Evidence Based Practice (EBP) is the use of systematic decision-making processes or provision of services which have been shown, through available scientific evidence, to consistently improve measurable client outcomes. Instead of tradition, gut reaction or single observations as the basis for making decisions, EBP relies on data collected through experimental research and accounts for individual client characteristics and clinician expertise.
Within the field of mental health, Evidence Based Practice is often discussed in terms of treatments or interventions which are effective. Evidence Based Treatments (EBTs) are interventions which have scientific findings to demonstrate their effectiveness or efficacy in improving client outcomes. Treatments are often placed along a continuum of support based on the rigorousness and amount of supporting research ranging from treatments which have strong support to those which are untested to those which have produced negative outcomes. Data sources used to make these evidence determinations include randomized experiments, which compare treatment with a control or placebo group or compare the treatment with another already established treatment; and single case design experiments which compare an individual subject’s baseline with their response to treatment.
There are four generally accepted evidence levels along the continuum of research support on which experts attempt to categorize practices, based on the body of evidence and outcomes indicated supporting each treatment method. Briefly, they are:
- Level 1: “Best Support”
- Level 2: “Good Support or Moderate Support”
- Level 3: “Promising Practice”
- Level 4: “Practices with Known Risks”
Specifically, evidence determinations are based on the following criteria:
Level 1: “Best Support” - Interventions receiving “best” support must have supporting research evidence obtained in one of the following ways:
- Two or more between-group design experiments demonstrating that treatment is superior to placebo or already established treatment
- Two or more between-group design experiments demonstrating that treatment is equivalent to an already established treatment
- Ten or more rigorous single case design experiments which demonstrate treatment efficacy In addition, all experiments must:
- Be conducted with treatment manuals
- Specify characteristics of client samples
- Have treatment effects demonstrated by at least two different investigators
Level 2: “Good Support or Moderate Support” - Interventions receiving “good or moderate” support must have supporting research evidence obtained in one of the following ways:
- Two or more experiments showing treatment is superior to a wait-list control group
- Treatment manuals, specification of the sample, and independent investigators is not required
- One between-group design experiment utilizing manuals and a specified sample which demonstrates treatment is superior to placebo or previously established treatment
- One between-group design experiment utilizing manuals and a specified sample which demonstrates treatment is equivalent to previously established treatment
- Four or more rigorous single case design experiments utilizing manuals and specifying sample clients which demonstrate treatment efficacy
Level 3: “Promising Practice” - “Promising practices” meet the following criteria:
- Sound theoretical basis in generally accepted psychological principles or has been demonstrated to be effective with another target behavior.
- Substantial clinical-anecdotal literature indicating treatment value with the target behavior
- Generally accepted in clinical practice as appropriate for use with the target behavior
- No clinical evidence indicating that the treatment constitutes a substantial risk of harm to those receiving it, compared to likely benefits
- Book, manual, or other available writings which specify components and describe administration of treatment
Level 4: “Practices with Known Risks” - “Practices with known risks” meet the following criteria:
- Interventions which have evidence demonstrating harmful effects of a treatment. This evidence need only be based on one study or review of the intervention.