Cognitive decline: telephone questionnaire yields scores close to in-person results in a small study

Someone close to the older person interviewed in two ways

The Informant Questionnaire on Cognitive Decline in the Elderly, or IQCODE, collects responses from a caregiver or family member. This study therefore compares two ways of interviewing that person, not two cognitive tests taken directly by the older adult.

The questionnaire was administered in person and by telephone, five to seven days apart. The order of administration was randomized. Two different researchers, who had received training beforehand, conducted the assessments without knowing the results of the other administration.

This design allows the methods to be compared while limiting the influence of administration order and knowledge of the first score. This is a study of agreement between methods, not a trial demonstrating a clinical benefit of remote assessment.

Overall agreement masks differences across questions

For the total score, the intraclass correlation coefficient, used to quantify agreement, was 0.74. The authors describe this agreement as high. They report no statistically significant difference between the two methods.

However, the absence of a statistically significant difference is not enough to prove that the methods are interchangeable. Crucially, coefficients for individual questions ranged from 0.17 to 0.98, indicating low to high agreement depending on the item. The overall score therefore does not capture all the variability observed.

A potential route to assessment access, not diagnostic validation

The authors consider telephone administration an option for expanding access to cognitive assessment when in-person assessment is difficult. The results support this possibility for the IQCODE, without establishing that telephone administration actually improves access to care.

The available abstract reports neither a comparison with a reference diagnosis nor any measurement of effects on care. It therefore does not support the conclusion that the two methods are equally good at identifying a cognitive disorder.

The small sample and the fact that information is available only in the abstract call for caution. Detailed participant characteristics, funding and conflicts of interest could not be verified.

What we can take away

Administering the IQCODE by telephone appears to produce an overall score fairly close to that obtained in person. This supports the possibility of remote assessment, without demonstrating diagnostic equivalence or allowing the findings to be generalized to all cognitive tests.