Impact of BMI and BMI change on future drug expenditures in adults: results from the MONICA/KORA cohort study

Table 4 Impact of baseline BMI and BMI change on drug expenditures – two-step regression analysis

	Model 1				Model 2
	Impact of baseline BMI only				Impact of baseline BMI and BMI change
	1. Probability		2. Amount of costs if >0		1. Probability		2. Amount of costs if >0
	n = 2,946		n = 2,003		n = 2,946		n = 2,003
Parameter	Odds ratio	p-value	Exp(est)^a	p-value	Odds ratio	p-value	Exp(est)^a	p-value
Intercept	0.164	0.055	37.462	<.0001	0.168	0.058	33.191	<.0001
Age	0.974	0.449	0.951	0.018	0.971	0.399	0.950	0.008
Age squared	1.001	0.010	1.001	0.004	1.001	0.007	1.001	0.001
Sex: female	2.179	<.0001	0.967	0.582	2.162	<.0001	0.960	0.462
SES ^b,d		0.347		0.112		0.324		0.045
Lower middle class	1.036	0.817	0.922	0.389	1.034	0.830	0.929	0.387
Middle middle class	0.984	0.913	0.894	0.256	0.987	0.930	0.914	0.314
Upper middle class	1.247	0.137	1.084	0.400	1.253	0.129	1.095	0.292
Upper class	1.108	0.514	1.079	0.476	1.120	0.472	1.123	0.233
BMI ^c,d		<.0001		0.001		<.0001		<0.0001
Overweight	1.188	0.087	1.132	0.089	1.189	0.086	1.126	0.073
Moderately obese	1.776	<.001	1.431	<.001	1.769	<.001	1.426	<.0001
Severely obese	4.068	<.0001	1.361	0.025	4.134	<.0001	1.417	0.005
BMI change	–	–	–	–	1.037	0.087	1.055	<.0001
Scale	2.718	–	1.839	–	2.718	–	2.099	–
AIC	3228		14385		3227		14322

^aExponents of regression estimates are reported, that can be interpreted as factors.
Reference: ^bLower social class; ^cNormal weight.
^dp-values for overall impact of group variable are given.

ISSN: 1472-6963