Table 1.
Frequency of overall error with devices (quantitative and qualitative estimates)
| Device | Cross-sectional studies | RCTs | ||||
|---|---|---|---|---|---|---|
| Number of studies | Pooled estimate [95% CI] (I-squared)* | Range of error frequency | Number of studies | Pooled estimate [95% CI] (I-squared)* | Range of error frequency | |
| MDI | 19 | 86.8% [79.4–91.95] (98.5%) | 45.7–100.0% | 2 | 89.5% [75.7–95.9] (60.0%) | 85.1–94.2% |
| MDI with spacer | 3 | 52.0% [24.3–78.6] (87.3%) | 31.6–78.0% | 0 | N/A‡ | N/A‡ |
| BA-MDI (Autohaler®) | 2 | 57.7% [52.9–62.5] (84.9%) | 54.9–59.9% | 0 | N/A‡ | N/A‡ |
| Turbuhaler® | 4 | 55.5% [24.5–82.7] (93.6%) | 22.0–88.7% | 3 | 73.6% [55.6–86.1] (92.1%) | 54.0–82.2% |
| Diskus® | 1 | N/A‡ | 49.0% | 3 | 62.4% [51.9–72.0] (61.6%) | 50.0–68.3% |
| Aerolizer® | 2 | 75.1% [31.0–95.3] (97.2%) | 54.0–89.1% | 0 | N/A‡ | N/A‡ |
| Handihaler® | 0 | N/A‡ | N/A‡ | 1 | N/A‡ | 66.7% |
N/A Not applicable
* I-squared statistic indicates the percentage of variance that is attributable to between-study heterogeneity. The higher the percentage, the greater the heterogeneity. ‡Not applicable as number of studies was ≤1