Table 4.
Time t (hr) | OFF/ON (N = 26), expect ΔS I < 0 | ON/OFF (N = 26), expect ΔS I > 0 | ||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
QI (%) | Q2 (%) | Q3 (%) | p value | ΔS I > 0 % (confounders) | Q1 (%) | Q2 (%) | Q3 (%) | p value | ΔS I < 0 % (confounders) | |||
−2 | −4 | 3 | 13 | 0.9 | 16 | 62 | −15 | −6 | 1 | 0.7 | 19 | 73 |
−1 | 0 | 0 | 0 | 1 | 0 | 0 | 0 | 0 | 0 | 1 | 0 | 0 |
0 | −25 | −10 | −2 | 0.4 | 4 | 15 | −5 | 3 | 9 | 0.8 | 10 | 28 |
1 | −44 | −23 | −4 | 0.1 | 3 | 12 | −7 | 8 | 17 | 0.8 | 11 | 42 |
2 | −56 | −30 | −6 | 0.09 | 2 | 8 | −8 | 13 | 32 | 0.6 | 9 | 35 |
3 | −55 | −19 | −2 | 0.09 | 6 | 23 | −4 | 18 | 42 | 0.3 | 9 | 35 |
4 | −53 | −24 | −4 | 0.1 | 4 | 15 | −13 | 14 | 36 | 0.3 | 10 | 28 |
5 | −53 | −29 | −8 | 0.08 | 4 | 15 | −12 | 22 | 44 | 0.2 | 10 | 28 |
6 | −44 | −25 | −8 | 0.1 | 4 | 15 | −6 | 23 | 47 | 0.1 | 8 | 31 |
7 | −40 | −21 | −9 | 0.07 | 3 | 12 | −5 | 18 | 50 | 0.07 | 7 | 27 |
8 | −51 | −25 | −10 | 0.04 | 2 | 8 | −5 | 19 | 39 | 0.07 | 7 | 27 |
9 | −47 | −18 | −7 | 0.04 | 1 | 4 | −6 | 22 | 38 | 0.07 | 8 | 31 |
10 | −45 | −12 | −3 | 0.09 | 5 | 19 | −4 | 22 | 41 | 0.08 | 8 | 31 |
The results show ΔS I quartiles and number of confounders with ΔS I in direction opposite to median trend hypothesized. Q2 = median = 50 % percentile result. * p values measured using Wilcoxon rank sum tests, Q1 = 25 % percentile, Q3 = 75 % percentile