB = 1.85% awake fraction • Deep mean 14.47% • Efficiency mean 98.15%. Every one of those is materially better than the trailing 7-day averages that end on 08-14 (AHI 1.75, OA/h 1.35, awake 22.4 min, efficiency 94.7%).
Six-night 08-12 → 08-17 obstructive composite: OA 48, H 10, CA 5, RERA 4 across 40.23 h = 1.19 OA/h. Compare to the 5-night 08-11 → 08-15 composite (52 OA in 34.35 h = 1.51 OA/h). Two clean nights pulled the trailing OA rate down about 20%; the cluster is now measurably regressing, not just paused.
Buckets on nightly leak average:
the user r (leak avg vs. deep %): −0.03 (n=30). the user r (leak 90 vs. deep %): −0.06. the user r (leak avg vs. awake min): −0.32 (nights with higher mean leak have slightly less awake time, likely a confounder — leak-noisy nights so far coincide with more consolidated sleep windows). the user r (AHI vs. deep %): +0.02. the user r (AHI vs. awake min): −0.01. the user r (OA vs. awake min): +0.17. the user r (leak avg vs. REM %): +0.28. the user r (pressure avg vs. deep %): −0.06. the user r (pressure 90th vs. deep %): +0.09.
Interpretation: In the 30-night matched sample, none of the CPAP-side metrics (leak, AHI, pressure, OA) have a meaningful directional relationship with next-morning deep-sleep %. The strongest observed correlation is a negative one between leak avg and awake minutes — meaning the leak-noisy nights in this dataset have been, if anything, more consolidated. Tonight fits that pattern (Leak avg 1.14, Awake 7.5 min) rather than contradicting it, and reinforces the earlier interpretation that the sleep-architecture channel is driven by non-CPAP factors (TIB length, bedtime, day-of-week rhythms) more than by respiratory noise.
Comparing tonight's r-values to the prior report (which used n=60 across the fuller dataset that includes health-only nights before matched CPAP data began): today's r(leak avg vs. deep %) is −0.03 vs. prior −0.25. The weaker coefficient here reflects the smaller and more-recent-only matched sample (30 vs. 60); over the recent 30 nights the leak-vs-architecture signal has been essentially flat. This matches the observation that the higher-leak stretches of the last month (08-05, 08-09) did not produce systematically worse deep %.
| Metric | Tonight (08-17) | 7-day avg (matched, through 08-16) | 30-day avg (matched, through 08-16) |
|---|---|---|---|
| AHI | 1.11 | 1.60 | 1.68 |
| OA count | 5 | 8.0 | 7.6 |
| OA per hour | 0.80 /h | ~1.20 /h | ~1.12 /h |
| Leak avg (L/min) | 1.14 | 0.74 | 0.76 |
| Leak 90th (L/min) | 1.2 | 2.74 | 2.32 |
| Pressure avg (cmH₂O) | 7.57 | 7.55 | 7.50 |
| Pressure 90th (cmH₂O) | 9.54 | 10.25 | 10.10 |
| Usage hours | 6.28 | 6.66 | 6.73 |
| Deep sleep % | 14.21% | 13.72% | 14.48% |
| REM sleep % | 25.35% | 23.43% | 23.76% |
| Awake time (min) | 7.5 | 20.6 | 20.2 |
| Efficiency | 97.95% | 94.92% | 95.07% |
| TIB (h) | 6.13 | 6.68 | 6.65 |
Verdict: tonight extends the 08-16 pattern of "everything on the CPAP+sleep-consolidation side outperforming trailing baseline, with leak avg the only up-arrow." AHI, OA, OA/h, pressure avg, pressure 90th are all below their trailing averages; awake, efficiency, and REM are the best in weeks; only leak avg has moved the wrong direction. This is the strongest 2-night reading in the dataset since the late-July window.
| Week (Wed–Tue) | AHI avg | Leak avg | P_avg | P_90 avg | OA avg | Hours | n |
|---|---|---|---|---|---|---|---|
| 07-15 to 07-21 | 1.62 | 0.46 | 8.18 | 10.78 | 6.7 | 6.38 | 6 |
| 07-22 to 07-28 | 1.68 | 0.87 | 7.41 | 10.22 | 8.6 | 6.76 | 7 |
| 07-29 to 08-04 | 1.74 | 0.70 | 7.03 | 9.27 | 7.3 | 7.12 | 7 |
| 08-05 to 08-11 | 1.83 | 1.16 | 7.51 | 10.13 | 7.6 | 6.37 | 7 |
| 08-12 to 08-18 (partial, n=6) | 1.56 | 0.51 | 7.80 | 10.21 | 8.0 | 6.70 | 6 |
Partial-week update: (1) weekly AHI 1.56 — lowest partial-week reading in ~5 weeks (prior lowest: 07-15 to 07-21 at 1.62). (2) Weekly OA average 8.0 — off its peak but still the highest of the trailing 5 weeks; the 08-11/12 spikes (12, 13 OA) continue to dominate the mean. If tonight and 08-16 were the only nights in the window, weekly OA would be 4.0. (3) Weekly leak avg 0.51 — well below the prior week's 1.16; tonight's uptick to 1.14 has not yet moved the weekly picture. (4) Weekly P_avg 7.80 — the pressure release of 08-16/17 has begun to pull this back from what would have been an ~8.0 average through 08-15.
The 5-week AHI trend (1.62 → 1.68 → 1.74 → 1.83 → 1.56) now shows a clear bend down after the 4-week climb. Two clean nights are doing most of the work — a third clean night on 08-18 would make it a durable improvement rather than a mid-cluster respite.
The obstructive cluster is resolving on data, not just on hope. Two consecutive nights of AHI ≈ 1.1 and OA ≤ 5, and — critically — the auto-titrator has pulled pressure back for two nights in a row (P_avg 8.35 → 7.92 → 7.57). That's the algorithm's own signal that it's not seeing sustained load. Tonight P_90 fell to 9.54 — the lowest 90th-percentile pressure since 08-13 — meaning even the tail of the pressure distribution isn't being pushed high. The 08-15 watch-condition (which would have prompted a manual minimum-pressure raise) is no longer in play. If 08-18 lands OA ≤ 6 and AHI ≤ 1.5, the 08-11 → 08-15 window can be closed out as a discrete episode.
Leak channel — the trend needs one more night to interpret cleanly. Three consecutive nights of rising leak avg (0.31 → 0.91 → 1.14) is a real signal, but tonight's leak-90 improvement (4.8 → 1.2) argues against a spike-driven problem. The most plausible read is a small-magnitude persistent seal drift — check strap tension and cushion position before tonight. If 08-18's leak avg is back below 0.5, this was a 2–3-night mask-fit adjustment and self-corrected. If leak avg holds ≥ 1.0 for a third consecutive night, the mask/cushion probably needs attention regardless of what respiratory data show, because "sustained > 1.0" is the boundary where ResMed's own guidance starts recommending seal-checks.
Sleep architecture: baseline restored, not just present. Two nights (08-16, 08-17) at Deep ≈ 14.5%, REM 20–25%, Awake < 10 min, Efficiency > 98% — this is what "normal" looks like on this dataset. The 08-13/08-14 low-deep readings (10.93%, 9.73%) look increasingly like a two-night dip rather than a regime shift. Whether the underlying rhythm can hold — or whether next week reverts to the mid-July pattern of higher awake minutes (24–33 min) — depends more on bedtime discipline than on CPAP factors, given that r(leak/AHI vs. deep %) remain near zero over the matched history.
What to watch next. (1) AHI/OA on 08-18 — a third clean night (AHI ≤ 1.5, OA ≤ 6) closes the 08-11–08-15 cluster as a discrete event. Any bounce back to AHI ≥ 2.5 or OA ≥ 10 would reopen the question of whether pressure is still under-set at the low end. (2) Leak avg on 08-18 — the key data point for the 3-night uptrend. Back under 0.5 = self-correcting drift; stays ≥ 1.0 = mask-fit action item. (3) Bedtime consistency — 23:02 / 23:07 back-to-back is healthy; drift back to 23:35+ risks another late-week TIB compression. (4) REM stability — 08-16 showed a REM dip to 19.78%, tonight recovered to 25.35%; REM is the most volatile stage in this dataset (range 19.11–29.20% over the last 10 nights) so single-night moves aren't diagnostic, but sustained REM < 20% for 3+ nights would be worth flagging.