Table 5.

Adjusted odds ratios (ORs) from logistic regression models predicting probability of death/refusal over the follow-up period among nonparticipants in the first biennial examination

VariableOR (death)*95% CIOR (refusal)*95% CI
* n = 1,604 + 77 deaths (1,681) between baseline and the 2-year follow-up; n = 1,604 + 117 (1,721) refusals between baseline and the 2-year follow-up. Smaller numbers are due to missing data in multivariable models.
† In the model together. All other variables were adjusted for age at baseline, sex, education, and baseline CASI score.
p ≤ 0.05.
§ p ≤ 0.01.
CASI = Cognitive Abilities Screening Instrument.
Age at baseline, y1.061.01–1.1§1.010.97–1.04
Female sex0.470.28–0.79§1.120.74–1.68
Education, y1.000.91–1.090.920.85–0.99
Baseline CASI, score0.970.93–1.000.940.91–0.96§
Olfaction (continual score)1.000.90–1.110.960.88–1.04
Normosmic at baseline1.00Reference1.00Reference
Microsmic at baseline0.980.56–1.711.681.06–2.67
Anosmic at baseline1.290.60–2.771.340.66–2.71
Smoking status (never/past/current)1.701.12–2.590.990.68–1.44
Pack-years of smoking1.010.99–1.021.000.99–1.01
Deteriorating sense of smell1.150.63–2.080.700.41–1.21