Evidence-based psychotherapies for children and adolescents by Weisz and Kazdin (third edition) – Chapter 29 summary

Some structured, manual-guided treatments may show reduced effects when they are implemented in more clinically representative contexts and compared to usual practice in those contexts. This may be because treatments tend to be developed away from the clinical practice. There are several potential mismatches between evidence-based practice and real-world clinical care:

1. Clinicians in practice settings typically carry diagnostically diverse caseloads but most evidence-based practices on one problem or a homogeneous cluster.
2. Clinically referred youth typically present with comorbidity.
3. Treatment may need to shift during treatment while evidence-based practice often does not take this into account.
4. Everyday clinical care often has an unpredictable course contrary to the design of evidence-based practices which are standardized.

The deployment-focused model of treatment development and testing includes a series of steps aimed at building and testing interventions with the clients and clinicians and within real-world contexts. It is also tested whether interventions improve on current practice in those contexts. Making evidence-based practice work well in practice mainly involves making the treatments fit variations in individual and family characteristics.

Personalized mental health interventions include strategies for selecting treatments, deciding whether and how to combine them, determining what problem to target first and with what techniques and using information about individual client characteristics and ongoing treatment response to inform clinical decision making. In other words, treatment should be personalized.

There are several methods of identifying an individual’s optimal treatment:

1. Meta-analyses comparing treatments for specific client characteristics
   This includes a meta-analysis which compares different treatments based on specific characteristics of the patient. This can help guide selection.
2. Individualized metrics (e.g. probability of treatment benefit)
   This quantifies the benefit each client is expected to receive from alternative interventions. This can facilitate consideration of multiple characteristics in selecting interventions for individuals.
3. Data-mining decision trees
   This includes treatment selection which accounts for multiple characteristics but is informed by decision trees based on detecting and interpreting patterns in data (i.e. data mining).

Further personalization after selecting a treatment may occur through monitoring client progress and using the resulting data to adjust interventions. This can be done in several ways:

1. Measurement feedback systems (MFSs)
   This is a tool which is used to obtain feedback about client progress and to guide treatment decisions. Contents include outcome and process measures taken at multiple times during the course of treatment (e.g. outcome questionnaires).
2. Personalized treatment goals
   This can be assessed and tracked by MFSs. Identifying the problems a client finds most important can allow for further personalization of treatment.
3. Sequential, multiple assignment, randomized trials (SMARTs)
   This can inform the construction of decision rules by dividing the treatment regimen into two or more stages. Participants are first randomized to a particular treatment and treatment response is assessed. Participants are then assigned to one of several next-stage treatments, based in part on their response on the first treatment. SMARTs can evaluate first-stage and next-stage interventions and evaluate optimal time-points for decision making.

It is also possible to modify treatment protocols and create new ones to increase opportunities for customizing. This can be done in four ways:

1. Therapies adapted for specific subgroups
   Therapies can be adapted to a specific subgroup which is expected to respond poorly to a treatment and this can facilitate personalization of treatment.
2. Therapies targeting youths’ environments
   Therapies could also attempt to alter a client’s everyday environment.
3. Modular therapies
   Therapies could be modified and personalized by organizing the content into self-contained modules that can be selected, combined and sequenced in individually tailored ways. This may be especially useful in youth with heterogeneous problems.
4. Principle-guided therapy
   Therapies could also be guided by principles rather than standardized manuals. This can allow for personalization within the principles and treatments.

There are eight ways in which personalized psychotherapy could be achieved faster:

1. Organize and evaluate personalizing approaches
   This includes evaluating personalized approaches against non-personalized versions of those treatments. The evaluated treatments then need to be organized according to strength of evidence.
2. Exploit existing RCT data
   This includes using existing RCT data to inform intervention selection and development based on personal characteristics.
3. Prioritize big-impact personalizing approaches
   This includes prioritizing personalized approaches which have the broadest reach (e.g. good potential because they can be used across a broad range of problems and disorders).
4. Conduct idiographic research (e.g. single-case experiment)
   This already focuses on individuals rather than groups and is thus well-suited to inform and test personalized interventions.
5. Study tailoring strategies in usual care
   This includes scrutinizing usual care to find tailoring strategies that can be tested empirically. Strategies associated with the most successful outcomes may be identified and tested in controlled research.
6. Investigate mechanisms of change
   Identifying mechanisms of change can inform the development of more potent treatments and can inform personalization by allowing change without altering the mechanism of change. Next, the change mechanism may vary across individuals and knowing more about this could allow for more personalization. Lastly, knowledge on the mechanisms of change could inform combination and sequencing of interventions.
7. Identify mediators within and across RCTs
   Identifying mediators could reveal potential mechanisms of change.
8. Focus on psychopathological processes
   This includes focusing on the processes that drive psychopathology as treatment aimed at this could be more effective.