Skip to main content
. 2019 Feb 11;16(3):504. doi: 10.3390/ijerph16030504

Table 1.

Cross-sectional regression parameter estimates using GEE 1 for EGFR = HIV infection, serum albumin and BUN in the Bronx WIHS.

Variable Working Correlation Structure
Independence Equicorrelation 2
Point Estimate 95% CI Z-Value (p) Point Estimate 95% CI Z-Value (p)
HIV Infection
HIV,CS)
−2.04 (−5.07, 0.98) −1.32 (<0.19) −3.96 (−6.90, −1.03) −2.65 (0.0081)
Albumin Per g/dL
ALB,CS)
−6.21 (−8.95, −3.47) −4.44 (<0.0001) −9.84 (−12.01, −7.68) −8.93 (<0.0001)
BUN Per mg/dL
BUN, CS)
−1.87 (−2.12, −1.62) −14.45 (<0.0001) −1.22 (−1.46, −0.99) −10.30 (<0.0001)
Quasi-Likelihood Information Criteria (QIC) 10,847.14 10,836.27

1 Mixed models gave essentially similar point estimates; see Appendix A. 2 Interclass correlation of residuals from GEE-E was 0.45 indicating non-independence correlation was structurally correct.