International Scholarly Research Notices

Clinical Study

Pharmacodynamics Must Inform Statistics: An Example from a Cocaine Dependence Pharmacotherapy Trial

Table 2

Systolic and diastolic blood pressure results as a function of inclusion of baseline as a covariate in the statistical model.


Parameter	Analysis not including baseline as a covariate^a					Analysis with baseline as a covariate^a
Parameter	Estimate	SE	95% confid. limits			Estimate	SE	95% confid. limits

Systolic blood pressure
Intercept	124.80	1.74	121.38–128.21	71.66	<0.0001	28.10	6.67	15.0–41.19	4.20	<0.0001
Baseline	NA	NA	NA	NA	NA	0.78	0.05	0.67–0.88	14.51	<0.0001
Medication	−3.54	2.24	−7.94–0.86	−1.58	0.11	−4.16	1.37	−6.84–1.48	−3.04	0.002
Week	0.01	0.13	−0.26–0.27	0.01	0.99	−0.0193	0.13	−0.28–0.24	−0.15	0.88
Med by week	0.02	0.19	−0.36–0.40	0.09	0.93	0.02	0.19	−0.37–0.39	0.07	0.94

Diastolic blood pressure
Intercept	79.33	1.28	76.82–81.84	61.84	<0.0001	20.59	4.26	12.25–28.94	4.84	<0.0001
Baseline	NA	NA	NA	NA	NA	0.74	0.05	0.64–0.84	13.99	<0.0001
Medication	−2.91	1.66	−6.16–0.34	−1.75	0.08	−3.03	1.06	−5.11– −0.96	−2.87	0.004
Week	−0.04	0.11	−0.25–0.17	−0.38	0.70	−0.04	0.11	−0.25–0.16	−0.42	0.67
Med by week	−0.03	0.16	−0.35–0.28	−0.18	0.85	−0.05	0.16	−0.36–0.26	−0.30	0.76

Data from study weeks 1–12 included in analysis.