Table 3.

Multinomial Logistic Regression NAMCS (2005–2016, 2018).

		SGA vs. no antipsychotic			FGA vs. no antipsychotic			SGA vs. FGA
		AOR	95% CI	Sig	AOR	95% CI	Sig	AOR	95% CI	Sig
Predisposing factors
Age group
65+ yrs. vs 18–64 yrs		0.435	[0.309,0.612]	***	0.484	[0.250,0.936]		0.900	[0.461,1.759]
Gender
Male vs. female		1.144	[0.879,1.489]		1.128	[0.749,1.700]		1.014	[0.717,1.435]
Race/Ethnicity
Hispanic vs. Black		1.414	[0.809,2.473]		1.538	[0.747,3.167]		0.919	[0.460,1.837]
White vs. Black		1.209	[0.824,1.775]		0.883	[0.523,1.493]		1.369	[0.847,2.213]
Enabling Factors
Area
Non-metro vs. Metro		1.085	[0.601,1.959]		1.920	[0.855,4.309]		0.565	[0.307,1.041]
Insurance
Private/vs. Public		0.850	[0.577,1.250]		0.542	[0.315,0.935]		1.566	[0.949,2.585]
Others vs. Public		0.934	[0.632,1.380]		0.917	[0.512,1.640]		1.019	[0.597,1.736]
Need Factors
Patient establishment
Established vs. New		0.881	[0.513,1.512]		1.747	[0.782,3.904]		0.504	[0.240,1.061]
Visit reason
Routine vs. not routine		2.670	[1.775,4.016]	***	2.066	[1.146,3.725]	*	1.292	[0.713,2.341]
Number of medications
≥5 meds vs. <5 meds		1.202	[1.006,1.436]	*	1.186	[0.997,1.410]		1.013	[0.968,1.061]
Number of chronic conditions
≥3 conditions vs. <3 conditions		0.984	[0.970,0.997]	*	0.976	[0.956,0.998]	*	1.007	[0.982,1.033]

Note: Based on 1932 adults (weighted N = 38,403) adults (age ≥ 18 years) with schizophrenia between 2005 and 2015, 2016, and 2018.

Asterisks represent statistically significant group differences by type of treatment compared to the reference group based on multinomial logistic regression at α level of 0.05. The reference group for the dependent variable in the multinomial logistic regression was “no antipsychotic” for the first two columns and “typical antipsychotic” for the last column. Statistical tests were not performed to evaluate the significance of groups with a sample size <30. Abbreviations: SGA: second-generation antipsychotic; FGA: first-generation antipsychotic.

***p < 0.001; **0.001 ≤ p < 0.01; *0.01 ≤ p < 0.05.