Table 2. Evidence quality for each outcome.
| Asthma outcomes | Quality assessment | Effect | Quality | Importance | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Mean follow-up duration (months) | No. of studies | Design | Risk of bias | Inconsistency | Indirectness | Imprecision | Relative risk (95% CI) | Strong association | ||||
| Asthma exacerbation | 86.4 | 6 | Observational | No serious | No serious | No serious | No serious | 1.06 (1.04–1.09) | No | Moderate | Critical1 | |
| Hospitalization because of asthma | 43.6 | 4 | Observational | No serious | No serious | No serious | No serious | 1.22 (1.12–1.34) | No | Moderate | Important | |
| Unscheduled doctor visit because of asthma | 12 | 1 | Observational | No serious | No serious | No serious | Serious2 | 4.26 (2.52–7.19) | Yes3 | Low | Important | |
| ED visit because of asthma | 23.6 | 5 | Observational | No serious | No serious | No serious | No serious | 1.06 (1.01–1.10) | No | Moderate | Important | |
| SCS use for asthma | 15 | 1 | Observational | No serious | No serious | No serious | Serious4 | 1.20 (0.93–1.56) | No | Low | Important | |
CI, confidence interval; ED, emergency department; SCS, systemic corticosteroid. 1, exacerbations are the most important outcome for patients with asthma (3); 2, sample size =287; 3, the relative risk for an unscheduled visit for asthma was 4.257; 4, the Optimal Information Size criterion was met, the 95% CI overlapped with no effect, and the CI failed to exclude important harm.