Results from the largest treatment study of ADHD ever conducted
Attention Research Update   
              David Rabiner, Ph.D.   Sr Research Scientist, Duke University                        

Over the past year you may have seen preliminary reports of
results from the Multimodal Treatment Study of Children with
ADHD (MTA).    This is the largest and most comprehensive
treatment study of ADHD that has ever been conducted.  Last
month, the initial papers reporting the results from this study
were published.  This is a landmark study with a number of
important implications.

The December issue of the Archives of General Psychiatry
includes 2 papers that are based on this study.  The first
paper titled "a 14 month randomized clinical trial of treatment
strategies for ADHD" presents the major findings from the
study.  The second paper, titled "Moderators and mediators
of treatment response for children with ADHD" presents more
fine-grained analyses in which factors that might have influenced
response to the different treatments evaluated in the study
were examined.  In this summary, I will try to combine the
findings that are reported in these two papers.


It is important to begin by providing an overview of how the
study was conducted and the questions that it was specifically
designed to address.  The study represented the combined
efforts of investigators at 6 different sites around the country
and included 579 children ages 7 to 9.9 years who were
diagnosed as having ADHD, Combined Type using state-of-
the-art diagnostic procedures.  (Children diagnosed with
the hyperactive/impulsive subtype or inattentive subtype
were excluded.  This decision was made because the
combined type is the most frequently diagnosed type of
ADHD).   Approximately 20% of the participants were
girls and about the same percentage was African American.

After participants had been identified they were randomly
assigned to 1 of 4 different treatment conditions.  Fourteen
months later, the participants were carefully evaluated
so that the impact of the different treatments could be
evaluated.  Each treatment condition is described below.

medication management condition received medication
treatment only.  This began with a 28-day, double-blind
placebo-controlled trial in which the effects of 4 different
doses of methylphenidate (the generic form of Ritalin)
were evaluated.  The doses tested were 5, 10, 15, and
20 mg.  Children received a full dose at breakfast and
lunch, and then a half-dose in the afternoon.  Parent and
teacher ratings of children's behavior on each dose were
compared by a team of experienced clinicians, and the
best dose for each child was selected by consensus.
(In a double-blind placebo-controlled trial, the child
is receiving real medication during some days and a
placebo during other days.  Neither the child, the teacher,
nor the parent know when real medication is being
received and when placebo is being received.  Because
of this, parent and teacher ratings of the child's behavior
are not biased by the knowledge that the child is on

For children not obtaining an adequate response to
methylphenidate during this initial trial, alternate
medications were tested using non-double-blind
procedures in the following order until a satisfactory
medication and dose for the child was found:
dextroamphetamine (the generic version of
dexedrine), pemoline (the generic version of Cylert),
and imipramine (a tricyclic antidepressant).

Note: This study was begun before Adderall was
available for use which is probably why it was not
one of the medication options.

Of 289 participants initially assigned to receive
medication in either the medication management
condition or the combined condition (see below),
256 (88.6%) successfully completed this initial titration
period used to select an effective medication.  (For the
remaining children, parents either refused to try their
child on medication, there were intolerable side effects,
or parents could not cooperate with the careful titration

For about 69% of the children completing the initial
medication trial, an adequate response was obtained with
at least one of the doses of methylphenidate, and they began
their treatment on this dose.  Twenty-six children  who did
not respond to methylphenidate were found to do well on
dextroamphetamine and began on this medication.  A final
32 did not begin on any medication because they had
such a strong placebo response that no clear benefits of
medication could be demonstrated.

In addition to this VERY CAREFUL initial trial to determine
the optimal medication and dose for each child, half-hour
monthly visits were scheduled during which the provider
for that child would review information about the child's
behavior over the past month that was provided by parents
and teachers.  After carefully reviewing this information,
dosage adjustments were made using predetermined guidelines.
Adjustments that involved increases or decreases of more
than 10 mg/dose needed to be approved by a cross-site
panel of experts.

By the end of the study - 14 months later - about 74% of
participants of the 289 in the medication or combined treatment
groups were being successfully maintained on methylphenidate,
10% on dextroamphetamine, and just over 1% on pemoline.
Only two children were on any other type of medication.  (As
noted above, some children who were assigned to one of
the medication groups never received medication either because
their parents refused or could not follow the initial trial
procedures.) Side effects had also been monitored monthly
and over 85% of the sample were reported to show either
no or mild side effects.

It is important to emphasize how different this approach
to medication management was from what often occurs
in community treatment.  The primary differences are
1) the use of a double-blind trial to establish the best initial
dose and medication for each child; and, 2) regular follow-up
visits to evaluate ongoing medication effectiveness based
on parent and teacher reports with systematic adjustments
made as needed.

It is also important to note that almost all children were
judged to be effectively managed on one of the standard
stimulants (either methylphenidate or dextroamphetamine)
and none were judged to require a combination of
medications to effectively manage their ADHD symptoms.
I think this underscores how rarely medications need to
be combined to treat ADHD when a careful procedure is
used to test out the different types of stimulants that
are available.  This is discussed in another article in
this issue of ADHD RESEARCH UPDATE that you
will find below.

BEHAVIORAL TREATMENT - Behavioral treatment
included parent training, child-focused treatment, and
a school-based intervention.  Parent training involved
27 group sessions and 8 individual sessions per family.
The focus was on teaching parents specific behavioral
strategies to deal with the challenges that children with
ADHD often present.

The child-focused treatment was a summer treatment
program that children attended for 8 weeks, 5 days a
week, during the summer.  This program employed
intensive behavioral interventions that were administered
by counselors/aides who were supervised by the
therapists conducting the parent training.  The basic
model was one in which children were able to earn
various rewards based on their ability to follow well-
defined rules and meet certain behavioral expectations.
Social skills training and specialized academic
instruction was also provided.

The school-based treatment had 2 components: 10 to
16 sessions of biweekly teacher consultation focused
on classroom behavior management strategies, and
12 weeks of a part-time paraprofessional aide who
worked directly in the classroom with the child.
Throughout the school year, a Daily Report Card
was used to link the child's behavior at school to
consequences at home.  The Daily Report Card was
a 1-page teacher-completed ratings of the the child's
success on specific behaviors.  This was brought
home daily by the child to be reviewed by parents
with rewards for a successful day provided as

Consistent with what occurs in actual clinical practice,
the family and child's involvement in behavioral
treatment was gradually tapered over the 14 month
period.  In most cases, contact had been reduced to once
monthly or stopped altogether by the end of this

The main point to take away from this brief summary of
the behavioral treatment that children received is that it
reflects absolute state-of-the-art practice that would be
virtually impossible to obtain in a typical community
setting.  Thus, if anything, one would expect that the benefits
of behavioral treatment as implemented in this study would
be likely to be greater than what would typically be

COMBINED TREATMENT - Children in the combined
treatment group received all of the treatments that are
outlined above.  Individuals supervising the child's
behavioral and medical treatments conferred regularly,
and this was used to guide overall treatment decisions.
Consistent with what has been found in prior studies,
by the end of the study, children in the combined group
were being maintained on lower daily doses of
methylphenidate than children who received medication
alone (Average doses were 31.2 mg/day for the combined
group and 37.7 mg/day for the medication only group).

COMMUNITY CARE - It clearly would not be ethical
to assign children with ADHD to a no-treatment control
group for a study that persisted for 14 months.  Instead,
some children were randomly assigned to a group that
received "community care".  Following their child's
diagnosis of ADHD, parents of these children were provided
with a list of community mental health resources and
made whatever treatment arrangements they preferred.

Most of the 97 children in this group (over 2/3s) received
medication from their own provider during the 14 months.
Several things are interesting about the medication these
children received compared to children who received
medication as part of the study.  First, community care children
received less medication each day.  For those treated
with methylphenidate, the average daily dose was 22.6 mg/day
compared to the average daily doses of 31.2 mg and
37.7 mg noted above.  In addition, community care children
received an average of 2.3 doses per day compared to
the 3 times/day dosing for children in the study.

Finally, while none of the children receiving medication in
the study were maintained on either clonidine or a
combination of medications, 4 children seen by community
physicians were treated with clonidine and 10 children
received more than one medication.  Thus, it appears
that physicians in these communities were in some ways
more conservative in their use of medication (i.e. prescribed
lower doses of methylphenidate) and in some way less
conservative (i.e. were more likely to use medications
other than the widely used stimulants).


The MTA Study was designed to address 3 fundamental
questions about the treatment of ADHD.  These questions are
as follows:

1. How do long-term medication and behavioral treatments
compare with one another?

2. Are there additional benefits when they are used together?

3. What is the effectiveness of systematic, carefully delivered
treatments vs. routine community care?


There is a tremendous amount of data presented in these
papers and it is really not possible to summarize it all.
Below, however, are what I found to be the most important

First, let me list the variety of different outcomes that
were assessed and reported.  These include:

* Primary ADHD symptoms - ratings provided by parents
and teachers;

* Aggressive and oppositional behavior - ratings provided
by parents, teachers, and classroom observers;

* Internalizing symptoms (e.g. anxiety and sadness) -
ratings provided by parents, teachers, and children;

* Social skills - ratings provided by parents, teachers,
and children;

* Parent-child relations - rated by parent;

* Academic achievement - assessed by standardized tests;
(It is unfortunate, I think, that more frequent measures of
academic performance in the classroom were not
collected.  These tend to be more sensitive to change than
scores on standardized achievement tests.  Thus, the
reliance on achievement tests alone as the measure of
academic performance may not have enabled important
changes in academic functioning to be captured).

In considering the results presented below it
is important to place them in this overall context:

Children in all 4 groups (i.e. medication only, behavioral
treatment only, combined treatment, and treatment in the
community as chosen by parents) showed significant
reductions in their level of symptoms over time in
most areas. Thus, even though some treatments were
clearly superior to others in certain domains, overall,
even children receiving the "least effective" treatment
tended to show important improvement.  Thus, these
data should not be interpreted in a framework of "what
worked" and "what did not work".  Instead, it is a
matter of what seemed to be most effective among treatments
that all showed some positive effects.

1. How do long-term medication and behavioral treatments
compare with one another?

For both parent and teacher ratings of primary ADHD
symptoms (i.e. inattention and hyperactivity/impulsivity),
medication management alone was clearly superior to
behavioral treatment alone.

On all the other outcome measures reported, medication
management and behavioral treatment did not differ

Thus, although medication was found to be superior to
behavioral treatment on core ADHD symptoms, this
did not extend to other important areas of children's
functioning such as oppositional behavior, peer
relations, and academic achievement.

2. Did participants receiving combined treatments show
higher levels of improvement than participants receiving
medication treatment alone or behavioral treatment alone?

Combined treatment and medication management treatment
did not differ significantly in any of the 6 domains.
This suggests that for most children with ADHD, adding
behavioral intervention on top of well-conducted
medication management is not likely to yield substantial
incremental gains.

As can often be true with statistical analyses, however,
this conclusion changes somewhat depending on how you
look at the data.

For example, when you look at the rank ordering on different
outcomes for children in the different groups, children in the
combined treatment group did best on 12 of 19 outcome measures
while those in the medication management group were best on
only 4.  In addition, when the individual outcome measures
are combined into composite measures, or when children's
outcomes are grouped into excellent response vs. less dramatic
response categories,  children receiving combined treatment
did modestly, but significantly, better.

Compared to behavioral treatment alone, combined treatment
was found to be superior on parent and teacher ratings of
primary ADHD symptoms, on parent ratings of aggressive/
oppositional behavior, on parent ratings of children's
internalizing symptoms, and on results of the standardized
reading assessment.  Thus, adding medication to the
treatment of a child already receiving behavioral intervention
is likely to yield substantial benefits for most children.

3. Did participants assigned to each of the 3 MTA treatments
(i.e. medication management, behavioral treatment, and
combined treatment) show greater improvement than children
receiving community care?

The answer to this question was clear and straight forward.
Both combined treatment and medication treatment were
superior to community care for parent and teacher reports of
primary ADHD symptoms while behavioral treatment
was not.

In general, parents and teachers tended to report a decline of
approximately 50% in inattentive and hyperactive/impulsive
symptoms for children in the medication and combined treatment
groups.  For children receiving community care, the declines
reported were in the 25% range and were comparable to those
reported for children receiving behavioral treatment.

In the non-ADHD domains, (e.g. oppositional behavior,
internalizing symptoms, social skills, and reading achievement)
combined treatment was always superior to community treatment,
with particularly dramatic differences in parent reports of oppositional/
aggressive behavior.  Medication management and behavioral treatment
were superior to community treatment on a single domain only.

Overall, these data indicate that although children treated in
the community made modest gains over the course of the
study, those receiving medication treatment in the MTA
study - either alone or in combination with behavioral
treatment - did significantly better.  This was especially
true for children receiving the combined treatments.
Possible reasons for this will be discussed in the summary
section below.


In addition to the analyses reported above, the MTA research
group was interested in whether the effect of the different
treatments may have differed depending on certain
characteristics of the children.  Thus, they also looked at
whether similar results were obtained:

1. for boys vs. girls - as noted above girls made up about
20% of the overall sample;

2. for children with and without an additional diagnosis of
either Oppositional Defiant Disorder (ODD) or Conduct
Disorder (CD);

3. for children with and without a co-occurring anxiety disorder;

In general, there were no substantial differences in the effectiveness
of the different treatments depending on these variables.  Thus,
similar treatment results were found for boys and girls and for
children with and without a co-occurring behavior disorder.  There
was some indication, however, that for children with a co-occurring
anxiety disorder, behavioral intervention alone was as effective
as both medication management and the combined treatment.  It
is also worth noting, however, that children with anxiety disorders
who received medication only did not have a poorer response to
medication than other children.  Thus, prior and less intensive
studies in which it has been reported that children with ADHD and
an anxiety disorder do not do as well on stimulant medication are
contradicted by these results.


In a final set of follow up analyses, the researchers also analyzed
the results according to how children and parents were able to
adhere to the prescribed treatments.  Thus, children assigned to
the medication management condition were divided into 2 groups
depending on whether they or not medication treatment was
implemented as recommended and whether the family attended at
least 80% of the scheduled follow-up visits where the ongoing
impact of the medication could be monitored.

For behavioral treatment, children were divided into 2 groups
depending on whether or not parents attended at least 75% of
the scheduled parent group meetings, the child attended at least
75% of the summer treatment program, and whether the child and
paraprofessional working with the child in the classroom were
both present for 75% of the intended days.  If any one of these
3 conditions were not met, the behavioral treatment was not
considered to have been implemented as intended.

For the combined treatment group, families had to adhere to the
guidelines for both medication management and behavioral
treatment to be placed in the "as intended" group.  Otherwise,
they were placed in a group that was judged to not have
adhered to treatment as recommended.


The first thing that is interesting to note is the percentage of
families in the 3 MTA study treatment conditions that were
able/willing to adhere to treatment as recommended.

Acceptance/attendance was higher for the medication management
treatment (78% of families completing treatment as intended)
than in behavioral treatment (63%) or combined treatment (61%).
Thus, even when state of the art behavioral treatment was
provided to families at NO CHARGE, almost 40% of families
were unable and/or unwilling to fully take advantage of it.

In terms of the impact of treatment adherence on child outcome,
significant effects were found only for the medication management
group.  Thus, for children where the recommended medication
management procedure was followed more closely, the outcomes
were significantly better.  For the behavioral and combined
treatment conditions, in contrast, no differences in child outcomes
depending on treatment adherence were found.  It seems reasonable
that the absence of an effect of adherence for the combined treatment
group is that most of the families in the non-adherent category
were there because they failed to comply with the behavioral
treatment procedure, and that these children did as well as the
"adherers" because of the benefits they derived from the medication.


There is a LOT here to digest.  Before trying to pull together what
seem to me to be some of the fundamentally important implications
of this study, it is important to note that many additional papers
will be emerging from this work.  In particular, although the children
in this study are no longer receiving their treatment as part of the study,
they do continue to be followed.  This will enable the researchers to
examine the sustained impact of different treatments beyond the 14
month outcome data that were presented in this initial paper.  Thus, it
is certainly possible that results based on 2 or 3 year outcomes may
look somewhat different from what was found after 14 months.

Several other caveats are important to note.  First, in this study children
with the inattentive subtype of ADHD were specifically excluded.
Thus, these results can not be generalized to children with this subtype
of ADHD.

Second, treatments investigated in this study were limited to those with
the greatest empirical support to date: medication and behavioral
treatment.  This study thus sheds no light on the effectiveness of other
types of treatment for ADHD such as dietary interventions, biofeedback,
etc.  Additional research on other treatment options that is as careful
and well conducted as this study is certainly needed.

That being said, what are some of the important conclusions to be drawn
from the data presented so far and what do these results mean for parents
and health care providers who are concerned about doing the best they
can for their child and their patients?  (Please note that these are my
opinions, and that other scientists, health care providers, and educators
might reach somewhat different conclusions from those I present.  Also,
it is important to stress that conclusions about treatment are predicated
on a careful evaluation of ADHD having been done in the first place, as
was the case in this study).

For many children with ADHD, Combined Type, medication alone is
likely to be an effective and perhaps even sufficient treatment when
care is taken to determine the optimal medication/dose for each child
and when the ongoing effectiveness of medication is carefully monitored.

I am aware that many people may find this conclusion to be
distasteful, but I think it is a reasonable one to draw from
these data.  Remember, I am a Ph.D. not an M.D., and thus do
not provide medication myself.

Although there was some indication for a mild to modest superiority
for combined treatment on some outcomes, overall, children who
received medication alone tended to do about as well as children who
received the combined treatment.  This was true even though the behavioral
treatment provided in this study was far more intensive than would be
routinely available in any community setting.  In fact, I think it is reasonable
to say that the behavioral treatment provided in the MTA setting could
simply not be duplicated in any other context.

This does not mean that there is no place for behavioral treatment in
the management of children with ADHD (see below).  To me, however,
it suggests that a reasonable approach may be to begin with carefully
conducted medication trial to be certain that the maximum possible benefits
from medication are being attained.

When this has been done, and there are still important difficulties in a child's behavioral, academic, and/or social functioning, adding behavioral or other psychosocial interventions that specifically target these residual problems should be pursued. These interventions can make an IMPORTANT difference for an individual child, even though the benefits at a group level are apparently not so dramatic.

It should also be noted that combining behavioral treatment with
medication management did enable children to be maintained on a
somewhat lower dose of medication.  The authors note, however,
that the actual significance of this difference is unclear.  Many parents and
physicians may regard  this as quite  important, however.  Thus, if
maintaining your child on the minimum dose of medication required to yield optimum results is important to you, than combining medication treatment with carefully executed behavioral interventions is  likely to be required.

Intensive and well-conducted behavioral treatment can also be an
effective option for treating children with ADHD.   For most
children it will probably be less effective than careful medication
treatment, however, and it may be hard for parents to implement
as directed.

Once again, I think it is very important to note that the behavioral
interventions implemented in this study were also associated with
significant reductions in ADHD symptoms and some improvement
in other domains.  The reductions in ADHD symptoms were not
as great as for the medication management group, but in other areas,
no statisically significant differences between these treatments were

There are, however, some important points to keep in mind here.  First,
as noted above, the intensitiy and quality of the behavioral treatment
provided to children in this study could probably not be matched in
any other context - it is just not available outside of a research setting.
Whether a less intensive behavioral treatment would also be shown
to produce significant gains over a 14-month period is thus unknown.
Chances are, however, that behavioral treatment as typically practiced
would probably not be as helpful as what was able to be provided in
the study.

Second, it is very difficult for parents to persist with the type of
behavioral treatment used in this study - about 40% were not able
to adhere to the treatment even though it was offered in the study
at no charge.

Finally, it should be noted that although not many statistically
significant differences between  behavioral treatment and medication
management were found, 26% of the parents whose child was
receiving behavioral treatment only as part of the study opted to
add medication to their child's treatment.  In contrast, only 2%
of parents whose child was receiving medication opted to add
behavioral treatment.  This certainly suggests that many parents of
children receiving behavioral treatment only were less likely to be
satisified with the results of their child's treatment.

Overall, I think a reasonable conclusion is that behavioral
intervention - when used in isolation - is likely to be less effective
than medication management, harder for parents to implement, and
more expensive.  To me, this suggests that the
most appropriate use of behavioral treatment for many children
may be not as the sole intervention, but as something that is
carefully incorporated into a child's treatment to address
problems that are not sufficiently helped by medication alone.

How medication is prescribed makes a difference and parents need
to insist that their child's physician have an objective procedure in
place to determine the optimum medication/dose for their child, and
to carefully monitor the ongoing effectiveness of medication treatment
for their child.

An inescapable conclusion from this study is that children who received
medication from the MTA staff did significantly better than children who
received medication from community physicians.  Although the reasons for
this can not be determined with certainty, it seems quite likely that this was
because of the care that was taken initially to determine the optimum dose
for each child, and to then carefully monitor how the child was doing and
to make adjustments as needed.  Parents need to insist that this be done for
their child.  Physicians need to begin using more objective procedures for
evaluating medication response on a routine basis.  This is not hard to do
but it does take a bit of time.  (Remember, you can easily use the ADHD
Monitoring System that you received when you subscribed to evaluate
the ongoing effectiveness of your child's treatment).

There are some differences in medication treatment in the MTA group and
the community care group that we do know with certainty.

1. Children treated by community physicians may be routinely under-

Children treated with medication alone in the MTA study who did
well on methylphenidate received an average of almost 38 mg/day
in 3 separate doses.  Children treated with methylphenidate in the
community received an average of about 23 mg/day - a dose reduction
of about 40% - spread over 2 doses per day.   Even though children
receiving medication as part of the combined treatment were on lower
doses then the medication only group, they still received a substantially
higher dose than the community treated participants.

Because MTA-treated children did much better, it seems reasonable to
conclude that many children treated in the community were not receiving
enough medication to obtain the maximum possible benefit.

Please do not interpret these data to mean that every child should be
on the average dose used in the MTA study.  Remember, some children
do better on lower doses and some on higher, and the best dose for each
child needs to be determined using a careful trial.

Also, it is important to remember that the daily total dose and 3 administration
per day figure noted above was for methylphenidate and would certainly
be different for other medications.  For example, recent data suggests
that Adderall - not used in this study because it was not available when
the study was conducted - can produce at least comparable benefits
to methylphenidate with fewer administrations per day.

2. Children treated by community physicians are often put on non-stimulant
medications and/or combinations of medications that are not necessary.

I think this is a really important point.

Recall that virtually every MTA participant receiving medication was
able to be managed effectively on either methyphenidate or the generic
version of dexedrine.

Very few needed to be prescribed a different class of medication like an
antidepressant and not a single child was prescribed a combination of meds
(e.g. methylphenidate and clonidine).  In contrast, over 10% of children
treated by community physicians were on multiple medications and over
16% were treated with an antidepressant.

What I conclude from these data is that when stimulant medication
is prescribed carefully, there will be VERY few cases where another
class of medication needs to be used and ALMOST NO CASES where
multiple medications are needed.

I think that what may often happen in the community is that physicians
give up on stimulants before an adequate dose has been tried, or before
alternative stimulants have been tried.  Instead, a switch is made to
a different type of drug or a new drug is combined with the stimulants.

This is problematic for several reasons.  First, no other class of drugs
has been shown to be as effective as stimulants for treating ADHD.
Second, despite the concerns that many people have about possible
adverse health consequences of stimulant medications, available
support for the long-term safety of these medications is greater than
for the other medications that are often switched to or added.

So, if I were a parent of a child with ADHD, I would ask LOTS of
questions of my child's physician before I switched him or her to
a non-stimulant medication or had my child take multiple medication.
(e.g. "Why don't we try a higher dose first?"  Why don't we test
the effect of another type of stimulant first?")

If you are a provider of medication, I think these data should be
carefully considered before such a switch is recommended.

Return to Recommended Articles