Table 1:

Transition probabilities according to modified Rankin Scale score^*

Time after stroke; mRS score^†	mRS score; 9-mo transition probabilities between different mRS states from 90 d to 1 yr after stroke^‡
90 d	0	1	2	3	4	5	6
Canada
0	0.450	0.427	0.089	0.014	0.002	0.000^§	0.017
1	0.159	0.575	0.196	0.045	0.008	0.000^§	0.017
2	0.057	0.333	0.371	0.174	0.046	0.002	0.017
3	0.007	0.064	0.147	0.467	0.210	0.015	0.090
4	0.001	0.012	0.039	0.212	0.336	0.052	0.349
5	0.000^§	0.001	0.004	0.035	0.115	0.444	0.401
China
0	0.445	0.422	0.088	0.014	0.002	0.000	0.029
1	0.157	0.568	0.193	0.044	0.008	0.000	0.029
2	0.056	0.328	0.367	0.172	0.045	0.002	0.029
3	0.007	0.063	0.145	0.461	0.207	0.015	0.102
4	0.001	0.011	0.038	0.208	0.329	0.051	0.362
5	0.000	0.001	0.004	0.034	0.113	0.435	0.413
	mRS score; annual transition probabilities between different mRS states beyond 1 yr after stroke
1 yr	0	1	2	3	4	5	6
Canada
0	0.782	0.188	0.013	0.000^§	0.000^§	0.000^§	0.017
1	0.046	0.819	0.108	0.005	0.005	0.000^§	0.017
2	0.002	0.082	0.746	0.073	0.070	0.002	0.024
3	0.000^§	0.001	0.021	0.367	0.351	0.021	0.239
4	0.000^§	0.001	0.021	0.367	0.351	0.021	0.239
5	0.000^§	0.000^§	0.000^§	0.000^§	0.000^§	0.733	0.267
China
0	0.772	0.185	0.013	0.000^§	0.000^§	0.000^§	0.029
1	0.045	0.808	0.106	0.005	0.005	0.000^§	0.029
2	0.002	0.080	0.737	0.072	0.069	0.002	0.036
3	0.000^§	0.001	0.020	0.361	0.346	0.021	0.251
4	0.000^§	0.001	0.020	0.361	0.346	0.021	0.251
5	0.000^§	0.000^§	0.000^§	0.000^§	0.000^§	0.721	0.279

Note: mRS = modified Rankin Scale.
↵* All the transition probabilities in the table were used for the base-case analysis. All probabilities in the table follow a Dirichlet distribution. We constructed the ranges of all transition probabilities by increasing and decreasing the base-case value by 20%.
↵† 0 = no symptoms, 1 = symptoms without clinical disability, 2 = slight disability, 3 = moderate disability, 4 = moderately severe disability, 5 = severe disability, 6 = death.
↵‡ Sources and assumptions: we initially estimated the transition probabilities for both arms, within 1 year and beyond 1 year, by fitting the data from the Oxford Vascular Study (38) into a multistate model (Appendix 2, available at www.cmajopen.ca/content/11/3/E443/suppl/DC1). (23)^, (39) We further adjusted the probabilities to account for the differences in background mortality between populations from different countries by substituting mortality data for the UK population in the Oxford Vascular Study with mortality rates for Canada and China and proportionally adjusting the other transition probabilities. (40)^– (44) We assumed that the transition probabilities within 1 year and beyond 1 year were the same in the 2 treatment arms.
↵§ The value is not 0 but, rather, smaller than 0.0005.