MTA Study Critique

From: Ian Goddard (
Date: Wed Jan 26 2000 - 20:54:44 MST

  The "MTA study" funded by the National Institute
  of Mental Health is being touted as the long-awaited
  research confirming that long-term (14 month) drugging
  of children with stimulants to control misbehavior is
  more effective than drug-free behavioral therapy. In
  the following report, however, Peter Breggin, MD,
  gives us a second opinion about the findings of this
  study. The first objection Breggin raises constitutes
  the complete elimination of the study: the study was
  not a placebo-controlled, double-blind clinical trial.
  Amazing! Both the study subjects and those evaluating
  the performance of the subjects knew which subjects
  were on drugs and which were not, which is known as
  an "open label" study. The reasons why double-blind
  placebo-controlled studies are essential for valid
  drug research are basic to established standards
  for scientific research -- standards that the much-
  touted MTA study fails to satisfy. Based on that fact
  alone, the MTA study is essentially junk science that
  amounts to a taxpayer-funded advertisement for drugs,
  yet that's just the beginning of Breggin's critique:

A Critical Analysis of the Multimodal Treatment Study for
Attention-Deficit/Hyperactivity Disorder (The MTA Study)

by Peter R. Breggin, M.D.
January 12, 2000

     After many months of positive publicity in the psychiatric
and the general media, the results of the Multimodal Treatment
Study for Attention-Deficit Hyperactivity Disorder (The MTA Study)
were finally published in December 1999 (MTA Cooperative Group,
1999a&b). The study was sponsored by the National Institute of
Mental Health (NIMH) at six separate sites in the United States.
At each site, the study compared four treatment conditions: (1)
medication management alone, (2) combined medication management
and behavioral therapy, (3) behavioral treatment, and (4)
community care. The average age of the children was eight and
80% were boys.
     The aim of the study was to "resolve controversies and
clinical quandaries about the relative value of medication and
behavioral treatments" (National Institute of Mental Health,
undated). The proponents of the study claim that it demonstrated
the superiority of stimulant treatment over behavioral treatments
and routine community treatment. However, an examination of the
MTA study reveals that it was so grossly flawed methodologically
that it lacks scientific validity. The following is a general
critique of the MTA study.

I. The MTA was not a placebo-controlled, double-blind clinical

     The MTA study fails to meet the commonly accepted criteria
for a scientific study of medication efficacy or effectiveness.
It was not a placebo-controlled double-blind clinical trial (MTA
Cooperative Group, 1999a). First, there was no placebo control
group and no non-treatment control group. Second, to reach their
conclusions, the investigators relied upon evaluations made by
teachers and parents who were not blind to the treatment. That
is, the raters whom the investigators relied upon knew whether
or not the children were taking medication.

     In short, the MTA was an "open label" study--the kind that
does not qualify as scientifically valid. It could not be used,
for example, for FDA approval of a drug. Rater, researcher, and
subject biases, unconscious or conscious, influence the outcome
of "open label" studies. Researchers commonly want to prove that
the treatment under investigation is effective, and in the case
of the MTA studies, all the principal investigators were staunch
advocates of medication. Evaluators of efficacy are often
unconsciously influenced to perceive what they anticipate
perceiving, such as improvement in subjects treated with drugs.
Also, the subjects themselves often respond positively to please
their doctors or as a result of their belief that the drug will
help them. Therefore, open label studies have been discredited
for purposes of studying efficacy. As Nies and Spielberg (1996,
p. 45) observe, "Placebo effects, which occur in a large percentage
of patients, can confound many studies--particularly those that
involve subject responses; controls must take this into account"
(for additional discussions of placebo, double-blind procedures,
and research standards, see Fisher and Greenberg, 1989).

II. The blind classroom raters found no difference in any of
the treatment groups, i.e., behavioral interventions were equal
to medication interventions.

     The MTA used one group of "blinded ratings of school-based
ADHD and oppositional/aggressive symptoms..." (MTA Cooperative
Group, 1999a, p. 1074). The blind raters observed the children
in the classroom only. The data from these raters are produced
in Table 5 (pp. 1082-3) of the study. The blind raters found no
difference between any of the treatment groups on any of the
variables involving ADHD or oppositional behavior. However, this
extremely important finding, that the only potentially objective
raters found no drug effect, was given no importance in the study
conclusions. Nonetheless, the finding of the blind classroom
raters is one of the most important findings of the study,
confirming that stimulant drugs have no long-term positive effect.

III. There was no control group of untreated children.
     There was no non-treatment control group. The MTA compared
various treatments, not treatment versus no treatment. Two-thirds
of the community-treated group received a variety of medications,
as well as other interventions, and therefore it cannot be
considered a non-treatment control group. Basically, the studied
up compared three drug conditions to behavioral interventions,
without comparing the drug treatment to placebo.

IV. Thirty-two percent of the Medication Management group was
already on medication for ADHD at the start of the MTA.

     Table 3 (p. 1079) shows that of 144 medication management
subjects, 46 (32%) were on medication for ADHD at the start of
the selection process. This is generally not acceptable in a
study of medication effects and corrupted the study. Since the
children were already receiving medication, it is highly probably
that their parents had already determined to their own satisfaction
that the drugs were helpful. Therefore they could not participate
objectively in a "random" drug study.

V. The Medication Management group was highly selective (in ways
that are not fully described) and probably not typical of children
who seek services for "ADHD."

     The study initially screened 4,541 children. These children
were referred from a variety of sources, such as public
advertisements, clinics, and schools. The aim was to draw from
a broad spectrum of the kind of children whose parents bring them
for services for "ADHD." But of these 4,541 children, only 579
subjects (12.8%) were selected to enter the trials.

     The authors do not clarify why so many children were rejected
from the study. The small number of subjects actually selected for
the trials leaves a suspicion that the children in the trials do
not reflect a representative community group. As already noted,
many of the children were already taking stimulant medications.
Overall, the group that entered the trials may have little relevance
to range of children who seek help in clinical practice.

VI. The Medication Management group was relatively small.

     The actual medical management group is much smaller than might
be suspected from the seemingly large scale of the study. Of the
579 who entered the trials, only 144 entered medication management,
i.e., received medication alone. Thirteen of these dropped out
before starting, limiting the actual start group to 131. Eight
more dropped out during the study for a total of only 123 finishers
in medication management.

     Thus, of the 4,541 children originally screened, only 2.7
percent (123) completed the medication management trial.

     In addition, many of the children were comorbid for other
psychiatric diagnoses, so that the group of children diagnosed
solely with ADHD was much smaller than the total of 123 finishers.

     In my experience reviewing scientific studies, it is unusual
for the trial cohort to be such a small sample of the original
screened group. Furthermore, the study was not nearly so large
as touted, with only 124 finishers in medication management and
even fewer children diagnosed solely with ADHD in the medication
management group.

VII. The children did not rate themselves improved.

     The children self-rated themselves on an anxiety scale (the
MASC, Table 5 in the study). They did not rate themselves
differently in any treatment category at any time. Furthermore,
I received inside information that the children also rated
themselves on a depression scale. This is confirmed by a handout
provided by the Columbia project (New York State Psychiatric
Institute and Columbia University Division of Child & Adolescent
Psychiatry, 1994). However, no data are reported in the study
concerning the depression scale.

     Stimulants commonly cause depression in children. This
raises the question, "Were the depression self-rating scales
dropped because they indicated a worsening of the children's

VIII. Most of the subjects were boys.

 Boys represent a disproportional number of the children who
are medicated with stimulants. One reason is the demonstrated
usefulness of stimulants in suppressing overall normal spontaneous
behavior, thus making boys easier to manage under highly controlled
circumstances (Breggin, 1999a&b). In part to counter the argument
that stimulants are used for behavioral control of boys who are
being over-controlled in school, attempts have been made to
include more girls in studies. Despite efforts to recruit more
girls, 80% of the subjects were boys, reconfirming that stimulants
are used to suppress the relatively higher activity rates of boys
compared to girls.

IX. Use of the Placebo Washout

     Although it is not clear how it was conducted, a placebo
washout was used, and subjects who responded to placebo were
dropped from the study. This technique skewed the study to
favor drug effects.

X. Drug treatment was continuous for fourteen months; behavioral
treatments were stopped earlier.

     The MTA is not a comparison study of drugs and behavioral
treatments at 14 months, since in the last few months, behavioral
treatments were spaced out to once a month or stopped. Continuing
the drug treatment while discontinuing the behavioral treatment
gives the drug treatment an unfair advantage.

XI. The behavioral treatments were flawed.

     The study utilized behavioral treatments developed by Russell
A. Barkley. Barkley has used these techniques for decades to try
to show that drugs are better than behavioral treatments. He does
this by comparing drug treatment to his own type of behavioral
treatments. Barkley's behavioral approach is doomed to failure
because it treats the child as a defective object suitable for
control by parents and teachers rather than as a sentient being
in conflict with adults at home and/or at school. It ignores
everything that is known about family systems and the necessity
of changing the overall patterns of relationship in the family,
starting with the parents. Nonetheless, these limited behavioral
approaches did as well as all the other treatments, according to
the only "blind" observers (see above).

XII. Most children suffered from adverse drug reactions (ADRs).

     Sixty-four percent of children were reported to have some
ADRs, 11.4% moderate, 2.9% severe. The authors of the study
dismiss the severe reactions because 6 of 11 were in the category
of "depression, worrying, irritability." They explain these
"could have been due to nonmedication factors." In reality,
placebo-controlled double-blind clinical trials show depression,
worrying, and irritability are common stimulant ADRs (trials
reviewed in Breggin 1999a&b).

XIII. There were no trained observers for ADRS.

     ADRs were recorded on a two page check list by teachers
and parents. There was no apparent training for this process.
In addition, parents and teachers were reassured in writing that
the drug was safe and that ADRs were not serious, thus creating
a bias. Furthermore, many ADRs--such as behavioral suppression,
loss of spontaneity, apathy, and increased obsessive behavior--
are seen as improvements by parents and teachers. The use of
aware, experienced professionals, rather than parents and
teachers, is absolutely necessary in order to determine the
frequency and severity of ADRs (Borcherding, Keysor, Rapoport,
Elia, and Amass, 1990; studies reviewed in Breggin, 1999a&b).

     In clinical practice, I have found that asking children if
they are having adverse drug effects is a very important part of
the assessment. Often the child is having drug-related problems,
such as headaches or "blah" feelings, without realizing their
source and without telling anyone. The MTA study made no
specific effort to ask the children what drug effects they might
be experiencing.

     Overall, from the opening statement in the paper to its
conclusion, it is obvious that the investigators do not intend
to evaluate the single most important issue surrounding the long-
term use of stimulant drugs--the risks they pose to the children.

XIV. There was no improvement in academic performance.

     In a note to Table 4 (MTA Cooperative Group, 1999a,) the
authors of the MTA study admit there was no improvement or
difference in academic performance in spelling or math. The
table itself seems to indicate no improvement in reading as well.
Overall, no academic improvement was found as a result of any

XV. There was very little effect on social skills.

     Social skill differences among the groups were limited to
a significant difference favoring combined treatment over standard
community care. Neither was better than behavioral or medical
management treatment (MTA Cooperative Group, 1999a).

XVI. All the principal investigators were well-known drug advocates.

     How could so many experienced professionals produce such an
unscientific study? The framer of the studies, Peter Jensen (then
at NIMH), and all the principal investigators are avid drug
advocates who have been touting the positive results of the study
even before the study was completed or published. The six principal
investigators included Laurence Greenhill, C. K. Conners, William
Pelham, Howard Abikoff, James Swanson, and Stephen Hinshaw (MTA
Cooperative Group, 1999am, p. 1077). They have devoted their lives
to encouraging the concept of ADHD and the drugging of children.
Some, like Conners, have been doing so for four decades.

     Laurence Greenhill of the New York State Psychiatric Institute
and Columbia University represents the kind of conflict of interest
that exists among MTA researchers. Although it was later removed
amid controversy over research on children at the institute, the
New York State Psychiatric Institute and Columbia University website
listed the funding of its researchers as of December 21, 1998 (NYSPI
Sponsored Research, 1998). Greenhill had research funds or other
financial support from six drug companies: Richwood, Bristol-Myers,
Solvay, Wyeth-Ayerst, Glaxo, and Eli Lilly.

XVII. The parents and teachers were exposed to prodrug propaganda.

     The families and teachers were exposed to the pro-drug biases
of these investigators in the materials given to them before they
enrolled in the study. The "Teacher Information" for the MTA study
presents the usual claims about how much harm ADHD causes children
(NIMH MTA Study, undated, a). It states the children will be
treated with a "safe and effective dose of medication..." [bold in
original]. This kind of built-in bias has no place in a study that
used the teachers to rate safety and efficacy.

     The "Information for Parents" handout had similar built-in
biases, including a reference to biochemical imbalances and genetic
factors in "ADHD" (NIMH MTA Study, undated, a). In fact, based on
the information handouts given out by Columbia and the NYSPI for
the MTA, the parents in this study were not given informed consent
for the risks posed to their children by the drugs.

XVIII. Use of Intent-to-treat Analyses

      In order to achieve an apparent superiority of for medication
management over behavioral treatment and community care, the study
employed "intent-to-treat analyses" (MTA Cooperative Group, 1999b).
This method uses data from patients who have dropped out of the
study. For example, a patient who drops out because of adverse
drug reactions could be counted as successful if, on his or her
last visit, a positive evaluation was reported. This misleading
method is used when an analysis of the actual finishers of a study
fails to obtain the result that is sought after by the investigators.


     The MTA study has been highly promoted by advocates of drug
therapy as a demonstration of the superiority of stimulant treatment
for ADHD. In fact, the study failed to meet the basic scientific
criteria for a drug trial. It was not placebo-controlled and in
fact had no non-treatment control group. It was not double blind.
Teachers and parents provided the ratings relied upon by the study,
but both groups knew whether or not the children were taking
medications. The MTA study was "open label" and would not have
qualified, for example, as a study for the FDA-approval process.
As research, it is scientifically unsound.

     Furthermore, the blind raters in the study who observed
behavior in the classroom found no differences over fourteen
months in any of the treatment conditions. In other words, the
observations generated by the most objective observers confirms
that there was the medicated children did not better than the
other children in the study. Unfortunately, the MTA study also
failed to examine the kind of interventions that, in actual
clinical practice, prove very effective in helping children
labeled with ADHD. These interventions include individualized
family counseling aimed at improving relationships in the family
and individualized educational approaches that inspire children
to engage themselves in school (Breggin, 1998, 2000).

     In summary, the MTA study failed to adhere to basic standards
for scientific studies of medication efficacy and cannot be used
to draw any valid conclusions. Furthermore, the data it generated
tends to indicate that stimulant medication produced no different
results than any of the other intervention. The MTA study does
not demonstrate the superiority, or even the usefulness, of
stimulant medication in the treatment of children labeled with
ADHD or any other presumed psychiatric disorder.


 Borcherding, B.V., Keysor, C.S., Rapoport, J.L., Elia, J., and
Amass, J. (1990). Motor/vocal tics and compulsive behaviors on
stimulant drugs: Is there a common vulnerability? Psychiatric
Research, 33, 83-94.
 Breggin, P.R. (1998). Talking back to Ritalin. Monroe, Maine:
Common Courage Press.
 Breggin, P.R. (1999a). Psychostimulants in the treatment of
children diagnosed with ADHD: Risks and mechanism of action."
International Journal of Risk and Safety in Medicine, 12 (1),
3-35, 1999.
 Breggin, P.R. (1999b). Psychostimulants in the treatment of
children diagnosed with ADHD: Part I: Acute risks and psychological
effects. Ethical Human Sciences and Services, 1, 13-33.
 Breggin, P.R. (2000). Reclaiming our children: A healing solution
for a nation in crisis. Cambridge, MA: Perseus Books.
 Fisher, S., and Greenberg, R. P. (Eds.) (1989). The limits of
biological treatments for psychological distress: Comparisons with
psychotherapy and placebo. Hillsdale, New Jersey: Lawrence Erlbaum
 MTA Cooperative Group. (1999a). A 14-Month randomized clinical
trial of treatment strategies for attention-deficit/hyperactivity
disorder. Archives of General Psychiatry, 56, 1073-1086.
 MTA Cooperative Group. (1999b). Moderators and mediators of
treatment response for children with attention-deficit/hyperactivity
disorder: the multimodal treatment study of children with attention-
deficit hyperactivity disorder. Archives of General Psychiatry,
56, 1088-1096.
 MTA side effects rating scale--parent. (undated). Obtained from on
March 1, 1999.
 MTA side effects rating scale--teacher. (undated). Obtained from on
March 1, 1999.
 National Institute of Mental Health. (undated). Multimodal
treatment study for attention-deficit hyperactivity disorder (the
MTA study). Obtained from
on March 1, 1999.
 New York State Psychiatric Institute and Columbia University
Division of Child & Adolescent Psychiatry. (1994). Grand Rounds:
The multimodal treatment study of attention deficit hyperactivity
disorder (MTA Study)--An NIMH cooperative agreement grant.
(unpublished; distributed on March 9, 1994).
 Nies, A.S., and Spielberg, S. P. (1996). Principles of
therapeutics. In Hardman, J.G., and L. E. Limbird (Eds.),
Goodman & Gilman’s the pharmacological basis of therapeutics,
ninth edition, pp. 43-62. New York: McGraw-Hill.
 NIMH MTA Study. (undated, a). Information for parents. Obtained
on March 1, 1999.
 NIMH MTA Study. (undated, b). Teacher information. Obtained
on March 1, 1999.
 NYSPI Sponsored Research. (last update, 1998, December 21).
Research foundation for Mental Hygiene, Inc., Psychiatric
Institute Division. Obtained at
on March 1, 1999.

     1. Peter R. Breggin, M.D., Director, International Center
for the Study of Psychiatry and Psychology, 4628 Chestnut
Street, Bethesda, MD 20814. A version of the report will be
made available at

     2. I wish to thank members of the International Center
for the Study of Psychiatry and Psychology (ICSPP) who took the
time to share their views with me. Ginger Breggin provided
critical background research. My assistant, Ian Goddard, made
many trenchant observations.

 Author's Note: In the public interest, this paper may be
reproduced, provided attribution to the author is given.

Peter Breggin's report posted:
  Amphetamine Use Linked To Cognitive Impairment:

  Ritalin Induces Schizophrenic Psychopathology:

  Peter Breggin's Letter To JAMA On Ritalin:

  Can Child-Ritalin-Use Violate Anti-Nazi Law?:

  Excerpts From Breggin's Book on Ritalin:



This archive was generated by hypermail 2b29 : Thu Jul 27 2000 - 14:02:45 MDT