Repeat HIV testers are more likely than first-time testers to report high-risk behaviour including unprotected sexual intercourse with an HIV-positive partner, multiple sexual partners and intravenous drug use [1–3]. As a consequence, HIV test counselling among repeat testers has tended to focus on strategies for risk reduction. However, repeat testers do not only report high-risk behaviour as their reason for seeking an HIV test. Other factors include testing as part of a regular health check, entering a new sexual relationship , talking to other people about testing  and believing that social norms favour HIV prevention [1,5]. Some people who seek a repeat HIV test may, therefore, do so as part of a personal risk-reduction strategy. The characteristics and behaviours of repeat testers have important implications for HIV test counselling and its role in prevention. Clearly HIV test counselling among repeat testers needs to be tailored according to their needs.
The aim of this study was to examine the characteristics of repeat HIV testers and consider their implications for risk appraisal, HIV test counselling and prevention.
The study was conducted in a same-day HIV testing clinic at the Royal Free Hampstead NHS Trust Hospital, London between September 1997 and July 1998. All clinic attenders (n = 2100) were invited to complete an anonymous self-administered questionnaire prior to their pre-test counselling and HIV test. Drawing upon previous studies [1,4], the questionnaire sought information on five groups of variables: (i) sociodemographics: age, sex, sexual orientation, residence, employment, status, ethnic group and place of birth; (ii) sexual risk behaviour in the previous 3 months: type, number and HIV status of partner(s) for unprotected penetrative sex (UPS); (iii) reasons for seeking the present HIV test: a list of 24 possible reasons allowed respondents to indicate all contributory factors and a main trigger; (iv) psychosocial variables: attitudes, emotional state and normative influences on testing, measured on Likert scales; and (v) HIV test history: number of previous HIV tests, date and place of last test (Royal Free/elsewhere).
HIV test results were recorded for all clinic attenders and were linked to the questionnaire by a number code that preserved confidentiality and anonymity.
Unprotected penetrative sex
A hierarchy of risk was created to distinguish those people reporting `higher-risk' UPS [i.e. with partner(s) who were HIV positive or whose status was unknown], from those reporting `lower-risk' UPS [i.e. with partner(s) who were HIV negative]. For those people who reported UPS with more than one partner, the highest level of risk was chosen. Only if all partners were known to be HIV negative were participants assigned to the lower-risk group. This hierarchy classified people according to the risk they were exposed to, and not the risk they presented to their sexual partner(s). Those reporting UPS with an HIV-positive partner were clearly engaging in a higher known risk than people whose partner's HIV status was unknown. However, for the purposes of HIV prevention, these two categories were combined to distinguish them from people engaging in lower-risk UPS. The small number of participants who did not report their sexual partner's HIV status (seven gay men, 16 heterosexual men and 18 heterosexual women) could not be included in the hierarchy of risk analysis.
Repeat testers were those people who had previously tested negative for HIV and were seeking another test. Participants who had only had one previous test and indicated that being `in the window period at my previous test' was a reason for their present HIV test were classified as first-time testers (n = 27). Repeat testers who did not provide information on the number of previous HIV tests (n = 65; 4.5%) were excluded from the analysis of the number of previous tests.
The sample was stratified by gender and self-reported sexual orientation . Non-parametric chi square, Fisher's exact and Mann–Whitney tests were used to examine differences between first-time and repeat HIV testers within each sexual orientation group.
Response rate and exclusions
Of 2100 questionnaires distributed in the clinic between September 1997 and June 1998, 1580 (75.2%) were completed and returned. Among the 1580 participants, 470 were gay men (29.7%; median age 30 years), 494 were heterosexual men (31.3%; median age 29 years) and 482 were heterosexual women (30.5%; median age 27 years); all were included in this analysis (n = 1446). Data from other sexual orientation groups were excluded because of small numbers (61 bisexual men, 36 bisexual women, 16 lesbians and 21 `other/not specified').
Previous testing history
Of the 1446 clinic attenders, 50.6% (731) were repeat HIV testers: gay men 71.7% (337/470), heterosexual men 42.1% (208/494) and heterosexual women 38.6% (186/482). The proportion of clinic attenders reporting one or two previous tests was similar across the three sexual orientation groups (29–36%); however, a greater percentage of gay men reported three or more previous tests (31.5%) than did heterosexual men (7.3%) and women (4.6%) (Table 1).
Of the 1446 clinic attenders, 2.4% (35) were HIV positive: gay men 6.6%, heterosexual men and women 0.4% (Table 1). HIV prevalence was higher among gay male first-time testers (9.8%) than among repeat testers (5.3%), although this difference was not significant (P = 0.1). There were no significant differences in HIV prevalence between heterosexual male or female repeat and first-time HIV testers (P = 1.0) nor between respondents to the study (2.4%, 35/1446) and non-respondents (3.2%, 16/504) (p = 0.3).
Within each sexual orientation group, there were no significant differences between repeat and first-time testers for the following variables: median age, residence, employment status, ethnic background, place of birth or number of children. (Full data are available from the authors on request.)
Unprotected penetrative sex
Among gay men, repeat testers were more likely to report UPS in the previous 3 months than first-time testers, but this differential was not statistically significant (P = 0.06). Nor were any significant differences found in the frequency of UPS between repeat and first-time testing heterosexual men (P = 0.3) or women (P = 0.7). Furthermore, there were no significant differences between repeat and first-time testers in the frequency of higher-risk UPS for gay men (P = 0.1), heterosexual men (P = 0.1) or heterosexual women (P = 1.0) (Table 2).
Other factors associated with repeat testing
In all sexual orientation groups, repeat testers were more likely to have had a sexually transmitted disease (STD), to know someone with HIV/AIDS and to know someone who had tested for HIV. However, repeat testers were no more likely than first-time testers to be seeking an HIV test in preparation for a new relationship. Among gay men, but not heterosexual men and women, repeat testers were significantly more likely to report having their current HIV test as part of a regular health check (Table 3).
There were no significant differences between repeat and first-time testers on emotive state variables with one exception. Heterosexual male repeat testers were more likely to report feeling sad to some degree than first-time testers (70% versus 56%;P = 0.001).
Number of previous HIV tests
Gay men reporting three or more previous HIV tests were more likely to report UPS than those reporting one–two or no previous tests (49.7, 35.3 and 32.3%, respectively;P = 0.01). No such difference was found among heterosexual men (63.9, 56.8 and 62.7%, respectively;P = 0.5) or women (59.1, 54.1 and 58.0%, respectively;P = 0.7). In particular, gay men who reported three or more previous tests were significantly more likely to report higher-risk UPS than those reporting one–two or no previous tests (42.2, 25.3 and 25.4%, respectively;P = 0.002). No such association was seen for heterosexual men (P = 0.2) or women (P = 1.0). (Full data available from authors on request.)
Gay men reporting three or more previous HIV tests were also more likely to have had an STD (60.8%) and to be seeking the current HIV test as part of regular health check (53.1%) than gay men with one-two or no previous tests (STD: 39.1 and 34.1% for one–two and no previous tests, respectively;P < 0.001; health check: 45.3 and 32.3%, respectively;P = 0.002). These associations were not seen among heterosexual men or women. For all sexual orientation groups, there was no association between the number of previous HIV tests and seeking the current test in preparation for new relationship (full data available from authors on request).
Nearly three quarters of gay men and 40% of heterosexual men and women seeking an HIV test at this London clinic in 1997–1998 were repeat testers, i.e. people who had previously tested negative for HIV who were seeking a further test. The vast majority of repeat testers once again tested HIV negative. Regardless of sexual orientation, the frequency of unprotected sexual intercourse was not significantly different between repeat and first-time testers. This finding is consistent with a normative shift in testing behaviour [6–9] whereby repeat HIV testing has been integrated into a personal sexual health strategy by many people rather than being a consequence of high-risk behaviour alone.
There was, however, an important exception to this pattern, which emerged when the number of previous HIV tests was examined. Gay men who had had three or more previous tests were more likely to report higher-risk UPS than other gay men, a finding that is consistent with other research [1,2,4]. This recurrent finding highlights a subgroup of repeat testers who are at high-risk of exposure to HIV and has important implications for counselling both before and after HIV testing. When talking to gay male repeat testers, counsellors in HIV and genitourinary medicine clinics should pay attention to the number of previous HIV tests. This will help to identify gay male repeat testers who fall into the group with elevated risk. HIV test counselling among this group needs to be tailored accordingly . A negative test result may serve, inadvertently, to reassure those men reporting recent high-risk behaviour. Post-test counselling provides an opportunity to challenge their risk taking and to discuss personal strategies for risk reduction.
Seeking a test in preparation for a new relationship was reported by over half the heterosexual men and women and one third of gay men. This suggests that a growing number of heterosexual men and women as well as gay men have adopted negotiated safety [11–13] as a risk-reduction strategy within a regular relationship, i.e. establishing that their own HIV status is concordant with their partner's. Repeat testers were just as likely to cite this as a reason for seeking the current test as first-time testers. This finding has important implications as it identifies a second group of repeat testers who would benefit from tailored counselling: those who use the test as an HIV prevention strategy when starting a new relationship. In these instances, counsellors should encourage the maintenance of positive health behaviours. Indeed, there may be a case for family planning and gynaecology services to discuss HIV testing with people who wish to stop using condoms or who are embarking on a new relationship. Since the sexually active population changes partners on a regular basis, emphasizing the benefits of establishing HIV seroconcordance will reduce the likelihood of HIV transmission within a regular relationship, an area that is often neglected .
Concern has been expressed that testing for HIV alone may result in coexistent STD being missed. In fact, previous research in this free-standing HIV testing clinic has shown the prevalence of STD to be low among clinic attenders . Nonetheless, in that study, six (4%) women, none of whom had previously been screened for an STD, were diagnosed with chlamydial infection. Selective STD screening should, therefore, be offered to people seeking an HIV test who report never having been screened before.
In a 1995–1996 survey in this clinic, there was a threefold difference in HIV prevalence between gay men testing for the first time (11.2%) and those who were repeat testers (3.9%) (P < 0.01) . In the present survey, although HIV prevalence among gay male first-time testers (9.8%) was still higher than among repeat testers (5.3%), the differential was reduced and not significant (P = 0.1).
Our findings have important implications for counselling and the role of HIV testing in prevention. It is vital to consider the heterogeneous nature of the repeat testing population and the many factors that influence their decision to seek a test. Because of the diversity of this group, it cannot be assumed that all repeat testers are at high risk of contracting HIV, although some may be. The use of a `problem identification approach' in HIV test counselling may be appropriate as it is geared towards each individual's risks and their reasons for seeking an HIV test [10,15].
Previous research has highlighted the high costs of repeat testing and the lack of evidence for a reduction in risk behaviour following repeat negative tests . Our data suggest that for many people repeat testing has become part of a personal risk-reduction strategy. Nonetheless, a subgroup of gay men with a history of regular repeat testing was identified who may require additional input at the time of their HIV test to address their risk behaviour. By being responsive to the needs of repeat testers, HIV testing and counselling can play an important role in HIV prevention.
We would like to thank Amanda Jones (HIV Prevention Coordinator), clinic staff, HIV counsellors and those who completed a questionnaire for their support.
1. Phillips KA, Paul J, Kegeles S, Stall R, Hoff C, Coates TJ. Predictors of repeat HIV testing among gay and bisexual men.
AIDS 1995, 9: 769 –775.
2. McFarland W, Fishcer-Ponce L, Katz MH. Repeat negative human immunodeficiency virus (HIV) testing in San Francisco: magni- #Rtude and characteristics.
Am J Epidemiol 1995, 142: 719 –723.
3. McCusker J, Willis G, McDonald M, Sereti SM, Lewis BF, Sullivan JL. Community-wide HIV counselling and testing in central Massachusetts: who is retested and does their behaviour change?
J Community Health 1996, 21: 11 –22.
4. Norton J, Elford J, Sherr L, Miller R, Johnson M. Repeat HIV testers at a London same-day testing clinic.
AIDS 1997, 11: 773 –781.
5. Kalichman SC, Schaper PE, Belcher L. et al
. It's like regular part of gay life: repeat HIV antibody testing among gay and bisexual men.
AIDS Educ Prev 1997, 9 (Suppl. B): 41 –51.
6. Ajzen, Ajzen, I. The theory of planned behaviour.
Organisational Behav Hum Decision Processes 1991, 50: 179 –211.
7. Oskamp S, Thompson SC (eds). Understanding and Preventing HIV Risk Behaviour
. London: Sage; 1996.
8. Kelly, Kelly, JA. AIDS prevention: strategies that work.
AIDS Reader 1992, 2: 135 –141.
9. Wallston BS, Alagna SW, deVellis BMcE, deVellis RF. Social support and physical health.
Health Psychol 1983, 2: 367 –391.
10. Norton J, Miller R, Johnson MA. Promoting HIV prevention: a problem identification approach to interventions in post-HIV test counselling.
AIDS Care 1997, 9: 345 –353.
11. Kippax S, Crawford J, Davis M, Rodden P, Dowsett G. Sustaining safe sex: a longitudinal study of a sample of homosexual men.
AIDS 1993, 7: 257 –263.
12. Kippax S, Noble J, Prestage G. et al
. Sexual negotiation in the AIDS era: negotiated safety revisited.
AIDS 1997, 11: 191 –197.
13. Elford J, Bolding G, Maguire M, Sherr L. Sexual risk behaviour among gay men in a relationship.
AIDS 1999, 13: 1407 –1411.
14. Madge S, Elford J, Lipman MC, Mintz J, Johnson MA. Screening for sexually transmitted diseases in an HIV testing clinic: uptake and prevalence.
Genitourin Med 1996, 72: 347 –351.
15. Bor R, Miller R, Goldman E. Theory and Practice of HIV Counselling: a Systemic Approach
. London: Cassell; 1993.