Informed Consent and Stimulant Medication: Adolescents' and Parents' Ability to Understand Information About Benefits and Risks of Stimulant Medication for the Treatment of Attention-Deficit/Hyperactivity Disorder

Abstract

Background:

This study of informed consent examines understanding of information needed to consent to stimulant treatment for attention-deficit/hyperactivity disorder (ADHD). The understanding of adolescents with ADHD, their parents, control adolescents, and their parents is compared.

Method:

Fifty-eight ADHD and 64 control adolescents between the ages of 12 and 16 and their parents were studied. Baseline understanding of information was determined. Subjects received information relevant to informed consent for stimulation medication and afterward were evaluated on their recall understanding and their final understanding.

Results:

Knowledge was increased after the information session for all subjects. There was no significant difference between unadjusted baseline, recall, and final knowledge of control adolescents and parents. Although unadjusted baseline, recall, and final knowledge of ADHD adolescents is significantly less than that of parents, 78% of ADHD adolescents had final understanding scores within 2 standard deviations of parents' scores. After controlling for baseline understanding and cognitive variables, there was no significant difference between understanding of ADHD adolescents and ADHD parents, whereas control adolescents understanding scores were higher than that of their parents. Understanding was highly associated with mathematics achievement in all groups.

Conclusion:

The majority of adolescents with ADHD, both with and without a history of stimulant medication treatment, have understanding that is similar to their parents and their inclusion in the informed consent process should be encouraged. Extra care should be afforded to those adolescents with low numeracy or literacy to ensure their understanding.

Background

I nformed consent is obtained before embarking on a course of treatment. Informed consent to treatment occurs when the individual has the capacity to consent, has received a disclosure of the relevant treatment information, and provides consent to treatment voluntarily (Grisso and Appelbaum 1998). The relevant information consists of the nature of the treatment, the benefits and risks of the treatment, the course of the illness without treatment, and the alternative risks.

The terms capacity (legal term) and competence (ordinary language, bioethics) refer to the ability to understand and the ability to appreciate the information disclosed related to the proposed intervention. Capacity and/or competence are specific to the situation. There are several controversial issues related to capacity, including determination of capacity, the age of capacity, and the effect of psychiatric disorders on capacity (Schachter et al. 2005). First, although there are agreed upon theoretical definitions of capacity, there are no agreed upon gold standards or operational criteria by which to determine or assess capacity (Glass 1997). Second, there is debate about whether adolescents have the capacity to consent (Scott et al. 1995). Although there is a theoretical consensus in the literature that adolescents 14 years of age or older have understanding of information similar to adults, there is limited empirical data (Schachter et al. 2005). There is controversy regarding whether adolescents are sufficiently mature to understand and appreciate information and evaluate it with respect to stable personal long-term values and goals (Scott et al. 1995). There is debate about whether adolescents are sufficiently independent and mature to make autonomous medical decisions. It has been argued that society may wish to protect adolescents from decisions they may later regret. Finally, there is limited empirical data regarding the impact of psychiatric disorders on capacity among adolescents.

In the absence of a priori thresholds for capacity determination, most studies of capacity in adolescents have compared the abilities of adolescents to that of parents or adults. This review article focuses on the understanding component of capacity. Several empirical studies have compared, using hypothetical vignettes, adolescents' and parents' understanding of information. Studies among community subjects and among adolescents with mental health problems have shown that using hypothetical medical vignettes, adolescents' understanding resembled adults' understanding (Weithorn and Campbell 1982; Mulvey and Peeples 1996). Studies of hospitalized adolescents found that most subjects 12 years of age or older were competent and competence correlated with age, reading level, and intelligence (Billick et al. 2001). A series of studies examined knowledge about psychiatric medication among children and adolescent psychiatric inpatients with a variety of diagnoses (Bastiaens 1992a, 1992b, 1995). There was no difference between parents' and adolescents' knowledge after information sessions (Bastiaens 1995). Knowledge was greater after more intensive information interventions (Bastiaens 1992b). Children's knowledge correlated with verbal IQ but not with age (Bastiaens 1992a, 1992b). However, children did not learn about the serious side effects or treatment alternatives (Bastiaens 1992a, 1992b). In these studies, the effect of knowledge before the intervention was not examined.

However, in many of these studies the understanding component of the informed consent process is assessed immediately after disclosure. These studies may be assessing “recall” of information rather than “true understanding.” In clinical practice, patients receive information typically from a variety of sources, including verbally from the physician or healthcare team, and then review information and/or obtain additional information before making a healthcare decision. This study is unique in examining understanding of information when patients have the information. That is we try to assess true understanding rather than recall alone, thereby reflecting the clinical situation more accurately. Further, we wanted to assess understanding among adolescents with a chronic psychiatric disorder, such as attention-deficit/hyperactivity disorder (ADHD) that required ongoing treatment.

ADHD is a common chronic psychiatric disorder affecting children and adolescents (American Academy of Child and Adolescent Psychiatry 1997) and typically requires chronic treatment. The prevalence among youth with ADHD of coexisting reading, mathematics, and combined reading and mathematics disorders is 16%, 8%, and 10%, respectively (Capano et al. 2008), and these learning disorders may independently affect learning and/or understanding of medical information. Medication treatment with stimulant medication is the main treatment option. Despite the increased recognition of and treatment of ADHD among adolescents, few studies have examined adolescents' understanding of treatment options and their ability to participate in the informed consent process for stimulant treatment for ADHD.

In Canada and the United States, the age at which the law recognizes a child to be capable of making valid healthcare choices varies across the states, provinces, and territories (Boonstra and Nash 2000; Rozovsky 2003). Common law decision in Canada and Health Care Consent Act in Ontario state that there is no age below which individuals cannot give consent (Health Care Consent Act, 1996 ; Hopp v Lepp, 1980; Reibl v Hughes, 1980). In the United States, the age at which the statutes permit adolescents to make healthcare decision differs, depending on the jurisdiction, the adolescents' age, the nature of the healthcare decision and whether the adolescent is determined to be mature or emancipated (Boonstra and Nash 2000; Rozovsky 2003). In England, minors are determined to be competent based on understanding rather than age (Larcher and Hutchinson 2010). In most jurisdictions, laws relating to consent are found in several statutes, and the relationships between them can be complex (Rozovsky 2003).

Determining whether adolescents have the capacity to make these decisions is particularly important in jurisdictions that permit adolescents to give consent. However, even when adolescents are not capable to consent, or reside in jurisdictions where they are not legally permitted to give consent, it has been recommended that they be involved in the consent process to the extent possible and assent to treatment. From a practical point of view, in most situations it is very difficult to work with adolescents without their agreement or to expect them to adhere to a treatment regime if they do not assent to the process.

This project examines one aspect of adolescents' ability to give informed consent, specifically adolescents' understanding of information needed to consent to stimulant treatment for ADHD. Adolescents' ability to appreciate information, another component of capacity, is not explored in this article. In the absence of a gold standard for determining a threshold for adequate understanding, adolescents' understanding is compared with that of their parents. Two methods were used to compare adolescents' understanding with that of parents. First, overall understanding was compared, and second, to estimate the percentage of adolescents whose understanding scores was significantly lower than their parents, we dichotomized adolescents' understanding scores using a cutoff of 2 standard deviations (SD) below the mean parental understanding scores. To determine the effect of ADHD on understanding, ADHD adolescents' understanding is compared with that of control (non-ADHD) youth and their parents. To get a better understanding of the potential correlates of capacity, the cognitive and psychosocial correlates of understanding were also examined.

The objective of this study was to evaluate understanding of information related to the informed consent process for treatment of ADHD among adolescents with ADHD and those without ADHD. It is hypothesized that adolescents' understanding would be less than their parents after a structured information disclosure session. Adolescents with ADHD would have less understanding than control adolescents after a structured information disclosure session.

Methods

Sample

Adolescents 12–16 with ADHD (ADHD-A) and without ADHD (Control-A) and their parents (ADHD-P and Control-P) were recruited in Toronto from physician offices, ADHD Web sites, schools, community centers, the Centre for Addiction and Mental Health, and advertisements in newspapers. The clinical diagnosis of ADHD was confirmed using the parent module of the NIMH Diagnostic Interview Schedule for Children IV (DISC-IV) (Shaffer et al. 1996, 2000). Control-A did not meet criteria for ADHD on the DISC-IV (Shaffer et al. 1996, 2000) and scored in the nonclinical range on both the Child Behavior Checklist (CBCL) (Achenbach 1991a), and Youth Self-Report (YSR) (Achenbach 1991b).

A total of 153 parent/adolescent pairs were enrolled in the study. In the final analysis, 20 ADHD parent/adolescent dyads were eliminated due to intermediate or negative DISC-IV ADHD scores and 11 control pairs were eliminated due to positive DISC-IV ADHD scores or scoring in the clinical range on the CBCL or YSR.

Procedure

After a joint informed consent procedure for research participation, parents and adolescents were separated. Each parent and adolescent had understanding of the information needed to consent to stimulant treatment for ADHD evaluated by completing a Measure of Understanding (MUA). A structured disclosure process was used to inform subjects about the potential benefits and risks of stimulant treatment for ADHD. A research assistant read subjects an information form; subjects could also read the form at the same time. Subjects had the opportunity to ask questions regarding the information. The research assistants had a minimum of an honors bachelor's degree. Usually, adolescents had the information form disclosed by a research assistant with a master's degree. The MUA was administered to each subject three times: at baseline, immediately after receiving information (recall), and when they were given the information sheet (final knowledge).

Both parents and adolescents completed the Wide Range Achievement Test Revision 3 (WRAT 3) (Wilkinson 1993). Adolescents were administered 4 subtests of the Wechsler Intelligence Scale for Children-IV (WISC-IV) (Wechsler 2003), and parents provided demographic information and completed the Conners' Parent Rating Scale–Revised (L) (Conners 1997). Socioeconomic status (SES) was coded using the Blishen Index (Blishen et al. 1987), a Canadian measure of socioeconomic index. The procedure was administered by research assistants with either a minimum of an honors bachelor's degree or the author (D.S.). Research assistants who administered the protocol to adolescents had graduate training in administration of psychoeducational testing, including tests of intelligence. Since several research assistants initially participated in this study, scoring of answers to the structured interview was reviewed by a research assistant and discrepancies reviewed and rectified by the principal investigator.

Measures and instruments

Information form

An information form was developed in conjunction with physicians, parents, and adolescents that addressed the information necessary for informed consent for stimulant treatment of ADHD. The disclosure form addressed ADHD, the illness and complications, the benefits and risks of stimulant medication, and the benefits and risks of an alternative treatment such as behavior therapy for treating ADHD. Risks and benefits were presented using both quantitative and qualitative qualifiers. The disclosure was kept at a grade 6 readability level to facilitate comprehension. The information form was pretested before use in this study.

Measure of understanding

The MUA examines understanding of ADHD, the benefits and risks of stimulant medication, as well as the alternative treatment of behavior therapy or counseling. The MUA has a combination of 8 structured open-ended interview (SI) questions and 42 multiple choice questions (MC). Three additional questions were excluded. The MUA was developed by the investigators for this study. The structured open-ended questions were modified from the MacArthur Competence Assessment Understanding of Treatment Disclosures (Appelbaum and Grisso 1992). Participants received a score of 2, 1, or 0 on the structured interview questions. The multiple-choice questions were scored 2 (definitely true or false), 1 (probably true or false), or 0 (don't know or incorrect). The MUA questions could be classified into 5 broad categories relating to different facets of ADHD (Disorder, 9 items; Treatment, 5 items; Medical Benefits, 12 items; Side Effects, 17 items; Alternate Treatment, 7 items). A total composite score of understanding was derived using the SI and MC questions of the MUA. The maximum attainable understanding score is 100. Unanswered questions were assigned a value of zero; this was only seen in 21 participants. All questions could be answered based on the disclosure form. The MUA does not examine other aspects of abilities related to informed consent such as insight and appreciation of information.

Face validity of the disclosure form and questionnaire was determined in the following manner. The disclosure and questionnaire were initially reviewed and pretested on 15 participants (5 physicians, 5 parents, 3 adolescents, and 2 experts in bioethics) and subsequently revised. The revised MUA was then reviewed by and administered to 30 physicians (15 child psychiatrists and 15 pediatricians), 18 adolescents with ADHD, and 18 parents of children with ADHD, and further revised for the present study. Physician Total recall MUA score from this earlier version of the disclosure and MUA was 77.9 (SD = 9.7). In this phase of the study, baseline MUA and final MUA was not evaluated. Adolescent and parent participants from pretesting are excluded from the analysis in this report.

Intelligence

Adolescents completed 4 subtests (Vocabulary, Block Design, Similarities, Matrix Reasoning) of the WISC-IV (Wechsler 2003). A composite IQ estimate was derived (Sattler 2001) using the WISC subtest scores to calculate the deviation quotient using the following formula: Deviation quotient (IQ) = (15/Sc) (Xc − Mc) + 100, where Sc (standard deviation of composite score) = Ss(sqrt (n + 2∑r _jk), n = 4 subtests, Ss (subset standard deviation) = 3, ∑r_jk (sum of the correlation between the component subtests), Xc (composite score) = sum of subtest scaled scores in the short form, and Mc (normative mean) = 10 × number of component subtests = 10 × 4 = 40. The intercorrelations of the subtests and sums of scaled scores for composites were obtained from the WISC-IV manual.

Wide Range Achievement Test Revision 3

The WRAT 3 evaluates reading, spelling, and mathematic skills and has known reliability and validity (Wilkinson 1993). The WRAT 3 provides raw, age-standardized, and grade-equivalent scores for reading, spelling, and mathematics skills.

Conners' Parent Rating Scale

Parents completed the Conners' Parent Rating Scale–Revised (L), which assesses ADHD and other behavioral problems in adolescents (Conners 1997).

Child Behavior Checklist

Parents completed the CBCL, which assesses behavioral problems and social competencies of their children (Achenbach 1991a).

Youth Self-Report

Adolescents completed the YSR, which provides self-ratings on a range of behavioral items (Achenbach 1991b).

NIMH Diagnostic Interview Schedule for Children Version IV

Parents completed the disruptive behavior DISC-IV module, which was used to confirm clinical diagnosis of ADHD (Shaffer et al. 1996, 2000).

Ethics

The hospital's Research Ethics Board approved this project. Parental consent was obtained. Adolescents consented or assented to participate. A monetary honorarium was offered to participants.

Data analysis

Data are analyzed using the statistical packages on SAS version 8.0.

Descriptive information on the demographics and understanding variables was examined in the four groups independently. Parametric (t-test) and nonparametric (chi-square) tests were used to examine the difference in understanding between ADHD and control groups. Paired t-tests were used in examining the difference in understanding between the matched parent/adolescent dyads in ADHD and control groups. All tests of statistical significance were carried out at a level of 5% significance using two-tailed test procedure. Internal consistency of the items in the MUA was analyzed using Cronbach's alpha at each time period and among the different subject groups. Bivariate analysis (regressions and Pearson's correlation coefficient) was conducted to examine the relationship between understanding scores and potential predictors of understanding.

Modeling details

A hierarchical linear model was used to compare understanding across the four groups because of the clustered, hierarchical structure of the data. This analysis accounts for possible clustering of measurements from the same unit; in this case, the parent and adolescent from the same family. The data can be conceptualized as a two-level hierarchy, with diagnosis (ADHD or control) denoting the higher level (level 2) and subjects (parent or adolescent) denoting the lower level (level 1). A parent and adolescent in the same family can be thought of as belonging to a cluster (i.e., the family). The follow-up assessment of understanding is the outcome of interest, and baseline understanding is controlled by including it as a covariate in our model.

The effect of a number of predictors on final understanding of informed consent is estimated by fitting a linear model. The predictors of primary interest are diagnosis (ADHD or control), measured at level 2, and subject (parent or adolescent), measured at level 1. The diagnosis × subject interaction term is included. Potential predictor variables included as covariates in the model include income, SES, WRAT raw reading, spelling, and math scores and baseline understanding score. Income, SES, and diagnosis are characteristics of the family and are measured at level 2. Reading, spelling, math, and baseline understanding scores are characteristics of the parent or adolescent, and are included at level 1. Age of adolescent or parent was not considered in this model because the ranges were so different.

To examine predictors of understanding within the adolescent subjects, a general linear model was modeled using final understanding score as the outcome variable. Potential predictor variables included diagnosis (ADHD/Control), gender (male/female), age (12, 13, 14, 15, or 16), IQ, baseline understanding score, raw reading score, raw spelling score, and raw math score. Diagnosis and gender were modeled as categorical variables. Due to the limited age range of adolescents in this study (12–16), age was modeled as a categorical variable with five levels, with each age constituting a distinct level. IQ, baseline understanding, raw reading, spelling, and math scores were modeled as continuous variables. Each of the two-way interactions consisting of diagnosis was also tested for significance.

Results

Characteristics of subjects

Demographic information for the 58 ADHD and 64 control adolescent/parent dyads is detailed in Table 1. Although there were more males among the ADHD-A than Control-A sample (74% and 59%, respectively); this did not reach conventional levels of statistical significance. The age distribution among the adolescents was similar with the exception of fewer 16 year olds in the ADHD (10%) group compared with controls (20%). Sixty-nine (n = 40) percent of ADHD-A subjects were on stimulant medication. Based on the DISC-IV, among the ADHD-A 17% had conduct disorder and 62% had oppositional defiant disorder and among control A, none had conduct disorder and 8% had oppositional defiant disorder.

Table 1.

Demographic Characteristics of Attention-Deficit/Hyperactivity Disorder and Control Adolescents/Parents

	ADHD (n = 58)	Controls (n = 64)
Adolescent gender (%, n)
Male	74 (43)	59 (38)
Adolescent age (mean, SD)	13.7 (1.3)	14.0 (1.4)
CBCL score	64.0 (8.0)	43.9 (9.1)
YSR	59.8 (10.9)	50.1 (8.7)
C-DISC CD^a (%, n)
Positive	17 (9)	–
Intermediate	23 (12)	5 (3)
Negative	60 (31)	95 (61)
C-DISC ODD^b (%, n)
Positive	62 (33)	8 (5)
Intermediate	23 (12)	17 (11)
Negative	15 (8)	75 (48)
Conners
Inattentive (DSM-IV)–T Score^c	75.6 (11.5)	45.7 (6.3)
Hyperactive-Impulsive (DSM-IV)–T Score^c	76.6 (17.2)	46.1 (5.0)
Total DSM-IV–T Score^c	76.4 (11.7)	45.8 (4.8)
WISC Converted IQ^d (mean, SD)	102.6 (15.6)	109.3 (12.2)
Adolescent WRAT standard scores (mean, SD)
Reading^e	104.2 (12.0)	108.7 (10.0)
Spelling^f	102.2 (13.8)	110.3 (10.4)
Math^g	89.7 (14.4)	99.3 (14.2)
Adolescents on medication (%, n)	69 (40)	–
Social Class (Blishen Index) (mean, SD)	48.4 (12.3) (n = 54)	51.2 (14.7) (n = 55)
Parent gender (%, n)
Female	95 (55)	84 (54)
Parent age (mean, SD)	45.6 (6.0)	46.5 (4.9)
Parent high school completion (%, n)
YES	95 (55)	98 (63)
Parent number of years of postsecondary (mean, SD)	3.9 (2.4)	4.1 (2.4)
Parent WRAT standard scores (mean, SD)
Reading	101.5 (9.7)	102.5 (10.1)
Spelling	102.8 (9.8)	104.0 (12.0)
Math	94.3 (10.7)	98.4 (14.8)
Age difference between parents and children	31.9 (5.9)	32.5 (4.8)

6 missing.

5 missing.

p < 0.001.

t = −2.62, p < 0.05.

t = −2.24, p < 0.05.

t = −3.66, p < 0.001.

t = −3.68, p ≤ 0.001.

ADHD = attention-deficit/hyperactivity disorder; CBCL = Child Behavior Checklist; DISC = Diagnostic Interview Schedule for Children; SD =standard deviation; WISC = Wechsler Intelligence Scale for Children; WRAT = Wide Range Achievement Test; YSR = Youth Self-Report; CD = conduct disorder; ODD = oppositional defiant disorder.

ADHD-A had significantly lower estimated IQ, standardized reading, spelling, and math scores than Control-A. There were no statistically significant differences between the characteristics of parents of ADHD-A and control-A, including their standardized WRAT scores. Parental participants were primarily mothers (ADHD-P, 95%; Control-P, 84%) with a mean age of 46. Across all groups, raw mathematics scores were significantly correlated with raw spelling (p < 0.05) and raw reading scores (p < 0.05). Among adolescents the correlations between the raw reading, spelling and math scores, and IQ were examined: The highest correlation is observed between math scores and IQ (r = 0.64); the correlation between reading score and IQ (r = 0.41) and spelling scores and IQ are both a little bit lower (r = 0.41).

Measure of understanding

Internal consistency of the questionnaire, determined using Cronbach's alpha, revealed, in general, high internal consistency (α>0.8) seen in the overall, SI and MC questions, with internal consistency increasing from baseline to final understanding (Table 2).

Table 2.

Internal Consistency of Measure of Understanding

	Internal Consistency (Cronbach α)
	All Subjects	Parents	Adolescents
Total questionnaire (Baseline)	0.93	0.94	0.88
Total questionnaire (Final)	0.95	0.95	0.95
Structured interview (Baseline)	0.79	0.79	0.67
Structured interview (Final)	0.87	0.93	0.79
Multiple choice (Baseline)	0.92	0.94	0.87
Multiple choice (Final)	0.95	0.95	0.94

Understanding scores among the 4 groups at the 3 different time periods are presented in Table 3. Overall participants' scores increased from baseline after information disclosure. The ADHD-P had the highest baseline and final understanding scores. The ADHD-A had intermediate baseline understanding and lowest final understanding. Paired t-tests show significant differences (p < 0.0001) between the mean baseline and final understanding scores within each of the four groups. Over three administrations of the questionnaire, understanding increases in all groups; the least increase in understanding is seen in the ADHD-P, who have the higher baseline understanding score. Among the ADHD parent/adolescent dyads, there is a significant difference between parents and adolescents in scores for baseline, recall, and final understanding; however, this difference was not present in the control parent/adolescent dyads.

Table 3.

Total Understanding Among Attention-Deficit/Hyperactivity Disorder and Control Adolescents and Parents

Understanding mean (SD)	ADHD parents	ADHD adolescents	Difference (P−A)	Control parents	Control adolescents	Difference (P−A)
Total measure of understanding
Baseline	68.5 (12.9)	50.1 (12.6)	t = 8.68, df = 57, p < 0.0001	44.9 (13.8)	41.4 (12.2)	NS^a
Recall	84.3 (9.1)	67.2 (15.0)	t = 7.60, df = 57, p < 0.0001	75.7 (15.5)	76.3 (10.5)	NS^a
Final	86.7 (13.5)	74.3 (17.4)	t = 4.49, df = 57, p < 0.0001	83.3 (13.9)	87.8 (9.5)	t = −2.05, df = 63, p < 0.05
Difference (Final- Baseline)	t = 8.67, df = 57, p < 0.0001	t = 11.5, df = 57, p < 0.0001		t = 19.0, df = 63, p < 0.0001	t = 26.0, df = 63, p < 0.0001
Multiple choice
Baseline	55.9 (11.7)	41.7 (11.1)	t = 7.5, df = 57, p < 0.0001	35.6 (11.4)	35.1 (10.5)	NS^a
Final	71.9 (11.5)	60.6 (15.2)	t = 4.9, df = 57, p < 0.0001	68.2 (12.9)	72.3 (9.2)	t = −2.05, df = 63, p < 0.05
Structured interview
Baseline	12.5 (2.2)	8.5 (2.9)	t = 9.7, df = 57, p < 0.0001	9.3 (3.6)	6.4 (3.0)	t = 6.2, df = 63, p < 0.0001
Final	14.8 (3.6)	13.6 (3.0)	NS	15.2 (1.5)	15.4 (1.0)	NS^a

NS = Not significant at p < 0.05. ADHD = attention/deficit/hyperactivity disorder; P−A = parents−adolescents.

Among the ADHD-A, 78% had a final understanding score within 2 SD of the mean ADHD-P final scores. In the Control-A group, 100% of the adolescents had an understanding score within the mean Control-P scores. Combining together the ADHD and control groups, ∼89% of the adolescents are within 2 SD of the parental mean understanding score.

The scores on the multiple choice questions among the 4 groups follow a similar pattern to the overall understanding scores (Table 3), with ADHD-A subscale scores significantly lower than their parents. However, this is not seen for control dyads. In contrast, the scores on the final SI questions show very little variation among the four groups.

Correlation with understanding scores (parents/adolescents)

There is no significant correlation (p < 0.05), using the Pearson correlation coefficient between parents' and adolescents' understanding scores, in either the ADHD or control groups at all three points of the MUA administration.

Hierarchical linear modeling results

A model is fitted that included all the predictors of interest: Diagnosis, subject, the diagnosis × subject two-way interaction, income, social class, baseline understanding score, and the raw WRAT reading, spelling, and math scores. Raw scores rather than standardized scores were used in this model as scores were comparable across groups. In this model, diagnosis (p = 0.002), subject (p = 0.03), baseline understanding score (p < 0.0001), the diagnosis × subject interaction (p = 0.0008), and raw math score (p = 0.0002) were significant. Income (p = 0.24), social class (p = 0.32), reading score (p = 0.53), and spelling score (p = 0.12) were not significant. Given the nonsignificance of income and social class, the model was refitted with these predictors removed. The resulting model was selected as our final model.

This final model used 121 out of the 122 families, and 238 out of a possible 244 subjects. Five subjects (including both members of one ADHD family) were excluded because of missing WRAT scores. One other subject was excluded because of an unknown understanding score. The significance of the predictors included in the final model is reported in Table 4.

Table 4.

Mixed Model for Final Understanding

Predictor variables	F	df	Δ Follow-up understanding Score ^a (SE)	95% CI for Δ follow-up understanding score	p
Baseline understanding score	26.3	1, 111	2.90 (0.57)	(1.78, 4.03)	<0.0001
Diagnosis	11.5	1, 111			<0.001
Subject	4.0	1, 111			0.05
Diagnosis × subject	17.4	1, 111			<0.0001
Reading score	0.36	1, 111	1.49 (2.49)	(−3.44, 6.43)	0.55
Spelling score	0.58	1, 111	1.70 (2.24)	(−2.73, 6.13)	0.45
Math score	27.7	1, 111	7.70 (1.46)	(4.80, 10.6)	<0.0001

Change in understanding score for every 10-point increase in predictor variable. Δ = change.

The results of these analyses suggest that baseline understanding score and raw mathematics scores are independently associated with final understanding score; a 10-point increase in baseline understanding score is associated with a 2.90 increase in mean follow-up understanding score, and this increase is statistically significant (95% CI: [1.78, 4.03]; p < 0.0001). Raw math score is associated with follow-up understanding score; a 10-point increase in raw math score is associated with a 7.70 increase in mean follow-up understanding score, and this increase is statistically significant (95% CI: [4.80, 10.6]; p < 0.0001). When mathematics score is included in the predictors, the reading and spelling scores are not significantly associated with final understanding.

However, because mathematics, reading and spelling are highly correlated, we wondered whether the inclusion of mathematics achievement in the regression equation would obscure the potential independent contributions of reading or spelling achievement. Hence, we repeated the regression equations including only 1 academic score at a time. When mathematics and spelling are excluded from the model, reading is a predictor of understanding(F_{1, 114} = 14.2; p = 0.0003); a 10-point increase in reading score is associated with a 6.15 increase in mean final understanding score, and this increase is statistically significant (95% CI: [2.91, 9.39]; p = 0.0003). When spelling alone is a predictor (F_{1, 114} = 19.0; p < 0.0001); a 10-point increase in spelling score is associated with a 6.07 increase in mean final understanding score (95% CI: [3.30, 8.83]; p < 0.0001).

The diagnosis × subject interaction term is significant; therefore, both diagnosis and subject are associated with final understanding scores. Least-squares mean final understanding scores (i.e., adjusted for covariates) are shown in Table 5 and are examined across levels of the interaction (Table 6). The tables explain the significance of the two-way diagnosis × subject interaction (F_{1, 111} = 17.4; p < 0.0001). Diagnosis and subject are both significantly associated with final understanding, but the effect of subject (adolescent versus parent) differs depending on the diagnosis, whether the dyad is ADHD or control. In the ADHD group, adolescents' understanding is not significantly different from parents' understanding after controlling for baseline understanding, and reading, spelling, and mathematics scores (Table 6). In the control group, adolescents' mean understanding score is greater than that of parents. Pair-wise comparisons show that control adolescents have significantly higher scores than their parents, ADHD adolescents, and ADHD parents after controlling for cognitive variables and baseline understanding. The pair-wise comparisons that do not involve the control adolescents are not statistically significant. All p values reported in Table 6 are Bonferroni adjusted.

Table 5.

Least-Squares Mean Understanding Score (SE) at Follow-Up, for Adolescents and Parents in Each Diagnosis Group

	ADHD	Control
Adolescents	78.9 (1.6)	91.1 (1.5)
Parents	81.6 (1.8)	81.7 (1.5)

Table 6.

Pair-Wise Comparisons of Understanding Least-Squares Means

A	B	Diff. (A–B)	SE (Diff.)	95% CI (Diff.)	df	t	adj. p
ADHD Adolescents	Control Adolescents	−12.1	2.2	(−16.4, −7.8)	111	−5.61	<0.0001
ADHD Adolescents	ADHD Parents	−2.7	2.4	(−7.49, 2.12)	111	−1.11	1.000
ADHD Adolescents	Control Parents	−2.7	2.4	(−7.52, 2.06)	111	−1.13	1.000
Control Adolescents	ADHD Parents	9.5	2.5	(4.49, 14.4)	111	3.77	0.0016
Control Adolescents	Control Parents	9.4	2.0	(5.44, 13.4)	111	4.70	<0.0001
ADHD Parents	Control Parents	−0.05	2.4	(−4.89, 4.79)	111	−0.02	1.000

Diff = difference; adj = adjusted.

Predictors of understanding among adolescents

The model for predicting final understanding among adolescents is shown in Table 7. After adjusting for other variables in the model diagnosis, baseline understanding and mathematics were significantly associated with follow-up understanding score. Age, IQ, reading, and spelling were not significantly associated with follow-up understanding after controlling for other variables in the model. None of the 2-way interactions of diagnosis and other variables was significant. Mean follow-up understanding score is 11.4 points higher in the control group than it is in the ADHD group (95% CI: [6.27, 16.6]). A 10-point increase in baseline understanding score is associated with a 2.79-point increase in mean follow-up understanding score, and this increase is statistically significant (95% CI: [0.71, 4.86]). A 10-point increase in math score is associated with an increase of 5.92 points in mean follow-up understanding score, and this increase is statistically significant (95% CI: [0.20, 11.6]).

Table 7.

Adolescent Predictors of Understanding

Predictor of understanding	F	df	p
Diagnosis	19.3	1, 107	<0.0001
Gender	1.05	1, 107	0.31
Age	0.58	4, 107	0.68
IQ	2.83	1, 107	0.095
Baseline understanding score	7.09	1, 107	<0.01
Reading score	0.01	1, 107	0.99
Spelling score	0.10	1, 107	0.75
Maths score	4.21	1, 107	<0.05

To examine the impact of the inclusion of IQ in the model, we also fit a model that excluded IQ. In this model, the significance of the other variables remained unchanged. Math score became more strongly significant (F = 13.6, df = 1, 109; p = 0.0004) in the absence of IQ. In this model, a 10-point increase in math score is associated with an increase of 8.77 points in mean follow-up understanding score, and this increase is statistically significant (95% CI: [4.06, 13.5]).

Similarly, in the absence of math score, IQ becomes more strongly significant (F = 11.8, df = 1, 109; p = 0.0008), whereas the significance of the other variables remains unchanged. A 10-point increase in IQ is associated with an increase of 3.39 points in mean follow-up understanding score, and this increase is statistically significant (95% CI: [1.44, 5.35]).

Discussion

A careful structured information session where the benefits and risks of stimulant treatment for ADHD are presented at a grade 6 reading level and where attention is given to providing quantitative information using both qualitative and quantitative qualifiers can increase adolescents' and parents' knowledge of the information needed to consent to stimulant medication. This is true for ADHD adolescents, their parents, control adolescents, and their parents.

There is no significant difference between unadjusted baseline, recall, and final knowledge of control adolescents and control parents, suggesting that control adolescents' understanding is similar to their parents'. There was also surprisingly no correlation between dyads of parents' and adolescents' understanding at any point in time. We had hypothesized that their understanding would be correlated at all time points but especially at baseline. The reason for this is unclear and may reflect the ADHD adolescents' ability to understand or the amount of information they have received or sought out, and/or their retention of information. In contrast, unadjusted baseline, recall, and final knowledge of ADHD adolescents is significantly less than their parents. Approximately 80% of ADHD adolescents attained final knowledge scores within 2 SD of parental scores, suggesting that while their scores on average are lower than parents, the majority have knowledge scores in the range of parental scores. It is difficult, however, to evaluate the clinical significance of this as there is no absolute level below which an individual is deemed to have insufficient information on which to make a decision. In addition, using more commonly accepted measures of understanding, such as the structured interview, where there are fewer questions, there were no significant differences between the ADHD adolescents' and their parents' final scores on understanding scores. Most of the differences that emerged were on responses to the multiple choice questions, which reflect specific knowledge.

To further delineate the differences in knowledge that ADHD adolescents versus their parents retained, we statistically controlled for cognitive variables such as reading, mathematics, and spelling achievement and baseline understanding and found that the differences across the ADHD parent and ADHD adolescent groups were no longer significant. This suggests that the differences observed between ADHD parents and adolescents are related to their baseline understanding and cognitive variables rather than to whether they are an adolescent or parent member of a dyad.

In contrast, after controlling for cognitive variables and baseline understanding, control adolescents had higher understanding scores than parents. Across all groups reading, spelling, and mathematics independently predicted final knowledge across the groups, with the relationship strongest for mathematics.

The results of the linear regression for adolescents reaffirm the importance of mathematics and/or intelligence, and baseline understanding in the prediction of final understanding in the presence of other variables. The relationship between literacy, reading, and understanding of information needed for consent has been documented in several studies (Billick et al. 1998, 2001) as has the relationship between understanding of information and measures of intelligence (Bastiaens 1992a, 1992b; Casimir and Billick 1994).

However, the results of our study suggest that mathematics skills appear to influence understanding of information for consent to treatment more strongly than reading or spelling skills. This is particularly important for youth with ADHD, since the prevalence of mathematics disabilities among youth with ADHD is estimated to be 18.1%, which is approximately three times the rate in the nonclinical population (Capano et al. 2008). It may be that the mathematical concepts in the information text were too complicated for those with weaker mathematic skills. Although we tried to present the risk concepts as simply as possible using both qualitative and quantitative qualifiers, it may be that additional techniques are required to increase understanding among those with weaker mathematics skills. Alternative or pictorial representations of these concepts may be superior for individuals with weak mathematics skills (Paling 2003). It may be that number concepts confuse those with weaker mathematics skills and prevent them from understanding the other concepts. In Ontario, simple probability concepts are introduced into the curriculum in grade 1, but it is not until grade 6 that the student is expected to understand probability and its' representation (Ministry of Education (Ontario) 2005). Thus, the educational system and the timing of the teaching of these concepts may place another lower bound on the ability of youth to participate in the consent process and may suggest that even pictorial representations may be difficult for those with weaker mathematics skills. Another explanation for the association may reflect the test measures. Mathematics learning disabilities and weaknesses are hypothesized to be related to disruptive retrieval of information, executive dysfunctions, including problems with attention and inhibition and deficits in number sense (Geary 2004, 2010). Therefore, the association of understanding with mathematics skills may be mediated through attention. Although attention control is associated with reading, mathematics, and spelling achievement in subjects with ADHD, sustained attention predicts only mathematics achievement (Preston et al. 2009). This suggests that there may be unique attention factors associated with mathematics achievement that may be related to the underlying cognitive processes associated with understanding consent information as examined in this study. To conclude, this study confirms the important role of numeracy in understanding treatment information, informed consent, and medical decision making (Reyna and Brainerd 2007). Future studies will need to examine the role of various types of attention in addition to numeracy and literacy in informed consent.

This study also adds to the empirical literature on adolescents' ability to consent to treatment. First, this study shows control adolescents to be similar to their parents in terms of their ability to understand new information related to health and psychiatric disorders, supporting the literature that healthy adolescents older than 14 have abilities similar to parents. Second, among the adolescents, age was not a significant predictor of understanding; the literature is extended by showing that this is true even when adolescents approaching 12 years of age are included. These findings are similar to studies that noted that subjects with reading levels greater than grade 5 were deemed competent to consent to psychiatric inpatient hospitalization (Billick et al. 1998), and those of Mulvey and Peeples (1996), who observed similar levels of understanding of hypothetical vignettes between adolescents 14 years of age with psychiatric disorders and control adolescents. Bastieans (1995) also found no significant difference between parents' and adolescents' knowledge of psychiatric medications after an intervention study with hospitalized adolescents. This study supports the growing literature that suggests that the majority of adult subjects with mental health disorders can understand the information needed for consent to psychiatric treatment, but there may be reduced understanding for some subjects (Grisso and Appelbaum 1995).

One limitation of this study is that the majority of youth were not medication naïve and, in fact, had higher baseline knowledge than did controls. It is unclear whether the gain in scores with adolescents with ADHD would have been more or less had they been medication naïve. Our hope, initially, had been to recruit medication-naïve adolescents, but this was not possible. It is also possible that nonmedication-naïve adolescents may be more receptive to learning about ADHD and its pharmacological treatment than other adolescents and this may in turn also impact on learning.

Conclusions

The findings of this study add to the empirical literature on the ability of adolescents with mental health disorders to participate in the consent process. The findings support, at least for adolescents as young as 12, the common law decisions in Canada, which do not stipulate an age below which individuals are capable of understanding the information component of informed consent. This study suggests that with a properly devised information session, where care is taken to ensure that the information is conveyed in a simple but thorough process, the majority of adolescents with ADHD will understand the information. In addition, extra care should be taken to ensure understanding for subjects with reduced literacy and numeracy. This is particularly important since rates of mathematics learning disorders in clinical populations of school age children with ADHD is 18% (Capano et al. 2008). Since the majority of ADHD youth are typically started on medications before adolescence, it is important that they are always involved in the informed consent process to the fullest extent possible. As the youth gets older, the information should be revisited since they may be able to understand more than they had previously understood.

In clinical situations where the adolescent can have legal authority to consent, the clinician will need to ensure that the client has actually understood the information. When this is in doubt, it might be useful to ask some simple questions to ensure understanding (Appelbaum 2007; Larcher and Hutchinson 2010; Michaud et al. 2010). Even when adolescents do understand information needed to consent, their appreciation of information and ability to give voluntary consent also need to be examined. Even in jurisdictions where adolescents do not have the legal authority to consent, their understanding of the informed consent process can be increased using simple techniques such as breaking the process down and using appropriate language (Larcher and Hutchinson 2010). Enhanced understanding and participation in the decision-making process may also have a role to play in better decision making, treatment adherence, and treatment outcome.

Finally, while this article focused on understanding of information necessary to consent to treatment, the present results do not address other areas of importance in the informed consent process, including appreciation of information and the ability to make a make a voluntary decision free of coercion from others. For example, adolescents and parents may understand information about ADHD and the treatment of ADHD, but may not appreciate the diagnosis and its treatment. These other abilities associated with capacity can be assessed through a variety of questionnaires and/or questions (Appelbaum 2007), although there is also no gold standard for their determination. It has been suggested that the threshold for competency determination could vary with the risk benefit ratio of the proposed intervention and/or the seriousness of the consequences of the decision (Appelbaum 2007). An unresolved and related issue is whether a similar threshold for competency determination should be applied to both adolescents and adults or whether a higher threshold should be advocated for adolescents who may be less mature than adults. This study cannot answer those questions. Given the increased use of psychoactive medications within the adolescent population (Thomas et al. 2006), and the serious risks associated with some of them, it is particularly important that adolescents are involved in the consent process to the fullest extent possible. Future studies will need to address these other aspects of the informed consent process and extend this work by examining understanding among adolescents with other psychiatric disorders.

Clinical Significance

In summary, this study has examined the understanding of information necessary to consent to treatment. This study has shown that both adolescents' and parents' understanding of the benefits and risks of stimulant medication for treatment of ADHD can be increased using a careful disclosure process. In addition, understanding is associated with baseline knowledge and cognitive skills, particularly mathematics achievement. Therefore, for individuals with weaker numeracy and literacy skills, additional care is recommended to ensure understanding. Clinicians should involve adolescents in the informed consent process to the extent possible.

Footnotes

Disclosures

Ms. Tharmalingam is currently employed with the Health Council of Canada. Ms. Tharmalingam has no financial relationships with any pharmaceutical companies.

Dr. Debbie Schachter has mutual funds. Dr. Schachter does not know whether any of these mutual funds have pharmaceutical stocks that might pose a conflict. She has no other relationships with any pharmaceutical companies.

Dr. Irwin Kleinman has mutual funds. Dr. Kleinman does not know whether any of these mutual funds have pharmaceutical stocks that might pose a conflict. Dr. Kleinman owned stock in Johnson and Johnson for a few months.

Acknowledgments

This project was supported by the Canadian Institute for Health Research, Grant #64471.

The authors appreciate the statistical assistance of Tamara Arenovich and Gautam Sajeev with the Hierarchical Linear Modeling Analysis and regression analysis and Dr. Jack Williams with the methodological aspects of this project.

References

Achenbach

. Manual for the Child Behaviour Checklist/4–18, 1991 Profile. Burlington, VT: University of Vermont, Department of Psychiatry, 1991a.

Achenbach

. Manual for the Youth Self-Report and 1991 Profile. Burlington, VT: University of Vermont, Department of Psychiatry, 1991b.

American Academy of Child and Adolescent Psychiatry. Practice parameters for the assessment, treatment of children, adolescents, adults with attention-deficit/hyperactivity disorder. J Am Acad Child Adolesc Psychiatry, 36:85S. 1997.

Appelbaum

. Clinical practice: Assessment of patients' competence to consent to treatment. N Engl J Med, 357:1834–1840. 2007.

Appelbaum

, Grisso

. Understanding Treatment Disclosures. Worcester, MA: University of Massachusetts Medical Center, 1992.

Bastiaens

. Knowledge, expectations, and attitudes of hospitalized children and adolescents in psychopharmacological treatment. J Child Adolesc Psychopharmacol, 2:157–171. 1992a.

Bastiaens

. The impact of an intensive educational program on knowledge, attitudes, and side effects of psychotropic medications among adolescent inpatients. J Child Adolesc Psychopharmacol, 2:249–258. 1992b.

Bastiaens

. Compliance with pharmacotherapy in adolescents: Effects of patients' and parents' knowledge and attitudes toward treatment. J Child Adolesc Psychopharmacol, 5:39–48. 1995.

Billick

, Burgert

, Friberg

, Downer

, Bruni-Solhkhah

. A clinical study of competency to consent to treatment in pediatrics. J Am Acad Psychiatry Law, 29:298–302. 2001.

10.

Billick

, Edwards

, Burgert

, Serlen

, Bruni

. A clinical study of competency in child psychiatric inpatients. J Am Acad Psychiatry Law, 26:587–594. 1998.

11.

Blishen

, Carroll

, Moore

. The 1981 socioeconomic index for occupations in Canada. Can Rev Sociol Anthropol, 24:465–488. 1987.

12.

Boonstra

, Nash

. Minors and the right to consent to health care. The Guttmacher Report on Public Policy, 3:4–8. 2000.

13.

Capano

, Minden

, Chen

, Schachar

, Ickowicz

. Mathematical learning disorder in school-age children with attention-deficit hyperactivity disorder. Can J Psychiatry, 53:392–399. 2008.

14.

Casimir

, Billick

. Competency in adolescent inpatients. Bull Am Acad Psychiatry Law, 22:19–29. 1994.

15.

Conners

. Conners' Rating Scales-Revised: User's Manual. Toronto, ON: Multi- Health Systems Inc., 1997.

16.

Geary

. Mathematics and learning disabilities. J Learn Disabil, 37:4–15. 2004.

17.

Geary

. Mathematical disabilities: Reflections on cognitive, neuropsychological, genetic components. Learn Individ Differ, 20:130. 2010.

18.

Glass

. Refining definitions and devising instruments: Two decades of assessing mental competence. Int J Law Psychiatry, 20:5–23. 1997.

19.

Grisso

, Appelbaum

. Assessing Competence to Consent to Treatment: A Guide for Physicians and Other Health Professionals. New York: Oxford University Press, 1998.

20.

Grisso

, Appelbaum

. The MacArthur treatment competence study. III: Abilities of patients to consent to psychiatric and medical treatments. Law Hum Behav, 19:149–174. 1995.

21.

Health care consent act, 1996, S.O. C. 2, sch. A. a. 1996.

22.

Hopp v lepp, [1980] 2 S.C.R. 192. c, 1980.

23.

Larcher

, Hutchinson

. How should paediatricians assess gillick competence? Arch Dis Child, 95:307–311. 2010.

24.

Michaud

, Berg-Kelly

, Macfarlane

, Benaroyo

. Ethics and adolescent care: An international perspective. Curr Opin Pediatr, 22:418–422. 2010.

25.

Ministry of Education: The Ontario curriculum Grades 1–8: Mathematics. http://edu.gov.on.ca/eng/curriculum/elementary/math18curr.pdf 2005.

26.

Mulvey

, Peeples

. Are disturbed and normal adolescents equally competent to make decisions about mental health treatments? Law Hum Behav, 20:273–287. 1996.

27.

Paling

. Strategies to help patients understand risks. BMJ, 327:745–748. 2003.

28.

Preston

, Heaton

, McCann

, Watson

, Selke

. The role of multidimensional attentional abilities in academic skills of children with ADHD. J Learn Disabil, 42:240–249. 2009.

29.

Reibl v hughes, [1980] 2 S.C.R. 880. b, 1980.

30.

Reyna

, Brainerd

. The importance of mathematics in health and human judgment: Numeracy, risk communication, and medical decision making. Learn Individ Differ, 17:147–159. 2007.

31.

Rozovsky

. The Canadian Law of Consent to Treatment, 3rd. Toronto, ON: LexisNexis Butterworths, 2003.

32.

Sattler

. Assessment of Children: Cognitive Applications, Fourth. La mesa, CA: Sattler, 2001.

33.

Schachter

, Kleinman

, Harvey

. Informed consent and adolescents. Can J Psychiatry, 50:534–540. 2005.

34.

Scott

, Reppucci

, Woolard

. Evaluating adolescent decision making in legal contexts. Law Hum Behav, 19:221–244. 1995.

35.

Shaffer

, Fisher

, Dulcan

, Davies

, Piacentini

, Schwab-Stone

, Lahey

, Bourdon

, Jensen

, Bird

, Canino

, Regier

. The NIMH Diagnostic Interview Schedule for Children version 2.3 (DISC-2.3): Description, acceptability, prevalence rates, and performance in the MECA study. J Am Acad Child Adolesc Psychiatry, 35:865–877. 1996.

36.

Shaffer

, Fisher

, Lucas

, Dulcan

, Schwab-Stone

. NIMH Diagnostic Interview Schedule for Children version IV (NIMH DISC-IV): Description, differences from previous versions, and reliability of some common diagnoses. J Am Acad Child Adolesc Psychiatry, 39:28–38. 2000.

37.

Thomas

, Conrad

, Casler

, Goodman

. Trends in the use of psychotropic medications among adolescents, 1994 to 2001. Psychiatr Serv, 57:63–69. 2006.

38.

Wechsler

. Wechsler Intelligence Scale for Children—Fourth edition. San Antonio, TX: Psychological Corporation, 2003.

39.

Weithorn

, Campbell

. The competency of children and adolescents to make informed treatment decisions. Child Dev, 53:1589–1598. 1982.

40.

Wilkinson

. Wide Range Achievement Test 3. Wilmington, DE: Wide Range Inc., 1993.