Treatment of Attention-Deficit/Hyperactivity Disorder in Preschool-Age Children: Child and Adolescent Psychiatrists' Adherence to Clinical Practice Guidelines

Questionnaire design

For this study, the investigators created the Preschool ADHD Treatment Questionnaire (PATQ). This instrument was designed to obtain information about physician demographic characteristics—gender, board-certified specialty, type of practice, years in practice, and practice location—and physicians' practice patterns, including the number of child/adolescent patients seen per week, frequency with which they make a diagnosis of ADHD in preschool patients, their first-line medication of choice, and how often they use medication and/or behavioral management as part of their initial treatment plan. The PATQ also asked respondents to rate 19 clinical factors that might influence their decision to start medication. Respondents rated each factor on a four-point Likert scale: “not important,” “somewhat important,” “important,” and “very important.” The factors were grouped into five categories: physical safety (PS), educational concerns (EC), emotional distress (ED), social issues (SI), and historical risks factors (RF) for ADHD.

Subjects

Contact information for CAPs was obtained from a commercial list broker, medical marketing service (MMS), Inc. MMS, Inc. provided names and mailing addresses for 6141 CAPs. From this total cohort, questionnaires were mailed to 2200 randomly chosen CAPs who were practicing across the United States. Analyses were limited to responses from CAPs who indicated they were currently board certified in child and adolescent psychiatry.

Survey administration

The survey was mailed with a cover letter, postage-paid return envelope, and a complimentary copy of an ADHD Medication Guide (www.ADHDMedicationGuide.com) developed at our center. The cover letter explained the purpose and voluntary nature of the study. It clearly stated that the respondent's agreement to participate in the study was implied by his/her completing and returning the survey. To minimize potential risk to confidentiality, the survey did not ask for any individual identifying information. The cover letter included an explicit statement not to include any individual identifying information on the survey or on the envelope. In addition, to maintain anonymity, the return envelope did not have a return address. Each completed survey was given a unique identifier code upon receipt, before data entry. The data were entered into an Excel spreadsheet.

Statistical analysis

Chi-square tests were used to compare respondents' answer choice frequencies across specialties for each question. Significant p-values for chi-square tests of multiple comparisons were determined by using the Bonferroni correction. The alpha was set at p < 0.003 (p < 0.05/19). Effect sizes (φ) were reported for significant chi-square values.

Respondent demographics

Of the ∼2200 surveys that were mailed out to CAPs, a total of 356 surveys were returned, giving an overall raw response rate of 16.2%. Data from 11 respondents who were not board certified in child and adolescent psychiatry were excluded from the analysis, and six more respondents were excluded because they indicated that they never diagnose ADHD in preschool-age children. Thus, data from 339 CAPs were included in the analyses. The demographic and practice characteristics of the final sample are shown in Table 1. Of the 339 respondents, 53.4% were female. More than half (56.1%) had been in practice for more than 16 years and only 8.3% for fewer than 5 years. The vast majority of the sample (91.0%) reported seeing at least six children or adolescents with ADHD per week.

Preschool ADHD treatment recommendations

Tables 2 and 3 provide a detailed breakdown of the percentage of CAPs who would recommend parent training in behavior management and/or medication as treatment for a 4–5-year-old child with ADHD. Figure 1 and Supplementary Data (Supplementary Data are available online at www.liebertpub.com/cap) are a schematic showing the actions the respondents reported they would take at each decision point in their treatment of P-ADHD.

OPEN IN VIEWER

FIG. 1.

Schematic of responses by child psychiatrists regarding management of a preschool child with ADHD. ADHD, attention-deficit/hyperactivity disorder.

To determine how often physicians follow the AACAP guidelines for the treatment of preschool children with ADHD, we first calculated the percentage of CAPs who met each of the three requisite criteria individually: (1) “often” or “very often” recommend parent training in behavior management therapy (Recommendation #1), (2) “never” or “rarely” recommend medication as part of an initial treatment plan (Recommendation #2), and (3) choose MPH as the first choice medication (Recommendation #3) (Fig. 2 and Table 4).

OPEN IN VIEWER

FIG. 2.

Adherence criteria for preschool ADHD treatment. ADHD, attention-deficit/hyperactivity disorder.

Factors influencing the decision to treat preschool children with ADHD with medication

Although behavior therapy is recommended as a first-line treatment for P-ADHD, many communities may have limited professional resources to provide this form of therapy. To evaluate whether availability of behavioral treatments influenced the clinical approach of CAPs, respondents were asked “In your area, how available are behavioral counseling services for parents of preschoolers with ADHD?” One-fourth (25.7%) indicated it was “very available,” 53.7% “somewhat available,” and 20.6% “not very available.”As expected, CAPs were significantly more likely to “often” or “very often” recommend parent training in behavior management if it was perceived as being available in their local community. Behavior therapy was recommended by 96.5% of CAPs in communities, in which it was described as “very available,” versus 93.9% and 82.6%, in which it was described as “somewhat available” or “not very available,” respectively (χ ² = 9.89; p < 0.007).

CAPs were also asked, “In addition to recommending parent counseling in behavior management, how important are each of …[nineteen] factors in deciding to treat [p-ADHD] with medication?” These clinical factors were then grouped into five domains as displayed in Table 5. CAPs viewed PS and EC as most important for initiating pharmacotherapy, followed by ED, SI, and ADHD risk factors.

A child's being in “danger of hurting self” and “danger of hurting others” was considered to be the two most important factors in their deciding to use medication. Seventy-six percent (255/337) of CAPs rated “danger of hurting self” as “very important.” Seventy-one percent (239/337) rated “danger of hurting others” as “very important.” Fifty-four percent (181/338) included “recent expulsion” as the next most important factor, while 48.7% (165/339) identified “requires adult supervision” and 48.8% (165/338) identified “risk of expulsion” as the next equally important factors. Forty-two percent (141/339) identified “history of maternal tobacco use during pregnancy” as “not important,” making it the least important factor in deciding to use medication. The other factors identified as least important, in descending order, included “child adopted from an overseas orphanage” (40.2%, 136/338), “premature birth” (34.8%, 118/339), “history of maternal alcohol use during pregnancy” (27.7%, 94/339), and “history of maternal cocaine use during pregnancy” (25.1%, 85/338).

Of the 333 respondents who answered the question about medication as an initial treatment for P-ADHD (Table 2), 107 (32.1%) said they never or rarely prescribe medication as an initial treatment, whereas 149 (44.8%) sometimes, and 77 (23.1%) “often” or “very often” do so. Analyses were done to determine if there were relationships between how physicians rated the importance of each clinical factor as being important and their likelihood to prescribe medication as an initial treatment. When comparing the never/rarely group with the often/very often group, no differences were noted with respect to the three most important factors (“physical safety”) or least four important factors (ADHD risk factors). Significant differences between the groups were noted for EC, ED, and SI. Family history of ADHD was the only risk factor for which differences were noted. There were also significant differences noted between the ratings of importance of factors in the five categories (Table 5): PS was significantly more likely than EC to be rated “important” or “very important” by CAPs (χ ² = 92.84; p < 0.00001), EC significantly more likely than ED (χ ² = 37.01; p < 0.00001), ED significantly more likely than SI (χ ² = 48.92; p < 0.00001), and SI significantly more likely than ADHD risk factors (χ ² = 136.98; p < 0.00001). Accordingly, PS was significantly more likely than RF to be rated as “important” or “very important” by CAPs (χ ² = 1156.84; p < 0.00001). There was also a very significant difference in CAPs who rated factors in each of the five categories as “important” or “very important” versus other ratings (χ ² = 1679.71; p < 0.00001).

Analyses were also done to determine if there was an association between the stated importance of each of the 19 clinical factors and the frequency with which CAPs reported using medication as an initial treatment for P-ADHD (Table 6). Of the 19 factors, 10 were identified as “important” or “very important” for CAPs in their treatment decisions regarding medication. Table 6 reports effect sizes only for significant chi-squares—all of which were moderate in size (φ = 0.20 to <0.30). Clinical factors that are identified in Table 5 as most important (Physical Safety) and least important (ADHD Risk Factors) did not significantly influence decisions by CAPs to recommend medication as an initial treatment for P-ADHD. The rated importance of clinical factors in the categories of educational, emotional, or social concerns did differ between those who indicated that they “often” or “very often” use medication as a first-line treatment compared to those who do so “never” or “rarely.”

Adherence to guidelines

This study is the first to assess in a national sample to what extent board-certified child and adolescent psychiatrists follow the 2007 clinical guidelines for the treatment of P-ADHD. On one hand, our findings confirm that almost all CAPs recommend parent training in behavior therapy as a treatment for P-ADHD. On the other hand, our results suggest that the overwhelming majority of CAPs who manage ADHD often deviate from the AACAP guidelines when treating 4–5-year-olds with ADHD; only 7.4% of all responding CAPs fully complied with AACAP guidelines.

Admittedly, this analysis may have been unduly rigid in that clinicians who indicated they prescribe medication as part of an initial treatment plan did not necessarily rule out behavior therapy as an adjunct or concurrent therapy. For this reason, we also analyzed our response data with a less strict definition of adherence. With these liberalized criteria, the nonadherence rate was drastically reduced from 92% to 46%; nonetheless, this still suggests that almost half of CAPs do not follow guidelines when treating P-ADHD.

In trying to identify predictors of guideline adherence, post hoc analyses were done examining the volume of ADHD patients seen weekly and number of years in practice. Interestingly, CAPs who saw the lightest or the heaviest volume of patients were less likely to follow guidelines. This is not unexpected in that those who treat ADHD infrequently may be not aware of AACAP guidelines and those who treat very large number of cases may feel that their experience and expertise justify deviation from published guidelines on a case by case basis. In hindsight, it would have been helpful to assess respondents' familiarity with the AACAP guidelines, although it is worth noting that physician responses were obtained in 2012, many years following the publication of AACAP's general practice parameters for ADHD (Dulcan et al. 1997) and subsequent guidelines for psychopharmacologic treatment for very young children (Gleason et al. 2007).

It is important to note that AACAP's guidelines are not intended as mandates to clinicians. When published, the Academy emphasized that its algorithm was “not intended to suggest certainty, where none exists,” and that clinicians may “determine that an alternative approach is indicated in a particular clinical situation” (Gleason et al. 2007). Although AACAP clearly recognized and acknowledged the need for clinicians to deviate from their published guidelines, it is hard to imagine that they would countenance clinicians' deviating from guidelines in as much as 46%–92% of the time.

Use of ADHD medication in preschoolers

Although it was still low, the reported rate of adherence to each individual guideline was considerably higher than was adherence to all of the guidelines combined. Current clinical guidelines recommend behavior therapy as a first-line intervention before initiating pharmacologic treatment for preschoolers with ADHD (Gleason et al. 2007; Subcommittee on Attention-Deficit/Hyperactivity Disorder et al. 2011). The bases for this guideline is the demonstrated efficacy of behavior therapy in reducing ADHD behaviors (Pelham et al. 1998; Greenhill et al. 2003) and the greater potential for side effects of medication in this age group (Wigal et al. 2006). A recent review study by Charach et al. (2013) examined eight previous studies on parent behavior training and determined that there is “high strength of evidence” for its improving ADHD symptoms in preschoolers. The review concluded that parent behavior training is the “best first-line treatment” for ADHD in preschool children. Although the benefit of behavior interventions has not been directly compared to pharmacotherapy in this population, it is important that physicians understand its benefits and empirical evidence supporting it as parent behavior training as the first line of treatment.

While a significant majority of responding physicians (92.3%) reported that they “often” or “very often” recommend parent training in behavior management as part of initial treatment, only 31.6% of the sample reported “never” or “rarely” prescribing medication as part of initial treatment. One possible explanation is that the P-ADHD patients referred to these CAPs had such severe symptoms or impaired function that there was perceived clinical urgency in initiating pharmacotherapy. Another possibility is that many of these preschool children had, before evaluation by these CAPs, failed to benefit from behavior therapy and, thus, prescribing medication was indeed not a deviation from AACAP guidelines.

While pharmacotherapy is not recommended as initial treatment, treatment with medication is indeed recommended for 4–5-year-olds with moderate to severe functional disruption in the home and other settings, despite behavior therapy (Gleason et al. 2007; Subcommittee on Attention-Deficit/Hyperactivity Disorder et al. 2011). However, 71.3% of the surveyed CAPs “often” or “very often” recommended medication when behavior therapy had failed. Thus, despite proven safety and efficacy, a considerable number of CAPs are still reluctant to medicate young children.

ACAAP guidelines recommend MPH as the first-line medication for preschoolers (Subcommittee on Attention-Deficit/Hyperactivity Disorder et al. 2011). However, only 63.3% of CAPs selected MPH as their first-line medication. The efficacy and safety of stimulant medication (MPH and amphetamine preparations) in children 6 years of age and older is well established. Two nonstimulants, guanfacine and atomoxetine, also have demonstrated efficacy and safety in this age group. Historically, the evidence from experience with older children had been extrapolated to guide treatment of younger (preschool age) children. The rationale for recommending MPH as the preferred stimulant treatment for P-ADHD is that it is the only medication that has been studied adequately in this age group (Greenhill et al. 2006). The largest study, involving 165 children, was the PATS, which was the first randomized multicenter clinical trial designed to systematically assess the safety and efficacy of immediate-release MPH in preschool children with ADHD (Greenhill et al. 2006). The findings indicated that MPH is generally effective, safe, and tolerable in treating P-ADHD and should be used as the preferred pharmacological treatment (Wigal et al. 2006; Ghuman et al. 2008). However, because of the small number of quality studies available, Charach et al. (2013) found that there is “low strength of evidence” for the ability of MPH to improve ADHD symptoms in preschool children. In addition, preschool children tend to be more susceptible to adverse effects of MPH, with more frequent rates of appetite loss, trouble sleeping, stomachaches, social withdrawal, and lethargy compared to placebo (Charach et al. 2013).

There is insufficient evidence for the safety and efficacy in this age group of amphetamines or any of the nonstimulant ADHD medications to recommend them (Subcommittee on Attention-Deficit/Hyperactivity Disorder et al. 2011). However, despite the paucity of evidence, short-acting amphetamine products (Adderall^®, Zenzedi^®, ProCentra^®, and Evekeo™) are approved by the Food and Drug Administration (FDA) for use in children as young as 3 years, which may be confusing to physicians. At present, none of the long-acting stimulants is approved for use in children below age 6. The FDA recognizes that there is lack of clinical evidence to support use of amphetamines (short- or long-acting formulations) in young children and has requested that the manufacturer of lisdexamfetamine conducts studies to document safety and efficacy in 4- and 5-year-old children (Media, News, and Press Release 2014).

It is of interest that CAPs chose a nonstimulant as their first-line medication choice in as many as 1/5 cases. Although their rationale for doing so was not a focus of this study, considerations might have been that they were targeting other than core ADHD symptoms (e.g., aggression) or that many of these were children with an autism spectrum disorder. There have been other reports that appear to corroborate this phenomenon. For example, in an analysis of prescription data for preschoolers with emotional and behavioral disorders on Midwestern state Medicaid, mid-Atlantic state Medicaid, and a Health Maintenance Organization setting in the Northwest, investigators found that there was a 6.8- to 28.2-fold increase from 1991 to 1995 in the prevalence of clonidine use, as opposed to a 1.8- to 3.1-fold increase in stimulant use (Zito 2000).

There remains limited empirical evidence for ADHD medication efficacy and safety in preschoolers, and there is little doubt that guidelines will continue to evolve over time. Similarly, the availability of liquid, chewable, and transdermal stimulant formulations not only make it easier to administer stimulant medications to young children but they also allow these medications to be titrated more precisely (although off-label)—even at doses below current FDA recommendations. It is possible that specific medication prescribing trends will evolve as clinicians gain greater familiarity and experience with these various new formulations.

Factors influencing the initiation of pharmacotherapy in preschool children with ADHD

Since a large majority of CAPs use medication as initial therapy, it is important to examine which factors are important to them when making the decision to do so. The present study found that educational and PS concerns were the most important factors in considering whether to prescribe medication for preschool children with ADHD. In contrast, factors that confer an increased risk on children of having ADHD (especially maternal tobacco use during pregnancy) were reported to be the least important. It appears that clinicians will deliberately deviate from published guidelines in considering pharmacotherapy for preschool children with ADHD when, in their judgment, a child's physical well-being and/or developmental functioning are significantly at risk.

Although this study has many strengths and the findings are revealing, several methodological weaknesses and limitations must be acknowledged. First, as is typical of survey-based research involving clinicians, our overall response rate (16.2%) was low. Since we do not have demographic data for the sample to whom questionnaires were mailed, we cannot rule out that there are some demographic differences between our sample and the larger population of CAPs and that this may introduce some potential skew or bias in our results. In contrast, with responses from 339 CAPs, our sample was more than adequate for statistical analysis and the cohort was demographically varied in terms of age, gender, clinical practice setting, and volume of ADHD patients treated.

A second notable limitation of our study relates to the wording of the question about recommending medication as part of an initial treatment regimen. Although there was unintended ambiguity regarding whether or not behavior therapy had been previously implemented, this was not a significant factor. For example, the fact that 92.3% of respondents indicated that they “often” or “very often” recommend parent training makes it clear that the overwhelming majority of CAPs understood the intent of the question—that this was a treatment naive family (i.e., the family did not have prior behavioral therapy). Concern about ambiguity may only apply to the very small number of respondents (24/339) who indicated that they “never” or “rarely” recommend parent training. It is further reassuring that 6 of these 24 CAPs indicated they would only recommend medication if behavior therapy had failed.

Availability of parent training in behavior management significantly influenced whether or not CAPs recommended it as a first-line treatment. We cannot verify the accuracy of CAPs' perceptions regarding availability. If the CAPs' assessment of availability was indeed accurate, then one could argue that deviation from guidelines regarding behavior therapy was perhaps justified. This approach is formally suggested by the American Academy of Pediatrics. In their recent guidelines for ADHD, the AAP states that “In areas where evidence-based behavioral treatments are not available, the clinician needs to weigh the risks of starting medication at an early age against the harm of delaying diagnosis and treatment” (Subcommittee on Attention-Deficit/Hyperactivity Disorder et al. 2011). Of course, it is possible that CAPs may not have an accurate assessment of the availability of behavioral therapists and that their predilection to pharmacotherapy as a first-line treatment may influence their perception of resource access.