Comparing factor structure of the Mood Disorder Questionnaire (MDQ): In Italy sexual behavior is euphoric but in Asia mysterious and forbidden

Abstract

Background

The introduction of screening questionnaires, such as the Mood Disorder Questionnaire (MDQ), has stimulated clinical and epidemiological studies on bipolar disorders. In this work, we studied the item response pattern of the MDQ in the Italian population and compared the results with those of the validation of the MDQ in Asian studies (Chinese and Korean), analyzing similarities and differences among the populations studied.

Methods

The sample was made up of 2278 participants, distributed as follows: 56.6% females, 50.8% living in the north-central Italy, and 33.7% living in rural areas.

The factor analysis was run on the matrix of tetrachoric correlations. The psychometric properties of the MDQ were also studied using the Rasch logistic model.

Results

The parallel analysis found two significant components. The first includes symptoms referring to acceleration, danger and irritability as risky behaviors, social interaction problems and mental flow. The second includes symptoms referring to self-confidence and energy. With respect to the Korean/Chinese results, the Italian sample, item 11 (“much more sex”), appears related to self-confidence and energy, while in Asia it is connected with items expressing risky behaviors and irritability.

Limitations

Differences in the frequency of comorbid disorders in Asian and Italian populations should be considered. The results should be confirmed and compared with those of other populations.

Conclusions

Cultural differences appear to be associated with a different symptomatic expression of bipolar spectrum disorders. Future research will investigate the role of gene–environment interaction in the genesis of these differences.

Introduction

Bipolar disorders are widespread and severe mental illnesses with a lifetime prevalence ranging from 1% to 2.6% in different surveys worldwide (Merikangas et al., 2011). Bipolar disorders onset at an early age, show high rates of recurrence, and often result in chronic morbidity, functional impairment, high suicide rates and are a burden on relatives and other caregivers (Fajutrao et al., 2009).

Currently, some methodological issues in the epidemiology of bipolar disorders are controversial (Carta and Angst, 2005): medical practice and screening studies often underestimate the problem. The episodes of hypomania and sub-threshold hypomania may be unrecognized as a pathological entity and/or forgotten by patients who are referred to health care facilities for depressive episodes.

The introduction of screening questionnaires such as the Mood Disorder Questionnaire (Hirschfeld et al., 2000, Hirschfeld et al., 2003) or the Hypomania/Mania Symptom Checklist-32 (Angst et al., 2010, Carta et al., 2006) has enhanced the assessment of bipolar disorders.

The positivity prevalence in the MDQ was quite similar in three community surveys carried out in western settings: 3.7% in the United States (Hirschfeld et al., 2003); 3% in Italy, (Carta et al., 2012a) and 3.5% in France (Carta et al., 2013). But some doubts have been raised about the accuracy of MDQ in detecting Bipolar Disorders because it showed low sensitivity in clinical settings in the United States (Zimmerman et al., 2009) and its low positive predictive value of MDQ was also found in imprisoned people (Kemp et al., 2008). In contrast, the cross-national validation studies of MDQ in Europe reported more encouraging findings: MDQ was found to have good accuracy in clinical psychiatric settings in the United Kingdom (Twiss et al., 2008), in Turkey (Konuk et al., 2007), in Spain (Sanchez-Moreno et al., 2008, de Dios et al., 2008), and in Italy and France both in clinical (Hardoy et al., 2005, Weber Rouget et al., 2005) and community samples (Carta et al., 2010, Rouillon et al., 2011).

In conclusion, the MDQ is an instrument whose value is currently recognized specifically to screen for a lifetime history of a manic or hypomanic syndrome. It is not considered a diagnostic tool and some criticisms remain concerning the use of MDQ as a case-finder for bipolar disorder (Zimmerman, 2012).

The original version of the MDQ consists of three questions. The first one includes thirteen yes/no items to evaluate the bipolar syndrome according to DSM-IV criteria. The other questions go deeper into the simultaneous presence of additional symptoms and their clinical severity. In any case, the main core of the instrument is the stem question: from here on, when referring to the MDQ, we refer only to these thirteen items.

Generally speaking, the MDQ has shown good internal consistency.

Furthermore, some studies have been conducted to identify a multifactorial structure. In these studies, Exploratory Factor Analysis was performed using the Varimax rotation, assuming orthogonal (i.e. independent) components.

Three surveys were conducted in western countries on psychiatric patients in Poland (Kiejna et al., 2010), Spain (Sanchez-Moreno et al., 2008), Italy and the United States (Benazzi and Akiskal, 2003). All these studies found a two-factor structure.

In the Chinese MDQ, validated on a psychiatric population (Chung et al., 2008, Lin et al., 2011) two factors were identified, whereas three factors were identified in the general population (Chung et al., 2009). But another Chinese study in a psychiatric population (Yang et al., 2011) found a three-factor structure. Three components are also found in a Korean outpatient sample (Jon et al., 2009).

Thus there may be some differences: in western contexts two-factor structures have been found in all cases, in China and Korea the results are not always univocal and have often found a structure with three factors. These differences may be linked to specific cultural variables since in China mood disorders have been described as presenting with specific symptoms (Kleinman, 1982), while more generally it has been suggested that the mode of expression of bipolar disorders can be influenced by the cultural context (Carta, 2013).

Hardoy et al. (2005) proposed an Italian version of the MDQ. The questionnaire was validated in an Italian (Sardinian) sample assumed to have clinical pathologies. The Italian MDQ showed a fairly good performance, and different cut-off points were identified for several uses of the questionnaire. The paper of Hardoy and colleagues provides important information about the accuracy against a diagnosis carried out by a clinician by means of a psychiatric interview as a gold standard. However, further analyses and in-depth studies are required. Furthermore, details about the factorial structure and the discriminant capability of items are necessary.

In this paper, the MDQ was tested on the general Italian population, using a larger sample than that of the study by Hardoy and colleagues, with a wider geographic and socio-demographic base. Different from related research, we do not discuss cut-off criteria or sensitivity/specificity of the instrument. We studied the differences in item response pattern of the MDQ, comparing the results on the Italian population against the validation of the instrument on Asian studies (Chinese and Korean), analyzing similarities and differences between the two populations.

Exploratory and confirmatory factor analysis was used to assess the dimensionality of the instrument, and the Rasch model was applied to assess the properties of the thirteen items of the first question of the MDQ. We studied the reliability of the scale, the calibration of the item difficulties and the differential item functioning across different subgroups of individuals.