JMIR Publications

ABSTRACT

Background. Attention-Deficit/Hyperactivity Disorder (ADHD) affects approximately 3-5% of adults globally, characterized by inattention, hyperactivity, and impulsivity, causing substantial functional impairment across occupational, academic, and social domains. Associations between lifestyle factors, including physical activity patterns, sleep quality and duration, screen time behaviors, dietary intake patterns, anthropometric characteristics, and substance use, remain significantly underexplored in North African populations and require comprehensive international investigation, given the severe limitations in epidemiological data. Understanding these modifiable factors could inform evidence-based interventions to manage symptoms and improve function. Objectives. To (i) estimate adult ADHD prevalence in Tunisia and internationally stratified by presentation type and demographics, (ii) examine associations between comprehensive lifestyle factors and symptom severity across multiple domains, and (iii) employ machine-learning (ML) algorithms to identify complex non-linear patterns and interaction effects between lifestyle variables and ADHD symptomatology across diverse populations. Methods. This two-phase quantitative cross-sectional study will recruit approximately 5,000 Tunisian adults aged 18-65 years in Phase I, followed by 50,000 international participants across North Africa, the Middle East, Europe, and North America in Phase II. Data collection employs a dual-mode approach: Google Forms for digital administration and paper-based questionnaires for participants with limited internet connectivity, with mode selection determined by availability at the time of distribution. The assessment battery comprises validated instruments totaling approximately 130-132 items requiring approximately 28-32 minutes of completion time, including the Adult ADHD Self-Report Scale, the International Physical Activity Questionnaire-Short Form, the Pittsburgh Sleep Quality Index, the Smartphone Addiction Scale-Short Version, the Bergen Social Media Addiction Scale, the Screen Time Questionnaire, Short Food Frequency Questionnaire, and the novel Substance Use Assessment Scale. To accommodate the international sample, all instruments will be offered in English, French, and Arabic, allowing participants to choose their preferred language. Officially validated translations will be used where available. For instruments lacking a validated version, a standardized translation will be employed for this study, with subsequent psychometric validation planned. ML algorithms, including random forests, gradient boosting, and neural networks, represent the primary analytical approach, complemented by multivariable regression for association examination. Expected Outcomes. This protocol provides the first comprehensive adult ADHD prevalence estimates for Tunisia and establishes international baseline cross-cultural data enabling systematic comparisons across geographic regions and healthcare systems. ML identification of complex interaction patterns between lifestyle factors and symptom presentations represents the primary methodological contribution, revealing non-linear relationships and distinct phenotypic subgroups. Findings will inform the development of targeted lifestyle-based interventions addressing modifiable risk factors.