SCNR. A heretic alternative. [BE/BA News]
couldn’t resist to try an alternative based on the Hodges-Lehmann estimates instead of medians. Same setup like yesterday.
Simulation settings: 2,500 studies with 16 subjects
Sampling every 5 minutes up to 2 × tmax of R (2.50 h),
exponentially increasing intervals to tlast (16 h) = 38 samples:
0, 5, 10, 15, 20, 25, 30, 35, 40, 45, 50, 55, 60, 65, 70,
75, 80, 85, 90, 95, 100, 105, 110, 115, 120, 125, 130, 135,
140, 145, 150 min, 3.5, 4, 5.5, 7, 9.5, 12.5, 16 h.
Seven formulations
L = lower limit: tmax = 1.000 h, ka = 2.190 / h, t½,a = 19.0 min
T1 = fastest : tmax = 1.125 h, ka = 1.822 / h, t½,a = 22.8 min
T2 = fast : tmax = 1.188 h, ka = 1.672 / h, t½,a = 24.9 min
R = Reference : tmax = 1.250 h, ka = 1.539 / h, t½,a = 27.0 min
T3 = slow : tmax = 1.316 h, ka = 1.417 / h, t½,a = 29.4 min
T4 = slowest : tmax = 1.389 h, ka = 1.296 / h, t½,a = 32.1 min
U = upper limit: tmax = 1.562 h, ka = 1.065 / h, t½,a = 39.0 min
Simulation results:
L = lower limit
Median : 1.0833 h (Bias: +0.0833 h, Range: 0.7500 - 1.4583 h)
HL : 1.1146 h (Bias: +0.1146 h, Range: 0.8333 - 1.4583 h)
Skewness: +0.4452
T1 = fastest
Median : 1.2083 h (Bias: +0.0833 h, Range: 0.8750 - 1.6250 h)
HL : 1.2500 h (Bias: +0.1250 h, Range: 0.9167 - 1.5833 h)
Skewness: +0.4093
T2 = fast
Median : 1.2917 h (Bias: +0.1042 h, Range: 0.9167 - 1.7917 h)
HL : 1.3125 h (Bias: +0.1250 h, Range: 0.9583 - 1.6667 h)
Skewness: +0.3858
R = Reference
Median : 1.3333 h (Bias: +0.0833 h, Range: 1.0000 - 1.7917 h)
HL : 1.3750 h (Bias: +0.1250 h, Range: 1.0417 - 1.7500 h)
Skewness: +0.3505
T3 = slow
Median : 1.4167 h (Bias: +0.1009 h, Range: 1.0000 - 1.8750 h)
HL : 1.4375 h (Bias: +0.1217 h, Range: 1.1250 - 1.7500 h)
Skewness: +0.2960
T4 = slowest
Median : 1.5000 h (Bias: +0.1111 h, Range: 1.1250 - 2.0000 h)
HL : 1.5208 h (Bias: +0.1319 h, Range: 1.1250 - 1.9583 h)
Skewness: +0.2594
U = upper limit
Median : 1.6667 h (Bias: +0.1042 h, Range: 1.2500 - 2.2500 h)
HL : 1.6771 h (Bias: +0.1146 h, Range: 1.3333 - 2.1250 h)
Skewness: +0.0878
HL: Hodges-Lehmann estimate
Comparisons:
According to the guidance (based on medians)
passed ‘±20% criterion’ (80.00-125.00%)
L = lower limit: 52.1%
T1 = fastest : 80.9%
T2 = fast : 90.4%
T3 = slow : 91.2%
T3 = slowest : 84.8%
U = upper limit: 57.0%
Heretic alternative (based on Hodges-Lehmann estimates)
passed ‘±20% criterion’ (80.00-125.00%)
L = lower limit: 53.2%
T1 = fastest : 87.2%
T2 = fast : 94.9%
T3 = slow : 95.2%
T3 = slowest : 89.5%
U = upper limit: 61.1%
Seemingly more powerful.
Dif-tor heh smusma 🖖🏼 Довге життя Україна!
Helmut Schütz
The quality of responses received is directly proportional to the quality of the question asked. 🚮
Science Quotes
Complete thread:
- EMA: New product-specific guidances Helmut 2022-04-08 15:17 [BE/BA News]
- How would you implement it? ElMaestro 2022-04-09 11:45
- Confuse a Cat Inc. Helmut 2022-04-09 18:40
- Confuse a Cat Inc. ElMaestro 2022-04-09 21:47
- Confuse a Cat Inc. Ohlbe 2022-04-11 11:29
- Confuse a Cat Inc. Helmut 2022-04-11 13:59
- So many questions, so few answers Helmut 2022-04-11 13:03
- Preliminary simulations Helmut 2022-04-30 14:59
- Preliminary simulations ElMaestro 2022-04-30 19:10
- Preliminary simulations Helmut 2022-05-01 15:56
- Revisions of the PSGLs final Helmut 2023-06-23 13:29
- Revisions of the PSGLs final dshah 2023-06-28 14:43
- EMA: No problems with many sampling time points… Helmut 2023-06-28 15:59
- New simulations & some desultory thoughts Helmut 2023-06-29 11:34
- SCNR. A heretic alternative.Helmut 2023-06-30 11:50
- Revisions of the PSGLs final dshah 2023-06-28 14:43
- Preliminary simulations ElMaestro 2022-04-30 19:10
- Simulated distributions Helmut 2022-05-02 13:43
- Confuse a Cat Inc. Ohlbe 2022-04-11 11:29
- Confuse a Cat Inc. ElMaestro 2022-04-09 21:47
- Confuse a Cat Inc. Helmut 2022-04-09 18:40
- How would you implement it? ElMaestro 2022-04-09 11:45