SINGLE SYSTEM RESEARCH DESIGN
Kishon
SingleSystemDesign-PowerPoint.pptxSingle-System Designs
What Are Single-System Designs?
Studies involving outcomes with a single client
Uses the same logic as a time series group design
Not the same as a case study
Useful when group designs are not feasible or for client monitoring
Cheaper than group designs
No worries about obtaining a large sample
Easily replicated
Understanding Design Diagrams
A = non-intervention phase (baseline if before intervention; withdrawal if after intervention)
B = intervention phase
All subsequent letters represent different interventions (e.g., C = intervention phase that differs from B)
Example:
| A | B | A |
| Instances of self-harm (pre-treatment) | Dialectical Behavioral Therapy (treatment) | Instances of self-harm (post-treatment) |
General Rules
Aim is to compare client’s baseline outcomes to their outcomes during treatment (and after treatment if feasible)
The more measurements the better
Baseline measurements should be stable
Intervention phase measurements should be trending in some direction (up or down depending on your hypothesis/goals)
Interpretation of withdrawal phase measurements depends on design
Change can be measured statistically or visually (ideally both)
Common Baselines
A-B Design
Simplest single-system design with 1 baseline and 1 intervention
Ideal vs Non-ideal
A-B-A and A-B-A-B Designs
More complex, commonly called withdrawal/reversal designs
Outcome should go back to baseline (reverse) once treatment is withdrawn and improve again if/when treatment is resumed
Withdrawal
A-B-A and A-B-A-B Designs
Sometimes outcomes don’t reverse at withdrawal
Can indicate lasting change (or a confounding variable)
Problems with Baseline and Withdrawal
In the case of SW (esp. mental health), collection of baseline data is often unethical
Clients often in crisis and need services NOW
Even if ethical, insurance reimbursement often strictly limits # of sessions covered
Similar issues for withdrawal—unethical with client in crisis and difficult to keep them coming back for post-intervention measures
Possible solution—retrospective baseline
Example: Asking client to self-report previous self-harm instances in one session rather than have them go home and keep a journal over long period
B-A-B Design
Special reversal design used when baseline measurement not possible or unethical (crisis)
Possibility of ethical concerns with withdrawal remain
Causality is tricky due to lack of baseline
B-C-B-C Design
Single-system design with no baseline/withdrawal phases and 2 interventions
Useful for testing the effectiveness of 1 intervention over another
A-B-C-D Design
1 baseline phase and multiple interventions
Typically used when 1st intervention (B) is not working as intended
