Editors' note: Ordinal vs dichotomous analyses of modified Rankin Scale, 5-year outcome, and cost of stroke

Abstract

In “Ordinal vs dichotomous analyses of modified Rankin Scale (mRS), 5-year outcome, and cost of stroke,” Ganesh et al. reported that the ordinal mRS correlates better with 5-year mortality/disability/cost of care than the dichotomized mRS (using either 0–1 or 0–2 to measure good outcome). To optimize the utility of future stroke trials, these findings should be considered when determining how to compare poststroke mRS scores. One problem the authors noted with dichotomization is that patients with premorbid disability are likely to automatically be characterized as having a poor outcome because they cannot get better than their premorbid status poststroke. On a related note, Bruno comments that determining the prestroke mRS is subjective and that there are no guidelines for doing so. He recommends abandoning the use of the mRS to assess prestroke functional status and, instead, using a simple dichotomization of all patients as able or unable to complete activities of daily living prestroke. Ganesh et al. agree that there is a need for prestroke disability to be uniformly designated, but they caution that their poststroke disability data suggest that an ordinal approach may be higher yield than a dichotomized one. Notably, although Ganesh et al. demonstrate that the ordinal mRS correlates with 5-year mortality/disability/cost of care, the scale has been criticized for its dependence on the ability or inability to walk resulting in the automatic classification of all patients who are unable to walk, regardless of their cognitive status, as an mRS score 4 or 5. In “Ordinal vs dichotomous analyses of modified Rankin Scale (mRS), 5-year outcome, and cost of stroke,” Ganesh et al. reported that the ordinal mRS correlates better with 5-year mortality/disability/cost of care than the dichotomized mRS (using either 0–1 or 0–2 to measure good outcome). To optimize the utility of future stroke trials, these findings should be considered when determining how to compare poststroke mRS scores. One problem the authors noted with dichotomization is that patients with premorbid disability are likely to automatically be characterized as having a poor outcome because they cannot get better than their premorbid status poststroke. On a related note, Bruno comments that determining the prestroke mRS is subjective and that there are no guidelines for doing so. He recommends abandoning the use of the mRS to assess prestroke functional status and, instead, using a simple dichotomization of all patients as able or unable to complete activities of daily living prestroke. Ganesh et al. agree that there is a need for prestroke disability to be uniformly designated, but they caution that their poststroke disability data suggest that an ordinal approach may be higher yield than a dichotomized one. Notably, although Ganesh et al. demonstrate that the ordinal mRS correlates with 5-year mortality/disability/cost of care, the scale has been criticized for its dependence on the ability or inability to walk resulting in the automatic classification of all patients who are unable to walk, regardless of their cognitive status, as an mRS score 4 or 5.