October 5, 2010

The BRIEFER project - capturing the solution-focused process in an expert system

One of the interesting projects that was done by the pioneers of the solution-focused approach was the BRIEFER project which tried to formalize the therapy process and capture it in an expert system. In the interview I recently did with Wally Gingerich, he tells about how he looks back on BRIEFER and about what it was about and what it amounted to.

"I had been interested in expert systems for some time, and Steve was also interested in computers, so we hit on this idea of developing an expert system that would serve the advising function of the team behind the mirror. The therapist would conduct the first part of the session, then come behind the mirror and ask the computer what kind of task to give the client. The computer would ask questions about what happened in the session and the therapist would answer. After about half a dozen interactions the computer would make its recommendation and the therapist would go back into the room and give the task. This is a simplified version, of course, but that was the idea.

This project evolved in the context at the time of us wanting to understand more about how skilled therapists did their work – how they thought about their cases and decided what intervention to use. We found that skilled therapists were not very adept at giving clear explanations of their reasoning that would enable another therapist to replicate their work. This is typical of experts in general, by the way; true experts tend to function in an intuitive and un-self-conscious way, knowing “in their bones” what to do. We were unsatisfied with this, however, because we wanted to understand better what made SFBT work, and to be able to describe it in such a way that it could be taught and learned more easily.

Expert system methodology at the time paid a lot of attention to the “knowledge mining” process – how the programmer could extract (or construct) the rule-based knowledge the expert used to arrive at her expert advice. We thought of this as another methodology with the potential to help us figure out what expert SFBT therapists did. We had the good fortune about that time of coming into contact with Hannah Goodman, a graduate student who was working on her masters degree in computer science and needed a project for her thesis. She was interested in expert systems and we talked her into doing her project with us. She would use methods she had learned in class to “mine” the knowledge Steve used to decide what intervention to give the client. We were thinking that, if successful, this would reveal more clearly what expert therapists did, that it could be used as teaching too, and that it could even potentially take the place of the team behind the mirror providing solo therapists the advantages of a team without actually having one. We knew this was a reach, but that was what we were thinking.

We met many times with Hannah asking questions to find out what Steve did, constructing rules that she thought embodied that knowledge, and then trying the rules on real cases to see if they led to the intervention the therapist had actually used. Although what came out had probably existed in Steve’s head somewhere, it often seemed like a revelation to us. The finished expert system consisted of about thirty IF-THEN rules and in most cases it did a pretty good job of replicating the intervention advice the actual therapist had given. We called the system BRIEFER, because we thought it should make the therapy process more efficient and consistent, and briefer, of course, which was one of the objectives of the solution-focused approach. BRIEFER also conveyed the idea that the computer would “brief” the therapist about her cases, and we liked that little play on words.

We never subjected BRIEFER to rigorous testing and did not release it outside the clinic because we were afraid people would actually use it if anything happened we could be held liable, so BRIEFER, the computer program, eventually passed into oblivion. We often used BRIEFER in training and presentations, however, and I remember fondly John Weakland’s half-serious comment at one conference that we had better not let it get out that SFBT consisted of only 32 rules – it could put us out of a job and would let the secret out that anyone could do it!

Even though we did not use BRIEFER as an expert system, we felt the project was a success because it helped us describe with considerably more clarity the solution-focused therapy process. It made crystal clear just what information the therapist needed to gather in the initial interview, and how to put it together to figure out what direction to take with a case and what kind of intervention to give. This resulted, among other things, in the flow chart that appeared in the Family Process article which I use to this day in my teaching. That flow chart also helped me stay on track when I interviewed clients in those days, as I did not have the same level of expertise as Steve and found the chart a useful rubric. In effect, BRIEFER helped me interview and decide what intervention to give, even though I wasn’t using the computer version of the program, only the flow chart.

I don’t think the BRIEFER project significantly changed how solution-focused therapy was done at the clinic; it made clearer and more systematic what we were already doing. On the other hand, Steve might say that describing the process in a new language and structure did actually change it. That would be an interesting conversation to have…"

~Wally Gingerich, taken from Interview with Wally Gingerich

2 comments:

  1. Coert,

    Thanks for posting this. What are the 32 rules? Also, how can I get a copy of the Family process article he mentions in the interview?

    ReplyDelete
  2. Hi Rodney, perhaps someone who reads this can, I haven't got it.

    ReplyDelete

Enter your email address:

Delivered by FeedBurner