Table 4: Logistic regression predicting depression at baseline and followup.

Source of variationDepression at baseline ( 𝑁 = 2 7 9 9 )Depression at followup ( 𝑁 = 1 8 4 1 )
BS.EOdds ratio 𝑃 BS.EOdds ratio 𝑃

Base model:
Age0.030.011.030.0040.020.021.030.25
Gender (0 = male, 1 = female)−0.140.170.870.44−0.330.360.720.35
Race
 Chinese0.500.811.640.54−0.581.130.560.61
 Malay−0.510.920.600.58−0.321.280.730.81
 Indian0.390.921.480.67−0.851.430.430.55
Education−0.100.080.900.22−0.020.170.980.90
Marital status
 Single0.450.331.570.170.860.612.350.16
 Married0.260.211.290.240.450.401.570.26
 Divorced, separated0.460.391.580.280.640.741.890.39
Number of medical problems0.170.051.180.0010.170.091.180.06
Baseline ADL−0.040.040.970.08−0.140.060.870.01
Baseline MMSE−0.090.020.92<0.001−0.0020.041.000.96
Social contact frequency score0.170.091.180.080.100.181.110.57
Social activities score−0.040.040.960.25−0.110.080.890.16
Production activities score−0.080.050.930.08−0.070.090.970.40
Fitness activities score0.070.041.070.12−0.040.090.790.70
Health activities score−0.110.040.900.008−0.240.080.930.004
Baseline GDS0.290.041.33<0.001
 Living alone versus with others (1,0)0.210.341.230.050.110.711.110.10
 Lonely versus not lonely (1,0)0.440.161.550.0010.330.361.390.03
 Living arrangements*Loneliness0.310.521.360.040.160.331.170.4