Systematic review of quality of life and functional outcomes in randomized placebo-controlled studies of medications for attention-deficit/hyperactivity disorder

Attention-deficit/hyperactivity disorder (ADHD) affects approximately 6.8% of children, 2.8% of adolescents and 2.5% of adults worldwide [75, 76, 83]. The disorder is defined as persistent and developmentally inappropriate symptoms of inattention and/or hyperactivity–impulsivity that interfere with a patient’s social, academic and/or occupational functioning [8, 9]. The nature of functional impairment varies from patient to patient and with age and may encompass diverse outcomes such as underperformance at school or at work, unemployment or low income, substance abuse, smoking, teenage pregnancy, arrest, divorce or acquiring sexually transmitted disease [61]. Health-related quality of life (HRQoL) is generally acknowledged to represent the impact of ill-health on an individual’s ‘perception of their position in life, in the context of culture and value systems in which they live, and in relation to their goals, expectations, standards and concerns’ [2, 104]. It has become clear that ADHD has a negative impact on patients’ HRQoL and that this may be further exacerbated by, or may increase the risk of, other psychiatric conditions such as anxiety and depression [31, 34, 54, 93]. Impaired day-to-day functioning in domains such as educational achievement and interpersonal relations is the reason that most often underlies a patient’s or parent’s decision to seek medical advice [72].

Pharmacotherapies for ADHD include the psychostimulants methylphenidate (MPH) and amphetamines [including the prodrug lisdexamfetamine (LDX)], and the non-stimulants atomoxetine (ATX) and guanfacine. Guidelines recommend that drug therapy is used as part of a multi-modal treatment plan, which should include psychoeducation or other non-pharmacological interventions such as parent training and cognitive behavioural therapy [13, 102]. Randomized controlled trials of pharmacotherapies in children, adolescents and adults with ADHD have typically used a clinician-rated symptom scale as the primary efficacy outcome measure, and meta-analyses of these studies have confirmed that ADHD medications are very effective in relieving patients’ symptoms [33, 77, 88, 89]. As recognized in the recent European Medicines Agency guidance on inclusion of functional outcomes in clinical studies of ADHD medications [44], it is now widely acknowledged that treatment of ADHD should aim not only to relieve patients’ symptoms, but also to improve their functioning and HRQoL.

Assessing functional impairment or HRQoL in a randomized controlled trial typically relies on completion of one or more of the many available ADHD-specific or generic questionnaires by patients themselves or by proxy raters (physicians, parents, teachers or family members) [30]. Use of these measures in clinical trials of ADHD medications has grown rapidly in recent years [30]. Because ADHD is a behavioural disorder that affects multiple aspects of patients’ lives, the 18 well-defined symptoms of ADHD interact and partially intersect with the constructs of functional impairment and HRQoL (Fig. 1). Although measures of functional impairment may share many similarities with measures of HRQoL, functional impairment is usually considered to be objective and ideally assessed by unbiased methods, whereas HRQoL is usually considered to be subjective and ideally assessed by the patients themselves. What precisely is being measured by a patient-rated functional impairment instrument or a proxy-rated HRQoL instrument may therefore be open to debate and discussion [2]. Another important aspect of measuring functional impairment and HRQoL is that the instruments should sample domains of impairment that are commonly affected by patients’ symptoms, but should not merely serve as surrogate measures of symptoms. Moderate, but not strong, correlations between measures of symptoms, functional impairment and HRQoL suggest that related, but distinct, constructs are being assessed [18, 19, 21, 32, 47, 81, 92, 95].

These factors present challenges for the development and selection of instruments to measure HRQoL and functional impairment. Generic instruments have the advantage of capturing a broad picture of health status, but the potential disadvantage of poor sensitivity to the deficits characteristic of ADHD. Disorder-specific instruments may offer improved sensitivity by specifying particular impairments, but may miss others. When any single patient typically reports impairment in only a small proportion of the specified items within a domain, the instrument may suffer from poor sensitivity if the scoring system averages out this potentially severe and clinically relevant impairment across all the specified items within that domain. US Food and Drug Administration guidance on patient-reported outcome measures states that instruments should be psychometrically validated and standardized to a reference population [46]. Standardization may present problems for ADHD-specific instruments if they are not designed for use in people who do not have ADHD. An advantage of generic instruments that are standardized to a reference population is that they enable direct comparison of the burden of disease in patients with ADHD with that of community norms and patients with other physical or mental health problems.

Results from clinical trials and observational studies using questionnaire-based instruments indicate that patients with ADHD experience severely compromised HRQoL [34]. In a 2010 systematic review of clinical trials, open-label studies and post hoc analyses, evidence supporting a positive short-term effect of medication on quality of life was limited mainly to studies of ATX in children and adolescents [29]. Recent evidence from observational and registry studies indicates that pharmacological treatment of ADHD is associated with increased achievement and decreased absenteeism at school [16], a reduced risk of trauma-related emergency hospital visits [58], reduced risks of suicide and attempted suicide [28], and decreased rates of substance abuse [27] and criminality [56].

This systematic review focuses on the effect of pharmacological treatment of ADHD on functional impairment and HRQoL outcomes in randomized placebo-controlled studies. We aim to: survey the instruments used in these studies; assess the severity of baseline impairment relative to controls, when possible; collate the most robust evidence about the impact of medications on HRQoL and functional impairment outcomes; and assess the relationship of treatment-related changes in these measures with changes in ADHD symptom scales.

A systematic literature searching and screening strategy was designed to identify articles published in peer-reviewed journals that reported findings of randomized placebo-controlled clinical trials of pharmacotherapy for ADHD using measures of functional impairment or HRQoL as outcomes. We extracted these outcomes for qualitative review and analysis, together with the principal ADHD symptom-based outcomes from the same studies.

This systematic review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines and was registered with the PROSPERO international prospective register of systematic reviews (identifier: CRD42015027595).

This systematic review aimed to evaluate HRQoL and functional outcomes in published randomized placebo-controlled clinical trials of medications for ADHD. Tandem extraction of primary or principal symptom-based efficacy outcomes alongside HRQoL and functional outcomes allowed the results of interest to be considered in the context of the effect of study drugs on ADHD symptoms.

Baseline data from the included studies were consistent with published observational studies in showing that children and adolescents with ADHD have substantially impoverished parent-rated HRQoL compared with population norms [32, 80]. Adult HRQoL instruments used in the studies were nearly all ADHD specific, and baseline values are therefore not comparable with reference populations. Baseline deficits on all multi-domain HRQoL and functional impairment instruments, whether generic or ADHD specific, or self- or proxy-rated, were more pronounced in domains relating to achievement/productivity, risk-taking and interpersonal relations than in other psychosocial domains. Little or no impairment was evident in domains relating to physical functioning on the generic instruments. Participants in the included studies were not selected based on HRQoL assessments, but were usually required to have moderate to severe ADHD symptoms (e.g. ADHD-RS-IV total score ≥28), indicating that the poor HRQoL observed at baseline is a reflection of the patients’ characteristics rather than the study inclusion criteria. Overall, these findings support the notion that day-to-day functional impairments and HRQoL deficits are typical in patients with ADHD.

Pharmacotherapy relieved ADHD symptoms to a significantly greater extent than did placebo in all but two of the included studies, as would be expected from the large body of published clinical trials and recent meta-analyses [33, 77, 88, 89]. The two studies in which no significant benefit over placebo was found were both small studies in adults (N ≤ 50), one of reboxetine and one of dextroamphetamine. Effect sizes of medication versus placebo for ADHD symptom-based outcomes (when available) were above 0.5 in nearly all of the included studies in children and adolescents, and in most of the studies in adults. We used an effect size of 0.5 as an approximate but universally applicable threshold for a clinically meaningful difference. This is supported by a study showing that nearly all estimates of minimum important differences for HRQoL instruments were close to 0.5 standard deviations and that this may correspond to the limit of people’s ability to discriminate differences over a range of criteria [71].

Functional impairment and poor HRQoL can be viewed as constructs that relate to, but are distinct from, ADHD symptoms and that reflect the impact of ADHD on patients’ day-to-day lives (Fig. 1). In support of this conceptualization, severity of or improvement in ADHD symptoms correlated moderately to strongly (but not perfectly) with severity of or improvement in functional impairment and HRQoL deficits, both in the included studies for which this was reported [42, 62, 66, 86] and elsewhere in the literature [18, 19, 21, 32, 47, 81, 92, 95]. Furthermore, effect sizes (when available) of active medications versus placebo were smaller for functional or HRQoL outcomes than for symptom-based outcomes in nearly all of the included studies (Tables 4, 5). None of this evidence, however, reveals whether medications affect functioning and HRQoL directly, or only through the medium of symptom relief, or both.

In children and adolescents, large effect sizes (≥0.8) were observed for HRQoL and/or functional outcomes in short-term studies of long-acting stimulant medications (LDX, OROS-MPH and TD-MPH). Large effect sizes were not observed in any included studies of non-stimulants (ATX and GXR) and were below 0.5 in many of these. Treatment effect sizes in the included studies suggest that placebo-adjusted improvements in some measures of functioning and of HRQoL are likely to be of sufficient magnitude to be considered potentially clinically relevant, especially with stimulant treatment. Large effect sizes in an individual domain of an instrument may not necessarily reflect large overall effects on HRQoL if some domains have less impact on overall HRQoL than others according to the instrument’s factor structure. Nevertheless, effect sizes on multi-domain instruments were largest in domains relating to achievement/school, risk-taking and interpersonal relations, reflecting the domains with the greatest deficits observed at baseline. In adults, stimulant and non-stimulant effect sizes for functional or HRQoL outcomes were below 0.5 in all included studies except the single adult LDX study. Before drawing any conclusions about the effectiveness of medications or the nature of impairments and deficits in different age groups, however, it should be noted that almost all HRQoL and functional outcomes were parent-rated in child/adolescent studies and self-rated in adult studies. If HRQoL is a subjective experience that can only be judged by patients themselves, then it is questionable whether HRQoL can really be rated by parents, teachers or investigators acting as proxies. Proxy-rated HRQoL assessments may therefore be more appropriately regarded as measures of functional impairment than of HRQoL.

The diversity of instruments used to assess HRQoL and functional impairment in the clinical trials of ADHD medications included in this review suggests a continuing lack of consensus in the field, and perhaps more generally, about which non-symptom-based outcomes should be included and how they should best be assessed. For example, the only HRQoL or functional impairment data identified in this systematic review for patients receiving OROS-MPH were from studies of LDX or ATX in which OROS-MPH was a reference treatment; and the only functional impairment data for patients receiving ATX were from a study of GXR in which ATX was a reference treatment. While some of the differences in choice of HRQoL or functional impairment outcome measure may reflect the changing demands of regulators, it is also the case that the design of a clinical trial involves balancing inclusion of additional outcomes of interest with the feasibility of conducting the study and analysing the results. For example, the use of parent-rated instruments in child/adolescent studies may reflect concerns that some young patients would be unable to complete questionnaires [31, 34], while the use of ADHD-specific HRQoL instruments in adults may reflect concerns that generic instruments would not be sensitive to the impairments characteristic of ADHD in adults. The use of these instruments may also reflect a lack of suitable alternatives: this systematic review focussed only on instruments that have been used in published randomized placebo-controlled studies in patients with ADHD. These complexities reflect the difficulty of sampling the key domains of impairment in individual patients from different perspectives, without merely sampling symptoms. It will be interesting to see whether recent work to develop a standardized nomenclature and toolkit to describe and code functional impairment in people with ADHD based on the World Health Organization’s International Classification of Functioning, Disability and Health will result in a more cohesive and structured approach to assessing HRQoL and functional impairment in clinical trials than at present [35, 36].

This systematic review was limited to randomized, double-blind, placebo-controlled studies published in peer-reviewed journals. The limitations of this approach should be considered when interpreting the results. One potential source of bias is that secondary outcomes or entire studies that do not favour the ADHD medication being tested may be less likely to be published than those that indicate beneficial effects. To limit this bias, we excluded post hoc analyses because they may be more likely than pre-specified analyses to report results favouring the investigational product. Another potential source of bias is that the review did not aim to assess the safety of ADHD medications. Side effects, adverse events or poor tolerability may themselves negatively affect present or future functioning and HRQoL, in addition to other potentially undesirable effects. Finally, the review included studies with different enrolled populations: although all studies enrolled patients with diagnosed ADHD, recruitment criteria differed among studies. For example, some studies included patients with specific psychiatric comorbidities or inadequate responses to previous ADHD medication. This could have affected the qualitative synthesis if study population factors are responsible for differential effects of medication across studies. Despite these limitations, and the limitations of the included studies, the results presented here comprise the most robust available evidence to date that ADHD medications not only provide effective relief of symptoms, but may also reduce functional impairments and improve HRQoL in children, adolescents and adults with ADHD. Furthermore, real-world evidence from registry studies suggests that this may be the case for patients in clinical practice [16, 56, 57], as well as for those enrolled in clinical trials.

Whether pharmacological therapy is appropriate, and if so which of the treatment options and which accompanying behavioural or psychological intervention is most suitable for each patient, are decisions for healthcare professionals, parents and patients. The evidence presented here should encourage everyone involved in a patient’s treatment to aim for reduced functional impairment and improved HRQoL, as well as relief of symptoms, for all patients with ADHD.