Table 3.

Intervention vs. Implementation Trial Design Perspectives: A Hypothetical Example of the Use of Motivational Interviewing (MI) for Substance Use Disorders in the Homeless Population

	Efficacy Design Principles	Effectiveness Design Principles	Implementation Design Principles
Hypothesis	MI beats control	MI beats control	MI will be adopted and sustained
Population & setting	Exclude psychosis, bipolar, anxiety; any setting with cooperative patients	Include most comorbidities; typical setting is nonspecialized practice sites	Unit of observation may be patients, providers, or clinics; typical setting is nonspecialized practice sites
Outcome measures	Health outcomes, many: "just in case…"	Health outcomes, short & sweet	Emphasize MI adoption measures
Intervention: clinicians	PhDs, MSWs hired & trained by PI	Addiction counselors hired as study staff	Endogenous addiction counselors
Intervention: fidelity	Trained to criterion, audiotaped for fidelity	Trained to criterion, QI-type monitoring as in clinical system	Formative evaluation the focus
Context	Make sure that the trial is successful, at all costs	Work within “typical” conditions	Maintain typical conditions
Research support	Crypto-case management	Research support, but “firewalled”	Research support limited; e.g., only for training
Validity emphasis	Internal > > external	External > internal	Plan to optimize protocol in real time using formative evaluation, in violation of “traditional” considerations of internal validity, while systematically documenting adaptations