Table 3.

Estimates of the Risk Ratio Comparing Recurrence of Ocular Herpes Simplex Virus Between Participants Randomized to Acyclovir or Placebo From Various Models (n = 308), Multicenter Herpetic Eye Disease Study, 1992–1998

Model	No. of Outcomes	No. at Risk	Adjusted RRa	95% CI	SE for ln(RR)
Complete data, physician- diagnosed recurrence
Acyclovir group	26	155	0.68	0.44, 1.07	0.23
Placebo group	38	153	1
Total	64	308
Naive analysis
Acyclovir group	22	155	0.93	0.55, 1.59	0.27
Placebo group	23	153	1
Total	45	308
Validation subgroupb
Acyclovir group	8	46	0.61	0.27, 1.35	0.41
Placebo group	12	45	1
Total	20	91
Direct maximum likelihood (nondifferential)			0.68	0.34, 1.38	0.36
Direct maximum likelihood (differential)			0.65	0.31, 1.40	0.39
Multiple imputation (nondifferential)			0.69	0.35, 1.36	0.35
Multiple imputation (differential)			0.69	0.34, 1.41	0.36

Abbreviations: CI, confidence interval; ln(RR), natural log of the risk ratio; RR, risk ratio; SE, standard error.

^a All models were adjusted for race, sex, age, and number of previous recurrences.

^b Validation subgroup included 91 participants.