Main outcomes, patients with RRMS at baseline: subgroup of patients with RRMS at baseline (n=4217; average follow-up 8.9 years)
| Outcome measure | Model | Actual progression (95% CI) | Predicted progression (natural history)(95% CI) | Absolute treatment effect (predicted natural history progression less actual) (95% CI) | Relative rate of disease progression (actual divided by predicted natural history)(95% CI) |
| (1) | (2) | (3)=(2−1) | (4)=(1)/(2) | ||
| Rounded EDSS* | Markov | 1.55 (1.49 to 1.61) | 1.80 (1.78 to 1.83) | 0.25 (0.19 to 0.31) | 86% (83% to 89%) |
| EDSS* | MLM | 1.55 (1.49 to 1.61) | 2.23 (2.21 to 2.25) | 0.68 (0.62 to 0.74) | 69% (67% to 72%) |
| Utility | Markov | 0.113 (0.108 to 0.119) | 0.164 (0.163 to 0.166) | 0.051 (0.046 to 0.056) | 69% (66% to 72%) |
| MLM | 0.113 (0.108 to 0.118) | 0.150 (0.148 to 0.152) | 0.037 (0.031 to 0.042) | 76% (72% to 79%) |
*For the Markov model, half-integer EDSS states are combined with the next lower integer EDSS state. Calculations with the MLM using rounded EDSS values gave almost identical results to those using the full EDSS scale.
EDSS, Expanded Disability Status Scale; MLM, multilevel model; RRMS, relapsing remitting multiple sclerosis.