Taking feedback to the next level: Efficacy of expected treatment response feedback to therapists

Nowadays, many mental health care organizations measure their patient's
progress through routine outcome monitoring. The way that progress is measured
and the way therapists are provided with feedback on their patient's progress
strongly differs between organizations. The question is whether all methods are
as effective, and there is surprisingly little research on the effectiveness of
routine outcome monitoring available.
Research in the United States shows that providing feedback to therapists based
on a prediction model, can improve patient outcome, especially for those
patients that are not progressing well in treatment (o.a. Lambert, 2007). The
prediction model that was used in the studies by Lambert was calculated based
on the initial severity in patient disfunctioning, measured with the Outcome
Questionnaire (OQ). Lambert and colleagues conducted five randomized controlled
trials on the effectiveness of feedback based on the prediction model en found
that in the 'not on track' group (patients that were not progressing well)
treatment outcomes were significantly better in the experimental group than in
the no feedback control condition. In the 'on track' group, the feedback did
not have an effect on outcome.
In the Netherland, a prediction model for Dutch patients was predicted, based
on almost 2000 patients in four mental health care organizations. Our model
uses the initial severity of patients disfunction, as well as patients'
expectancies on treatment outcome as predictors for progress. Results showed
that this was also a significant predictor for progess in the data that was
collected.

In this study, two forms of feedback and one control condition are compared:

* Control group: the therapist gets no feedback.
* Outcome monitoring feedback: the therapist gets feedback on the patients
progress in a progress chart.
* Complex feedback based on the Dutch prediction model: the therapist gets a
progress chart that compares the patient's actual progress with the expected
treatment response.

Main questions:
1. Does providing feedback to therapists improve treatment outcome?
2. Does providing feedback based on the prediction model lead to better
outcomes than progress feedback alone?

Secundary questions:
1. How well can therapists predict their patient's progress?
2. Are there differences between therapists in treatment outcomes and is this
related to the way they use the feedback?

A 2-year randomized controlled clinical trial on the effectiveness of feedback
interventions for therapists

The intervention consists of providing feedback to therapists on their
patients' progress.

In the outcome monitoring feedback condition the therapist gets feedback on
patient progress in progress charts and tables. The progress of the patient can
be viewed by the therapist at all times, by logging on to the feedback system
(RequestXL), but is also actively provided by e-mail at session 1, 3, 5, 10 and
15.

In the complex feedback condition with prediction model the actual treatment
course (based on the OQ scores) of the patient is compared with the predicted
treatment course. The expected treatment course is calculated by a formula. The
progress of the patient can be viewed by the therapist at all times, by logging
on to the feedback system (RequestXL), but is also actively provided by e-mail
when the patient is not progressing well. The therapist then receives an e-mail
with high urgency.
In the complex feedback condition feedback is also provided on the ASC. The ASC
is administered when the patient goes off track (through the 75% negative bound
of the confidence interval around the predicted treatment course) and measures
the therapeutic alliance, motivation, social support and life events. The ASC
is combined with so called Clinical Support Tools, a set of Microsoft Word
documents that provide practical tips on improving the therapeutic alliance,
motivation and social support. The practical tips are based on a literature
review on these topics.

Burden:

The burden for the patient consists of completing a 5 minute questionnaire
before each treatment session, for a maximum of 15 times. After the research
period (end of treatment or after 15 session) they will get two follow-up
measures after 3 and 6 months. The average treatment duration for outpatient
indivual treatment was around 9 sessions in a previous study at one of
treatment settings. If patients go off track, they are asked to complete an
addition questionnaire, that also takes about 5 minutes to complete.
In addition to the self-report questionnaires, patients are getting a
diagnostic interview that takes between 1,5 to 3 hours, with an average of 2
hours per patient.

The burden for the therapist consist of completing brief questionnaires during
the treatment of a participating patient (total burden around 10 minutes) and
completing a longer questionnaire once every 6 months (about 20 minutes per
adminstration, total of 5 administrations).

Risks:
The risks for both the patient and the therapist are minimal.