Table 1.

Type of model selected for each outcome measure.

		Generalized linear models
Outcome	Measure	Distribution	Link function
Number of falls in the previous year	Count	Poisson	Log
Fallers: non-Fallers	Binary	Binomial	Logistic
Number of medical conditions $\ge$ 4	Binary	Binomial	Logistic
Number of medications $\ge$ 4	Binary	Binomial	Logistic
Normal eyesight: Abnormal eyesight, $n$ (%)	Binary	Binomial	Logistic
Using bifocal or multifocal eyeglasses, $n$ (%)	Binary	Binomial	Logistic
Other eye conditionsa (treatment: non-treatment), (%non-treatment)	Nominal	Multinomial	Cumulative Logit
Appropriate footwear: Inappropriate footwear,b $n$ (%inappropriate footwear)	Binary	Binomial	Logistic
Having $\ge$ 4 home hazard environments, $n$ (%)	Binary	Binomial	Logistic
Regularly go to toilet at night $\ge$ 2 times, $n$ (%)	Binary	Binomial	Logistic
Regularly go to toilet at night $\ge$ 4 times, $n$ (%)	Binary	Binomial	Logistic
Number of medical conditions	Quantitative	Gamma	Identity
Number of medications	Quantitative	Gamma	Identity
Functional Reach test	Quantitative	Normal	Identity
Step Test (worst side)	Quantitative	Normal	Identity
Timed Up and Go test (TUG)	Quantitative	Gamma	Identity
TUG (secondary manual task)	Quantitative	Gamma	Identity
TUG (secondary cognitive task)	Quantitative	Gamma	Identity
Timed Chair Stand	Quantitative	Gamma	Identity
Hand reaction time	Quantitative	Gamma	Identity
Modified PASEc score	Quantitative	Gamma	Identity
Exercise frequency (times/week)	Quantitative	Gamma	Identity
Total exercise time (hours/week)	Quantitative	Gamma	Identity
Modified Falls Efficacy Scale	Quantitative	Gamma	Identity
MMSEd	Quantitative	Gamma	Identity
Number of home environmental hazards	Quantitative	Gamma	Identity

^aOther eye conditions including cataract, glaucoma, Pterygium, Pinguecula; ^bInappropriate footwear including poorly fitted shoes/slippers, unstable shoes, slippery shoes, shoes with $>$ 1 inch high-heel, worn-out footwear; ^cModified PASE (hours/week); ^dMini-Mental State Examination (Thai version).