# Reference‐Group Adjusted Behavioural Dysfunction Questionnaire Score Discriminates Highly Behavioural‐Variant Frontotemporal Dementia From Major Depressive Disorder and Alzheimer's Disease Dementia

## TL;DR

A new scoring method for a questionnaire improves the ability to distinguish behavioral variant frontotemporal dementia from other brain disorders.

## Contribution

A reference-group adjusted scoring method for the BDQ is introduced to enhance its diagnostic accuracy for bvFTD.

## Key findings

- Adjusted BDQ scores showed higher or similar discrimination between bvFTD and MDD or ADD compared to unadjusted scores.
- The adjusted method achieved at least 90% sensitivity and specificity with smaller cut-off differences.
- Machine learning techniques were used to validate the adjusted scoring method via fivefold cross-validation.

## Abstract

Given the fact that behavioural variant frontotemporal dementia (bvFTD) is characterised by behavioural disorders, the assessment of these disorders is essential for early diagnosis of bvFTD. In this regard, the recently developed Behavioural Dysfunction Questionnaire (BDQ) that captures the bvFTD‐specific behavioural disorders is promising in discriminating mild‐stage bvFTD from other neurodegenerative and psychiatric disorders. In this study, we aimed to increase the discriminatory power of the BDQ by adaptation of its scoring depending on the reference group to bvFTD.

In this combined prospective and retrospective cross‐sectional study, data of 241 patients [i.e., 50 patients with mild‐stage bvFTD, 71 patients with major depressive disorder (MDD) and 120 patients with mild‐stage Alzheimer's disease dementia (ADD)] were analysed. We calculated the BDQ score in two ways: (1) as the average score of the domains' mean scores and (2) by adjusting the scoring depending on the reference group by using machine learning techniques, validated by fivefold cross‐validation.

The adjusted BDQ score showed a higher (bvFTD vs. MDD) or similar (bvFTD vs. ADD) discriminatory power than the unadjusted BDQ score, with a considerably smaller difference between cut‐offs with at least 90% sensitivity and at least 90% specificity.

We recommend using adjusted BDQ scores when MDD or ADD are the reference groups to bvFTD. Similar approaches should be taken for other reference groups to bvFTD to best reflect the thinking of clinicians who have specific reference groups in mind as differential diagnoses to bvFTD.