Recruitment of pre-dementia participants: main enrollment barriers in a longitudinal amyloid-PET study

Table 2 Binary logistic regression model including sociodemographic and clinical predictors for refusing participation in the AMYPAD-PNHS

	Research cohorts					Clinical cohorts
	B	SE(B)	Wald	p-value	Exp(B) [95%-CI]	B	SE(B)	Wald	p-value	Exp(B) [95%-CI]
Age (years)	0.01	0.01	2.05	0.15	1.01 [0.92–1.18]	− 0.01	0.02	0.16	0.69	0.99 [0.95–1.04]
Sex (female = 1)	0.14	0.14	1.00	0.32	1.15 [0.87–1.52]	− 0.44	0.51	0.72	0.39	0.65 [0.24–1.76]
Education (years)	-0.01	0.02	0.68	0.41	0.99 [0.95–1.02]	− 0.07	0.06	1.24	0.27	0.94 [0.83–1.05]
MMSE (score)	-0.06	0.04	1.99	0.16	0.94 [0.86–1.02]	− 0.22	0.11	3.89	0.048*	0.80 [0.64–1.00]
Family history (none of the parents = 1)	-0.01	0.15	0.01	0.93	0.99 [0.73–1.33]	2.08	0.70	8.79	0.003*	8.02 [2.03–31.8]
Number of prior visits	0.04	0.07	0.39	0.53	1.04 [0.92–1.18]	− 0.93	0.19	23.3	0.000*	0.40 [0.27–0.58]

Logistic models included participant status (consent = 0, decline = 1) as dependent variable and age, sex, years of education, MMSE score, dementia in family history (none or ≥ 1 of the parents), and the number of prior visits as independent variables. As external factors (Fig. 1) are expected to universally affect participants irrespective of individual characteristics, participants assigned to this reason of decline were excluded from these analyses (n = 95). Research cohorts: n = 1263 (n consent = 976, n decline = 287). Clinical cohorts: n = 243 (n consent = 215, n decline = 31). Asterisks indicate predictors that add significantly to the prediction (p < .05)

ISSN: 1758-9193