To the Editor:—Authors of “Hearing Impairment as a Predictor of Cognitive Decline in Dementia”1 are to be commended for their interest in the relationship between sensory and cognitive deficits, and their attempt to extend previous cross-sectional and longitudinal research on this topic. Unfortunately, inadequate and inappropriate measures, methods and analyses flaw their effort, and may have led them to draw incorrect inferences from their findings. First, the Mini-Mental State Examination (MMSE)2 is demonstrably affected by communication deficits, whether due to language or hearing. While a useful clinical tool, and frequently the instrument of practical choice in studies involving a large number of subjects, given the present sample size and involvement of a neuropsychologist, measures less sensitive to hearing impairment would have been more appropriate. Second, the variability in follow-up time and changes in assessment procedures (ie, both site and administrator), while likely driven by practical concerns, introduced additional sources of variance. Although the second administrator was said to have been “trained by” the first, this unfortunately does not substitute for evidence of high inter-rater reliability across groups. Further, there was evidence from the improved scores within the nonhearing-impaired groups that the change in site of administration from clinic to home did affect performance. The unaddressed issue is whether effects on exam performance were differential across groups, due to factors related or unrelated to hearing impairment. It is also unfortunate that audiometry was not repeated on follow-up, as a check on the reliability of repeat measure, even given the lack of predicted change within the follow-up period. Third, in general, multiple t-tests are not a preferred method of analyzing data of this type. Nonparametric analyses might have been more appropriate, given the nature and likely distributions of these data; however, if parametric approaches were defensible these should have been multivariate. As the authors report using ANCOVA to “adjust” for age effects in change scores, presumably these were conducted but not reported. An odd feature of the reported “age-adjusted” change scores was the resultant increase in group differences. Perhaps this was an error in the construction of Tables 1 and 2, with transposition of the bottom two lines of each? As the hearing-impaired and intact differed significantly in age, and both frequency and severity of hearing impairment have been associated consistently with scores on measures like the MMSE,3,4 this factor should have been taken into account in all between-group comparisons of MMSE scores, rather than solely the change measure. Finally, the methodological and analytical flaws noted may have led the authors to erroneous conclusions regarding the relationship between hearing and cognitive function. Previous cross-sectional studies, using audiometric techniques and the MMSE or similar measures of cognitive function, also found significant correlations between pure tone and speech discrimination scores and performance on the cognitive measures.3,4 Adjusting for age effects, however, reduced this association to a nonsignificant level. The above critical comments are not in contradiction of the clinical importance of hearing impairment in the assessment and management of patients with dementia syndromes, whether SDAT or of other etiologies. As risk for both hearing impairment and dementia increase with age, they are likely to be coincident in an elderly population. Moreover, the global deterioration in brain function associated with a progressive CNS neurodegenerative disease like SDAT, does sometimes appear to affect association areas at early stages. The result may be impaired speech discrimination, even with intact peripheral sensory function. To the degree that hearing impairment, whether due to sensori-neural damage or impacted cerumen, independently interferes with information processing. It can exacerbate the dysfunctions associated with dementia. One needn't look to sophisticated arguments for support of an awareness and prophyaxis of sensory deficits for patients with dementia. As with dentures and spectacles, every attempt to minimize functional deficits, including the use of hearing aids as appropriate, is certainly to be recommended.5 This letter was referred to the authors of the original paper, but no response was received.
Search from 250 M+ research papersSearch from 250 M+ research papers
Feb 1, 1995
E William Yund + 1
Open Access