Table 2.

Comparison of results from pooled individual-level data analysis and distributed regression analysis based on data from 34 healthcare organizations, organized into 27 distinct network partners (or distinct databases), in the PCORnet Antibiotics and Childhood Growth Study. The results were from a linear regression model that considered antibiotic use as a binary variable (any use vs. no use) and body mass index z-score as the continuous outcome variable among patients without complex chronic conditions. The model included all the covariates in the table plus 26 indicator variables for network partners.

Variable	Parameter estimate		Standard error
	Pooled individual-level data analysis	Distributed regression	Pooled individual-level data analysis	Distributed regression
Any antibiotic use (vs. no use)a	0.03419	0.03419	0.00478	0.00478
Female (yes vs. no)	0.01466	0.01466	0.00418	0.00418
Age at outcome (in months)b	0.00489	0.00489	0.00040	0.00040
Race
Asian	−0.20892	−0.20892	0.01070	0.01070
Black	0.05863	0.05863	0.00522	0.00522
Other	0.03098	0.03098	0.00890	0.00890
Unknown	0.03669	0.03669	0.00794	0.00794
White	REF	REF	REF	REF
Hispanic ethnicity (yes vs. no)	0.33664	0.33664	0.00679	0.00679
Preterm birth status (yes vs. no)	−0.22002	−0.22002	0.00928	0.00928
Asthma diagnosis (yes vs. no)	0.15297	0.15297	0.00694	0.00694
No. of infection episodesa,b,c	0.02184	0.02184	0.00195	0.00195
No. of corticosteroid prescription episodesa,b, c	0.06124	0.06124	0.00394	0.00394
No. of healthcare encountersa,b,d	−0.01568	−0.01568	0.00214	0.00214

^aMeasured before 24 months of age

^bAdjusted for as a continuous variable in the model

^cRe-coded as 0, 1, 2, 3, 4+

^dLog-transformed