Dose-Dependent Effects of Stimulant Medication Treatment on Criminal Convictions in Individuals with Comorbid ADHD and Substance Use Disorders: A Within-Individual Analysis Using Swedish National Registers

Attention-deficit/hyperactivity disorder (ADHD) and substance use disorders (SUD) frequently co-occur, with up to 25% of adults with persisting ADHD symptoms meeting criteria for SUD, and approximately 25% of individuals with SUD meeting diagnostic criteria for ADHD. This comorbidity is associated with worse outcomes than either condition alone, including higher rates of psychiatric hospitalization, premature death, and criminal behavior. Stimulant medication treatment (SMT) holds a dominant position in the pharmacological treatment of ADHD and has demonstrated good efficacy on core symptoms in both children and adults. In patients with comorbid SUD, however, the evidence base is limited. A 2015 meta-analysis found only a small effect of standard-dose SMT on ADHD symptoms and no effect on substance use in this group. A small number of studies have examined higher dosage ranges, with some promising results — notably a Swedish trial using methylphenidate up to 180 mg/day and an American trial using extended-release mixed amphetamine salts equivalent to approximately 200 mg lisdexamfetamine/day — but these were hampered by high dropout rates and short follow-up periods. In 2024, the Swedish Agency for Health Technology Assessment (SBU) concluded that the available evidence is insufficient to determine whether SMT benefits ADHD symptoms or substance use in this population. SMT has also been proposed to reduce criminal behavior in individuals with ADHD, through its effects on impulsivity and executive functioning. A 2012 Swedish register study found lower conviction rates during treatment periods among ADHD patients, but a 2024 systematic review found limited evidence for this association. Importantly, neither of these specifically examined patients with concurrent SUD, nor did they investigate dose-response relationships. The purpose of this study is to examine dose-dependent associations between SMT and criminal convictions in individuals with comorbid ADHD and SUD, using Swedish national register data from 2005–2024. By employing a within-individual design, the study eliminates confounding from time-invariant individual characteristics — such as genetic predisposition, stable personality traits, and socioeconomic background — that would bias between-individual comparisons. Dispensed prescriptions from the Swedish Prescribed Drug Register will be used to construct treatment episodes. Different stimulant preparations will be converted to dexamphetamine equivalents to allow unified analysis across medication types. Treatment periods and dosage levels will be ascertained using the PRE2DUP (From Prescription Drug Purchases to Drug Use Periods) method, a validated algorithm for estimating drug use periods from dispensing data. Dosage levels will be categorized into three tiers: normal dose, high dose, and very high dose. Periods of imprisonment and psychiatric inpatient care will be censored, as these represent time during which individuals are not at liberty to commit offences in the community. The primary analysis will use stratified Cox regression to compare the hazard of criminal conviction across dosage categories (no treatment, normal dose, high dose, and very high dose), using within-individual variation. Normal dosage will serve as the reference category, allowing simultaneous assessment of whether higher doses confer additional benefit and whether any treatment is superior to no treatment. Secondary analyses will examine associations with specific crime types. The study population comprises all individuals in Sweden who received diagnoses of both ADHD (ICD-10: F90) and SUD (ICD-10: F10–F19) in specialized care during 2005–2024, estimated at approximately 15,000 individuals based on preliminary data from Region Skåne. Expected outcomes include clarifying whether dose-response relationships exist between SMT and criminal behavior in this population, and whether higher-than-standard doses — which are used in clinical practice but lack an evidence base — are associated with differential risk of criminal convictions. The findings are expected to contribute to the development of evidence-based treatment guidelines for this patient group, where current clinical practice varies substantially between providers and regions.