Sensitivity analysis to evaluate the effect of quality of randomised controlled trials on IOP reduction between latanoprost and timolol
| Difference in percentage IOP reduction | 95% Confidence intervals | |
| Design of trial | ||
| All trials | 6.0 | 4.3, 7.8 |
| Double blind, parallel | 6.0 | 4.2, 7.8 |
| Double blind, crossover | 5.6 | −9.9, 21.1 |
| Single blind, parallel | 6.8 | −1.0, 14.7 |
| Withdrawals | ||
| All trials | 6.0 | 4.3, 7.8 |
| Withdrawals ⩽10% | 5.4 | 3.3, 7.5 |
| Withdrawals >10% | 7.8 | 4.4, 11.2 |
All pooling was undertaken using fixed effect model as no heterogeneity was detected by Q test.