Many continuous variables should be analyzed using the relative scale: a case study of β2-agonists for preventing exercise-induced bronchoconstriction

Table 3 Analysis of β₂-agonist effects on FEV₁ decline in categories of placebo-test FEV₁ declines

FEV₁ decline after placebo (range)	Mean FEV₁ decline after placebo	Absolute effect on FEV₁ decline (pp) (95% CI)	Relative effect (95% CI)	N
10–19%	15.5%	15.2 (10.2–20.3)	0.99 (0.79–1.24)	33
20–29%	24.7%	23.6 (17.7–29.9)	0.92 (0.74–1.14)	29
30–39%	34.5%	33.0 (25.0–40.9)	0.95 (0.78–1.15)	34
40–49%	44.5%	39.7 (30.6–49.3)	0.83 (0.69–1.00)	29
50–83%	59.9%	44.3 (34.8–54.6)	0.71 (0.60–0.84)	34

Footnote: this table is restricted to 159 participants who had exercise-induced FEV₁ decline of ≥10% in the placebo test. The 95% CI for the relative effect was calculated with the approach described in [5], see Additional file 1. Abbreviations: pp., percentage points

ISSN: 2046-4053