Back to Search Start Over

Poster 8: Reliability of the Problem Behaviours Assessment for HD (short version)–Data From the TRACK-HD Study

Authors :
Miranda J. Say
Alexandra Durr
N. Arran
Marie-Françoise Boissé
R. A. C. Roos
S.J.A. van den Bogaard
J. Read
E. 't Hart
Sarah J. Tabrizi
J. Callaghan
Damian Justo
Eve M. Dumas
A. Coleman
Blair R. Leavitt
E. J. van Duijn
Track-Hd Investigators
David Craufurd
Catherine Bourdet
R. Dar Santos
Source :
Neurotherapeutics. 7:140-140
Publication Year :
2010
Publisher :
Springer Science and Business Media LLC, 2010.

Abstract

Background TRACK-HD is a multi-center, longitudinal study investigating a range of motor, cognitive, neuropsychiatric, and imaging measures as potential biomarkers for clinical trials in Huntington's disease (HD). Psychiatric assessments were conducted using a 10-item semi-structured interview, the short version of the Problem Behaviours Assessment for HD (PBA-s). Interviews were video recorded for quality assurance purposes, and a proportion randomly selected for re-scoring at another site. The study therefore provided an opportunity to examine the reliability of the measure between multiple pairs of raters in different cultural and linguistic contexts. Methods The study included 360 subjects (120 affected, 120 premanifest gene carriers, 120 controls), and 185 interviews were selected for re-scoring by an expert rater. Digital video recordings were uploaded to the TRACK-HD Web portal and selected for re-scoring by an independent investigator blind to the subjects' diagnosis. A further 73 interviews were obtained from the REGISTRY clinic in Manchester and used for training the expert raters. Results Complete agreement between raters was achieved for 86% of scores (κ = 0.733), and 97.6% were within one point of the expert rating, the standard for quality assurance purposes. Agreement levels for individual pairs of raters ranged from 75% to 94% (κ = 0.59–0.86); agreement within one point ranged from 92% to 99%. Complete agreement for the 20 individual ratings (10 symptoms * frequency and severity) ranged from 67% to 100% (κ = 0.56–0.97); agreement within one point ranged from 94% to 100%. Conclusions The overall results indicate substantial agreement between all raters, whereas the κ-statistic for individual rater pairs ranged from ‘moderate' to ‘almost perfect' agreement, with almost all raters achieving at least a ‘substantial' level of agreement. These data suggest that the PBA-s, with appropriate training, is a reliable measure of psychiatric symptoms in the HD population.

Details

ISSN :
19337213
Volume :
7
Database :
OpenAIRE
Journal :
Neurotherapeutics
Accession number :
edsair.doi...........5c9c949c53fb9546cf4d780bb19154dc
Full Text :
https://doi.org/10.1016/j.nurt.2009.09.010