The Problematic Behaviour Scale (PBS-10): psychometric characteristics and cut-offs using Swiss national survey data

Background and relevance:

The Assessment of Criteria for Specific Internet-use Disorders (ACSID-11) introduced and tested by Müller et al. (2022; Journal of Behavioral Addictions 11: 427-450) in Germany offers a valuable 11-item instrument to screen non-substance-bound behavioural disorders in population-wide surveys based on ICD-11 criteria. Given the heterogeneity of screening instruments in the field, ACSID-11 reacts to the need to develop a harmonized instrument bridging different behavioural disorders, which guarantees comparability. To be applied in the Swiss context and other countries, the scale needs to be further tested.

Data and methods:

Hence, we included an adapted version of ACSID-11 – the Problematic Behaviour Scale (PBS-10) – in the Swiss nationally representative cross-sectional survey study “Health and Lifestyle” 2023 (n=5995). We generalized the scale for offline and online behaviour (social media use, shopping, gaming, gambling, and pornography use). The survey was conducted in German, French, and Italian. We assessed the factorial structure of the scale using exploratory varimax-rotated principal component factor analysis (PCA). Internal consistency was analyzed using Cronbach’s alpha. To evaluate the external validity of PBS-10, we analyzed its association with the Patient Health Questionnaire-4 (PHQ-4). To determine cut-offs of the scores of PBS-10 for each behaviour, we conducted multiple Receiver Operator Characteristic (ROC) regressions of PHQ-4 on PBS-10 controlling for sex, age, education, and survey language.

Results:

The PCA generally retains the two factors impaired control/increased priority and escalation/continuation for the three survey languages and for all cases. Internal consistency of the PBS-10 is very high, and varies from α=.90 for social media to α=.97 for gambling. The external validation analysis shows that for each 1sd of PBS-10 PHQ-4 is between .27sd for shopping and .42sd for gambling higher (p<.001). ROC-regression shows that the highest areas under the curve (AUC) for the tested cut-offs of the normalized PBS-10 scores (0-100) vary between AUC=.56 for shopping to AUC=.67 for pornography use. Based on these cut-offs we estimate the following prevalences of problematic behaviour as a percentage of regular users in Switzerland: social media use: 26.3%, shopping: 10.2%, gaming: 21.6%, gambling: 6.0%, pornography use: 5.3%.

Discussion and conclusion:

In all, PBS-10 has proven to serve as a suitable screening tool for problematic behavioural conduct for the activities social media use, shopping, gaming, gambling, and pornography use (online and offline) in German, French and Italian in the Swiss context. However, the external validity with regard to self-reported mental disorders is low to modest and the ability to maximize sensitivity and specificity is limited. Hence, future research may address clinical validation of the PBS-10 and its applicability for further behaviours.