The results of the analysis of the probability of complete pain relief at 2 days
| Number of studies | Range of proportion* | Meta-analysis of proportion (95% CI) | Heterogeneity | |
|---|---|---|---|---|
| Intention to have cordotomy | 6 [Refs. 14 17 21 23 24 Antrobus, unpublished data, 2011] | 0.20–1.00 | Fixed-effect model† 0.79 (0.68 to 0.87) | I2=56.9% (95% CI 0% to 82.67%) τ2=0.80 |
| Random-effects model‡ 0.75 (0.52 to 0.89) | ||||
| Had heat lesion | 6 [Refs. 14 17 21 23 24 Antrobus, unpublished data, 2011] | 0.33–1.00 | Fixed-effect model 0.83 (0.72 to 0.90) | I2=32.3% (95% CI 0% to 72.7%) τ2=0.34 |
| Random-effects model 0.81 (0.66 to 0.91) |
*The studies reported data on pain relief as ordinal data; these were dichotomised to complete pain relief versus partial or no pain relief.
†Inverse-variance weighted method.
‡DerSimonian and Laird method.