Table 4.

Association between body weight, BMI, and route and frequency of substance administration among African American females

		Weight (kg)	BMI (kg/m2)
Routea
Injecting speedball or cocaine (vs. no)	β coefficient	−7.084	1.563
	SE	(7.254)	(3.171)
	P value	0.330	0.622
Injecting heroin (vs. no)	β coefficient	−8.478	−2.800
	SE	(6.533)	(2.871)
	P value	0.196	0.330
Snorting or sniffing cocaine (vs. no)	β coefficient	5.335	2.694
	SE	(5.008)	(1.646)
	P value	0.288	0.103
Snorting or sniffing heroin (vs. no)	β coefficient	−4.201	−2.479
	SE	(3.695)	(1.315)
	P value	0.257	0.061
Smoking crack or cocaine (vs. no)	β coefficient	−2.832	−1.734
	SE	(3.243)	(1.212)
	P value	0.383	0.154
Smoking heroin (vs. no)	β coefficient	−15.344	−5.303
	SE	(5.079)	(2.148)
	P value	0.003	0.014
Frequencyb
Active marijuana use
<1/week vs. no	β coefficient	−2.523	−0.945
	SE	(3.702)	(1.354)
	P value	0.496	0.486
≥1/week vs. no	β coefficient	−5.619	−1.458
	SE	(3.473)	(1.349)
	P value	0.107	0.281
Active opiate (other than heroin) use (vs. no)
<1/week vs. no	β coefficient	−6.192	−2.724
	SE	(4.573)	(1.756)
	P value	0.177	0.122
≥1/week vs. no	β coefficient	−6.125	0.772
	SE	(8.500)	(3.261)
	P value	0.472	0.813
Active speedball, crack or cocaine use (vs. no)
<1/day vs. no	β coefficient	−2.669	−1.403
	SE	(3.448)	(1.269)
	P value	0.439	0.270
≥1/day vs. no	β coefficient	−5.032	−3.291
	SE	(5.445)	(1.767)
	P value	0.356	0.064
Active heroin use
<1/day vs. no	β coefficient	−4.685	−2.343
	SE	(3.872)	(1.368)
	P value	0.227	0.088
≥1/day vs. no	β coefficient	−13.893	−4.499
	SE	(4.634)	(1.885)
	P value	0.003	0.018
Active other drug use (e.g., stimulants, sleeping pills, buprenorphine)
<1/week vs. no	β coefficient	2.779	−0.251
	SE	(5.853)	(1.860)
	P value	0.635	0.893
≥1/week vs. no	β coefficient	9.220	4.950
	SE	(7.858)	(3.557)
	P value	0.242	0.165
Active alcohol use
<2/week vs. no	β coefficient	−4.964	−1.559
	SE	(3.098)	(1.089)
	P value	0.110	0.153
≥2/week vs. no	β coefficient	−0.403	0.485
	SE	(3.767)	(1.415)
	P value	0.915	0.732

^aLinear regression models were adjusted for age, unemployment, income, marital status, current pregnancy, depression, HIV infection, current smoking, active alcohol use, active marijuana use, active opiate use, active other drug use, and different cocaine/heroin routes other than the route of interest.

^bLinear regression models were adjusted for age, unemployment, income, marital status, current pregnancy, depression, HIV infection, current smoking, active alcohol use, and different types of drug use other than the drug of interest.