Measures and methods for outcome assessment (ie, cognitive function)
| Publication | Domain of cognitive function assessed (name of measure) | Data collection procedure | Analyses used | Validity and reliability |
|---|---|---|---|---|
| Cohort designs | ||||
| Hamer and Stamatakis23 | Processing speed (immediate word recall), memory (delayed word recall) and executive function (verbal fluency)31 | Immediate word recall: read 10 words and recalled as many words as possible immediately after. Delayed word recall: using same list, recalled words after they completed other cognitive function tests. Verbal fluency: named as many animals as possible in 1 min. A global cognitive function score calculated from the sum of standardised scores on each test | Type of regression: linear mixed models with random effect intercept; Covariates and confounders: age sex, smoking, alcohol, physical activity, social status, disability, chronic illness and body mass index (BMI) | Unknown |
| Kesse-Guyot et al24 | Lexical-semantic memory (phonemic fluency and semantic fluency36), episodic memory (RI-48 test35), working memory (Digit span forward and backward33), Executive function (Delis-Kaplan Trail Making Test34) | Phonemic fluency: cited as many words as possible in 2 min beginning with the letter ‘p’. Semantic fluency: named as many animals as possible in 2 min. RI-48 test: delayed cued recall test. Digit span forward and backward: repeated sequence of seven digits, forward and backward. Trail Making Test: connecting numbers and letters alternating between the two series | Type of regression: structural equation modelling; Covariates and confounders: age, gender, education, time lag between baseline and cognitive evaluation, occupation, energy intake, number of 24-hour records, BMI, depressive symptoms, memory issues, diabetes, hypertension and cardiovascular disease | Phonemic and semantic fluency: test–retest reliability r=0.82.51 RI-48: classified 88% of people with mild cognitive impairment (MCI) or Alzheimer's disease (AD) correctly.35 Digit span: B=0.64 in confirmatory factor analysis.52 Trail Making: r=−0.38 compared to Stroop Color Word Score53 |
| Publication | Domain of cognitive function assessed | Data collection procedure | Analyses used | Validity and reliability |
| Kesse-Guyot et al25 | Lexical-semantic memory (phonemic fluency and semantic fluency36), episodic memory (RI-48 test35), working memory (digit span forward and backward33), executive function (Delis-Kaplan Trail Making Test34) | Phonemic fluency: cited as many words as possible in 2 min beginning with ‘p’. Semantic fluency: named as many animals as possible in 2 min. RI-48 test: delayed cued recall test. Digit span forward and backward: repeated sequence of 7 digits, forward and backward. Trail Making Test: connecting numbers and letters alternating between the two series | Type of regression: principal component analysis | Phonemic and semantic fluency: test–retest reliability r=0.82.51 RI-48: classified 88% of people with MCI or AD correctly.35 Digit span: B=0.64 in confirmatory factor analysis.52 Trail Making: r= −0.38 compared to Stroop Color Word Score53 |
| Case-control designs | ||||
| Kivipelto et al26 | Screened for cognitive impairment and all-cause dementia | Mini-Mental State Examination (MMSE)37 at screening phase and for those who scored <24, neuropsychological examinations conducted to screen for all-cause dementia, according to the Diagnostic and Statistical Manual of Mental Disorders38 | Type of regression: multiple logistic regressions | MMSE: test–re-est reliability: r=0.89; validity: r=0.78 compared to Verbal IQ and r=0.66 compared to performance IQ54 |
| Lindstrom et al27 | Screened for AD | Cases evaluated by neuropsychological, laboratory and neurological examinations | Type of regression: unconditional logistic regression model | Unknown |
| Cross-sectional designs | ||||
| Rosenberg et al28 | Executive function and processing speed (Trail Making Test41) | Trails A was completed first, followed by Trails B. Both items were scored using completion time in seconds, and scores for participants who were unable to complete the examination were set to the maximum value (300 s). Executive function was estimated by subtracting time of Trails A from Trails B | Type of regression: linear mixed-effects models Covariates and confounders: age, gender, marital and educational status. | Trail Making Test: Reliability r=0.60–0.90;41 discriminant validity: t=16.20 (p<0.001).54 |
| Steinberg et al29 | CogState computerised battery.43 Cognitive tests assessed: (1) processing speed; (2) visual attention, recognition and memory; (3) verbal learning and memory; (4) immediate recall; (5) delayed recall; (6) working memory and (7) problem-solving and reasoning | Participants administered the CogState tests over a 35 min period | Type of regression: linear regression analyses Covariates and confounders: age, sex, race and education | Validity: r=0.49–0.8355 |
| Vance et al30 | Memory (Benton Visual Retention Test45); processing speed and executive function (Trail Making Test41); visual memory, perceptual organisation and planning (Rey-Osterrieth Complex Figure Copy and Recall Tests46) | Benton Visual Retention Test: shown a series of geometric designs and then draw from memory. Trail Making Test: connected 25 alternating number and letter circles in sequence as quickly as possible. Rey-Osterrieth Complex Figure and Recall Tests: reproduced complex figure while present and then from memory | Types of regression: latent growth model; Covariates and confounders: age, depression and physical activity (PA) | Benton Visual Retention Test: inter-rater reliability r=0.80–0.90.54 Trail Making Test: reliability r=0.60–0.90;41 discriminant validity: t=16.20 (p<0.001).54 Rey-Osterrieth Complex Figure Copy and Recall Tests: reliability r=0.9046 |