Measures of clinical effect in vestibular rehabilitation

Abstract

Full text

Dear Editor,

Peripheral vestibular hypofunction is a condition caused by the involvement of specific regions of the inner ear or the vestibular nerve, characterized by the functional loss of one or both peripheral vestibular systems, which can cause symptoms such as dizziness, vertigo, postural instability, oscillopsia, nausea, loss of balance, and proprioceptive deficit. 1 These symptoms can interfere with the quality of life of patients, making it difficult to perform daily activities. This dysfunction can increase the risk of falls, due to the impairment of balance and body perception, reinforcing the need for targeted strategies that promote functional recovery and prevent secondary complications. 1

Proprioceptive training has emerged as an alternative for the treatment of individuals with vestibular hypofunction, since it consists of using mechanical and somatosensory stimuli to improve skills that depend on sensorimotor integration, through exercises that challenge postural control and body perception. 2 This process is influenced by neural factors that are important for the consolidation of motor learning, such as the function of the cerebellum and primary cortex, which, when combined with different somatosensory stimuli, produce proprioceptive adaptations. 3

In this context, the study by Özaltın et al. 4 presented relevant findings regarding the vestibular rehabilitation of individuals with peripheral vestibular hypofunction. However, some issues related to the clinical usefullness of the results and their applicability, in addition to statistical significance, require further exploration to strengthen their practical implications.

When analyzing the data and calculating the measures of clinical effect, 5 it was observed that groups 1 and 2 present statistical similarity in some variables but generate distinct clinical effects. This may be attributed to a sampling problem, resulting in type-II error.

Measures of clinical effect were calculated, including Cohen's d, Cohen's U3, number needed to treat (NNT), and probability of superiority (PS). 5 In dynamic balance, assessed by the Tinneti Gait and Timed Up and Go (TUG) tests, Cohen's d values were 0.93 and −0.99, respectively, indicating a large clinical effect of group 1, compared to group 2. These results show that 82.4% (Cohen's U3) of the individuals in group 1 surpassed the mean of group 2 in the Tinneti Gait test, with a probability of superiority of 74.5% and an NNT of 2.98. Similar findings were observed for the TUG test ( Table 1 ).

In the sensation avoiding item, the clinical effect between the groups was also high (d = −0.95). This implies that 82.9% of the individuals in group 1 are above the mean of group 2, with 63.5% overlap between the groups, and a 74.9% probability of superiority for group 1. Furthermore, to achieve a more favorable outcome in group 1 compared with group 2, it is necessary to treat, on average, 2.9 individuals ( Table 1 ).

The greatest clinical effects of group 1 over group 2 were observed in items shoulder posture (d = −4.22), quality of Life (d = −3.57), and neck tilt (d = −2.84). This hierarchy of clinical effects is relevant for the selection of therapeutic approaches and for the planning of interventions in the continuum of care.