In this issue...

The effects of reward and punishment on the performance and motivation of children with ADHD

A behavioral strategy to control impulsivity

Does ADHD predict poor driving outcomes?

One of the more frustrating and confusing aspects of ADHD for parents is how their child's performance and motivation for tasks can seem to fluctuate so dramatically from one day to the next.  Homework can go relatively smoothly on one night and then seem close to a total nightmare the next.  One way of trying to deal with this issue is through a combination of rewards (i.e. receiving tangible notice for doing something well) and response cost (i.e. taking something away for the failure to perform as expected - this is akin to a mild form of punishement). What is the impact of such procedures on the motivation and performance of children with ADHD?

This interesting and important question was addressed in a study published recently in the Journal of Consulting and Clinical Psychology  (Carlson, C., & Tamm, L. (2000). Responsiveness of children with ADHD to reward and response cost: Differential impact on performance and motivation. JCCP, 68, 73-83.)  In this study, the authors were interested in examining 3 fundamental questions that have important implications for the use of behavioral interventions with ADHD children.  These questions were:

1. Compared to children without ADHD, how is the performance of children with ADHD affected by reward (R) and response cost (RC)?

2. Do the affects of reward or response cost vary for children with ADHD depending on whether they are engaged in a high- or low-interest task?

3. Does the use of  either reward or response cost negatively affect the motivation of children with ADHD?  

Participants in this study were 44 children between the ages of 8 and 10.  Twenty-two had been diagnosed with ADHD, Combined Type (i.e. they had both inattentive and hyperactive/ impulsive symptoms) and 22 did not have ADHD and served as control subjects.  Approximately 30% of the children in each group were girls.  Children diagnosed with ADHD for this study had been carefully and systematically evaluated using a variety of standardized behavior rating scales and clinical interviews.  Although many of the ADHD children were being treated with medication, all tasks completed during the study occurred after an 18-hour medication "wash out" period. Because the effects of stimulant medications typically do not persist beyond 4-8 hours, one can safely assume that the results reported below would not have been influenced by a child's being on medication.  There were no differences in IQ between children with ADHD and control subjects.

In this study, children participated in 2 separate computer tasks.  The first, called Crazy Symbols, was purposely designed to be of low interest.  In this task, the child was presented with a set of 5 different individual symbols on the computer screen and was required to match this to one of eight choices of figure sets positioned circularly around the target figure.  The child made a selection by clicking the mouse on the selected figure set and received immediate auditory feedback from the computer about whether his or her choice was correct.  After each choice, the child pressed a key to initiate the next trial.   As noted above, this task was purposely designed to be relatively boring to children and it was expected that most participants would find it to be uninteresting.

The second computer task was called Jet Pack.  In this task the child used the keyboard direction buttons to maneuver a figure around a maze, with the goal of "picking up" the correct key from a display of keys of various colors.  For each trial, the child was presented with a particular colored key set at the bottom of the screen to indicate which color key was supposed to be picked up.  Once again, the child received immediate auditory feedback from the computer when they picked up the correct key.  In contrast to the task described above, the Jet Pack task was designed to be interesting and engaging to children in the study.

During the study, children participated in these tasks under several different conditions.  In the reward condition, they were given the opportunity to earn dimes for every trial they completed successfully. Forty different trials were allowed providing the opportunity to earn up to $4.00.  In the response cost condition they were given $4.00 in dimes to begin with and a dime was taken away each time the child solved a problem incorrectly or failed to solve it correctly within the allotted time.  Finally, there was also a no contingency condition in which children completed the same tasks without receiving any feedback from the computer or reward for their performance.  These tasks were completed during separate sessions and the order - of both the tasks and the contingency condition - was carefully balanced to insure that any differences found could not be attributed to order of administration.

By setting up these contingencies for the two different computer tasks, the experimenters created conditions that enabled them to carefully examine the impact of rewards and response cost on the problem-solving performance of children with ADHD during high- and low-interest tasks.  At the end of each task, children were also asked to provide ratings of how much they enjoyed each task, how well they thought they had performed, and how much they would like to do the task again.

Results

The major findings of this study were as follows:

The use of both reward and response cost contingencies had a significant and positive impact on the performance of children with ADHD.  In contrast, the performance of children without ADHD was not similarly affected.

The impact of either contingency on ADHD children was really quite dramatic.  For example, during the Crazy Symbol task, the % correct response for ADHD children were just over 90% in either the reward or response cost condition and below 70% when neither contingency was in place.  They also made an average of less than 2 mistakes during the reward or response cost condition compared to 10 mistakes during the no contingency condition.  Similar differences were found for the Jet Pack task, although these were not as pronounced, probably because this was a more inherently interesting task for the children.  For the control children, in contrast, performance during the reward, response cost, and no contingency condition was essentially identical.

There were some differences in the impact of reward and response cost contingencies on ADHD children that are worth noting.  Children with ADHD got fewer problems incorrect in the response cost condition than in the reward condition.  Thus, when faced with the loss of a dime for an incorrect response, they made significantly fewer errors than when they would simply gain a dime for a correct response.

Overall, children with ADHD performed significantly worse than control children when no contingency was in place. During the reward or response cost condition, however, their performance on the two computer tasks was essentially equal to that of the control children.  As noted above, however, the response cost contingency appeared to produce the greatest gains in the performance of children with ADHD.

The use of either reward or response cost contingency did not have an adverse impact on the motivation of either children with ADHD or control children.

When the idea of using either reward or response cost programs to help manage the behavior of children with ADHD is introduced, some parents and educators are concerned about providing this "artificial" source of motivation.  A frequent question is why should a child be rewarded for something he or she should be doing anyway? A flip-side of this concern is that by introducing any form or "artificial" motivation (i.e. rewards contingent on the child's performance), the child's natural or inherent motivation for the task will be adversely affected.

In this study, however, there was no indication at all that the use of either contingency had an adverse affect on the child's motivation to engage in the task.  In fact, all children tended to report that they enjoyed the games more when they had the opportunity to earn rewards; this seemed to be especially true for the children with ADHD.  Of  course, one can not generalize from this relatively short experimental task to what would happen over an extended period in the real world, and one does need to pay attention to whether the use of reward contingencies adversely affect a child's own motivation for a task.

There is, however, an important qualifier to this general conclusion that is important to note.  When the authors looked at the motivation and task liking for the ADHD children alone, they did find evidence that the response cost procedure may reduce motivation and task liking in certain conditions.  Specifically, during the less-interesting computer task (i.e. Crazy Symbols), although response cost was associated with slightly better performance for children with ADHD, it also led to less enjoyment of the task and reduced motivation to perform it again.

Summary and Implications

Overall, the results of this study provide a careful demonstration of how the systematic use of contingencies can enhance the task performance of children with ADHD.  In fact, in this study, the use of either reward or response cost improved the performance of children with ADHD to the point that it no longer differed significantly from children without ADHD.  This does not mean, of course, that findings from an experimental study such as this can be translated immediately into the classroom or home environment (i.e. that such procedures will "normalize" the behavior and performance of children with ADHD), but they do emphasize what an important and powerful tool they can be.

What does this study tells us about the use of reward and response cost contingencies in the behavioral treatment of children with ADHD?   First, it is important to be clear about exactly what these two contingencies involve: providing a reward for a desired response vs. removing something the child values when he or she behaves in a non-desired way.

Recall that in this study, response cost was found to enhance the performance of children with ADHD (i.e. they made fewer errors) relative to the reward condition. During the low interest task, however, this came at the expense of reducing liking of the task and motivation to engage in it again relative to the reward condition. Thus, in some circumstances, the beneficial affects of response cost on performance may need to be weighed against the adverse affects - relative to a reward contingency - on enjoyment and motivation.

This has potentially important clinical implications as it suggests that response cost - although more effective - may be less palatable to children with ADHD than reward. If behavioral interventions involving the use of response cost are less acceptable to children, they may be less consistently implemented by parents and teachers and be decreasingly effective over time.  This may vary, of course, from child to child, but it does argue for the importance of considering both effectiveness and desirability in designing treatment programs.

For example, imagine setting up a behavior plan designed to reduce the likelihood of careless mistakes on school assignments.  The reward approach would simply involve providing an incentive for every problem answered correctly. The response cost approach would involve taking something away for every problem on which a careless error was made. A combination approach would involve providing a reward for each problem answered correctly, and taking something away for each careless error.  For children without ADHD, doing well on the assignements may be a sufficient "reward" to promote careful and conscientious work.  For a child with ADHD, however, it is often necessary - as the results of this study demonstrate - to include more tangible rewards to promote good performance.  (Note: These incentives do not necessarily have to be something tangible like the dimes that were used in this study.  Simply rewarding points for pre-specified behaviors that can be accumulated and later traded in for more "concrete" rewards - e.g. "buying" access to television or computer time - can work very well for many children.)

The data from this study also suggest that the response cost approach will actually yield better performance for most children with ADHD (i.e. fewer careless mistakes).  If, however, a child is especially prone to become upset when a desired object is removed, then this added benefit may not be worth it.  This is something to carefully consider when thinking about developing a behavioral intervention for your child and designing such an intervention in consultation with a child mental health professional.

---

Although not every child or adult with ADHD is impulsive (i.e. those with the predominantly inattentive subtype), impulsive behavior is one of the most problematic aspects of this condition for many individuals.  For children, this frequently takes the form of talking out of turn and blurting out answers without first raising one's hand.  This can be an enormous source of frustration to parents and teachers alike.

A study appearing in a recent issue of Behavior Modification describes a procedure for dealing with this specific symptom - to get children to raise their hands before talking (Posavac et. al., (1999). A cueing procedure to control impulsivity in children with ADHD. Behavior Modification, 23, 234-253). Although this is a small-scale study - actually almost a case study involving 4 different individuals - I wanted to present it because it gives a good overview of behavioral treatment procedures and exemplifies the assets and liabilities inherent in this approach.  As such, it has important implications for understanding the use of behavioral interventions in general.

Participants in this study were four 8-year-olds diagnosed with ADHD, Combined Type (i.e. they had both the inattentive and the hyperactive/impulsive symptoms) who were participating in an 8-week outpatient summer treatment program.  These particular participants were selected for the study because they all identified similar target behaviors (i.e. low rates of hand raising and high rates of talking out of turn).  As part of their summer treatment program, all boys participated in daily social skills training groups that contained a total of 9 boys.  It was in the context of these social skills groups, that the behavioral intervention to increase the boys' hand-raising behavior was implemented. (Although 9 boys were in the social skills group, the behavioral intervention to increase hand-raising behavior was implemented with just 4 of the boys).

Note: The notion of clearly identifying a behavior to change - i.e. the target behavior - is a important concept within behavioral therapy.  The idea is to identify a specific behavior, or behaviors, that one wants to enhance through the intervention and to be clear about exactly how that behavior is defined.  In this study, the target behavior was clearly defined as "raising one's hand in the group before talking". One can then observe how frequently the behavior is occurring prior to treatment (i.e. before the intervention was implemented, how often each boy raised his hand in the group before talking) and how often did it occur after the intervention began.  By comparing these two measures, one has an "objective" account of whether the intervention is affecting the target behavior in the desired way.

Also note that the target behavior was selected so that the intervention involved increasing something positive - i.e. raising one's hand - rather than decreasing something negative - i.e. talking out of turn.  In general, one tries to design behavioral interventions to increase the child's rate of engaging in desired behaviors rather than reducing the rate of negative behaviors.  The latter happens naturally as the more desirable behaviors increase in frequency.  The reason for designing interventions this way is that it requires the child to do something active and prosocial.  Not doing something negative, in contrast, does not necessarily require active prosocial behavior from the child.

Finally, the target behavior was selected to be one that has potentially important ramifications for the child's functioning more generally.  In this particular study, targeting "raising one's hand" was intended to have the effect of making the boys less impulsive, thus enabling them to participate in the social skills group more effectively overall.

One problem that parents can often have in trying to use behavioral plans themselves is they are not clear or specific enough about what the "target behavior" is.  For example, trying to get your child to "cooperate" is certainly a laudable goal, but can be problematic from the perspective of a behavioral treatment approach because "cooperate" is not specifically defined and can mean lots of different things. As a result, the parent may be somewhat inconsistent in the exact behaviors they try to target for reinforcement and the child may not know exactly what he or she is supposed to do in order to "cooperate".  Instead, if the target behavior is defined as "Complying with requests within 1 minute of when they are made" - a more specific although limited definition of what it means to cooperate - one can much more easily monitor whether the targeted behavior is being demonstrated.

In this study, the authors developed what they called a "cueing procedure" that was specifically designed to increase the frequency with which the boys raised their hands before speaking. Four measures of "cueing" were included in this procedure:

1. Visual reminder - During each social skills group, the boys wore a badge on their shirt on which the target behavior of raising one's hand before talking was recorded.  The logic behind this type of strategy is that children with ADHD have difficulty using internally generated rules to guide their behavior and thus need to be provided with as much external structure and reminders of those rules as possible.

(I remember an interesting example of this from my own practice.  An eight-year-old boy I was treating who had been doing quite well in school suddenly began to show a real increase in behavior problems and was having his name "put on the board" with a noticeable increase in frequency.  When asked if he had any idea for why he was starting to have more trouble, he replied that he did - the teacher had moved his seat so that the flag was now blocking his view of the poster that showed the class rules, so he was no longer able to look at the poster as a reminder of what to do. When his seat was moved again to provide him with an unobstructed view of the rule poster, his behavior improved immediately.)

2. Goal Evaluation - A timer was set to go off every 5 minutes and each child would then be individually evaluated as to whether the behavioral goal had been met.  Each child evaluated his own behavior, peers evaluated each other's behavior, and the group leader considered these determinations and then made the final evaluation as to whether or not each child had met the behavioral goal (i.e. raising hand before talking in the group).

Note: This also illustrates one of the fundamental aspects of behavioral interventions for children with ADHD - frequent feedback about how they are doing.  Children with ADHD tend to have difficulty using goals to help guide their behavior over long periods of time (e.g. telling a child he/she will get a reward for a good week in school is unlikely to be effective because a week is too long a period to expect most school-age children with ADHD to be able to focus on a goal and use it to direct their behavior).  So, time needs to be broken up into much shorter intervals with feedback about progress towards the goal provided at the end of each interval.  This helps to keep the goal fresh in the child's mind so it can be used to help direct his or her behavior during the relatively short time span.

Of course, this also illustrates why effective behavioral interventions can be so difficult to carry out.  Clearly, one can not provide feedback about the child's behavior and progress towards goals over such brief intervals for an extended period of time.  While this procedure may have been practical to implement in a 50-minute social skills training group, it would be next to impossible to implement over the course of an entire school day.  It does illustrate what some would consider to be the ideal, although modifications to make such a procedure practical for use in real world settings are obviously important and necessary.

3. Rewards for Goal Attainment - Rewards for desired behavior are a key aspect of virtually all behavioral interventions.  In this study, when the child was determined to have met the goal, he was rewarded with praise and with a large sticker that was publicly posted.  If the goal was not attained but progress was made, a smaller sticker was posted.  When no progress was made, the group leaders discussed with the child what needed to be altered in their behavior to attain the goal the next time.  

Results

As one might expect, this intervention had clear and dramatic effects on the frequency with which the boys raised their hands in the group before talking.  In fact, for each of the 4 boys, the rate of hand-raising before talking at least doubled - for some boys, the increases were even more dramatic.  As one would expect, the corresponding behavior of talking out of turn showed a large decrease when the intervention was in place.

Other aspects of these results are less encouraging, however, and illustrate the limitations inherent in programs such as these.  First, when the intervention was withdrawn, the rate of the desired behavior returned to pre-treatment levels, as did talking out of turn.  In other words, the intervention did not seem to produce any lasting change in the targeted behavior.  Of course, the same is true for medication - after it has cleared out a child's system, levels of symptoms typically return to their prior level.

Second, even when the treatment was in place and having the desired effect, it did not seem to generalize to other aspects of the boys' behavior.  That is, parents did not report any improvement in their child's behavior at home.  This is one of the most important limitations of such narrowly focused behavioral treatments - they generally have a strong effect on the specific behavior being targeted but tend to have little or no impact on behavior that is not a specific focus of the treatment.  In addition, even for the behavior being targeted, the results do not necessarily generalize to settings other than the one where the intervention was implemented.

For this reason, it is important to try and target behaviors that are really important for the child's functioning.  I usually suggest that parents think about 2-3 things that really matter - i.e. the kinds of behavior where an improvement would result in a meaningful difference for the child and family, even if lots of other things were still staying the same.  For example, if a child is showing high levels of aggression towards peers, than targeted and reducing this via behavioral treatment can make a real difference in a child - and parents' - life, even if he or she still does a number of annoying things like forgetting to turn out lights or put away clothes.  The bottom line is to focus on the stuff that really matters and learn to live with what is less important.  

Summary and Implications

As noted above, this study is useful because it provides a nice illustration of traditional behavioral treatment and highlights both the strengths and weaknesses of this approach.  It also highlights the care and effort that needs to be put into designing and implementing a behavioral intervention.  Such interventions can be enormously helpful in managing the behavior of children with ADHD, either alone or in conjunction with medication treatment. For children receiving medication, they can be essential in targeting problem areas that remain even if medication is providing a number of important benefits.

For a more general discussion of behavioral treatment for ADHD, click here.

---

I've been told that having one's child begin their career as a driver is a parent's nightmare under the best of circumstances. Fortunately, I don't have to deal with this myself for a few more years - can't wait.  For parents whose child has ADHD - or at least significant levels of attentional difficulties - these concerns can be even greater.  The thought of having their inattentive, impulsive child behind the wheel of a car is something that many parents I have worked with described as "nerve wracking".

What do we know about the relation between ADHD symptoms and driving difficulties in adolescents/young adults?  Is this as bleak a circumstance as many parents imagine?  This was the focus of a very interesting study that appeared recently in the Journal of the American Academy of Child and Adolescent Psychiatry  (Woodward, L.J., Fergusson, D.M., & Horwood, J. (2000). Driving outcomes of young people with attentional difficulties in adolescence. JAACP, 39, 627-634).

Participants in this study were 1265 New Zealand children who had been followed from birth into young adulthood.  (This is obviously an enormous undertaking and many, many issues besides the topic of the current report were examined using this cohort of children).  For this study, the researchers obtained parent and teacher reports of children's attentional difficulties and hyperactive behaviors when they were 13 years old and created a score for each participant by averaging the parent and teacher ratings.  Then, children were placed into one of five groups based on the severity of their "attention problems" score.  So, only a small percentage of the 1265 children would have been considered to have had ADHD.

Driving outcomes for these individuals were obtained when they were 21 based on detailed interviews with each participant. From the original sample of 1265 adolescents, 941 were able to be interviewed about their driving history at age 21.  Three different types of adverse driving outcomes were inquired about.

First, participants were asked whether they had been involved in an accident that produced injury (5% of the total sample had) and/or a non injury producing accident (38% of the sample had).

Second, participants were asked about whether they had engaged in drinking and driving.  This included questions about whether they had ever driven a motor vehicle while drunk (26% said they had), while seriously intoxicated/very drunk (10% said they had), or ever been arrested on a drunk driving charge (5% said they had).  In addition, 10 other questions related to drinking and driving were asked and an overall "drinking and driving" score was computed based on each participant's response to these items.

Finally, participants were asked about their history of traffic violations.  For each year between 18 and 21, participants were asked whether they had ever: 1) driven without a license (14%); 2) driven without a proper vehicle registration (40%); or 3) participated in street racing (10%).  In addition to these items, an additional set of items pertaining to dangerous driving behavior (e.g. speeding, driving through a red light) were also asked about, including questions about the frequency of such dangerous driving habits (choices ranged from never to nearly every day).

The basic question of this study is whether high levels of ADHD symptoms at age 13 - as indicated by the combined parent and teacher ratings - predicts more adverse driving outcomes during the 18-21 year age period.  Of course, because ADHD often occurs in the context of other difficulties (e.g. conduct problems), and these other factors could possibly account for any relationship between ADHD symptoms and poor driving outcomes that were found, it is important in a study like this to try and control for this possibility.

Fortunately, the authors were careful to do this.  Thus, in addition to ratings of ADHD symptoms on all participants at age 13, they also collected data on other child characteristics that could be related to young adult driving outcomes including conduct problems (also based on parent and teacher ratings), IQ, and, of course, gender.  In addition, data was collected on measures of parenting and family functioning (e.g. maternal disciplinary style; child's exposure to parental separation, divorce, or death).  Social background characteristics to provide an indication of the child and family's standard of living were also obtained.

The last piece of information collected dealt with each individual's driving experience at age 21.  Specifically, each young adult was asked about the number of months he/she had held a driver's license of any kind as well as their estimate of the total distance they had driven.  This is important because those who driven more would have greater opportunity to have an accident.

Results

So, do ADHD symptoms at age 13 predict more adverse driving outcomes in young adulthood?  The answer is not quite as clear-cut as you might expect.

The authors first compared driving outcomes for the sample divided into the 5 groups that ranged from very low to very high on the ADHD symptom ratings at age 13.  In this analysis, no attempt was made to control for the other factors (e.g. conduct problems, family environment) that they had thought might be important.

When examined in this way, the results are striking: on every single driving outcome measure, subjects in the highest ADHD symptom group look worse than those without such problems.  For example,  those in the highest group were:

* about 3 times as likely to  be involved in an accident than those in the lowest 2 groups;

* about 2.5 times more likely to report having driven when seriously intoxicated;

* about 3 times as likely to have been involved in street racing;

* to have an overall traffic violation score that was more than twice as high.

In general, the pattern seemed to be one of increasingly adverse driving outcomes across the 5 groups.  From this perspective, it would appear that ADHD symptoms do, in fact, predict a wide variety of problematic driving outcomes and that parents' concerns about their ADHD child and driving are well-founded.

Next, however, the authors examined these same outcomes across the 5 groups after all the other factors they had collected (e.g. conduct problems, IQ, family environment factors, driving experience) were taken into account.  What this type of analysis enables one to learn is what is the independent contribution of age 13 ADHD symptoms to young adult driving difficulties after all these other characteristics have been accounted for.  This tells you, in essence, whether poor driving outcomes can be explained by the independent contribution of ADHD symptoms to driving difficulties, or whether the predictive relationship that was found really occurred because of these other factors.  For example, we know that ADHD and conduct problems often co-occur. One can also imagine how children with serious conduct problems, regardless of whether they also have ADHD, might be destined to become dangerous drivers.  If this is true, than children with ADHD and conduct problems would be found to have poorer driving outcomes, but this might be entirely because of the conduct problems and have nothing really to do with the child's ADHD.  In such a case, concluding that ADHD predicts poor driving in young adulthood would be erroneous.

So, what did this more sophisticated analysis yield?

For the most, part, it appears that the independent contribution of ADHD symptoms to young adult driving difficulties is relatively modest.

Overall, for 8 of the 11 driving outcome measures considered, ADHD symptoms at age 13 no longer predicted poorer outcomes after controlling for other relevant factors.  The 3 outcomes that continued to be predicted by ADHD symptoms were driving without a license, general traffic violations, and risk of a motor vehicle accident involving injury.  In addition, it was only the latter outcome that remained significant after appropriate statistical corrections were made.  Involvement in a motor vehicle accident that results in injury is, of course,  a very important outcome.  In fact, one could reasonably argue that it is the most important outcome of all those that were considered.

If ADHD symptoms did not independently predict the vast majority of adverse driving outcomes, what factors did?  This varied somewhat across the different outcomes, but in general:

* boys had consistently worse outcomes than girls;

* children with more extensive conduct problems at age 13 tended to have poorer driving outcomes;

* driving longer distances predicted more adverse driving outcomes;

* having one's license for fewer months predicted more adverse outcomes;

Here is something I found particularly interesting. Overall, those with the highest ADHD scores at 13 reported having their license for the least amount of time - almost a year less on average than those in the low attention problem group - but also reported having driven the greatest distance - about 30% more than the low attention problem group.)

Summary and Implications

Although this study indicates that many of the apparent associations between ADHD and driving problems reflect the presence of associated difficulties, the sobering conclusion that remains is that ADHD symptoms during young adolescence does dramatically increase the risk of being involved in an injury producing motor vehicle accident during young adulthood.  In fact, after controlling for a number of possible confounding factors, those in the highest ADHD symptom group were over 4 times more likely to be in such an accident than those in the low group (14% vs. 3%) and twice as likely as those in the moderate group (14% vs. 7%).  Parental caution and concern is thus quite warranted.  When high levels of ADHD symptoms are combined with being male, having significant conduct problems, and spending extensive time driving despite relatively limited experience, the risks of involvement in a serious accident increase substantially.

No specific recommendations for dealing with this concern are provided in this paper.  Obviously, however, doing one's best to make sure that their young adult child's ADHD symptoms are being managed well should reduce the risk of driving difficulties. This can be quite a challenge, however, when one's "child" is an independent adult who may refuse and deny any appropriate interventions.  Wish I had some answers here, but I don't - this is just a difficult issue that probably depends to a great extent on how well a child's condition has been managed and discussed over many years.

One limitation of the study is the authors did not differentiate the driving risk that accrues from inattentive vs. hyperactive impulsive symptoms.  It is easy to imagine how both types of problems would contribute to serious motor vehicle accidents, but it would have been very nice if the authors had been able to evaluate the relative contribution of each type of symptom.

---

Neurofeedback treatment - also known as EEG biofeedback - is an alternative approach for treating ADHD and other types of attentional problems that many regard as a promising - but still unproven - treatment.  In this treatment approach, individuals are trained to produce the patterns of brain activity that are associated with better-sustained attention and lower impulsivity.  Over the years, a number of published reports have suggested that this approach has positive results for many children with ADHD, and that these benefits are sustained over time.  Critics contend that these studies have suffered from a variety of methodological problems, however, making it impossible to make any definitive statement about the utility of this treatment approach for ADHD.  This treatment approach has also been criticized because it is typically time consuming and expensive.

A study recently summarized in Clinical Psychiatry News (April, 1999) reports on a large-scale study of neurofeedback treatment that appears to address at least some of these concerns.  (Note: Clinical Psychiatry News is a trade publication for the psychiatric profession and not a peer reviewed journal.  In this publication, studies that have been published or presented at professional meetings are reviewed. The study reviewed in this article was conducted by Dr. Siegfried Othmer and presented at the recent annual meeting of the American Association for the Advancement of Science.)

In this study 1089 children and adults (726 children and 363 adults) with either formally diagnosed ADHD and/or other "attentional and behavioral disorders" received at least 20 sessions of neurofeedback therapy across 32 different psychology clinics and private practices.  Approximately two-thirds of the participants were male and just under 20% had received a formal diagnosis of ADHD.  Diagnostic information on the remaining participants was not specified, although it was noted that most had previously been treated for attentional disorders without success.  None of the participants were receiving stimulant medication or anti- depressant medication during the study. (Note: The lack of diagnostic specificity is one methodological problem with this study.)

Specific details of the neurofeedback training were not provided.  Typically, however, this involves training individuals so that they become able to produce patterns of brain activity that are associated with better sustained attention and reduced impulsivity.  The idea is that by teaching an individual with ADHD to be able to produce these patterns, they will become better able to regulate their own attention level and to be less impulsive.

Results

After receiving the 20 sessions of neurofeedback training, 85% of the subjects were reported to show improvement in their ADHD symptoms.  Improvement was measured by subjective reports (i.e. what the participants themselves felt), clinical evaluations, and computer-based tests of sustained attention. In addition, a subset of 157 participants who continued their neurofeedback treatment over an additional 20 sessions were reported to demonstrate continued improvement.  Information on how long the reported improvements were sustained was not provided in the study summary.

These results are impressive - especially considering that all the participants were reported to have previously tried - and failed to be helped - by more conventional treatments such as stimulant medication.  Given the extremely large and diverse nature of the sample, these results would appear to indicate that neurofeedback treatment has the potential to be a useful treatment approach - either alone or in conjunction with other treatments - for many individuals with ADHD or other types of attentional difficulties.

On the other hand, critics will rightly point out that absence of specific diagnostic information on participants, and the reliance on ONLY subjective reports and lab-based measures (i.e. computerized tests of sustained attention) to assess improvement is problematic.  Although documenting improvement on such outcome measures is important, it is no substitute for clearly demonstrating meaningful improvement in important real-world settings.  For example, it would be much more impressive to have shown that teachers - especially teachers who did not know the child was receiving treatment - reported significant improvement in the academic and behavioral functioning of children receiving the neurofeedback treatment. It is this type of data collected in carefully designed studies that is need to conclusively establish the efficacy of neuro- feedback treatment.  The author of this study appears to be well aware of this need, however, and will hopefully be providing such data down the road.

Summary and Implications

What can be concluded from a report such as this?  I think a balanced view is that the study contributes to a body of literature in which neurofeedback treatment emerges as a promising treatment approach for many individuals with ADHD.  This study is impressive because of its large and diverse sample, and the fact that an apparently positive response was obtained by individuals who had reportedly not responded well to prior treatment approaches.  It is unfortunate that some of the limitations noted above prevent more definitive conclusions from being reached, but one hopes that we will soon be seeing other studies of neurofeedback treatment in which such problems are addressed.

In the meantime, it is important to emphasize that should you consider neurofeedback treatment for your child, the key is to carefully monitor how your child is doing in school.  The fact that he or she may show improvement on computerized tests of attention doesn't really matter much if behavior and academic performance at school are not also changing for the better.  The need for this careful monitoring, with adjustments made to a child's treatment as needed, of course, is the key to promoting a child's long-term success regardless of what treatments a child is receiving.