. 2014 Oct 22;2(4):e43. doi: 10.2196/mhealth.3710

Table 3.

Multivariate logistic regression model for factors predicting the download of mobile health apps (N=848).^a

Variable		Adjusted OR	95% CI	P value
Age (years)		0.96	0.95-0.97	<.001
Race/ethnicity				.002^b
	Caucasian	reference
	Filipino	0.89	0.54-1.48	.66
	Korean	0.52	0.31-0.88	.02
	Latino	0.37	0.20-0.69	.002
Education				.005^b
	High school or some high school	reference
	College or some college	2.62	1.44-4.80	.002
	Graduate school	2.93	1.43-6.0	.003
Family member with MI^c		2.02	1.16-3.51	.013
Type of survey				.001
	Online survey	reference
	Paper survey	0.50	0.34-0.75

^aBackward elimination step-wise multiple logistic regression (Wald). Variables entered in initial model: race/ethnicity, age, gender, marital status, years lived in the United States, education, primary language is English, body mass index, smokes cigarettes, has high blood pressure, has high cholesterol, family member with diabetes, family member with heart attack, physical inactivity, perceived risk for MI, self-reported health status, discussed diabetes with provider, and survey type completed.

^bOverall P value.

^cMI: myocardial infarction.