Table 2.

IRR^† estimates and confidence intervals for Actinomyces counts after adjusting for baseline counts in sites that had an initial PD≥5 mm from the GEE NB model using the exchangeable correlation structure in Stata

		Part (c) Comparison made to counts after 3 months
	(b) IRR: 3 months	6 months	12 months	18 months	24 months
Treatment effect on Actinomyces counts	Marginal effects estimate 99% CI	IRR estimate 99% CI	IRR estimate 99% CI	IRR estimate 99% CI	IRR estimate 99% CI
(a) Reference treatment: SRP.		1.48 (0.69, 3.15)	1.43 (0.74, 2.77)	1.42 (0.54, 3.78)	1.06 (0.44, 2.54)
Compared with SRP
SURG+AMOX+MET	0.83 (0.39, 1.78)	1.03 (0.44, 2.44)	1.44 (0.64, 3.19)	1.26 (0.44, 3.60)	1.69 (0.65, 4.39)
SURG+TET	1.09 (0.51, 2.35)	0.95 (0.40, 2.23)	1.19 (0.53, 2.66)	1.07 (0.37, 3.06)	1.45 (0.54, 3.86)
SURG‡	0.84 (0.40, 1.77)	1.37 (0.60, 3.12)	1.50 (0.71, 3.20)	1.06 (0.38, 2.95)	1.53 (0.58, 3.99)
AMOX+MET+TET	0.83 (0.39, 1.77)	1.28 (0.55, 2.95)	1.34 (0.63, 2.86)	1.40 (0.49, 4.03)	0.96 (0.37, 2.46)
AMOX+MET	1.15 (0.55, 2.42)	0.90 (0.39, 2.06)	0.90 (0.39, 2.07)	0.91 (0.31, 2.64)	1.09 (0.39, 3.03)
TET	0.94 (0.42, 2.09)	1.05 (0.45, 2.46)	0.91 (0.40, 2.05)	0.86 (0.29, 2.53)	1.31 (0.46, 3.70)
SURG+AMOX+MET+TET	1.35 (0.62, 2.92)	0.93 (0.41, 2.09)	1.00 (0.48, 2.11)	0.86 (0.30, 2.45)	1.18 (0.45, 3.11)

There were no significant changes in counts of Actinomyces that were observed in all treatment groups compared to counts in the reference group, SRP.

^†Incidence rate ratio=exp (β).

^‡SURG was performed 3 months after the baseline visit. Results shown at the 3-month study period were for SRP.

SURG=surgery; AMOX=amoxicillin; MET=metronidazole; TET=tetracycline; IRR=incidence rate ratios.