Chambless & Hollon (1998). Defining empirically supported therapies.” - Article summary

Empirically supported treatments refer to clearly specified psychological treatments shown to be efficacious in controlled research with a delineated population. Treatment efficacy must be demonstrated in controlled research in which it is reasonable to conclude that benefits observed are due to the effects of the treatment and not to chance or confounds. This is best demonstrated in randomized clinical trials. Replication is critical in this case.

A treatment which has been found efficacious in at least two studies by independent research teams is an efficacious treatment. A treatment which only has one study supporting the efficacy or if all the research has been conducted by one team is possibly efficacious treatment.

The efficacy research has to be done with proper methods. There are different types of group design of efficacy research:

1. Comparison with no treatment
   This comparison demonstrates whether a treatment actually works.
2. Comparisons with other treatments or with placebo
   This comparison demonstrates whether a treatment effect is due to the treatment or due to another effect (e.g. placebo; receiving attention).
3. Comparisons with other rival interventions
   This comparison demonstrates whether a treatment effect is due to competing mechanisms (i.e. the mechanism of another treatment) or due to a unique mechanism. It can also compare relative benefits of competing interventions.

The more stringent the comparison is, the more value researchers should attach to the results. Treatments that are efficacious and specific are treatments which control for other factors that could explain the treatment effect (e.g. therapeutic alliance). This increases confidence in the theoretical explanatory model of the treatment.

Problems with comparisons with other rival interventions are interpretation (i.e. interpreting the treatment as being efficacious while it is a downgraded version of a well-established treatment) (1) and low statistical power (2).

In order to demonstrate that a treatment is equivalent to a well-established treatment, it should meet the following criteria:

1. Investigators have a sample size of 25-30 per condition.
2. The unproven treatment is not significantly inferior to the established efficacious treatments on tests of significance .
3. The pattern of data indicates no trends for the established efficacious treatment to be superior.

A psychological treatment plus placebo condition which is better than a placebo condition may not be due to the psychological intervention alone but due to the interaction between the placebo and the psychological treatment.

Efficacy needs to be described for a specific population or problem rather than in general. Factors that may determine the population (e.g. socioeconomic status) need to be described. The tools used in efficacy research need to have properly demonstrated reliability and validity. It is best to use multiple methods of assessment. Measures of life quality and functioning, rather than just symptoms need to be included as well, to properly evaluate a treatment.

It is important to know whether a treatment has an enduring effect and whether treatments differ on this. Return to treatment in the absence of documentable symptomatic events is not necessarily an index of underlying risk as there are a lot of reasons for a person to resume treatment.

Follow-up designs often do not assess treatment and symptom status in an ongoing fashion and they are susceptible to bias resulting from differential retention (e.g. patients must complete treatment and show improvement to be retained in the sample). This bias can thus be the result of the sample size being significantly different than the one at the start of treatment. It is not necessarily a good idea to retain all participants but neither is leaving some out of the analysis. The length of the follow-up required depends on the natural course of the disorder and the strength and stability of the treatment effect.

One method of assessing clinical significance is in terms of meaningful change (i.e. intersection between functional and dysfunctional populations) and reliable change (i.e. error of measurement). People who meet both criteria. The norm for clinical significance depends on the seriousness of the disorder.

Treatment manuals are extensive descriptions of the treatment approach therapists are to follow. These manuals usually need to be supplemented by additional training and supervision. Training is essential for therapists but it is difficult to assess therapist competence. Outcome variability of research into treatments may be associated with the preferences and expertise of the research teams involved (i.e. a therapy does better when it is given by people who are experts in it). Inferences regarding treatment efficacy can only be framed in the context of how the therapy was delivered and by whom.

There are several typical errors in data analysis:

1. Type I error.
2. Uncontrolled pre-test post-test comparisons are made (i.e. this does not necessarily indicate a trend in the data).
3. Differential attrition is not taken into account.
4. No test for therapist or site effects.

There are several issues specific to single-case experiments:

1. Establishing a stable baseline.
2. Typical acceptable designs.
3. Defining efficacy.
4. Interpretation of results

The ABAB design is a design in which there is a baseline period, a treatment period, another period of non-treatment and a period of treatment again. The multiple baseline design is a design in which there are multiple baselines across behaviours and settings.

In the case of conflicting results, the quality of the conflicting research needs to be examined (1), the direction of the majority research needs to be considered (2) and meta-analyses should be consulted (3). Meta-analyses could obscure qualitative differences in treatment execution.

Efficacy studies are often difficult to generalize (e.g. research only uses one ethnic group) and the interaction between personality characteristics and treatment characteristics is often not studied and known.

Treatment effectiveness refers to whether treatment actually works in clinical practice. Treatment efficacy refers to whether treatment actually works in controlled settings.

It is relevant to consider the extent to which evidence from efficacy trials is relevant to the kinds of patients actually seen in clinical practice. The generalizability across therapists and settings also needs to be taken into account as therapists in RCTs often have more training and supervision than the average therapist in a clinical setting and that these therapist can focus on a single problem (e.g. depression) using a single intervention (e.g. CBT), while this is not possible in daily clinical practice.

There is no consensus regarding whether the controlled nature of RCTs is beneficial or detrimental to treatment outcomes. The controlled nature does not allow the clinician to use his judgement to tailor the treatment to the patient. However, following protocol may yield better results.

A treatment needs to be feasible. This means that the patients need to be able to adhere to the treatment. Furthermore, in order for a treatment to be feasible in clinical practice, sufficient therapist need to be able to provide that treatment to patients. The cost-effectiveness of a treatment also needs to be evaluated. The costs and benefits of a treatment need to be evaluated both in the short-term and in the long-term.