MDQ ineffective for BD screening in clinical practice
MedWire News: Results from a US study do not support the use of Mood Disorders Questionnaire (MDQ) for bipolar disorder (BD) screening in psychiatric clinical practice because of low positive predictive values.
"Several self-report screening questionnaires have been developed to improve the detection of BD, the most widely studied being the MDQ, observe Mark Zimmerman (Brown Medical School, Providence, Rhode Island) and team.
However, they add: "Studies of the performance of the MDQ in heterogeneous samples of psychiatric outpatients presenting for treatment have raised concerns about the adequacy of the MDQ as a screening measure because of its relatively low sensitivity."
To investigate further, the researchers examined the operating characteristics of the MDQ at all cutoff scores to determine the cutoff point that would be appropriate for the purpose of BD screening.
They studied 752 psychiatric outpatients, aged a mean of 39.4 years, who were interviewed with the Structured Clinical Interview for DSM-IV, and completed the MDQ.
In total, 80 (10.6%) patients were diagnosed with a lifetime history of bipolar disorder, of whom 33 were diagnosed with bipolar I disorder, 27 with bipolar II disorder, 15 with bipolar disorder not otherwise specified, and five with cyclothymia.
Overall, the mean score on the MDQ symptom items was 4.7 out of 13, with BD patients scoring significantly higher than those without the disorder, at 9.5 versus 4.1.
The researchers found that when positivity of BD was based only on symptom score and not level of impairment then the cutoff score associated with 90% sensitivity was 5.0. At this cutoff, the specificity of the MDQ was 60.7%, positive predictive value was 22.1%, and negative predictive value was 98.8%, with an area under the receiver operating characteristic curve (AUC) score of 0.87.
When BD positivity required at least a mild level of impairment, then the cutoff score associated with 90% sensitivity was 1.0. At this cutoff, the specificity of the scale was 41.8%, positive predictive value was 15.4%, and negative predictive value was 96.9%. The AUC score was 0.83.
When BD positivity required at least a moderate level of impairment, then the scale did not achieve 90% sensitivity even at a cutoff of 1.0. At this cutoff, the sensitivity was 77.5%, specificity was 61.5%, positive predictive value was 19.3%, and negative predictive value was 95.8%. The AUC score was 0.80.
At the cutoff recommended by the developers of the MDQ to identify BD cases (≥7 symptoms and at least moderate impairment), the sensitivity was 67.5%, specificity was 84.5%, and its positive and negative predictive values were 34.2% and 95.6%, respectively.
Zimmerman and team conclude in the journal Comprehensive Psychiatry: "These findings indicate that when the cutoff to identify cases on the MDQ was set to achieve a desired level of sensitivity as a screening instrument most cases screening positive on the scale did not have BD.
They add: "Low positive predictive value does not support the use of the MDQ or any bipolar disorder screening scale in psychiatric clinical practice."
MedWire (www.medwire-news.md) is an independent clinical news service provided by Springer Healthcare Limited. © Springer Healthcare Ltd; 2011
By Mark Cowen