Issues in medication adherence for children and adolescents with attention-deficit hyperactivity disorder

Introduction

With the expansion of the diagnostic criteria in the DSM-IV (Diagnostic and Statistical Manual of Mental Disorders, 4th ed.) to include three subtypes of attention-deficit hyperactivity disorder (ADHD), the reported prevalence rate of ADHD has increased from 3% to 5% to about 12%.[1,2] Also there has been a 2.5-fold increase in stimulant medication prescriptions in adolescents in the USA between 1990 and 1995.[3]

While the short-term efficacy of stimulant medication with children has been adequately demonstrated in the literature,[4,5] and to a lesser degree with adolescents, [6] there remain difficulties in measuring long-term treatment efficacy with this population. One difficulty in measuring long-term treatment efficacy may be poor adherence to taking the medication. Poor adherence with methylphenidate was shown in one study [7] where 20% of families stopped taking the medication after 4 months and 44% after 10 months despite positive reports of effectiveness. The adolescents’ adherence in this study was worse than for the younger subjects. What may add to the poor adherence is the lack of the child’s or the adolescent’s awareness of any maladaptive behavior or awareness of change on the medication. In another study on how children feel about taking stimulants, 22 of 52 said they were definitely helpful, but 8 of these 22 disliked taking them and were eager to discontinue the medication.[8]

One review reports poor adherence in taking medication across all ages, and the authors cite one study that demonstrated less adherence among teenagers.[9] However, another recent study of adherence in 44 adolescents with ADHD symptoms but not with a formal diagnosis of ADHD showed adherence rates between 85% and 90%. These subjects were recruited from several centres and were offered an honorarium for participating.[10] Poor adherence in adolescents taking medication for psychiatric illness has also been reported.[11,12] For example, in a study of adolescents discharged from an adolescent psychiatry unit, a non-adherence rate of 33% was found using patient reports to assess adherence.[13]

Adherence to medication is traditionally measured by one of the following techniques:

• A pill count, where medication is dispensed and the patient returns all unused medications

• Serum, urine, or salivary levels of the medication

• A micro-chip in the pill bottle cap that measures the number of times the bottle has been opened

• Patient reports of how often the medication was missed [14]

While the biological collection methods offer an objective measure of adherence, there are major methodological difficulties. In addition to being costly and intrusive, blood serum levels do not necessarily correlate with clinical effectiveness in neuroleptics and most antidepressants. Measuring serum levels of stimulant medications are also problematic in that the breakdown products used to evaluate their presence or absence are very volatile. Additionally, patients may ensure the medication is administered only on the day a sample is taken.

Though patient recollection of adherence is considerably less intrusive, it has been considered to be less reliable [15] as patients tend to overestimate their adherence to taking medicines.[16] Some researchers, however, have found a positive correlation between direct measures (blood and urine samples) and patient reports.[17]

Given the finding that patient reports of adherence are correlated (but an overestimate) with the more intrusive direct measurements, the present study examined adherence rates to stimulant medication as part of telephone interviews with the adolescents and the parents of the children. Questions about satisfaction, symptom improvement, and adherence were asked.

Method

Subjects

The Attention-Deficit Assessment Clinic of British Columbia’s Children’s Hospital serves the province of British Columbia and reflects a mix of straightforward and tertiary consultations. Patients in the ADHD clinic are referred between the ages of 6 and 18 from family physicians, schools, pediatricians, and psychiatrists. The assessment procedures of the clinic include a full psychiatric assessment with a full five-axis DSM-IV diagnosis. The clinician collates information from family and individual interviews, school records, teacher reports, rating scales completed by parents and teachers, child observation, a developmental history, a family history, a full mental status exam, and psychological testing (when available), to arrive at a diagnosis.

Interviews

Letters were sent to 80 consecutive adolescent patients (aged 13 to 18 years) and 40 parents of consecutive admissions of children (aged 6 to 12 years) seen in the ADHD clinic in 1998. These were sent to patients seen for their initial consultation 6 to 12 months prior to the beginning of this study.

The letters requested that the recipients complete a 20-minute telephone interview with a research nurse about their contact with the ADHD clinic and the recommendations that were made. The nurse was independently hired and was not a member of the ADHD clinic team.

A semi-structured interview was devised to answer questions about symptom improvement (yes or no); whether the participants were diagnosed with ADHD; whether they were placed on medication and if so which one; whether they were still taking the medication as prescribed at the time of the interview; and on a scale of 1 to 7 (where 1 is poor and 7 is excellent) how they would rate the experience of being at the clinic. Two open-ended questions were used to identify what was helpful or not helpful at the clinic. The interviews took place within 4 weeks of mailing the letter.

Results

Chi-square (χ2) tests were used for comparing dichotomous variables and t tests for continuous variables. The Eta test of variance allows comparisons between continuous and dichotomous variables.

Of the 80 adolescents, 57 agreed to interviewed. (Eight families could not be traced, 12 adolescents did not want to do the interview, and three said they would but no suitable time could be arranged.) Of the 40 parents of the 12-year-olds and under, 35 agreed to be interviewed. (Three of the families could not be found and two refused an interview.) See the Table for a summary of the medication and adherence data.

While the adolescents in this sample were as likely as children to receive a diagnosis of ADHD χ2(1, n = 89) = 3.44, p = .064 they were significantly more likely to be placed on medications χ2(1, n = 23) = 6.78, p = .009. There was no difference in the use of methylphenidate χ2(1, n = 90) = 0.522, p = .470 or dextroamphetamine χ2(1, n = 90) = 1.39, p = .238 between the groups. Adolescents had been on medication for significantly more time than the child group t(1,44) = 6.89, p = .012. However, given the age differences between the two samples, this is not surprising.

Of those patients receiving medications, there was no difference in reported proportions of adherence χ2(1, n = 69) = 1.71, p = .191. There was also no difference between the two groups on patient satisfaction χ2(1, n = 98) = 1.59, p = .452 or ratings of symptom improvement t(1,90) = .554, p = .701.

In an effort to more clearly delineate factors associated with adherence, correlations with ratings of symptom improvement and patient satisfaction were computed with the total sample, as there were no differences on these measures between the groups. Patient satisfaction was not significantly correlated with adherence (Eta p = .299). Symptom improvement, on the other hand, was significantly correlated with adherence (p = .015) even when the length of time on medication was controlled for (p = .046).

Discussion

We found that by including a straightforward adherence question in a semi-structured interview, 19% of the total sample reported non-adherence to their stimulant medication regimen. This rate is within the range reported in the ADHD literature. [7,18,19] One other study cited found a much higher rate of adherence, but their sample was recruited by advertisements in several sites and they were paid to participate, so they may have been a more highly motivated and more compliant sample than ours.[10] Although there was no statistically significant difference in reported rates of adherence between the two groups, the child sample rate of adherence was higher at 90%. The high degree of adherence in the younger sample supports earlier research finding age to be a predictor in degree of adherence.[6]

When the two samples were collapsed, reports of adherence were significantly correlated with overall symptom improvement, but not with satisfaction with the clinic. Satisfaction with the clinic was a more generic question that also took into account the patient’s attitude toward the staff and waitlists, so it is not surprising that no relationship was found.

Adherence can be affected by patient beliefs about the disorder, treatment, and medication side effects and, in adolescents, concerns about the stigma of the illness or taking medication, as they are particularly concerned about what the peer group accepts.[20,21] In surveys on what is acceptable to treat ADHD, behavior modification is preferred to medication.[22]

Lask and colleagues have outlined parameters to improve adherence to medication, physiotherapy, and exercise for children with cystic fibrosis.[23] The parameters—which include showing empathy and enthusiasm and encouraging mutual decision making with the patients and their families—may be relevant to patients with ADHD. Physicians also need to ask about the perceived side effects and find out whether these are side effects of the medication or recurrence of the ADHD symptoms after the medication wears off. [24]

There are two limitations to this pilot study. First, the adherence data were not collected through the use of a standardized questionnaire. However, there were no questionnaires with adequate validity and reliability. Second, the sample size was relatively small and is specific in that it is from a single source. A larger sample and larger representation would allow for greater generalization of the findings. Despite these limitations, the admission by 19% of the total sample that they were non-adherent is clinically significant, particularly in light of the fact that patient reporting of non-adherence has been considered to be an underestimate.

Conclusions

If non-adherence is a problem in 19% of adolescents and children taking stimulants for ADHD, and if this represents an underestimate of actual rates, it is important to identify and address the factors that contribute to these rates. The findings from this study illustrate that an estimate of child/adolescent adherence can be collected by patient self-reports.

Clinicians may need to arrange for the questioning about adherence. Scheduling brief appointments could do this. To maintain adherence the clinicians would need to ensure that they were flexible about discussing alterations to the regimen, especially about taking long-acting medication, which many patients prefer. Clinicians must also be careful to ask about missed dosages if the medication does not seem to work in the first few weeks of treatment. When treatment has been given for several months and there seems to be diminishing efficacy, poor adherence should be considered. Some of the findings of this study may be germane to other chronic disorders that require medication and treatment for long periods.

Note

Information in this article was presented in modified form to Canadian Academy of Child Psychiatry, Halifax, September 1998.

Medication and adherence data.

* Data available on 21 children and 48 adolescents
† Based on a 7-point Lickert scale

Stuart Fine, MB, FRCPC and David Worling, PhD

Dr Fine is a professor with the Department of Psychiatry, University of British Columbia, and head of the ADHD Clinic at BC’s Children’s Hospital. Dr Worling is a psychologist with British Columbia’s Children’s & Women’s Health Centre, research associate psychologist with the Division of Child Psychiatry, and an adjunct professor with the Department of Psychology, University of British Columbia.