TABLE 2.

Overview of challenges and lessons learned from Alzheimer’s disease (AD) trials with regard to clinical outcome assessments (COAs), as well as a recommendation framework for the evaluation of performance based cognitive tests based on the Consensus-Based Standards for the Selection of Health Measurement Instruments (COSMIN) methodology

Challenge	Whatwe have learned	Related validity aspects	Considerations	Recommendations (for examples of applications, see Section 3.3)
What to measure	Existing tests are less focused on the cognitive processes that are relevant in early (pre)clinical stages of AD.	Content validity (including face validity) Interpretability (including ecological validity)	All items refer to relevant aspects of the construct to be measured AND together comprehensively reflect the construct to be measured. All test aspects (items, Instructions, response options) are understood by the target population as Intended.	a. Test should have a clear theoretical framework with respect to the construct of Interest and target population. b. The target population and clinical experts should be involved In development of test content and material. c. Test should be shown to be suitable (i.e., comprehensible, relevant, comprehensive, culturally appropriate) for the target population.
How to measure	Tests are insufficiently sensitive to disease progression in early stages of AD. It remains unclear whether (changes in) test scores are clinically meaningful.	Construct validity (including structural validity) Interpretability (Including clinical meaningfulness)	The scoring algorithm (single score, weighted/unweighted composite score) is an adequate reflection of the dimensionality (one or multiple domains) being measured. Test captures the construct of interest and Is associated with related constructs and not associated with unrelated constructs. Test is sensitive to disease progression and stable In the absence of change.	a. A factor analysis should be performed to demonstrate adequate model fit and score use (single score or domain scores). b. Test score should be validated against relevant clinical (e.g., everyday functioning, quality of life) and/or biological measures (e.g., neurodegeneration, amyloid or tau biomarkers) and relationships are in line with hypotheses. c. An Independent validation study on the ability to capture cognitive change in the target population should be performed. d. It should be examined what (change In) scores patients and caregivers perceive as clinically meaningful.
Who to measure	Tests are insufficiently suitable across groups (e.g., different cultures). Tests are differentially sensitive to change across different stages of AD.	Cross-cultural validity (Including measurement Invariance) Longitudinal validity (Including responsiveness)	Test performance is not Influenced by cross-cultural differences OR a cross-cultural adaptation has been made. Test is equally sensitive to change across disease stages OR is known to be sensitive to change In the specific target population.	a. A representative and diverse target population should be Involved in test development and validation. b. Test should have no measurement Invariance across groups and show no Important differential item functioning when comparing groups based on demographic characteristics (such as age, sex, education) and cultural aspects. c. The selection of tests/items should be adapted to disease stage of Interest. d. Test should show no or limited range restrictions in scoring in the target population.