Evaluation
What the domain tests
Evaluation items ask whether you can move from a referral to a defensible baseline and a plan. Start with the occupational profile: who the client is, what they need and want to do, and the contexts that shape performance. Then analyze occupational performance by observing the occupation and measuring the performance skills, client factors, and activity demands that explain what you see. Top-down reasoning is the default; bottom-up testing (MMT, goniometry, cognitive screens) is used to explain a hypothesis, not to generate one.
Concepts to own
Know the difference between screening (flagging risk, not diagnosing), evaluation (comprehensive, standardized or structured), and reassessment. Pick tools using psychometrics: reliability (interrater for observational scales), validity for the population and setting, sensitivity versus specificity, responsiveness and minimal clinically important difference, and minimal detectable change. An interview (COPM, OCAIRS) yields the client's perspective and goals; a performance measure yields capacity.
Traps
Choosing the tool you know best rather than the one that answers the referral question; using an instrument outside its validated population or age range; treating a screen as a full evaluation; quoting a score without considering MDC; ignoring environment, culture, and caregiver context; reading "best" as "most thorough" when the stem asks for the most efficient or most occupation-based option.
Rule of thumb
Ask three questions before selecting any assessment: What question am I answering? Does this tool measure that in this population? Is its evidence strong enough for the decision I am about to make? If the stem concerns real-world safety or performance, prefer direct observation of the occupation in a realistic context over isolated component testing.
Worked example
A client three weeks after a left CVA is going home, and the team asks whether he can safely prepare a simple meal. The best answer is structured observation of meal preparation in a simulated kitchen, because it samples the actual domain of concern and predicts real performance. A manual muscle test or a paper-and-pencil cognitive screen describes components but does not answer the referral question.