. 2018 Mar 12;67(3):420–429. doi: 10.1093/cid/ciy051

Table 2.

Suicidal and Self-Injurious Behavior by Randomization Arm and Prespecified Antiretroviral Therapy Groups

	N	Immediate ART		Deferred ART		HR^a	95% Confidence Interval	P Value	HR Ratio^b	P-Value for Interaction^c
	N	Number of Participants With Events	Rate per 100 PY	Number of Participants With Events	Rate per 100 PY	HR^a	95% Confidence Interval	P Value	HR Ratio^b	P-Value for Interaction^c
ITT analysis, mean follow-up 3 years
EFV prespecified	3515	19	0.35	12	0.22	1.38	(0.7, 2.9)	.39	1.86	0.24
Other ART prespecified	1169	9	0.50	13	0.69	0.74	(0.3, 1.7)	.49
ITT analysis, follow-up truncated at 1 year after randomization
EFV prespecified	3515	9	0.52	2	0.11	3.74	(0.8, 17.5)	.09	3.67	0.15
Other ART prespecified	1169	7	1.25	7	1.19	1.02	(0.4, 2.9)	.96
Censoring deferred arm participants at ART initiation
EFV prespecified^d	3394	18	0.36	4	0.10	3.31	(1.1, 9.9)	.03	3.18	0.07
Other ART prespecified^e	1137	9	0.56	8	0.66	1.04	(0.4, 2.7)	.93

Cox proportional hazards models were used for this analysis.

Abbreviations: ART, antiretroviral therapy; EFV, efavirenz; HR, hazard ratio; ITT, intention to treat; PY, person-years.

^aHR (immediate/deferred), estimated in Cox proportional hazards models, stratified by history of psychiatric diagnosis.

^bRatio of HRs (immediate/deferred) within the EFV prespecified subgroup over the other ART prespecified subgroup.

^cP value for the interaction between indicators for treatment group and prespecified ART regimen; compares HRs (immediate/deferred) between subgroups by prespecified ART.

^dOf these events, 6 and 0, in the immediate vs deferred arms, respectively, occurred among 100 participants with prior psychiatric diagnoses. The immediate group excludes participants who did not start ART. Follow-up time starts at EFV start date. Of the 3515 participants with EFV in the prespecified regimen, 117 in the immediate arm were excluded (32 never started ART, 85 never used EFV), and 4 in the deferred arm were excluded (3 started ART at randomization, 1 participant was lost to follow-up at randomization).

^eOf these events, 5 and 2, in the immediate vs deferred arms, respectively, occurred among 161 participants with prior psychiatric diagnoses. Of the 1169 participants without EFV in the prespecified regimen, 32 were excluded (in the immediate group, 7 never started ART, and for 25, the first ART regimen contained EFV). The immediate group excludes participants who did not start ART. Follow-up time starts at ART start date and was censored at EFV start.