In the present study, adherence to OROS MPH treatment was generally high, as might be expected in a randomized, placebo-controlled trial. Self-reported adherence in real-life settings is, however, substantially lower [7, 8]. Factors associated with reduced adherence in the present study included score on the DUSI-R psychiatric disorder scale, completing university versus high school, and shorter time since diagnosis of ADHD; and men were more adherent than women. Adherence was also higher in subjects receiving placebo than in those receiving OROS MPH. In particular, the discontinuers in the MPH groups had a low adherence. In the patients who completed the trial, there was no difference in adherence between the treatment groups. In previous studies that looked at factors associated with adherence in adults with ADHD, adherence was significantly correlated with ADHD symptoms, substance abuse and younger age [7, 8]. These studies had small sample sizes, however, and were based on patient reports of adherence, so direct comparison with the present study is difficult.
It is possible that some of the factors identified as contributing to non-adherence (e.g. education, time since diagnosis) may, in fact, reflect some degree of misdiagnosis of ADHD, with subsequent impact on adherence and study discontinuation. For example, the disorder is highly genetic, is lifelong, and is likely to interfere with completion of education. Furthermore, some of its symptoms may be mimicked by other psychiatric disorders (e.g. anxiety or depression may produce difficulty concentrating or maintaining focus on current activities). It is important to note, however, that a short time since diagnosis is common in adults with ADHD, as they are unlikely to have been diagnosed during childhood in Europe. The impact of short time since diagnosis on adherence may therefore reflect lower acceptance of the disorder and less motivation for treatment. Similarly, lack of family history could suggest a subgroup who have ADHD that is less heritable, or where the patient does not know or realise that ADHD is present in the family (not uncommon when one is not yet familiar with the clinical presentation of ADHD). Psychiatric comorbidity in patients with ADHD may lead to low adherence through impact of MPH on other symptoms, such as anxiety or depression, while substance abuse is clearly associated with non-attendance, stagnation of treatment and interference with evaluation of medication. Similarly, completion of higher education does not necessarily point to an incorrect diagnosis; more intelligent individuals with ADHD tend to achieve higher levels of education than those of lower intelligence, yet still experience impairments relative to their potential.
The analysis of change in CAARS:O-SV showed that lower adherence was significantly associated with less improvement in ADHD symptoms, particularly in subjects receiving OROS MPH. The other notable predictor of treatment efficacy was the presence of a family history of ADHD; this may be because subjects with affected family members may have had access to greater support. As expected, baseline CAARS:O-SV score was also significantly associated with improvement in CAARS:O-SV, as was time since ADHD diagnosis. As non-adherence is related to study discontinuation, the analysis was repeated on study completers and it was again shown that non-adherence significantly affected the level of improvement in CAARS:O-SV. There were, however differences between the main analysis set and the completer population, which may be related to the impact of discontinuation in the main analysis set, and may also have been affected by the decreased statistical power in the smaller population or the lower variability in adherence in the completer population.
A main strength of the present study is the size of the cohort, although it should be noted that this was a post-hoc analysis and the study was not designed to assess adherence. In addition, identification of factors that are statistically associated with adherence does not always translate into clinical relevance . A randomized, controlled trial may not be the best setting in which to evaluate adherence, which may be higher than in the ‘real world,’ as suggested by previous small studies in adults with ADHD [7, 8]. The generalizability of the study findings may therefore be limited to clinical trials in which patient inclusion is strictly controlled, excluding patients with, for example, psychiatric co-morbidity. It should also be noted that there are numerous possible methods of assessing adherence, including patient reports, carer reports, pill counts, pharmacy data, electronic measurement and measurement of plasma drug levels; but none of these methods is ideal. In the present study, it is possible that measurements of adherence were affected by patients deliberately returning fewer capsules to give the impression that their adherence was higher . Furthermore, one important predictor of adherence—number of medication doses—could not be analysed in the present study, as all patients received a single daily dose of OROS MPH or placebo. Finally, the potential impact of selecting a cut-off for the definition of ‘adherent’ other than 95% was not investigated.
Concepts of adherence are continuing to evolve, and the factors involved in adherence can be extended to a broader concept of ‘concordance,’ as recently introduced by David Coghill (J. J. S. Kooij; personal communication). This describes something beyond adherence, namely mutual agreement between patient and physician with the need for medication and the correct way to take it. Patient education is a key part of achieving concordance.