POOGI for ABA

Suggested Criteria for the Treatment of Problem Behaviors

I believe that one of the hardest problems in the treatment of problem behaviors is identifying the criteria for success. On the one hand, we want to ensure the long-term success of our clients. That implies we should set very high standards for defining success. On the other hand, should we expect perfection? Don’t typically developing children have a tantrum once in a while? Are we being unreasonable? I’m skeptical of many of the criteria that are frequently used. Many of them are necessary, but not sufficient. In my view, you need to achieve the following seven outcomes in order to demonstrate successful treatment that is likely to make a real difference in a person’s life:

1. The behavior problems have been substantially reduced: OK, maybe it doesn’t have to be perfect, but if we didn’t reduce the behavior problems at all, it is hard to argue that we did very much. In addition, we need to collect these types of data to direct daily treatment decision making. Some research articles define “successful” treatment as a reduction in problem behaviors by a certain percent. As a practitioner, that doesn’t make any sense to me. We know it is possible to dramatically reduce problem behaviors and still have huge problems.

2. Appropriate alternative behaviors have been increased: Sure, it is great that the problem behavior is reduced. But if you haven’t increased appropriate alternative behaviors, it is very hard to see how it would last over time without maintaining the program forever. In addition, no practitioner decides that people doing nothing is an acceptable goal. If all that happened is the adults have been shaped to avoid problems, like the parent who tells his child he doesn’t have to do any homework, most people wouldn’t consider that very successful.

3. The treatment is occurring under practical conditions: Sure, it is probably necessary to treat problem behavior under specialized conditions some of the time. But eventually, we have to return to practical conditions. If the behavior change has only been demonstrated for in 5-15-minute sessions or under very special conditions, you aren’t there yet.

4. The behavior change maintains over time: Everyone is happy that the severe problem behavior has been successfully treated. But the hard part is maintaining the change over time. This is a frequent problem. Who cares if the behavior was substantially reduced, only to come back later?

5. The program has a high level of social validity: It is essential that the stakeholders have high levels of support for the goals, procedures, and outcomes of the program. If not, they are not like provide the necessary support over time. Of course, it is possible to have high social validity with all of the above criteria met, yet still have significant problems.

6. The appropriate alternative behaviors are occurring under natural communities of reinforcement: It is wonderful if we have achieved a high level of success with behavioral intervention. But if the reinforcement doesn’t occur under natural conditions, it is likely a short-term success. No one can maintain a behavior change forever with contrived contingencies. Eventually, it almost always becomes impossible to maintain behavior changes if they don’t occur naturally. I believe the only long-term possibility is natural reinforcement.

7. The person is flexible enough to change their own behavior when the situation changes– As frustrating as it is, sometimes even a natural contingency is not enough. In the real world, things change all the time. If the person can’t change with them, the behavior problems are likely to come back.

Of course, there are some clients with some behavior problems where all of these criteria are not possible with the current state of knowledge in the field. But if you don’t have a good idea of what success looks like, it is hard to Poogi.

If fast food workers are taught to say “would you like fries with that?” and reinforced for doing so, people really do buy more fries. The staff are satisfied with the program, people buy more food, and the business makes more money.

There is a lot of evidence that reinforcement systems can make a large difference in staff performance. Certainly, if you provide reinforcement for staff to engage in certain behaviors, you will likely see increases in those behaviors. If those behaviors are carefully targeted to important results, the whole organization will improve. The logic is impeccable.

Many argue that these types of systems can be used in any type of organization for any type of job. It doesn’t matter how complex the job is. Just implement the research-based procedures, and you will get great results. Maybe that is true, I’m open to the idea. I’ve worked in lots of organizations and seen many types of systems like this. Sometimes they seem to work, but other times there are unintended consequences.

For example, I once worked in an organization that evaluated staff partially based on the percentage of behavior change programs moving in the desired direction. I was talking to the supervisor about the fact that one student didn’t seem to have enough to do during the day. I gave the supervisor a list of ideas of things we could teach that might be beneficial for the student. The supervisor suggested we wait three weeks until after the evaluations were completed, and the percentage of behavior change programs in the desired direction was measured. That way, if any of the new programs didn’t work out, we wouldn’t be penalized on the measurements.

The benefits of behavior change programs moving in the desired direction are obvious. We can see that. But by using that as a measure to evaluate staff, the implementation of new programs was delayed. That response to the reinforcement system is not immediately obvious. What about things we decide not to even attempt because it might mess up the percentage of behavior changes in the desired direction?

The problem, as I see it, is that the positive results of these types of interventions are obvious. The negative results are hard to measure. Many managers I’ve talked to about this concern tend to downplay the importance of this problem—”It doesn’t happen to me.” But I’ve seen it here, and here, and also here, so I’m a bit skeptical.

Sure, reinforcement systems for staff behavior are often important. Just realize that they have to be very carefully designed. One thing that I think is often overlooked is figuring out what types of negative effects the systems might be causing, and taking preventative action when needed.

When I was a young behavior analyst, an experienced researcher warned me about being too impressed with the results of research:

“Did the people in those studies really live better lives after the experiment was over?”

Research on the treatment of problem behavior in individuals with disabilities has made tremendous gains in the last fifty years or so. The improvements in safety, ethics, effectiveness, training, measurement, and practicality are evident.

But if you read the research carefully, you will notice that the people involved in many of those studies have a long way to go before the severe problem behavior is effectively treated. That’s probably in part because many studies are not designed to document effectiveness, but to look at just one small aspect of treatment–part of the assessment process, what type of communication to teach, or 1000 other tiny details. Those studies are extremely important as they can teach us how to Poogi every aspect of the process towards effective treatment. We need those studies.

My main concern is that if you aren’t a careful reader, from some of the wording, you can think that there are no longer any concerns regarding problem behavior for these individuals after these studies are over. Sometimes, authors add statements like, “the problem behavior was eliminated” or “the problem behavior was successfully treated.” Yet, the only evidence provided is data collected during sessions between 5 and 15 minutes in length. I suggest thinking about research articles as a story of someone’s life. After this treatment, what is Joe or Sally or Bob’s life like now? Is it any better? In some cases, the answer is almost certainly yes. But often, I doubt it.

There are some, but very few, studies that document that the treatment was successful throughout the day, and that it lasted over the long-term. Perhaps, that isn’t surprising as it would be extremely expensive research, very hard to maintain experimental control, and just damn difficult to do. Those of us that are concerned on a daily basis with the long-term successful treatment over longer periods of time need to study that research. It can definitely help improve practice. Just be skeptical–not all the success stories are really that successful.