Background: Balance dysfunction is common in stroke patients. The Berg Balance Scale (BBS) is useful for evaluating the balance function of stroke patients, and it can estimate the minimal clinically important difference (MCID) in balance. BBS scores differ among stroke patients depending on whether they require walking assistance. The MCID should thus be estimated separately for patients who require assistance and those who do not.
Objectives: To estimate the MCID of individuals who have had an early subacute stroke and require a walking aid and those who do not, to assist the clinical determination of the effectiveness of therapy.
Methods: This was a retrospective clinical analysis of 80 early subacute stroke patients. We estimated the MCID by using the Functional Ambulation Categories (FAC) as anchors for changes in BBS scores during a 1-month period. The MCID was estimated based on a cutoff score for separating the patients who achieved a FAC change ≥1 point on receiver operator characteristic curves. The area under the curve (AUC) was used to measure the discrimination accuracy. The MCID was estimated for the patients who needed walking assistance and those who did not.
Results: The estimated MCID of BBS scores in the assisted-walking group was 5 points and the AUC was 0.84 (p < .01); the corresponding values in the unassisted-walking group were 4 points and 0.62 (p = .26).
Conclusions: For early subacute stroke patients who require assistance to walk, a 5-point improvement in the BBS score is a useful indicator for reducing the amount of assistance.
Keywords: Minimal clinically important difference; balance; berg balance scale; early subacute stroke; outcome measure.