treatment-integrity

# Treatment Integrity Treatment integrity (also called procedural fidelity or implementation fidelity) measures the degree to which an intervention is implemented as designed. Without treatment integrity data, you cannot attribute behavior change to the independent variable. ## Why Treatment Integrity Matters - **Internal validity**: If procedures are not implemented correctly, you cannot conclude that the intervention caused the observed change. - **Replication**: Other practitioners cannot replicate your results if the actual procedures differ from the documented procedures. - **Ethical obligation**: Clients deserve to receive the intervention that was designed for them, not a degraded version. - **Troubleshooting**: Low treatment integrity is often the reason an effective intervention appears to fail. ## Components of Treatment Integrity ### Antecedent Components - Correct discriminative stimuli presented (instructions, materials, prompts). - Correct timing and sequence of antecedent events. - Appropriate environmental arrangement in place. - Correct session structure (number of trials, inter-trial interval, session duration). ### Prompt Components - Correct prompt type used (verbal, gestural, model, physical). - Correct prompt level within the hierarchy. - Correct timing of prompts (e.g., within 3 seconds of the SD). - Correct fading steps followed. ### Consequence Components - Correct reinforcer delivered for correct responses. - Correct schedule of reinforcement maintained. - Correct error correction procedure implemented. - Appropriate consequence for problem behavior (e.g., planned ignoring, response cost). - Correct timing of consequence delivery. ### Session Structure - Correct number of opportunities/trials provided. - Correct inter-trial intervals maintained. - Correct session duration. - Correct data collection procedures followed. ## Designing Treatment Integrity Checklists ### Steps 1. Decompose the intervention into discrete, observable implementer behaviors. 2. Sequence the behaviors in the order they should occur. 3. Write each step as a clear, binary (yes/no) observable action. 4. Include a "not applicable" option for steps that may not occur in every session (e.g., error correction when no errors occur). 5. Pilot the checklist with two independent observers and revise for clarity. ### Example Checklist: Discrete Trial Teaching | Step | Y/N | |------|-----| | Secures client attention before presenting SD | | | Presents SD clearly and consistently | | | Waits appropriate interval for response (3–5 seconds) | | | Delivers correct prompt at correct level (if needed) | | | Delivers reinforcer within 3 seconds of correct response | | | Implements error correction for incorrect responses | | | Records data accurately for each trial | | | Maintains inter-trial interval (3–5 seconds) | | | Completes specified number of trials | | ## Calculating Treatment Integrity **Formula**: (Steps implemented correctly / Total steps applicable) × 100 ### Example If 18 of 20 applicable steps were implemented correctly: (18/20) × 100 = 90% ### Acceptable Standards - **Minimum**: 80% integrity, though higher is always preferred. - **Target**: 90%+ for most interventions. - **For restrictive procedures**: Near 100% integrity is essential. ## Assessment Schedule - Assess treatment integrity on a minimum of 20–33% of sessions. - Assess across all implementers (not just the most experienced). - Assess across all conditions or treatment phases. - Assess at the beginning of implementation (to catch training gaps early), periodically throughout, and whenever behavior change stalls. ## Relationship to IOA Treatment integrity and IOA are different constructs: - **IOA** assesses whether observers agree on the measurement of the dependent variable (the behavior). - **Treatment integrity** assesses whether the independent variable (the intervention) was implemented correctly. - Both are necessary for rigorous practice. You need reliable measurement (IOA) of the behavior and faithful implementation (treatment integrity) of the intervention. ## Addressing Low Treatment Integrity ### Diagnostic Questions 1. Does the implementer understand the procedure? (Knowledge deficit) 2. Can the implementer perform the procedure? (Skill deficit) 3. Does the implementer have the resources and environmental support? (Environmental barrier) 4. Does the implementer receive feedback on their performance? (Feedback deficit) 5. Is the procedure too complex for the current context? (Design problem) ### Corrective Strategies - **Knowledge deficit**: Provide written protocols, review the rationale, answer questions. - **Skill deficit**: Use behavioral skills training (BST)—instruction, modeling, rehearsal, feedback. - **Environmental barrier**: Simplify materials, adjust staffing, reduce competing demands. - **Feedback deficit**: Provide immediate, specific performance feedback with a positive-to-corrective ratio of at least 3:1. - **Design problem**: Simplify the procedure. If a procedure requires 15 precise steps in a fast-paced environment, consider whether simplification is possible without compromising effectiveness. ## Independent Variable Integrity vs. Dependent Variable Reliability | Concept | What It Assesses | Example | |---------|-----------------|---------| | Treatment integrity | Was the intervention delivered correctly? | Did the teacher use the prompt hierarchy as written? | | IOA | Was the behavior measured reliably? | Did two observers agree on the number of tantrums? | | Both together | Can we trust the functional relationship? | The intervention was delivered as designed AND the behavior change was measured reliably. | ## Reporting Treatment Integrity ### In Clinical Records - Document the percentage of sessions assessed for integrity. - Report mean integrity percentage and range. - Note any corrective actions taken when integrity fell below criterion. ### In Research Publications - Describe the integrity checklist and how it was developed. - Report the percentage of sessions assessed. - Report mean integrity and range across all conditions. - Describe how assessors were trained and their reliability (IOA on integrity ratings). **Example**: "Treatment integrity was assessed on 25% of sessions across all phases by an independent observer. Mean integrity was 95% (range: 88%–100%). The observer was trained to criterion (≥90% agreement with an expert) before data collection." ## Key References - Peterson, L., Homer, A. L., & Wonderlich, S. A. (1982). The integrity of independent variables in behavior analysis. *JABA*, 15, 477–492. - Cooper, J. O., Heron, T. E., & Heward, W. L. (2020). *Applied Behavior Analysis* (3rd ed.). Pearson. - Gresham, F. M. (1989). Assessment of treatment integrity in school consultation and prereferral intervention. *School Psychology Review*, 18, 37–50. - DiGennaro Reed, F. D., & Codding, R. S. (2014). Advancements in procedural fidelity assessment and intervention. *JABA*, 47, 1–12.