临床与咨询心理学导论 8 - Research Methods 2
L8 Research Methods 2: Evaluating Treatments & Establishing Evidence-Based Practices
参考文献/图片来源：Pomerantz, A. (2013). Clinical psychology: science, practice, and culture (3rd ed.). Thousand Oaks, CA: SAGE Publications.
8.1 Treatment Efficacy vs. Effectiveness
Treatment Evaluation
What treatment works for a particular disorder or problem?
What do we mean by a treatment “working”? What types of changes are observed?
What specific procedures do therapists need to follow to ensure that clients benefit from the treatment?
Are there any negative outcomes or side effects of treatment? What exactly should clients who participate in this treatment expect?
How can we know if one treatment works better than another?
Treatment Efficacy vs Treatment Effectiveness
• Treatment efficacy studies
- Methodology maximizes internal validity
• Ensuring that participants have a specific diagnosis/engaging in specific recruitment of participants - multiple problems’ people excluded in the research
• Randomly assigning participants to treatment, determined by chance
• Usually university or hospital settings, specific parts of treatment (eg scale)
• Ensuring therapist competence in treatment delivery
Research determine - select sample - randomization - treatment group & control group
• Treatment effectiveness studies
- Methodology maximizes external validity and maintains adequate internal validity
• Including patients referred to the actual clinical setting
• Administering treatment within “real world” services
• Using clinicians who routinely provide psychological services
• No strict protocols used
Usually treatment first evaluated through efficacy studies, if works, effectiveness studies can show if the treatment works in the less controlled circumstances.
8.2 Research Designs for Evaluating Treatments
8.2.1 Traditional Case Studies
• Case Studies:
- Involve a thorough and detailed examination of one person or situation
- Description, observation, and interpretation - Without statistical analyses
- Important part of the history of the field - Works by Witmer, Freud, etc.
- Limitations in conclusions that can be drawn from them because of lack of experimental design, concerns both internal and external validity
8.2.2 Single-Subject Experimental Designs
Focus on an individual’s change rather than on average change between groups
Participant acts as their own control
Small sample sizes may be adequate to determine causality
Helpful when it is not feasible to recruit a large sample
Cost-effective
• Rizvi & Nock (2008) - Provides an explanation of using single-case experimental design for clients with suicidal and self-harm behaviors
• Types of Single-Subject Experimental Designs:
- AB Design: Baseline Period (A), Treatment Period (B)
Ideally the baseline period is very stable.
Due to internal validity - weakest in all single-subject designs, no causal
- ABAB Design: Baseline, Treatment, Baseline (treat stop), Treatment
Stronger than AB design
Ethical issue to remove the treatment, especially if it’s working
- Multiple-Baseline Design:
• Treatment is never removed (i.e., one baseline per person, behavior, setting)
• Treatment is applied sequentially across different:
Individuals (person 1 starts, then person 2)
Behaviors (suicidal behavior targeted first, then ideation)
Settings (in session, at home, at school/work)
• Treatment is successful if change occurs when, and only when, the treatment is directed at the behavior, setting, or individual in question.
- Changing-Criterion Design:
• Treatment is never removed
• Only one true baseline
• Change reinforcement schedule for positive/desired behavior over time to require increasing behavior change
• Benefit = Focus on high-rate (i.e., more frequent) positive skills instead of a low-rate (i.e., uncommon) self-harm behavior
• If not achieving the next level, reconsider or lower the criteria.
8.2.3 Randomized Controlled Trials
• RCTs: maximum internal validity
- Gold standard for psychotherapy research
- Randomly assign participants to treatment group or a control condition
- Measure symptoms before and after treatment (both groups)
- eg This is an experimental design
• Independent variable: Treatment or Control
• Dependent variable: Symptoms
• Placebo: Medically inert treatment (e.g., sugar pill)
• The Placebo Effect: Tendency for people to show improvement when they believe that they are receiving treatment - Related to expectations for improvement, paying greater attention to health, behavior, and wellbeing - the placebo group should not experience the change if the treatment is effective.
• Types of Control Groups in RCTs
- No Treatment Control Group
• Controls for: Improvement/change over time
• Weaknesses: Participants know they are not receiving treatment, no expectation for change; ethical concerns (so not widely used).
- Waitlist Control Group - treatment after first several weeks without treatments
• Controls for: Improvement/change over time
• Weaknesses: Participants know they are not receiving treatment, no expectation for change
• More ethical than providing no treatment
- Attention-Placebo Control Group
• Controls for: Improvement/change over time and expectations for improvement
• Addresses the placebo effect, no actual treatment
- Treatment as Usual (TAU) Control
• Treatment that is already available and commonly used compared to a new treatment that might have more or different benefits
• Only possible when a well-established treatment exists, and develop a new
• Symptoms, level of functioning, and other data collected at at least 2 time points: Before treatment begins (Time 1) & After treatment ends (Time 2)
8.3 Evidence-Based Treatments
Path to Evidence-Based Treatment
• Eysenck (1952): Review of 19 psychotherapy outcome studies
- Less than half of people in psychodynamic treatment improved
- Compared to ~72% who improved/recovered when treated in hospitals and by general practitioners, only 44% identified from “improved” to “cured” in psychodynamic
Evidence-Based Treatments
• different from - Evidence-Based Practice in Psychology (EBPP)
- APA: “…the integration of the best available research with clinical expertise in the context of patient characteristics, culture, and preferences.”
- EBPP includes a broader range of activities (e.g., treatment, assesSment, case formulation)
- Distinct from Evidence-Based Treatments (EBTs)
• Increased efforts to establish evidence-based treatments and disseminate those treatments through the early 1990’s
• APA Task Force
- Developed specific criteria for EBTs
- First published report on evidence-based treatments in 1995, with a preliminary list of treatments
- An expanded list was published in 1996 and updated several times, most recently in 2008
• Criteria established by an APA working group in 1996
Well-Established or Efficacious Treatments if:
- 2 + RCTs showing superior to active control or equivalent to an established treatment
- OR 10 + single-subject experiments
Probably Efficacious Treatments if:
- Only 1 RCT
- OR 2 + experiments showing superior to waitlist control
- OR 3 + single-subject studies
• Debate and disagreement about exactly how evidence-based treatments should be defined - For example, some researchers have suggested more groups or focusing more on RCTs
eg - Nathan & Gorman (2002) suggested different criteria for treatment studies:
• Type 1 Studies: Most methodologically sound RCTs
• Type 2 Studies: RCTs with method flaws
• Type 3 Studies: Clinical trials with significant method flaws
• Continues to Type 6 Studies
• Using treatments that are not evidence-based (i.e., ineffective treatments) can:
- Waste the client’s resources (e.g., time, money)
- Cause clients to become hopeless or discouraged
- Lead clients to believe that there are no effective treatments at all
- Cause harm to clients:
• Worsening symptoms and decreased functioning
• Perpetuates false information and myths about a disorder
Clients - should be informed about EBTs and non-EBTs as choices
