Impulsive-Compulsive Behaviours in Belgian-Flemish Parkinson's Disease Patients: A Questionnaire-Based Study

Background Impulsive-compulsive behaviours (ICB) are a potentially harmful group of behavioural symptoms among the nonmotor aspects of Parkinson's disease (PD). Objective To develop and perform partial validation of a Belgian-Flemish version of the Questionnaire for Impulsive-Compulsive Disorders in Parkinson's Disease (QUIP) as a screening instrument for ICB in PD patients. Methods Using a translation-backtranslation method, we developed a Belgian-Flemish version of the QUIP, which was subsequently completed by 88 PD patients. QUIP-positive patients were invited for a semistructured diagnostic interview. Results A positive QUIP score for one or more ICB was observed in 37 patients (41%). In 15 patients (17%), a positive QUIP score for one or more impulse control disorders (ICD) was noted: pathological gambling in 1, hypersexuality in 8, compulsive shopping in 5, and compulsive eating in 8 patients. A positive QUIP score for punding, hobbyism, and/or walkabout was observed in 30 patients. The semistructured diagnostic interview was performed in 22 QUIP-positive patients. The diagnosis of ICB was confirmed in 6 patients, suggesting a positive predictive value of 27% for the Belgian-Flemish version of the QUIP. Conclusions We have developed a Belgian-Flemish version of the QUIP, which can be used as a screening questionnaire for ICB in PD patients. Our data suggest that sensitivity is high, specificity is low, and validity of the questionnaire is similar to the original version. We confirm the necessity of additional clinical assessment of QUIP-positive patients to ascertain a diagnosis of ICB.


Introduction
Parkinson's disease (PD) is the second most common neurodegenerative brain disorder and affects approximately 1% of the population above the age of 60 years in industrialized countries [1]. Besides motor symptoms, PD patients can experience several nonmotor problems. Impulse control disorders (ICD) are a subgroup of behavioural symptoms among the nonmotor aspects of PD. ese potentially devastating neuropsychiatric symptoms are characterized by failure to resist an impulse, drive, or temptation to perform an act that is harmful to the subject or to others [2]. Typical ICD in PD patients are pathological gambling (PG), hypersexuality (HS), compulsive eating (CE), and compulsive shopping (CS). In addition, compulsive behaviours, such as punding, hobbyism, and walkabout (PHW), can be observed. A particular form of behavioural disturbance is the dopamine dysregulation syndrome (DDS), in which patients develop an addictive behaviour towards dopaminergic medication. We will use the umbrella term of impulsive-compulsive behaviours (ICB) to cover the classical impulse control disorders (ICD) as well as other typical compulsive behaviours in PD patients (punding, hobbyism, and walkabout). e prevalence of ICD in PD patients ranges from 3.5% to 42.8% [2][3][4][5][6][7][8][9][10][11]. No significant difference in prevalence was observed between drug-naïve PD patients and healthy controls [3,12], suggesting that PD itself is not a major risk factor. e most important risk factor for developing ICD is probably the use of dopaminergic drugs, in particular dopamine agonists (DA). Higher DA dosage and longer treatment duration were associated with a higher prevalence of ICD in a longitudinal study in PD patients [13]. Additional risk factors for ICD have been reported, including male gender, younger age, levodopa therapy, personal history of smoking, preexistent history of ICD symptoms, personality profile characterized by impulsiveness and novelty-seeking, history of depression, history of substance abuse, and familial history of gambling [2,3,10,[14][15][16].
Weintraub and coworkers have validated the QUIP (Questionnaire for Impulsive-Compulsive Disorders in Parkinson's disease) as the screening instrument for ICB in PD patients [17]. We have translated the QUIP into Flemish in order to have a useful screening instrument for ICB in the Flemish-speaking patient population.

Methods
We have sent questionnaires by postal mail to 216 PD patients attending the outpatient clinic of the Antwerp University Hospital. All patients had a clinical diagnosis of Parkinson's disease according to the NINDS criteria [18] and absence of dementia was documented. e QUIP is a brief self-completed screening questionnaire. e first section assesses the following ICD: gambling, hypersexuality, and compulsive buying and eating behaviours. e second section assesses other compulsive behaviours (punding, hobbyism, and walkabout). e last section questions compulsive use of dopaminergic medication. Positive answers for each disorder are added, which leads to a maximum of five points for each of the ICD and DDS, and three points for punding, hobbyism, and walkabout. e QUIP was translated from English to Flemish by one member of the study team. Subsequently, two different team members independently translated this first Flemish version into English. e back-translated versions were compared to the original version and any discrepancies were modified after discussion among the team members. e patients were invited to complete the QUIP and an additional questionnaire to obtain information on sociodemographic status and personal and familial medical history. Both questionnaires were completed by the patient and, if available, with help of the caregiver. e QUIP-positive patients were invited for a more extensive interview, which was performed by one of the investigators. is semistructured interview was based on validated criteria, such as the DSM-IV criteria for gambling and binge eating [19], the Voon criteria for hypersexuality [20], and the Mc Elroy criteria for compulsive shopping [17,21], and a standardized semistructured questionnaire was used.
ree different investigators performed the diagnostic interviews, but the final interpretation was made by only 1 of those 3 investigators to reduce inter-rater bias.
Statistical analysis was performed using IBM SPSS version 20. Categorical variables were compared using Pearson's chi-squared test. Significant differences of continuous independent factors between groups were analysed using nonparametric tests. A receiver operating characteristic (ROC) curve was plotted for each ICB in the present study. We have compared the calculated area under the curve (AUC) in the present study with the analysis, which was reported for the original QUIP [17]. e sensitivity and specificity were also compared for each ICB, based on the validated cutoff values of the original questionnaire.

Compliance with Ethical Standards.
All activities involving human participants in this study were in accordance with the ethical standards of the institutional and national research committee and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards. e study was approved by the Antwerp University Hospital Ethics Committee (reference number 11/45/343). Informed consent was obtained from all individual participants included in the study for collection of questionnaire and medical data.

Results
Completed questionnaires were obtained from more than one-third of the patients (88/216 patients), resulting in a response rate of 40.7%. e demographic and clinical features of the patients are presented in Table 1.
QUIP scores, indicative for the probable presence of one or more ICB, were obtained in 37/88 patients (40.9%). In fifteen patients (17%), a positive QUIP score for one or more ICD was observed: PG in 1/88 (1.1%), HS in 8/88 (9.1%), CS in 5/88 (5.7%), and CE in 8/88 patients (1.1%). Punding, hobbyism, and/or walkabout (PHW) were present in 30/88 (34.1%) patients. We invited the QUIP-positive patients (37/ 88) for a more extensive interview. Fifteen (15/37) patients did not complete the confirmatory diagnostic interview and were excluded from further analysis. e positive predictive value of the Belgian-Flemish version of the QUIP was therefore calculated in the subgroup of 73 patients, excluding the 15 patients who did not complete the diagnostic interview. In this subgroup of 73 patients, a positive QUIP score was obtained for HS in 7/73 (9.6%), for CS in 3/73 patients (4.1%), and for CE in 7/73 patients (9.6%). In 18/73 patients (24.7%), a positive QUIP score for punding was observed, and in one patient, a positive score for DDS was documented. One patient (1/73; 1.4%) who scored negative on the QUIP (and thus was not invited for the interview) was diagnosed with PG on a follow-up visit with his treating neurologist.
Among the patients who completed both the QUIP and the diagnostic interview, a correct diagnosis of ICB was confirmed in 6/22 patients. e Belgian-Flemish translation of the QUIP therefore results in a positive predictive value (PPV) of 27% for the detection of ICB and a PPV of 33.4% for the detection of ICD (Table 2). e semistructured interview was only performed in QUIP-positive patients. erefore, we cannot make any statement concerning falsenegatives, and no valid conclusions can be inferred regarding sensitivity, specificity, and negative predictive value (NPV) of the Belgian-Flemish QUIP.

Discussion
We developed a Belgian-Flemish translation of the QUIP using a translation-backtranslation method. We conducted a study in 88 PD patients using this questionnaire. In 40.9% of the patients, the presence of an ICB was suggested, whereas in 17% of the patients, a positive QUIP score for an ICD was obtained. ese numbers are generally in line with the reported prevalence of ICB and ICD in previous studies, using the QUIP [2,9,14,[23][24][25][26][27][28].
Twenty-two QUIP-positive patients participated in a semistructured diagnostic interview. e diagnosis of ICB or ICD was confirmed in 6 of these patients. A positive predictive value of 27% was calculated for the detection of ICB by the QUIP in our study population, which is in accordance with studies using the English QUIP version. If we assume that all QUIP-negative patients would also have tested negative in the semistructured interview, we obtain a sensitivity of 100% and specificity of 76% for the Belgian-Flemish version of the QUIP. ese test characteristics are in accordance with the original English questionnaire [17]. We agree that in clinical practice, QUIP-positive patients should be further assessed in order to confirm the diagnosis of ICD/ICB [12,25]. e most prevalent ICD in our study population was hypersexuality, which was also the most frequently observed ICD in a recent review paper grouping 11 different studies [29]. e majority of QUIP false-positives were found in the punding/hobbyism/walkabout (PHW) group. According to the QUIP results, 34.1% of the patients suffered from PHW. e semistructured interview confirmed a diagnosis of punding in only 4.1% of the patients, whereas hobbyism and walkabout were confirmed in none of the patients. is observation could be partly explained by misinterpretation of the questions concerning hobbyism and walkabout. Furthermore, we should also keep in mind that the true ICD/ICB frequency in our study may remain underestimated, due to hesitation from patients and/or caregivers to acknowledge the presence of a potentially socially stigmatizing behavioural disorder.
ere was no systematic collection of information from caregivers or significant others, which also might contribute to this underestimation.
Our study has several other limitations, the most important one being the initial small sample size, partly due to a relatively low response rate (40.7%), as well as a low rate of patients completing the diagnostic interview (59.4%). e semistructured confirmatory diagnostic interview was only performed in QUIP-positive patients. erefore, we were not able to perform a reliable calculation of the sensitivity, specificity, and negative predictive value for the Flemish Our translation of the QUIP can be used as a screening instrument for ICB/ICD in the Dutch/Flemish-speaking PD patient population. e data of this pilot study suggest that the validity of the questionnaire is similar to the original version.

Data Availability
e data used to support the findings of this study are restricted by the Institutional Review Board of the Antwerp University Hospital according to local regulations, in order to protect patient privacy. Data are available, by contacting the corresponding author, for researchers who meet the criteria for access to confidential data.

Conflicts of Interest
e authors declare that they have no conflicts of interest.