Table 2. Clinical Outcomes Among 30 086 Children in the Retrospective Cohort.

	No. (%)		Stratified Analysisa		Full Matched Analysisb
	Broad-Spectrum Antibiotics (n = 4296)	Narrow-Spectrum Antibiotics (n = 25 790)	Risk Difference (95% CI), %c	P Value	Risk Difference (95% CI), %c	P Value
Primary Outcomes at 14 d
Treatment failured	147 (3.4)	809 (3.1)	0 (−0.6 to 0.7)	.88	0.3 (−0.4 to 0.9)	.39
Adverse eventse	157 (3.7)	695 (2.7)	0.9 (0.3 to 1.6)	.001	1.1 (0.4 to 1.8)	<.001
Secondary Outcomes at 30 d
Treatment failured	372 (8.7)	2082 (8.1)	0.3 (2.6 to 2.8)	.51	0.6 (−0.4 to 1.6)	.22
Adverse eventse	189 (4.4)	849 (3.3)	1.1 (0.4 to 1.8)	.001	1.2 (0.5 to 1.9)	<.001

^aComparisons were made within diagnosis and clinician strata. Risk differences were estimated using fixed-effects linear regression (first differencing method) and diagnosis and clinician strata were included as fixed effects.

^bChildren prescribed narrow-spectrum antibiotics were optimally matched to children prescribed broad-spectrum antibiotics based on a propensity score estimated from patient-level and clinical-level characteristics creating matched sets of difference sizes using all children. Risk differences were estimated using weighted logistic regression with marginal standardization.

^cA risk difference of more than 0 means that broad-spectrum antibiotics had a higher risk of the adverse outcome than narrow-spectrum antibiotics.

^dDefined as having the same acute respiratory tract infection diagnosis during the in-person or telephone encounter and a new prescription for a systemic antibiotic.

^eDefined as encounters for diarrhea, vomiting, candidiasis, noncandidal rash, other or unspecified allergic reaction, and other or unspecified adverse event.