Personalized Medication Response Prediction for Attention-Deficit Hyperactivity Disorder: Learning in the Model Space vs. Learning in the Data Space

Abstract

Attention-Deficit Hyperactive Disorder (ADHD) is one of the most common mental health disorders amongst school-aged children with an estimated prevalence of 5% in the global population (American Psychiatric Association, 2013). Stimulants, particularly methylphenidate (MPH), are the first-line option in the treatment of ADHD (Reeves and Schweitzer, 2004; Dopheide and Pliszka, 2009) and are prescribed to an increasing number of children and adolescents in the US and the UK every year (Safer et al., 1996; McCarthy et al., 2009), though recent studies suggest that this is tailing off, e.g., Holden et al. (2013). Around 70% of children demonstrate a clinically significant treatment response to stimulant medication (Spencer et al., 1996; Schachter et al., 2001; Swanson et al., 2001; Barbaresi et al., 2006). However, it is unclear which patient characteristics may moderate treatment effectiveness. As such, most existing research has focused on investigating univariate or multivariate correlations between a set of patient characteristics and the treatment outcome, with respect to dosage of one or several types of medication. The results of such studies are often contradictory and inconclusive due to a combination of small sample sizes, low-quality data, or a lack of available information on covariates. In this paper, feature extraction techniques such as latent trait analysis were applied to reduce the dimension of on a large dataset of patient characteristics, including the responses to symptom-based questionnaires, developmental health factors, demographic variables such as age and gender, and socioeconomic factors such as parental income. We introduce a Bayesian modeling approach in a “learning in the model space” framework that combines existing knowledge in the literature on factors that may potentially affect treatment response, with constraints imposed by a treatment response model. The model is personalized such that the variability among subjects is accounted for by a set of subject-specific parameters. For remission classification, this approach compares favorably with conventional methods such as support vector machines and mixed effect models on a range of performance measures. For instance, the proposed approach achieved an area under receiver operator characteristic curve of 82–84%, compared to 75–77% obtained from conventional regression or machine learning (“learning in the data space”) methods.

Is an irritable ADHD profile traceable using personality dimensions? Replicability, stability, and predictive value over time of data-driven profiles

“…Personality measures are easily administered in the clinical setting, and our finding indicates that personality might be relevant to predict treatment response. Researchers trying to predict treatment responders vs. non-responders in ADHD based on machine learning could, therefore, potentially benefit from including a personality questionnaire among their measures [54]. Whether personality measures will also predict different responses to different types of treatments could further be investigated.…”

Neurobiological support to the diagnosis of ADHD in stimulant‐naïve adults: pattern recognition analyses of MRI data

“…The possibility of assessing stimulant‐naïve ADHD patients on an individual basis gives us the opportunity of capturing information about the pathophysiology underlying the symptoms of ADHD without the bias of treatment‐induced brain alterations. Because all previous ML‐based MRI studies of adults never treated for ADHD symptoms were carried out using child and adolescent samples , the findings reported herein represent the first neurobiological evidence supporting the individual diagnosis of stimulant‐naïve ADHD patients specifically in adulthood. The diagnostic performance achieved in our study is similar to that reported in recent, large‐scale ML‐based MRI studies investigating potentially severe psychiatric disorders in adults such as schizophrenia and better than figures reported in studies of other major mental conditions such as bipolar disorder .…”

Individual Variability in Reaction Time and Prediction of Clinical Response to Methylphenidate in Adult ADHD: A Prospective Open Label Study Using Conners’ Continuous Performance Test II

“…The literature presents variable and partly conflicting results concerning the use of clinical data to predict response to MPH. Age, weight, gender, comorbid anxiety or oppositional defiant disorder, IQ, academic achievement, disease duration, severity of symptoms, level of maternal ADHD symptoms, and prior atomoxetine use have all been shown to have some predictive effect (Buitelaar et al, 2011; Buitelaar, Van Der Gaag, Swaab-barnkvhld, & Kuipkr, 1995; Chazan et al, 2011; Fredriksen, Dahl, Martinsen, Klungsoyr, Haavik, et al, 2014; Ishii-Takahashi et al, 2015; Johnston et al, 2015; Setyawan et al, 2015; Treuer et al, 2014; Wong et al, 2017). It has been difficult to predict treatment response to stimulants using neuropsychological tests in ADHD patient as well, and results from these studies have also been conflicting (Coghill et al, 2007; Wong et al, 2017).…”

“…A number of studies have examined response to medication in ADHD patients, to see whether baseline measures can be used to predict treatment response, primarily using clinical variables and neuroimaging markers (Buitelaar et al, 2011; Coghill, Rhodes, & Matthews, 2007; Ishii-Takahashi et al, 2015; Johnston et al, 2015; Kim, Sharma, & Ryan, 2015). In such studies, different statistical methods and machine learning approaches have been employed (Ishii-Takahashi et al, 2015; Wong et al, 2017).…”