Using population attributable risk to choose HIV prevention strategies in men who have sex with men

BMC Public Health

Table 2 PAR of HIV seroconversion risk factors including UAI by HIV status of sexual partners

Risk factor	PAR (95% CI) Crude model	PAR (95% CI) Adjusted model	Average annual number of HIV cases	Average lifetime HIV costs (range)⁽¹⁾AUD$million, 2010 dollars
				1% discounting	3% discounting	5% discounting
All risk factors	-	0.91 (0.81,0.97)	637	434 (403-471)	274 (249-304)	180 (160-205)
UAI by HIV status of sexual partners ⁽²⁾	0.85 (0.75,0.94)	0.77 (0.66,0.89)	539	367 (341-399)	232 (211-257)	152 (136-174)
UAI with HIV-negative only (serosorting)	0.15 (0.11,0.22)	0.10 (0.07,0.19)	70	48 (44-52)	30 (27-33)	20 (18-23)
UAI with some unknown HIV status	0.34 (0.27,0.43)	0.33 (0.26,0.42)	231	157 (146-171)	99 (90-110)	65 (58-74)
UAI with any HIV-positive	0.36 (0.26,0.45)	0.34 (0.24,0.44)	238	162 (151-176)	102 (93-114)	67 (60-77)
Circumcision status ⁽³⁾
Uncircumcision	0.08 (0.04,0.16)	0.07 (0.03,0.14)	49	33 (31-36)	21 (19-23)	14 (12-16)
Number of sexual partners
10+ casual sexual partners in the last 6 m	0.26 (0.18,0.36)	0.19 (0.09,0.34)	133	91 (84-98)	57 (52-64)	38 (33-43)
STIs ⁽⁴⁾
Anal warts between study visits	0.14 (0.10,0.18)	0.13 (0.09,0.19)	91	62 (58-67)	39 (36-43)	26 (23-29)
Anal gonorrhoea at study visit	0.06 (0.04,0.07)	0.02 (0.01,0.03)	14	10 (9-10)	6 (5-7)	4 (4-5)

CI = Confidence interval, PAR = population attributable risk, UAI = Unprotected anal intercourse, m = months
1. The average healthcare costs incurred from HIV seroconversions associated with specific risk factors were calculated by multiplying the PAR percent for each specific risk factor, by the 700 HIV infections associated with male homosexual exposure, by the average healthcare costs per HIV-infected person. We performed sensitivity analyses by rate of discounting and according to the bounds in delays in durations of time between infection and diagnoses and commencement of antiretroviral treatment but calculations were based on the best estimated PAR of the risk factors and not the 95% CI of the PAR.
2. Adapted from Jin et al [8]
3. Adapted from Templeton et al [9]
4. Adapted from Jin et al [21]

ISSN: 1471-2458