About the shape of research in sft, whether it should be done and where it might lead.

Yes - to sell it; give scientific credibility; give corporate identity; control 'pseudo-sft' publications; to ask how it works; to protect something we think is beautiful; because it will be done anyway and we want to shape the process.

No - because results are an illusion; because it limits what we can do; because it treats sft as a 'pill'.

If research is to be done who is the audience? Ourselves? Psychotherapists? Doctors? Scientists interested in self-organising systems? Society? Health care purchasing organisations?

PROCESS OF SFT

Sft is defined in the conversation between therapist and client. Asking the miracle question only becomes goal-oriented when the conversation or the client begins to construct goals. Should we use broad menu-driven definitions of sft or tight, operationally defined criteria? Or both, conducting studies on subgroups whose therapy conformed to tight sft criteria, abstracted from bigger samples treated with 'broad definition' sft.

Steve and Insoo (briefftc@AOL.com)core conditions - miracle question, scaling, break, compliments. Broad definition - sense of: client input and collaboration; client generated solutions via future-oriented questioning; client self-assessment - 'how well am I doing?' via scaling.

Matrix - note specific therapist techniques, therapeutic conversations, client responses. These will differ according to what type of therapy is being done eg sft, narrative. Certain features should always be present eg problem-free talk, positive regard for client, empowerment. Some should always be absent eg problem focus, giving direct advice. Some will occur only on occasion eg specific issues related to confidentiality and local child abuse rules. Potential observers should be able to recognise the appropriate behaviours and elements according to what style of therapy is allegedly being used. We would like further proposals as to what other elements should be included in this matrix. (Not laid out here in tabular form because of email limitations.)

POSSIBLE OUTCOME MEASURES

Quality of life MLDL (German)

Personal resources EMI-B (German)

Coping Resources Inventory (USA/UK)

Sense of Coherence (Antonowski) (Swedish; English translation available)

Symptom scales SCL-90 (USA/UK and other versions available); OQ-45 (USA)

Diagnosis ICD-10 (Europe); DSM-IV (USA/Europe)

Clinical Global Impression (all languages; called CGI to make it look more 'scientific')

Functional / social adaptation measures (GAS, CGAS (children), GARF)

Public health measures (SF-36, SF-12; USA/Europe versions available)

Clients' own scaling responses

Allen Wade (awade@UVIC.CA):

ECHO (Schaefer, Bavelas and Bavelas (1978/79) Teaching of Psychology).

A random sample of undergraduates were asked to write what questions that they would like for course evaluation. Second group sorted those questions and grouped them so that like questions were together; the same group then picked the best prototype of each group. Prototype questions given to third group as a questionnaire and ranked against other standard questionnaires written by faculty. The group

ranked the student-generated questionnaire (not identified as such) best for letting them express what they wanted to say. There were interesting differences between the student generated and faculty generated questionnaires. The faculty questionnaires never had items like 'Gives clear directions for a paper' or 'Does the professor respect the students?'. Technique could be applied similarly for clients of therapy.

Gunter Schiepek (uf341am@sunmail.lrz-muenchen.de):

23 elements of sft behaviour, some negative. Videotapes: sft behaviours present in other therapies but proportion differs (1 minute time blocks for analysis). Coded joint behaviours + look at 10 sec samples. Reliability of items>0.8. Factor analysis of data set gives 8 factor solution.

Comparison with psychoanalytic interviews showed less problem focus in sft. Differences linked to patterns of interaction rather than individual components of therapy. Chaos theory shows connection between client and therapist behaviours but unfortunately there was not enough time to explore this further.

(Recent publications by Gunter and colleagues in Psychotherapy Research give detailed expositions)

Other considerations:

Hypotheses to test?

Pilot projects or full outcome project (which should be randomised and rated blindly)?

Need for new measures vs. using existing instruments.

CONCLUSIONS

For ourselves it was suggested that the distrust of reductionist thinking meant that qualitative research is more likely. From an anthropological standpoint Michael Houseman and Marika Moisseff suggested measuring changes in therapist behaviour after adopting sft. They offered to design a simple self report questionnaire for therapists (looking at changes in conferences, reading etc in recent years).

For the critical audience of mental health professionals suitable research is the right language to gain their ears. It was recognised that quantitative measures are preferred, using 'objective' (doctor) measures rather than 'subjective' (patient) measures. However, it was felt that whilst compromises were necessary in order to 'join' with a critical readership, there should not be so many compromises that we ceased doing brief therapy. It seems that we need to have the process and outcome issues clarified for ourselves before we can embark on controlled studies of sft vs. other therapies or placeboes.

Jon Prosser (J.G.S.Prosser@newcastle.ac.uk):

A pilot project to establish credibility for outcome research methodology before attempting a fullsize outcome study or comparative study. 'Quality of life' outcome measures would satisfy sft therapists and are approaching respectability in the biomedical sciences. Quality of life is generally held to have two components: a subjective sense of wellbeing (equivalent to an increase in scale scores) and an 'objective' measure of fulfilment of potential. The manual for DSM-IV includes clinician-rated scales of functional adaptation. These are already validated and are increasingly accepted as an adjunct measure of psychiatric morbidity. They aim to answer the question 'Whatever your problems are, to what extent is your life affected or to what extent are you achieving your potential?'

A possible pilot project would be to collect scale scores and functional adaptation scores from clients before and after therapy to test the hypothesis that there will be acceptable levels of agreement between the two measures. Other studies have already shown links between the severity of mental health problems and functional adaptation so we need not address the issue of problem definition / diagnosis directly.

Harry Korman (hkorman@sbbs.se):

Colleague and he interested in aspects of therapy relationship process. They will pursue, keeping us informed.

For consumption by therapy customers and the public at large we agreed that research is not the best medium for spreading information. Rather we are looking at advertising a product. We could franchise it like CocaCola but imitators will soon appear. Also, market issues vary in different countries eg in Belgium you can choose your own therapist but in Australia doctors refer to Michael Durrant because he is a nice guy and 'reliable', not because they know about his good outcomes in objective terms. Possible methods could include TV, professional education, articles in the popular press. EBTA meetings and therapist training workshops should perhaps include advice on marketing the skills learned. There is already enough published evaluation research on sft for a 'digest' to be handed to potential insurers or employers. Michael Hjerth (michael.hjerth@taby.mail.melia.com)is currently advising FKC-Stockholm on aspects of their 'marketing' and may be willing to make some of his ideas available to the rest of us.

My thanks to Jon Prosser for collecting all these ideas during the meeting. As a final note one participant suggested that maybe research is neither neutral nor anti-therapeutic but might actually increase the effectiveness of therapy.

Alasdair Macdonald (ajmacdonald@compuserve.com) December 1997

~