Comparison between the biased event-based NNT (per year) and the appropriate ‘instantaneous’ event-based NNT (per day) for three trials of the fluticasone-salmeterol combination inhaler (ICS) reporting the NNT to prevent a COPD exacerbation
| Biased event based (rate per year) | Instantaneous event based (rate per day) | ||||||
|---|---|---|---|---|---|---|---|
| Study | Time span for NNT | ICS | No ICS | NNT | ICS | No ICS | NNT |
| TORCH1 | 3 years | 0.85 | 1.13* | 4 | 0.0023 | 0.0031* | 1250 |
| Kardos3 | 44 weeks | 0.92 | 1.40 | 2 | 0.0025 | 0.0038 | 769 |
| Anzueto6 | 1 year | 1.10 | 1.59 | 2 | 0.0030 | 0.0044 | 714 |
COPD, chronic obstructive pulmonary disease; ICS, inhaled corticosteroid; NNT, number needed to treat.
*In this Table, “No ICS” refers to the placebo group of the TORCH trial.