Sources: health-report-2026-08-23-to-2026-08-24.json (Apple Watch sleep stages, exported 05:41 on 08-24) + OSCAR_Greg_by_Day_2026-08-23.csv (ResMed AirSense/AirFlow 11 via OSCAR). Session start 23:04 on 08-23 — 31 min earlier than 08-22's 23:35, but still ~35–50 min past the 22:15–22:30 target. CPAP hours 6.48 vs. Watch TIB 6.41 (Δ ≈ 4.2 min, CPAP longer — tight alignment).
CLUSTER FULLY RELEASED — AHI 3rd-LOWEST OF MATCHED HISTORY. AHI dropped 1.54 → 0.77 (−50%) for the second straight night of decline off the 08-21 peak of 2.82. That's a 3-night arc of 2.82 → 1.54 → 0.77 (−73% peak-to-tonight). AHI rank 3/51 (3rd-lowest ever recorded in matched window) — only two nights in the entire matched history have registered lower. OA dropped 8 → 3 (−63%), also rank 3/51 (3rd-lowest ever). Total obstructive count (OA+H) went 21 → 10 → 5 (−76%). OA per hour = 0.46/h, also rank 3/51, and less than half the 30-day baseline of 1.23/h. The escalation trigger from the 08-22 report — "AHI ≥ 2.0 for 3 consecutive nights" — is definitively NOT met (last 3 nights: 2.82, 1.54, 0.77 — only one above 2.0). Treat the 08-18 → 08-21 cluster as a fully released, self-limiting event. No urgent clinical action indicated on the obstructive-event channel. Continue to package the 5-night arc for the next scheduled check-in.
RERA PATTERN FIRMING UP — 2 CONSECUTIVE NIGHTS AT ≥ 2 RERAs. RERA count 3 → 2 — first back-to-back nights of RERA ≥ 2 in the matched window. Rank 42/51 tonight (7th-highest ever). Before 08-22, the matched-history mean was 0.67/night and only 4 of 49 nights registered ≥ 2 RERAs. Now 2 of the last 2. This is NOT yet a formal escalation trigger (that requires RERA ≥ 3 on 2+ nights in a 7-night window — currently just 1 night at ≥ 3), but the directional signal is real and distinct from the discrete-obstruction pattern that defined the 08-18/21 cluster. RDI 1.08 tonight vs. AHI 0.77 means RERAs contribute ~29% of RDI (up from 0% across the entire escalation window). The airway-resistance signature the 08-22 report flagged is now a 2-night trend, not a one-off. Watch 08-24: if RERA ≥ 2 for a 3rd straight night, this becomes a sustained resistance signature worth mentioning to the clinician independent of AHI (which may keep sliding as it did tonight).
REM RECOVERED — 5th-HIGHEST OF MATCHED HISTORY. REM % rebounded from 17.51% → 30.09% (+12.58 pts) — the largest one-night REM increase in the matched history. REM % rank 47/51 (5th-highest ever). REM absolute minutes went 64.2 → 113.4 min (+77%). This is a striking mirror-image of 08-22's REM collapse. Two things worth noting: (1) it fully validates the 08-22 read that the REM drop was TIB/truncation-driven (tonight TIB was essentially identical — 6.41 h — yet REM completely recovered, which rules out TIB as the sole cause; the more likely joint driver was early-night micro-arousals from the RERA cluster on 08-22 pushing REM into a later, then-truncated cycle); (2) it means the "REM decoupling" concern the 08-22 report opened is closed on a single-night basis — REM is not tracking event or RERA load. But the RERA-count remained at 2 tonight and REM was robust — so the two channels look genuinely independent rather than sequentially linked. Reassuring for the decoupling thesis established through 08-21.
LEAK 90th PERCENTILE POPPED FROM FLOOR — MINOR MASK WOBBLE. Leak_avg 0.74 → 0.94 (still under the 1.0 mechanical-issue threshold, but by only 0.06). Leak_90 jumped 0.0 → 3.6 L/min — from the absolute floor (where it had been for 2 nights) to rank 35/51. Leak_max 20.4 (rank 25/51, at median). This does NOT re-open the mask channel — leak_avg is still sub-threshold, and 3-of-4 recent nights (08-20, 08-21, 08-22) were at or near the leak-90 floor, so tonight's 3.6 could be a single-cycle mask shift or brief position change. Threshold to re-open the mask concern (per 08-22 report exit criteria): leak_avg > 1.0 for 2+ consecutive nights OR leak_90 > 4.0 for 2+ consecutive nights. Neither met tonight. Monitor 08-24; if leak_avg pops above 1.0 or leak_90 stays elevated, revisit.
PRESSURE RETURNED TO MEDIAN — 08-22 P_AVG WAS TRANSIENT. P_avg dropped 8.00 → 7.53 (−0.47), landing at rank 17/51 (below matched-history median of 7.80). P_90 barely moved (10.72 → 10.66, rank 29/51). This exactly matches the 08-22 report's alternate prediction: "if P_avg drops back to the 7.4–7.6 range, the auto-titrator will have concluded tonight was a one-off." Prediction MET. The auto-titrator saw fewer discrete obstructive events and fewer RERAs and reduced its response accordingly. Pressure channel is off active concern — track normally.
BEDTIME IMPROVED BUT STILL LATE — TARGET NOT MET. 23:04 bedtime (vs. 08-22's 23:35 — 31 min earlier). Still ~35–50 min past the 22:15–22:30 target the 08-21/22 reports both recommended. TIB 6.41 h (essentially unchanged from 08-22's 6.41). TST 6.28 h (+9.6 min vs. 08-22). Efficiency 98.04% (rank 42/51) — up from 95.31% because awake dropped from 18.06 → 7.53 min (rank 9/51). Improvement direction is right, magnitude is inadequate. Three consecutive nights past midnight would establish a bedtime-drift trend; two isn't there yet, but 08-24 is decisive — if it's again in the 23:00–23:30 window, this is a new baseline that will need to be surfaced explicitly.
| Metric | Value |
|---|---|
| Bedtime (session start) | 23:04 on 08-23 (−31 min vs. 08-22; still +35–50 min past 22:15–22:30 target) |
| Time in bed | 6.41 h (384.4 min) — rank 25/51; unchanged vs. 08-22; −16.9 min vs. 30-day (401.7) |
| Total sleep duration | 6.28 h (376.8 min) — rank 30/51; +9.6 min vs. 08-22 |
| Sleep efficiency (TST/TIB) | 98.04% — rank 42/51; +2.73 pts vs. 08-22 (95.31%) |
| Awake time | 7.53 min — rank 9/51 (9th-lowest); −10.5 min vs. 08-22 (18.06) |
| Core sleep | 3.46 h (207.7 min) — 55.13% — below 30-day 62.26% |
| Deep sleep | 0.93 h (55.7 min) — 14.78% — near 30-day baseline (13.90%); rank 29/51 |
| REM sleep | 1.89 h (113.4 min) — 30.09% — rank 47/51 (5th-highest ever); +12.58 pts vs. 08-22; +6.27 pts vs. 30-day (23.82%) |
| CPAP usage | 6.48 h — rank 22/51; unchanged vs. 08-22; −4.3% vs. 30-day (6.77) |
| AHI | 0.77 — rank 3/51 (3rd-lowest ever); −50% vs. 08-22 (1.54); −57% vs. 30-day (1.79) |
| RDI | 1.08 — rank 5/51; RERA contributes 2 of ~7 events (~29% of RDI, up from ~23% on 08-22) |
| OA / H / CA / UA / RERA | 3 / 2 / 0 / 0 / 2 — OA rank 3/51 (3rd-lowest ever); H rank 15/51; RERA rank 42/51 (7th-highest, 2nd consecutive night ≥ 2) |
| OA per hour | 0.46 /h — rank 3/51; −62% vs. 30-day baseline (1.23/h) |
| Pressure avg | 7.53 cmH₂O — rank 17/51 (below median); −0.47 vs. 08-22 (8.00); +0.07 vs. 30-day (7.46) |
| Pressure 90th %ile | 10.66 cmH₂O — rank 29/51 (at median); −0.06 vs. 08-22 (10.72); +0.59 vs. 30-day (10.07) |
| Pressure max | 11.94 cmH₂O |
| Leak avg | 0.94 L/min — rank 35/51; +0.20 vs. 08-22 (0.74); +0.09 vs. 30-day (0.85) |
| Leak 90th %ile | 3.6 L/min — rank 35/51; +3.6 vs. 08-22 (0.0); +1.0 vs. 30-day (2.56) |
| Leak max | 20.4 L/min — rank 25/51 (at median) |
| Compliance flag | 1 (compliant) |
| Date | AHI | OA | H | CA | RERA | Leak avg | Leak 90 | P_avg | P_90 | Hours | Deep % | REM % | Awake (min) | Eff |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 08-11 | 2.57 | 12 | 4 | 0 | 0 | 0.42 | 1.2 | 7.28 | 10.02 | 6.62 | 13.57 | 25.24 | 21.07 | 94.61% |
| 08-12 | 2.68 | 13 | 3 | 2 | 1 | 0.28 | 1.2 | 7.99 | 10.00 | 6.72 | 20.78 | 19.11 | 28.10 | 92.80% |
| 08-13 | 1.29 | 8 | 0 | 2 | 0 | 0.20 | 1.2 | 7.01 | 9.54 | 7.75 | 10.93 | 26.70 | 20.07 | 95.60% |
| 08-14 | 1.63 | 11 | 1 | 0 | 1 | 0.22 | 0.0 | 7.94 | 10.86 | 7.38 | 9.73 | 29.20 | 33.62 | 92.19% |
| 08-16 | 1.13 | 3 | 3 | 1 | 2 | 0.91 | 4.8 | 7.92 | 10.94 | 6.22 | 14.73 | 19.78 | 6.02 | 98.34% |
| 08-17 | 1.11 | 5 | 2 | 0 | 0 | 1.14 | 1.2 | 7.57 | 9.54 | 6.28 | 14.21 | 25.35 | 7.53 | 97.95% |
| 08-18 | 1.63 | 7 | 2 | 1 | 1 | 1.19 | 3.6 | 7.55 | 11.11 | 6.12 | 8.86 | 18.71 | 13.55 | 96.34% |
| 08-19 | 2.20 | 10 | 5 | 0 | 0 | 1.92 | 4.8 | 8.18 | 10.34 | 6.83 | 16.50 | 24.11 | 7.02 | 98.25% |
| 08-20 | 2.55 | 14 | 2 | 0 | 0 | 0.69 | 1.2 | 7.80 | 11.16 | 6.27 | 7.68 | 20.91 | 13.05 | 96.43% |
| 08-21 | 2.82 | 15 | 6 | 1 | 0 | 0.20 | 0.0 | 7.68 | 10.14 | 7.80 | 14.18 | 25.49 | 2.51 | 99.45% |
| 08-22 | 1.54 | 8 | 2 | 0 | 3 | 0.74 | 0.0 | 8.00 | 10.72 | 6.48 | 13.82 | 17.51 | 18.06 | 95.31% |
| 08-23 | 0.77 | 3 | 2 | 0 | 2 | 0.94 | 3.6 | 7.53 | 10.66 | 6.48 | 14.78 | 30.09 | 7.53 | 98.04% |
(08-15 not in table: no matched Watch sleep-stage session.)
6-night cluster arc + release (08-18 → 08-23):
| Date | AHI | OA | H | OA+H | OA/h | Leak_avg | P_90 | RERA | REM % | Interpretation |
|---|---|---|---|---|---|---|---|---|---|---|
| 08-18 | 1.63 | 7 | 2 | 9 | 1.14 | 1.19 | 11.11 | 1 | 18.71 | Leaky, moderate events |
| 08-19 | 2.20 | 10 | 5 | 15 | 1.46 | 1.92 | 10.34 | 0 | 24.11 | Peak leak, rising events |
| 08-20 | 2.55 | 14 | 2 | 16 | 2.23 | 0.69 | 11.16 | 0 | 20.91 | Leak-fixed, events still rising |
| 08-21 | 2.82 | 15 | 6 | 21 | 1.92 | 0.20 | 10.14 | 0 | 25.49 | Leak floor, event peak |
| 08-22 | 1.54 | 8 | 2 | 10 | 1.23 | 0.74 | 10.72 | 3 | 17.51 | Reversion + RERA emergence |
| 08-23 | 0.77 | 3 | 2 | 5 | 0.46 | 0.94 | 10.66 | 2 | 30.09 | Full release + RERA persists |
Interpretation. The obstructive-event channel is now fully released — AHI, OA, and OA/h are all in the bottom quartile of matched history, at levels last seen on 08-10 (AHI 0.58, OA 3). The 6-night arc traces a clean rise-and-fall: peaked on 08-21, halved on 08-22, halved again on 08-23. But the release didn't come clean — RERA count stayed elevated (2 for a 2nd straight night, vs. 0.67 baseline), which means the airway is still under some form of resistance-driven pressure even as the discrete-obstruction pattern collapses. Two possible readings: (a) the underlying stressor that drove the cluster is dissipating but hasn't fully cleared, and the residue is showing up as RERAs rather than OAs; (b) the RERA emergence is a genuinely new pattern that started 08-22, coincident with but not caused by the OA cluster's release. Right now, both are plausible — 08-24 will discriminate: if RERAs return to 0–1, reading (a) wins and the whole 6-night episode is a single closed event; if RERAs stay ≥ 2 for a 3rd night, reading (b) wins and a new resistance-driven pattern is establishing.
| 08-22 report prediction for 08-23 | Actual 08-23 outcome | Verdict |
|---|---|---|
| Sustained AHI ≤ 2.0 for 3 nights closes cluster | AHI 0.77 (2nd sub-2.0 night) | MET twice — cluster closed |
| Any rebound AHI ≥ 2.5 re-opens cluster | AHI 0.77 (far below) | MET — no rebound |
| RERA return to 0–1 = 08-22's 3 was a one-off | RERA = 2 (2nd elevated night) | MISSED — resistance signature persists |
| RERA ≥ 2 (2nd night in row) establishes resistance-driven signature | RERA = 2 (2nd night) | MET — pattern establishing |
| P_avg ≥ 7.9 = persistent resistance signal | P_avg = 7.53 (dropped 0.47) | NOT MET — no persistent pressure elevation |
| P_avg drops to 7.4–7.6 = 08-22's pressure was transient | P_avg = 7.53 (in the band) | MET — 08-22 pressure was a transient response |
| REM ≥ 20% confirms 08-22 drop was TIB-driven | REM % = 30.09 | MET dramatically — REM 5th-highest ever |
| Bedtime 22:15–22:30 | Bedtime 23:04 | MISSED — still 30–50 min late |
Net verdict: 6 met, 2 missed (RERA persistence, bedtime). The primary framing — "cluster released, but airway-resistance signature to watch" — is confirmed. The auto-titrator concurred by reducing pressure; the sleep architecture concurred by fully restoring REM.
Buckets on nightly leak average:
the user correlations (n=51, tonight included):
| Pair | r | Reading |
|---|---|---|
| leak_avg vs. deep_pct | −0.18 | Weak negative — high leak slightly costs deep |
| leak_avg vs. awake_min | −0.09 | Essentially zero |
| leak_avg vs. rem_pct | +0.02 | Zero |
| ahi vs. deep_pct | +0.18 | Weak positive — decoupling holds |
| ahi vs. awake_min | +0.03 | Zero |
| ahi vs. rem_pct | +0.04 | Zero |
| p_avg vs. awake_min | −0.22 | Weak negative — higher pressure → less awake time |
| p_avg vs. rem_pct | +0.00 | Zero |
| oa vs. deep_pct | +0.11 | Weak positive |
| oa vs. awake_min | +0.08 | Weak positive |
| rera vs. rem_pct | −0.02 | Zero — supports 2-channel independence |
| rera vs. awake_min | −0.23 | Weak negative (unexpected direction; likely small-n artifact) |
| rera vs. deep_pct | +0.02 | Zero |
| tib_h vs. rem_pct | +0.09 | Weak positive — modest support for TIB-drives-REM |
| tib_h vs. deep_pct | −0.17 | Weak negative |
Reading. Tonight's data point (very low AHI, low OA, but elevated RERA, and record-high REM) further reinforces the CPAP-mechanic-to-architecture decoupling that has held through this entire cluster: no correlation exceeds ±0.25 in absolute value. The RERA-vs-REM correlation is dead flat (−0.02), which strengthens the reading that even a persistent RERA pattern is unlikely to compromise REM at the whole-night level — 08-22's REM drop was TIB/truncation, not RERAs, as tonight (same TIB, more RERAs, but full REM recovery) makes plain.
| Metric | Tonight (08-23) | 7-day avg (through 08-22) | 30-day avg (through 08-22) |
|---|---|---|---|
| AHI | 0.77 | 1.85 | 1.79 |
| OA count | 3 | 8.86 | 8.33 |
| OA per hour | 0.46 /h | ~1.33 /h | ~1.23 /h |
| H count | 2 | 3.14 | 2.87 |
| RERA count | 2 | 0.86 | 0.77 |
| Leak avg (L/min) | 0.94 | 0.97 | 0.85 |
| Leak 90th (L/min) | 3.60 | 2.23 | 2.56 |
| Pressure avg (cmH₂O) | 7.53 | 7.81 | 7.46 |
| Pressure 90th (cmH₂O) | 10.66 | 10.56 | 10.07 |
| Usage hours | 6.48 | 6.57 | 6.77 |
| Deep sleep % | 14.78% | 12.85% | 13.90% |
| REM sleep % | 30.09% | 21.69% | 23.82% |
| Awake time (min) | 7.53 | 9.68 | 17.88 |
| Efficiency | 98.04% | 97.44% | 95.67% |
| TIB (h) | 6.41 | 6.46 | 6.70 |
Verdict: strong reversion on the event/architecture channels, RERA remains the outlier. AHI is −57% vs. 30-day and −58% vs. 7-day — the lowest single-night AHI since 08-10. OA is at 34% of 30-day baseline. REM is +6.3 pts vs. 30-day and +8.4 pts vs. 7-day. Awake time is 42% of 30-day baseline. Efficiency is well above baseline. Only two channels sit above trailing averages: RERA (2x the 30-day) and leak_90 (1.4x the 30-day). Both are watchable but neither is at an escalation threshold. Pressure is essentially at the 30-day mean.
| 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 | 1.57 | 0.61 | 7.76 | 10.34 | 7.9 | 6.62 | 7 |
| 08-19 to 08-23 (partial, n=5) | 1.98 | 0.90 | 7.84 | 10.60 | 10.0 | 6.77 | 5 |
Tonight's clean night dragged the partial current week from 2.28 AHI / 11.75 OA (through 08-22) down to 1.98 / 10.0 (through 08-23). AHI 1.98 is now below the 2.0 line — no longer a weekly high on the AHI axis (08-05 to 08-11 was 1.83; 08-19 to 08-23 is 1.98, +0.15). OA/night = 10.0 is still the highest weekly average of the pressure-instrumented window (all prior weeks were 6.7–8.6). One more low-event night on 08-24 would push the week's OA average below 9 and effectively normalize this week. The elevated-week signal is fading but not gone — the OA-per-night figure is still the outlier.
PRIMARY FINDING: CLUSTER RELEASED CLEANLY ON THE EVENT AXIS; RERA REMAINS THE OPEN QUESTION. AHI 0.77 and OA 3 are 3rd-lowest of matched history — the obstructive channel has not just returned to baseline but is comfortably below it. Two consecutive sub-2.0 AHI nights formally close the 4-night escalation cluster per the 08-22 report's exit criteria. The RERA count of 2 for a 2nd consecutive night is the one thread still to pull. RERAs at that level are not clinically alarming in absolute terms (RDI 1.08 tonight, well below any pathological cutoff), but the 2-night persistence is a new pattern that didn't exist before 08-22. The mechanistic story: the auto-titrator moved pressure down tonight (8.00 → 7.53) despite the persistent RERAs — meaning the machine now reads the airway as easier to keep open than it was on 08-22, but with some residual resistance still triggering micro-arousals. That's a mildly-improving-but-not-clean picture.
REVISED CLINICIAN PLAN. Both existing escalation triggers from the 08-22 report — (a) AHI ≥ 2.5 any single night; (b) AHI ≥ 2.0 for 3 consecutive nights — are moving away from being tripped. Trigger (c) — RERA ≥ 3 on 2+ nights in a 7-day window — is at 1/2. Trigger (d) — AHI ≥ 2.0 AND REM ≤ 15% on same night — is not close. No urgent clinical action this week. Continue to save the 6-night arc + RERA pattern for the next scheduled check-in. Add a new soft-watch trigger: RERA ≥ 2 for 3 consecutive nights would establish a genuinely sustained resistance signature worth mentioning explicitly (currently 2 of 3 nights in that criterion — 08-24 is the tiebreaker).
LEAK CHANNEL: MICRO-WOBBLE, NOT RE-OPENED. Leak_avg 0.94 is still under the 1.0 threshold; leak_90 popped from 0.0 to 3.6 in one night. That's a single-cycle event, not a pattern. Threshold to re-open: leak_avg > 1.0 for 2+ consecutive nights OR leak_90 > 4.0 for 2+ consecutive nights. Not met. Continue current setup.
PRESSURE CHANNEL: NORMALIZED. P_avg 7.53 is at the 30-day mean. The 08-22 spike was a transient response, as predicted. Off active concern.
ARCHITECTURE CHANNEL: NORMAL, TENDING FAVORABLE. REM at 30.09% is one of the best readings of the entire matched window. Deep at 14.78% is at 30-day baseline. Awake time at 7.5 min is 9th-lowest of matched history. Efficiency 98.04%. The whole architecture stack came in above baseline tonight — the best result in nearly two weeks. The 08-22 report's TIB-truncation-plus-RERA-fragmentation explanation for the previous night's REM drop is fully validated: tonight had the same TIB (6.41 h) and similar RERA count (2 vs. 3), yet REM completely recovered. That decouples RERA from REM even more decisively than the 51-night correlation (−0.02).
BEDTIME CHANNEL: PARTIAL IMPROVEMENT, TARGET NOT MET. 23:04 vs. 08-22's 23:35 (+31 min earlier) is directionally right but still 35–50 min past target. Two nights in the 23:00–23:35 window makes this the shape of a habit forming. The 08-24 bedtime is diagnostic: if it's again after 23:00, treat this as a new baseline and either explicitly adjust the target or investigate the driver (evening obligations, screen time, wind-down failure). If it drops to 22:30 or earlier, tonight was slippage remnant.
What to watch on 08-24.
2026-08-23T23:04:38) marks the start calendar date, which lines up with the OSCAR "Period" for 2026-08-23. Convention continues to hold.