Table 3.

Regression models predicting incident Dementia in participants who were diagnosed with MCI at Baseline. Model 1 included demographic, cardiovascular risk factors and baseline MCI etiology; Model 2 adds recruitment factors.

	Model 1				Model 2
Coefficient	HR	HR 95% CI	p value		HR	HR 95% CI	p value
African American (reference: White)	0.66	0.56 — 0.77	<0.0001		0.71	0.61 — 0.84	<0.0001
Female (reference: Male)	1.01	0.91 — 1.11	0.8642		1.03	0.94 — 1.14	0.5172
Education: (reference: HS education)
under HS	1.02	0.81 — 1.28	0.8859		0.99	0.79 — 1.25	0.9542
over HS	0.79	0.68 — 0.93	0.0035		0.80	0.68 — 0.93	0.0049
Bachelor	0.93	0.81 — 1.08	0.3359		0.94	0.82 — 1.09	0.4149
Masters	0.85	0.73 — 0.98	0.0298		0.85	0.74 — 0.99	0.0343
Doctorate	0.81	0.68 — 0.97	0.0236		0.79	0.66 — 0.95	0.0112
unknown	2.14	0.79 — 5.77	0.1332		2.13	0.79 — 5.77	0.1350
Diabetes: (reference: non-diabetic)
recent/active	1.04	0.89 — 1.22	0.5856		1.07	0.91 — 1.25	0.4100
remote/inactive	1.77	1.12 — 2.79	0.0148		1.96	1.24 — 3.10	0.0040
unknown	1.33	0.60 — 2.94	0.4785		1.33	0.59 — 2.98	0.4885
Hypertension: (reference: non hypertensive)
recent/active	0.98	0.89 — 1.09	0.7567		0.98	0.89 — 1.08	0.6952
remote/inactive	1.01	0.77 — 1.32	0.9447		0.97	0.74 — 1.28	0.8546
unknown	0.91	0.39 — 2.15	0.8351		0.95	0.40 — 2.28	0.9086
Cardiac Event/Condition:* (reference: no event/condition)
recent/active	0.95	0.80 — 1.13	0.5712		0.95	0.80 — 1.14	0.5954
unknown	0.88	0.73 — 1.06	0.1770		0.87	0.72 — 1.05	0.1374
Primary etiology for MCI (reference: Alzheimer’s disease)
Vascular disease	0.60	0.46 — 0.79	0.0003		0.63	0.48 — 0.83	9.00E-04
Lewy body	1.17	0.93 — 1.47	0.1716		1.10	0.88 — 1.38	0.4079
other	0.64	0.52 — 0.79	<0.0001		0.63	0.51 — 0.77	<0.0001
missing/unknown	0.59	0.54 — 0.66	<0.0001		0.63	0.57 — 0.70	<0.0001
Referral: (reference: referred by self/relative/friend)
health professional					1.46	1.29 — 1.64	<0.0001
other†					0.88	0.77 — 1.00	0.0574
unknown					0.80	0.61 — 1.06	0.1225
Family History of Dementia: (reference: no family history)
≥1 1st degree relative					1.12	1.01 — 1.25	0.0256
unknown					1.00	0.83 — 1.21	0.9727

Reference for Baseline cognitively unimpaired: White, Male, High School education, non-diabetic, non-hypertensive, no history of cardiac event/condition, referred by self/relative/friend, and no family history of dementia.

The designation of “recent/active” was given if the condition occurred within the year prior to the study visit or required active management in the case of chronic conditions and the report was consistent with information obtained during the subject and study partner interview. The label “remote/inactive” was indicated if the condition existed or occurred in the past or more than one year ago, but had resolved or no longer needed treatment or management. As implied, if a condition was reported by either the participant or the study partner or indicated from other records, it was designated as “absent”. A label of “unknown” selected if there was insufficient data to determine the status of the condition.

^*Cardiac event/condition includes history of myocardial infarction (MI), atrial fibrillation (AFib), and congestive heart failure (CHF). “Recent/active” selected if any of the three conditions were “recent/active”; categorization “absent” if all three were “absent” or “remote/inactive”, and “unknown” for combinations of “unknown”, “absent”, and “remote/inactive”.

^†The category of “Other” for Referral Source represents a mixture of categories, including Alzheimer’s Disease Center (ADC) solicitation, non-ADC study, population sample, or non-ADC media appeal.

NOTE: As is evident in the table, variables of interest included a number of “unknown” responses. Sensitivity analyses were performed, wherein an “unknown” response for any of the covariates in the table resulted in the subject being excluded from analyses. Although sample size was markedly reduced, race results were practically unchanged.