Prescribing Multiple Neurostimulants during Rehabilitation for Severe Brain Injury
Table 3
Univariate optimal data analysis (UniODA) results, neurostimulants as explanatory variable.
Outcome variable
UniODA model
% satisfied
ESS
Recovery of full consciousness during rehabilitation
If received one neurostimulant, then predict recovery of consciousness during rehabilitation
30
43.3
1.000
0.8
If received multiple neurostimulants, then predict no recovery of consciousness during rehabilitation
66
42.4
1.000†
−30.9†
Recovery of full consciousness within one year
If received one neurostimulant, then predict no recovery of consciousness within one year
30
56.7
0.506
6.8
If received multiple neurostimulants, then predict recovery of consciousness within one year
70
64.3
Change in total DOCS-25 score above MDC90 of 4.7
If received one neurostimulant, predict clinically detectable change in DOCS-25 score
22
40.9
0.800
4.1
If received multiple neurostimulants, predict clinically nondetectable change in DOCS-25 score
70
35.7
Change in total DOCS-25 score above MCID of 2.58
If received one neurostimulant, predict clinically detectable change in DOCS-25 score
22
50.0
0.463
7.4
If received multiple neurostimulants, predict clinically nondetectable change in DOCS-25 score
70
40.0
indicates number of observations in a given predicted class category. % satisfied indicates percentage of observations in a given predicted class category. † indicates LOO and ESS values reported due to instability. That is, if LOO ESS is lower than training ESS, then the results are not LOO stable.