Skip to main content
. Author manuscript; available in PMC: 2014 Mar 1.
Published in final edited form as: Biometrics. 2013 Jan 22;69(1):197–205. doi: 10.1111/j.1541-0420.2012.01812.x

Table 1.

Estimates of regression coefficients and standard errors from GEE fits for: Epo (epoetin alpha); and Hemo (hematocrit). The covariates selected are based on those that are the major predictors identified using LASSO. The regression model also includes demographic and comorbidity variables for adjustment (not shown). For this analysis quadratic terms are consider for only the first lag, t − 1.

Transition Model LASSO-based Model

Outcome: EPO(t)

Variable Coef. (std. err.) Z Coef. (std. err.) Z
Hemo(t − 1) −0.0717 (0.0015) −48.66 −0.0591 (0.0151) −3.90
Hemo(t − 1)2 −0.0017 (0.0002) −11.06

Hemo(t − 2) 0.0025 (0.0014) 1.76
Hemo(t − 3) −0.0014 (0.0015) −0.96
Hemo(t − 4) 0.0024 (0.0013) 1.84
Hemo(t − 5) 0.0020 (0.0010) 1.97

Epo(t − 1) 0.5320 (0.0115) 46.33 0.1517 (0.0764) 1.98
Epo(t − 1)2 0.1280 (0.0061) 21.09

Epo(t − 2) 0.1282 (0.0102) 12.61 −0.3380 (0.0854) −3.95
Epo(t − 3) 0.0840 (0.0091) 9.24 0.2520 (0.0611) 4.12
Epo(t − 4) 0.0800 (0.0088) 9.14
Epo(t − 5) 0.0610 (0.0075) 8.14

Epo(t − 1):Epo(t − 2) −0.0946 (0.0102) −9.29
Epo(t − 1):Epo(t − 3) −0.0971 (0.0095) −10.21
Epo(t − 2):Epo(t − 3) 0.0832 (0.0077) 10.75
Hemo(t − 1):Epo(t − 2) 0.0142 (0.0012) 12.13

Outcome: Hemo(t)

Epo(t) −0.5070 (0.0567) −8.95

Hemo(t − 1) 0.5253 (0.0108) 48.50 0.2336 (0.0719) 3.25
Hemo(t − 1)2

Hemo(t − 2) 0.0783 (0.0096) 8.19
Hemo(t − 3) 0.0150 (0.0086) 1.76
Hemo(t − 4) 0.0309 (0.0084) 3.67
Hemo(t − 5) 0.0482 (0.0068) 7.07

Epo(t − 1) 2.0111 (0.0670) 30.04 −3.3692 (0.4720) −7.14
Epo(t − 1)2 0.2389 (0.0260) 9.19

Epo(t − 2) −0.1989 (0.0529) −3.76 3.4010 (0.4630) 7.35
Epo(t − 3) −0.3302 (0.0478) −6.92
Epo(t − 4) −0.4149 (0.0485) −8.56
Epo(t − 5) −0.2126 (0.0422) −5.04

Epo(t − 1):Epo(t − 2) −0.3183 (0.0320) −9.95
Hemo(t − 1):Epo(t − 1) 0.1022 (0.0102) 9.98
Hemo(t − 1):Epo(t − 2) −0.0535 (0.0106) −5.05