Table 4.

Estimated coefficients of the fitted logistic regression model and the final model for potential risk factors for postherpetic neuralgia.

Model	Variable	Coefficient	Standard error	P-Value	OR	95% CI for OR
Saturated model	Age group:
	60–69 years vs 50–59 yearsa	0.336	0.6474	0.6041	1.399	0.393–4.976
	70–79 years vs 50–59 yearsa	0.777	0.6409	0.2253	2.175	0.619–7.640
	≥80 years vs 50–59 yearsa	0.953	0.6420	0.1378	2.593	0.737–9.125
	Gender Female vs Malea	−0.165	0.4055	0.6835	0.848	0.383–1.876
	HZ related complications:  Yes vs Noa	−0.210	0.5319	0.6926	0.810	0.286–2.298
	Current immunosuppressive therapy  Yes vs Noa	0.590	0.7250	0.4155	1.805	0.436–7.474
	Pre-existing medical condition  Yes vs Noa	0.412	0.4063	0.3104	1.510	0.681–3.348
	HZ severity at initial visit
	Moderate pain vs No or mild paina	0.747	1.0937	0.4947	2.110	0.247–18.000
	Severe pain vs No or mild paina	2.175	1.0512	0.0386	8.801	1.121–69.083
	Timing of treatment with antiviral agentb ≥3 days vs <3 daysa	0.300	0.3934	0.4454	1.350	0.624–2.919
Final model	HZ severity at initial visit
	Moderate pain vs No or mild paina	0.608	1.0849	0.5752	1.837	0.219–15.400
	Severe pain vs No or mild paina	1.986	1.0368	0.0555	7.283	0.955–55.567

HZ = herpes zoster; OR = odds ratio; 95% CI for OR = Wald 95% confidence interval; Note: Final model is by considering backward model building strategy with SLSTAY = 0.05 and saturated model is by not considering any model building strategy.

^aReference category.

^bDifference between the date of rash onset and the date of first prescription of treatment of antivirals.