Predicting Step 2 CK Performance Using Automated Feature Selection and Nested Cross-Validation.

Attention Deficit/Hyperactivity Disorder (ADHD) is a neurodevelopmental disorder with inattentive, disorganized, hyperactive, and impulsive behaviors as its main clinical manifestations (). An umbrella review of 13 meta-analyses, including 588 studies with 3,277,590 patients from Africa, China, India, the US, Canada, Spain and so on, found a pooled prevalence of ADHD of 8.0% (95%CI 6–10%); however, ranging from different countries (). Comorbidity of psychiatric disorders is considered common in ADHD, the most common comorbidity was Oppositional Defiant Disorder (ODD, 34.7%), followed by Anxiety Disorders (18.4%), Specific phobias (11.0%), Enuresis (10.8%), and Conduct Disorder (CD) (10.7%) (). In Australia, a study revealed a total cost of AUD 20.57 billion in 2019, indicating high costs and a heavy burden at the individual, family, and societal level (; ). Medication treatment is one of the most common treatments for ADHD, which is categorized into stimulants (methylphenidate, amphetamines, etc.) and non-stimulants (atomoxetine, guanfacine, clonidine, etc.) (; ; ). The treatment effects will be influenced by medication adherence. Medication adherence, which is defined as the extent to which patients take their medications as prescribed by their healthcare provider (), is very improtant for medication treatment. There have been a lot of medication adherence measurements. However, no single measurement strategy has been deemed optimal, and it still remains global challenge because of the absence of gold standard methods for adherence measurement (; ). Medication adherence measurements could be categorized into subjective measurements (e.g., Medication Adherence Rating Scale, MARS), and objective measurements (e.g., medication possession ratio, MPR; proportion of days covered, PDC, etc.) (). Self-management during medication treatment significantly influences medication adherence. However, children and adolescents are often faced with shared management among multiple participants, with the primary role switching from parent-directed in childhood to self-directed in adolescence during the development (; ). Thus, it will be challenging to achieve good medication adherence in children and adolescents. For children and adolescents with other diseases, medication non-adherence may be associated with more emergency department visits, increased risk for hospitalization, increased healthcare use, and need for treatment escalation (; ). In children and adolescents with ADHD, a cohort study found discontinuation of medication in children with ADHD will lead to poor treatment outcomes whether medication is resumed or not, and children who stopped medication (162/166, 97.5%) had at least one impairment at the follow-up assessment time point after the last reported stop (). Another study found that good adherence may be protective against the risk of co-morbidities, and ADHD patients with good adherence have a significantly lower risk of oppositional defiant disorder (ODD) and conduct disorder (CD) (). Therefore, it is necessary to understand the current status and explore the risk factors of medication adherence in children and adolescents with ADHD.