The Impact of Voluntary Counselling and Testing: A global review of the benefits and challenges (UNAIDS, 2000, 96 p.): 2. Prevention of HIV transmission in special populations following VCT: 2.1 Men who have sex with men (MSM)

The Impact of Voluntary Counselling and Testing: A global review of the benefits and challenges (UNAIDS, 2000, 96 p.)

2. Prevention of HIV transmission in special populations following VCT

		2.1 Men who have sex with men (MSM)
		2.2 Sex workers
		2.3 STI clinic attenders
		2.4 Injecting drug users (IDUs)
		2.5 Prison populations
		2.6 Truck drivers
		2.7 Blood donors
		2.8 Young people

2.1 Men who have sex with men (MSM)

The group that has been most widely studied in relation to sexual behaviour following VCT is MSM. These studies have been carried out largely in industrialized countries (North America, Australia and Europe). Reviews of VCT interventions among MSM are available (Higgins et al., 1991⁵⁵, Wolitski et al., 1997⁵⁶, Weinhardt et al., 1999⁵⁷). Although risky sexual behaviour among MSM has declined, it was seen to be declining prior to VCT being available. Since VCT has been available, some of the studies show that men have reduced risky behaviour but this is independent of testing. In those studies where VCT is linked to increased behaviour change, sexual behaviour change is more marked in people who are seropositive. Long-term studies have shown that maintaining safer sex behaviour is difficult and many MSM are reported to resume risk behaviours with time (Kelly, 1998⁵⁸, Imrie, 1999⁵⁹).

Box 8: Studies among men who have sex with men

Summary: Background

· There has been a great change in the sexual behaviour of MSM since the late 1980s. Much of this behaviour change was probably a result of the very high morbidity and mortality associated with HIV in the gay community at the start of the epidemic in Europe and the United States. Many gay men watched friends die from HIV and it has been proposed that this direct experience led to behaviour change at an early stage. These changes were often noted before HIV testing became widely available.

· There is however, recent evidence that in some countries these changes in safer sexual behaviour among gay men are being eroded because of complacency, as life-prolonging ARV therapy is now widely available. Having an ”undetectable viral load” is misconstrued as having a very low risk of HIV transmission and therefore not needing to practise safer sex.

Summary: Efficacy/sexual behaviour change

· Several studies show significant decreases in risk behaviour associated with VCT, with seropositive men showing greater reduction in risk behaviour than seronegative men or men who are unaware of their HIV status. This may suggest that learning about being seropositive had a greater impact on changing behaviour and that seronegative men are still putting themselves at risk from infection.

· Although there are several studies that indicate that VCT is important in promoting safer sexual behaviour, many studies have inconclusive findings and there is no consistent evidence to show that VCT is effective in promoting safer sexual behaviour among MSM.

Table 4: VCT outcome studies reporting sexual behaviour from studies of men who have sex with men

Author (year)	Site	N	Results
Coates (1987) *	San Francisco, USA	502	All groups ¯ unprotected AI and # sex partners T/UN=NS.
Fox (1987) ·	Baltimore, USA	1001	T+ve ¯ unprotected AI cf. T-ve and UN. Most reduced risk behaviour prior to VCT
Cohn (1998) ·	Denver, USA	269	T+ve condom use cf T-ve. Most reduced risk behaviour prior to VCT.
McCusker (1988) ·	Boston USA	270	T+ve less likely to have unprotected AI cf. UN or T-ve.
Schechter (1988) +	Vancouver, Canada	361	T+ve and T-ve ¯ # of sexual partners following VCT.
Frazer (1988) ·	Australia	318	T+ve significantly more likely to use condoms for AI cf. T-ve or UN. T+ve ¯ unprotected AI cf. T-ve and UN.
Ross (1988) ·	Australia	172	VCT groups safer sex significantly cf. Counselled only, tested only and no intervention.
Valdiserri (1988) ·	Pittsburgh, USA	955	T+ve ¯ unprotected AI cf. T-ve and UN.
Van Griensven (1989) *	Amsterdam, the Netherlands	307	T+ve significantly more likely to use condoms for AI cf. T-ve or UN.
Ostrow (1989) +	Chicago, USA	474	All groups ¯ sexual risk behaviour - N/S differences between T+ve, T-ve & UN.
McKusick (1990) ·	San Francisco, USA	508	T+ve less likely to report unprotected AI than T-ve for both monogamous and non-monogamous men.
Doll (1990) +	San Francisco, USA	309	All groups ¯ sexual risk behaviour - N/S differences between T+ve, T-ve & UN.
Zapka (1991) +	Boston, USA	249	All groups ¯ sexual risk behaviour -N/S differences between T+ve, T-ve & UN.
Huggins (1991) +	Pittsburgh, USA	155	All groups ¯ sexual risk behaviour -N/S differences between T+ve, T-ve & UN.
Dawson (1991) +	4 UK cities	502	N/S differences between T=ve, T-ve and UT.
Roffman (1995) +	16 USAcities	1395	T more sexually active, but condom use.

N/S = no significant differences
¯ = decreased
= increased
T = tested and aware (of HIV status)
UN = unaware of HIV status
AI = anal intercourse
+ = N/s / inconclusive studies
· = declines in risky behaviour associated with VCT
* = comparisons between groups before and after testing available