- Comparisons for refusal rates (df = 1, n = 1856) and the prevalence of a given reason for refusal (df = 1, n = 535) are performed for a single PC versus all other PCs (white rows) and research versus clinical PCs (gray rows). (A) Continuous variables are reported in mean ± sd and categorical variables in n (%). Not all PCs provided data for both consented and declined participants: missing cases are reported in supplementary Table S2. (B) Refusal rate (%) = n informed / n declined for research PCs (EPAD-LCS, ALFA + , F-PACK, EMIF-AD 60 + +) and clinical PCs (UCL-2010–412, EMIF-AD 90 + , AMYPAD DPMS, Microbiota, FACEHBI). Orange indicates an above-average refusal rate and green a below-average refusal rate. Asterisks indicate a significant difference (p < .05) between the refusal rate in research versus clinical PCs (darker shade) or for this main PC versus all other PCs (lighter shade) according to a chi-squared test (df = 1, n = 1856). (C) The n (%) is reported which shows the absolute and relative number of times a subcategory for decline was grouped under one of the main categories 1–5 described in Figs. 1 and 4. Asterisks indicate significant differences (p < .05) between the prevalence of a given reason between research versus clinical PCs (gray) or for this main PC versus all other PCs (white) according to a chi-squared test (df = 1, n = 535). Crosses indicate unreliable chi-squared testing based on violation of the assumptions (≥ 80% of the cells had an expected value ≤ 5 or one cell had an expected value of < 1) [53]. PC, parent cohort; PET, positron emission tomography; MMSE, Mini-Mental State Examination; NA, not applicable
Alzheimer's Research & Therapy