Synthesis, Recommendations, and Implementation
What this domain tests
This domain tests whether you can move from a body of evidence to a defensible recommendation and a real-world implementation plan. Expect questions on evidence tables, GRADE, guideline appraisal, implementation frameworks, and evaluation of adoption.
Core rules
- Synthesis is not a vote count. Weigh studies by risk of bias, consistency, directness, and precision. Use an evidence table to compare design, sample, outcomes, and limitations.
- GRADE separates quality of evidence (high, moderate, low, very low) from strength of recommendation (strong or conditional). Quality can be downgraded for risk of bias, inconsistency, indirectness, imprecision, and publication bias; it can be upgraded for large effects, dose-response, or plausible confounding.
- Recommendations must state the population, intervention, comparator, and context. “We recommend early mobilization” is incomplete without specifying patients and conditions.
- Guideline appraisal: AGREE II covers scope, stakeholder involvement, rigor, clarity, applicability, and editorial independence. A guideline is not automatically applicable to your setting.
- Implementation frameworks include PARIHS (evidence, context, facilitation), CFIR (inner and outer setting, intervention characteristics, individuals, process), RE-AIM (reach, effectiveness, adoption, implementation, maintenance), and the Iowa Model.
- Plan for barriers and facilitators, stakeholder engagement, fidelity monitoring, and sustainability from the start.
Worked example
A team finds three RCTs and one systematic review supporting early mobilization after hip fracture. GRADE quality is moderate because of inconsistency and imprecision. The recommendation is conditional: “In older adults after hip fracture, we suggest early mobilization within 24–48 hours, with individualized assessment, rather than delayed mobilization.” Implementation includes orthopedic and physiotherapy champions, a mobility protocol, daily audit, and balancing measures for falls and pain.
Traps that cost marks
- Equating low-quality evidence with “do not do it”; low quality can still support a conditional recommendation.
- Upgrading recommendation strength because a guideline says so.
- Ignoring local context, resources, and patient preferences.
- Implementing without fidelity or outcome measures.
- Confusing dissemination with implementation.