Cognitive Assessment Instruments

Cognitive instruments in this library cover four distinct tasks. Brief screening tests such as the Mini-Mental State Examination, the Addenbrooke's Cognitive Examination and the Rowland Universal Dementia Assessment Scale give a global score in a short sitting. Domain-specific tests such as the Frontal Assessment Battery, the Trail Making Test and the Clock Drawing Test probe particular functions, most often executive function and visuospatial construction. Delirium tools such as the Confusion Assessment Method and the Delirium Rating Scale detect an acute, fluctuating state that global screeners routinely miss. Staging scales such as the Clinical Dementia Rating, the Global Deterioration Scale and the Functional Assessment Staging Tool describe where someone is in a disease course rather than testing them directly.

Informant questionnaires such as the Informant Questionnaire on Cognitive Decline in the Elderly ask a person who knows the patient well what has changed, which is often more informative than a single test score because it establishes a trajectory rather than a point.

A screening score is not a diagnosis and was never meant to be. These tests are sensitive to education, first language, literacy, sensory impairment and cultural familiarity with the test material, and the same score can mean very different things in two people. The Rowland Universal Dementia Assessment Scale exists specifically because widely used screeners perform poorly across culturally and linguistically diverse populations.

In research, cognitive endpoints need particular care with practice effects. Repeated administration of the same test improves performance independently of any treatment, which can hide a real decline or manufacture an apparent improvement. Alternate forms, adequate intervals between administrations, and a run-in assessment are the usual mitigations, and the choice needs to be made at design time because it cannot be corrected afterwards.

Check the education and language norms. Cut-offs derived in one educational and linguistic population do not transfer, and applying an imported cut-off will systematically misclassify people. Where adjusted norms exist, use them and say which were used.

Check whether the test is licensed. Cognitive instruments are among the most likely in this library to be commercially controlled, and some restrict administration to qualified users. The position on the Mini-Mental State Examination in particular has changed over time and should be confirmed rather than assumed.

Check the mode of administration. Several of these tests involve drawing, object naming or physical manipulation, and they do not transfer to a self-completed digital form without changing what is being measured. Remote and telephone adaptations exist for some instruments and are separate instruments with their own validation, not the original delivered differently.

Is a low screening score enough to diagnose dementia?

No. Screening tests indicate that further assessment is warranted. Diagnosis requires history, an informant account, functional assessment, and exclusion of reversible causes including delirium, depression, medication effects and sensory impairment.

Which cognitive test works across languages and cultures?

The Rowland Universal Dementia Assessment Scale was designed for this and is the usual choice where a population is culturally and linguistically diverse. Even then, validated translations and locally derived cut-offs matter, because a translated test is not automatically a validated one.

How do I avoid practice effects in a longitudinal study?

Plan for them at design time. Use alternate forms where they exist, leave adequate intervals between administrations, include a practice or run-in assessment that is not analysed, and expect some improvement in the first repeat regardless of treatment.

Can cognitive tests be delivered remotely?

Some can. Instruments that depend on drawing, physical objects or examiner observation change substantially when moved to an unsupervised digital format. Where a validated remote or telephone version exists, use that specific version and report it as such rather than treating it as the original test.