Risk behaviours in HIV-positive men who have sex with men participating in an intervention in a primary care setting

Abstract

Men who have sex with men receiving HIV care reported their sexual behaviours and their intentions, classified according to the Transtheoretical Model of Change, to modify the following behaviours: (1) condom use by partner type and activity type; (2) reduction of partner number; and (3) disclosure of HIV serostatus to partners. Most participants were white (68.8%) or black (29.5%) and were more likely to report unprotected sex with HIV-positive than with serodiscordant partners for most activities. Whites reported more partners than black patients (mean 4.1 versus 2; P < 0.0001) and black participants reported fewer HIV-negative (P = 0.0084) and -unknown status partners (P = 0.00095) than whites. Cocaine/crack use was associated with more sexual partners (P = 0.001) and more frequent unprotected sex with HIV-negative or -unknown status partners (P = 0.036). Readiness to change risk behaviour varied by partner status and type of sexual activity. Understanding patients' risks and their readiness to change behaviours may help providers to promote sexual health.

Keywords

sexual risk behaviour men who have sex with men HIV transtheoretical model of change substance use

INTRODUCTION

Advances in medical management have resulted in HIV-infected individuals living longer, higher quality lives. Unfortunately, the Centers for Disease Control and Prevention (CDC) recently released a new estimate of the annual number of new HIV infections in the United States, which revealed that the HIV epidemic is worse than previously recognized, indicating that approximately 56,300 individuals in the United States were newly infected with HIV in 2006. Further analysis of CDC's surveillance data showed that men who have sex with men (MSM) represented 53% (n = 28,720) of new infections in 2006.^1,2

A sizeable proportion of these new cases of HIV in the USA are attributable to transmission from HIV-positive individuals who are aware of their status, yet do not regularly practise safer sex measures.^3,4 As the population of HIV-infected individuals continues to increase globally, interventions are needed to promote adoption and maintenance of safer sex practices by HIV-infected individuals. Integrating brief interventions into the primary care setting has great intuitive appeal and would provide recurring opportunities to assess and reinforce safer sex behaviours and address other behaviours that may negatively influence one's health (e.g. substance use). Provider-delivered interventions also capitalize on the unique doctor–patient relationship, an effective motivator for enhancing behaviour change including sexual risk behaviours.^5–16 However, successful interventions to reduce HIV-risk behaviours in the context of continuing care require a better understanding of the complexity and breadth of behaviours practised by the patient population. In this report, we describe selected characteristics of HIV-infected MSM receiving care at an HIV primary care facility in the southeastern United States.

The assessment used in the present study included evaluation of HIV transmission risk behaviours and of each participant's level of motivational readiness to modify risk-taking behaviours. We used a computerized risk assessment and preprogrammed algorithms based on the Transtheoretical Model of Change.^17,18 According to this model, the process of changing one's behaviour (adoption or cessation) occurs through a continuum of five stages: (1) having no intention to change (precontemplation); (2) seriously considering making a change sometime in the future, usually defined as ‘within the next six months’ (contemplation); (3) planning to change soon, within the next 30 days, and has taken some steps towards taking action, for example, using condoms but inconsistently (preparation); (4) taking action, i.e. 100% condom use (action); and (5) sustaining the behaviour change over time (maintenance).

METHODS

This study was conducted at one of 15 Health Resources and Services Administration-sponsored Special Projects of National Significance initiative sites. From July 2004 until June 2006, men receiving primary care at a university-based HIV clinic were approached for study enrolment prior to their scheduled appointment. Inclusion criteria were: (1) male gender; (2) age over 16 years; (3) recipient of primary care in the clinic; (4) acknowledgement of any type of sexual contact with another male during the six months prior to enrolment; and (5) willingness to follow-up for up to 21 months. In a private location, the study purpose and procedures were explained to each potential participant and brief screening questions were administered. Once a participant was determined to be eligible for participation and a signed informed consent was obtained, the individual would interact with an audio, computer-assisted self-interview (ACASI) survey.

The ACASI questions included demographic characteristics, medical history, HIV transmission risk behaviours and potential risk behaviour modifiers such as substance use (including alcohol) and depression. Participants were asked about their short- and long-term intentions to change one or more risk behaviours of interest in this study: (1) condom use by partner type and type of sexual activity; (2) reduction of numbers of concurrent sexual partners; and (3) disclosure of HIV serostatus to all sexual partners based on the stages of change. Following completion of the ACASI survey, participants were seen by their primary care physician.

From July 2004 to February 2005, only the assessment was administered. Beginning in March 2005, a theory-based provider advice sheet based on each patient's own responses to the assessment was computer-generated and printed for use by the provider during the clinical encounter. In addition, a separate substance abuse and depression ‘flag’ sheet was generated for the provider, along with a three-point behavioural prescription for the patient to take home.¹⁹ The provider advice sheet and the behaviour prescription focused on one targeted behaviour. The ‘target’ behaviour was the one behaviour that the patient was either most ready to change, or in instances when a patient was at the same stage of readiness to change for multiple behaviours, the one behaviour that the computer selected for targeting based on a programmed algorithm.¹⁹ Recruitment of new study participants ceased on 29 June 2006. The data described are from the first intervention visit. The study's protocol and procedures were approved by the University of Alabama at Birmingham Institutional Review Board.

DATA ANALYSIS

All data were directly entered by each participant into the ACASI, uploaded weekly to the national coordinating site and made available on the study website. Descriptive statistics for categorical data were expressed as n (%). For continuous data, the descriptive statistics were expressed as median (range). In Table 2 (disclosure and sexual behaviour), proportions among the serostatus categories were compared using repeated measures for a binary variable, i.e. general estimating equations (GEE) method, based on a compound symmetry covariance structure. The GEE addresses the issue that a participant may have more than one partner and, thus, may have multiple answers. In comparing the two groups for the number of unprotected sexual acts and male partners as presented in Table 2, the negative binomial distribution was assumed in a general linear model to account for over-dispersion in the count data due to a large number of zeros and some extremely large observations. Fisher's exact test was used to compare substance use behaviour between white and black subjects. All P values <0.05 were considered significant. The estimated number of new HIV infections transmitted by the study population over the six months prior to the study was calculated by applying reported risk behaviour to a previously published equation.²⁰

RESULTS

Study population

A total of 987 male patients were approached in the clinic for study participation; 522 were ineligible due to lack of sexual activity in the previous six months (387/522; 74.1%), or because their only recent sex partners were female (114/522; 21.8%). Of the remaining 465 men, 231 (49%) refused study participation, primarily due to lack of time (79/231; 34.2%) and lack of interest (75/231; 32.5%). This resulted in a final cohort of 234 male patients enrolled for whom baseline data were available.

Most participants were white (161/234; 68.8%) or black (69/234; 29.5%) with a median age of 40 years (Table 1). Most participants had more than high school education (72.2%) and over one-third (40.2%) were in a domestic partnership/marriage or a committed relationship. Most respondents self-identified as gay or bisexual (230/234; 98.3%). Participants had lived with their HIV diagnosis for a median of 8.5 years (range = 0.1–20.8) at the time of enrolment. Over 85% of participants were on antiretroviral therapy at the initial visit, and over three-quarters reported that they were in ‘good’ to ‘excellent’ health.

Table 1

Study population (n = 234)

Characteristic	No. (%)
Demographic race/ethnicity
Black/African American	69 (29.5%)
Hispanic/Latino	1 (0.4%)
White	161 (68.8%)
American Indian/Alaskan	1 (0.4%)
Other	2 (0.9%)
Age
Median, range	40 (18–60)
Education
≤High school	65 (27.8%)
Some college/college degree	141 (60.2%)
≥some graduate school	28 (12.0%)
Income
≤$20 k	134 (57.3%)
20–40 k	55 (23.5%)
40–80 k	34 (14.5%)
>80 k	11 (4.7%)
Health insurance*
Medicaid/medicare	85 (36.3%)
VA	1 (0.4%)
Private/HMO	104 (44.4%)
None	61 (26.1%)
Relationship status
Single, never married	117 (50%)
Committed relationship	19 (8.1%)
Domestic partnership/marriage	75 (32.1%)
Separated/divorced/widowed	23 (9.8%)
Sexual identity
Heterosexual	2 (0.9%)
Gay/homosexual	204 (87.2%)
Bisexual	26 (11.1%)
Don't know/other	2 (0.9%)
HIV-related
Mode of infection*
Sex with man	232 (99.2%)
Sex with woman	5 (2.1%)
IVDA	7 (3.0%)
Blood tx/occupational	14 (6.0%)
Years HIV positive
(Median, range)	8.5 (0.1–20.8)
Years receiving care at clinic
(Median, range)	4.0 (0–20.5)
Currently on ART? ^†
Yes	171 (85.5%)
Recent VL detectable?
Yes	100 (37.5%)
No	84 (44.6%)
Most recent CD4
(Median, range)	400 (0–1473)
Rates health as …
Poor	3 (1.3%)
Fair	53 (22.7%)
Good to excellent	178 (76%)
Sex with females only
Yes	0 (0%)
No	234 (100%)
Sex with males only
Yes	223 (95.3%)
No	11 (4.7%)
Sex with both
Yes	4 (1.7%)
No	230 (98.3%)

*More than one response possible

^† n = 200 responders

ART = antiretroviral therapy; IVDA = intravenous drug abuse; VL = viral load

Sexual risk behaviour

In the six months preceding enrolment, over 95% of participants reported oral and/or anal sex with males only (Table 1). The remainder of individuals engaged in oral, anal and/or vaginal sex with both males and females (1.7%) or mutual masturbation only. During the six months prior to study enrolment, participants reported a median of one (1–3) female and two (1–100) male sexual partners with whom they had engaged in oral, anal and/or vaginal sex, with over half (55.1%) of the participants reporting multiple partners (data not shown). Among the 227 individuals reporting sexual partners, 53.3% (n = 121) had one or more HIV-positive partners, 47.6% (n = 108) reported one or more HIV-negative partners and about one-third (31.7%; n = 72) reported at least one partner whose HIV status was unknown (data not shown). Given the small number of female partners (n = 26), the analyses focused on male partners.

Sexual activities and disclosure of HIV status varied by the serostatus of male sexual partners (Table 2). Participants were more likely to disclose their infection to some or all partners known to be HIV positive (96.7%) than to some or all of their HIV-negative male partners (89.8%), or some or all status unknown partners (52.8%). Similarly, sexual behaviours practised by the study population were diverse (Table 2). The differences in behaviour based on known or perceived partner HIV serostatus reached statistical significance for all activities, except for unprotected receptive anal sex. Based on the numbers of unprotected insertive or receptive anal sex acts with HIV-negative and HIV-unknown status partners, study participants transmitted HIV to about one uninfected (statistical estimate 0.7) individual during the six months prior to study enrolment.²⁰

Table 2

Disclosure status and specific unprotected sexual activity by HIV serostatus of male partner (n = 227)

	HIV+	HIV−	HIV?
Disclosure
None	4 (3.3%)	11 (10.2%)	34 (47.2)
Some/all	117 (96.7%)	97 (89.8%)*	38 (52.8%)^†‡
Specific sexual activities by partner serostatus
UIO	39/119 (32.8%)	11/108 (10.2%)^†	11/72 (15.3%)*
URO	55/123 (44.7%)	55/108 (50.9%)*	34/75 (45.3%)
URA	47/123 (38.2%)	25/108 (23.2%)	14/75 (18.7%)
UIA	46/123 (37.4%)	12/108 (11.1%)^†	12/75 (16.0%)

*Significantly different (P<0.05) than the proportion for HIV+

^†Significantly different (P<0.005) than the proportion for HIV+

^‡Significantly different (P<0.005) than the proportion for HIV−

UIO = unprotected insertive oral sex; URO = unprotected receptive oral sex; URA = unprotected receptive anal sex; UIA = unprotected insertive anal sex

Differences in sexual risk by race were noted when analyses compared white (n = 161) and black (n = 69) participants. The numbers of reported sexual partners varied significantly by racial group, with whites reporting significantly more partners than black patients (mean of 4.1 versus 2, respectively; P < 0.0001). While the numbers of HIV-infected partners did not vary by race, compared with whites, black participants reported significantly fewer HIV-negative (mean 0.7 versus 1.3; P = 0.0084) and -unknown status partners (mean 0.6 versus 1.5; P = 0.00095). Although white participants had greater numbers of HIV-negative partners, they were more likely to disclose their HIV status to all of their HIV-negative partners than black participants (83.1% versus 52.2%; P = 0.008). There were no significant differences by race for disclosure to HIV-positive and -unknown status partners.

Readiness to change specific risk behaviours

The following analyses are limited to condom use and partner reduction behaviours that were ‘targeted’ at the baseline intervention visit. Therefore, individuals who were not targeted for one of these behaviours at the baseline visit are not included in the analyses represented in Table 3. We examined condom use based on the type of sexual activity (receptive anal, insertive anal and insertive oral), partner type (primary or other) and HIV serostatus of a partner (positive, negative or unknown). The pattern found for consistent condom use was strikingly similar across partner type (not shown); thus, data were combined and are shown across type of sexual activity and partner's serostatus in Table 3. Those who reported not using condoms consistently were classified into one of the three preaction stages (precontemplation [PC], contemplation [C] and preparation [P]) for consistent condom use across type of sexual activity and partner HIV serostatus.

Table 3

Partner reduction, patterns of consistent condom use by sexual activity and partner's HIV serostatus and readiness for changing behaviours

Sexual activity Partner's HIV serostatus	Consistent use % (n)	Inconsistent use % (N)
Sexual activity Partner's HIV serostatus	Consistent use % (n)	PC	C	P
Receptive anal
HIV negative (n = 48)	63 (30)	28% (5)	50% (9)	22% (4)
HIV unknown (n = 22)	68 (15)	14% (1)	57% (4)	29% (2)
HIV positive (n = 57)	37 (21)	50% (18)	39% (14)	11% (4)
Insertive anal
HIV negative (n = 29)	79% (34)	33% (3)	67% (6)	0% (0)
HIV unknown (n = 10)	59% (10)	0% (0)	71% (5)	29% (2)
HIV positive (n = 13)	30% (17)	51% (20)	41% (16)	8% (3)
Insertive oral
HIV negative (n = 64)	27% (17)	64% (30)	30% (14)	6% (3)
HIV unknown (n = 36)	17% (6)	50% (15)	43% (13)	7% (2)
HIV positive (n = 82)	4% (3)	67% (53)	30% (24)	3% (2)
Male partner reduction (n = 92)	−	37% (34)	20% (18)	43% (40)

Table 3 shows that for receptive anal sex, close to two-thirds (63%) of participants reporting this behaviour with an HIV-positive partner were engaging in unprotected sex and of these, 50% reported no intention to change their condom use behaviour. Approximately one-third of respondents reported unprotected receptive anal sex with HIV-negative (37%) and HIV-unknown (32%) partners. Most (86%) of those who were not using condoms consistently with HIV-unknown partners were willing to consider consistent condom use with such partners in the future. For insertive sexual activity, a similar pattern was found. Of respondents having insertive anal sex with an HIV-positive partner, 70% were not using condoms regularly and of these, 51% stated they were not motivated to start using condoms anytime soon. With HIV-negative and HIV-unknown status partners, 21% and 41% of respondents, respectively, were not having protected insertive sex. Again, men who were not using condoms with HIV-unknown partners were more willing to consider condom use during insertive anal sex in the future. The majority of respondents did not use condoms for oral sex. While the majority did not plan to change this behaviour, 50% of men with an HIV-unknown serostatus partner were at least considering behaviour change in the future. Men with more than one partner were also assessed regarding their level of motivation to reduce the number of sexual partners. There was nearly a balance in the proportion of respondents who reported no intention to reduce their number of partners (37%) and those who reported being ready to do so (43%) (Table 3).

Substance use behaviour

The use of substances that might modify decision-making and risk-taking behaviour was also common among respondents. Over three-quarters (76.1%) reported alcohol use in the previous three months, with 44.9% reporting five or more drinks on one occasion (Table 4). Over one-third of participants (36.8%) reported using one or more other drugs over the previous three months, most commonly marijuana (77.9%), methamphetamine (18.6%), inhalants (16.1%) and cocaine (12.8%). Among respondents who reported any substance use, 56.3% of white men and 30% of black men (P = 0.07) reported using ‘party drugs’, defined as use of two or more of the following: cocaine, crack, methamphetamine, inhalants, hallucinogens and/or tranquilizers (data not shown). Intravenous drug use was rare (1.3% of the total population).

Table 4

Substance use patterns in the study population (n = 234)

Five or more drinks in one day ( n = 178)*
Once a week or more frequently	31 (17.4%)
< 1/month to 1–3 times/month	49 (27.5%)
Never	98 (55.1%)
Any substance use	86/232 (36.8%)
IV drug use	3/234 (1.3%)
Specific substances used (non-IV only; n = 86)
Marijuana	67/86 (77.9%)
Crack	10/86 (11.6%)
Cocaine	11/86 (12.8%)
Heroin	0
Methamphetamines	16/86 (18.6%)
Vitamin K (ketamine)	1/86 (1.2%)
Opiates	5/86 (5.8%)
Tranquilizers/barbiturates	11/86 (12.8%)
Hallucinogens	5/86 (5.8%)
Inhalants	14/86 (16.3%)
Viagra	13/86 (15.1%)

*Among the 178 reporting alcohol intake within the three months prior to enrolment

IV = intravenous

Predictors of sexual risk based on substance use

Among participants reporting cocaine or crack use, the number of male sexual partners reported was about 2.6 times higher than for non-users (P = 0.001). The total number of unprotected sexual acts was not significantly different (P = 0.15) between these two groups, but the crack and/or cocaine users engaged in more frequent unprotected anal sex with HIV-negative or -unknown status partners than non-users (P = 0.036). Similarly, the number of male partners reported by individuals using ‘party drugs’ was about 4.7 times greater than for non-users (P < 0.0001) and the total numbers of unprotected sexual acts reported by those who took ‘party drugs’ was about six times greater than for non-users (P = 0.008).

DISCUSSION

Substantial numbers of HIV-positive individuals receiving continuing care in this primary care clinic are sexually active, with considerable variation in partner number, types of sex acts and the likelihood of disclosing one's serostatus to sexual partners. For six months preceding enrolment, the median number of male sexual partners for oral or anal sex were two, and ranged from 0 (no oral, vaginal or anal sex and who had engaged in mutual masturbation only) to 100. Therefore, this study confirms findings from other studies demonstrating the complexity of sexual risk behaviours practised by MSM, including apparent harm reduction measures that vary based on strategic positioning and serosorting.^21–23

Previous studies have shown that MSM decision-making regarding whether or not to engage in unprotected anal sex are based on partner characteristics, including known or perceived partner HIV serostatus and relationship status (primary versus non-primary partners), with higher risks associated with primary HIV-positive partners.^21–23 Similarly, respondents in this study were more likely to report unprotected sexual activity with HIV-positive partners than with HIV-unknown and HIV-negative partners in most instances. The opposite pattern was seen with disclosure – most respondents reported that they had disclosed their HIV serostatus to some or all of their HIV-positive (96.6%) and HIV-negative (89.8%) male partners, while little more than one-half had disclosed to some or all of their HIV-unknown male partners. This may reflect the circumstance that many HIV-unknown status partners were casual, and perhaps encountered in venues that are not conducive to disclosure.

Almost half (47.6%) of participants reported one or more HIV-negative partners, and about one-third (31.7%; n = 72) reported at least one partner whose HIV status was unknown. Unprotected insertive and receptive anal sex with serodiscordant partners translated into the transmission of approximately 0.7 new HIV infections over the six months preceding study enrolment. When sensitivity analyses were performed in which the level of protection conferred by ART was varied from 0% to 100% and the probability of transmission based on the estimated transmission risk of receptive and/or insertive anal sex was halved and then doubled for each behaviour, the number of new HIV infections varied from 0.17 (best case scenario) to 0.91 (worse case scenario). Therefore, if behaviours remained unchanged, up to two infections a year could arise from this relatively small study population. Supporting the importance of behaviour change interventions for HIV-infected persons in primary care, Morin et al. ²⁰ found that, based on self-reported behaviour, 3145 sexually active HIV-infected patients receiving primary care would be expected to transmit HIV to 37 individuals over a six-month period, with the number of new infections varying from 19 to 71, based on sensitivity analyses.

Consistent with other reports, men were least likely to practise protected oral sex. While oral sex is thought to be ‘lower risk’, this activity has been implicated in syphilis outbreaks, particularly among MSM.^24,25 Therefore, while this activity is not extremely high risk for transmitting HIV, it certainly puts participants at risk for other sexually transmitted diseases (STDs) and is a worthy target for behaviour change if the goal is to protect the individual's health.

This study also found that HIV-positive black MSM reported fewer numbers of total sex partners, as well as fewer HIV- and HIV-unknown status partners, compared with their white counterparts. However, when black MSM did report having HIV-negative male partners, they were less likely to report disclosure of their HIV status to all of their HIV-negative male partners compared with white MSM (52.2% versus 83.1%, respectively). These findings are supported by data from other studies that demonstrate lower HIV transmission risk behaviours among black MSM^26,27 and deserve further investigation.

Alcohol use was moderate in this population, with 44.9% of participants who drank in the previous three months reporting the consumption of five or more alcoholic beverages during one sitting, one or more times. Over one-third (36.8%) of respondents also reported drug use, most often marijuana. Our findings reinforce the correlation between substance use, particularly ‘party drugs’, with risky sexual behaviour among MSM, which continues to be an important influence in defining sexual risk in this population.²⁸ Finally, participants who reported cocaine and/or crack use were more likely to have unprotected sex with partners of discordant serostatus.

The complexity of the sexual behaviours reported by this cohort represent a challenge for behavioural intervention strategies to reduce HIV and STD transmission behaviours. Nonetheless, whether the outcome behaviour of interest is substance use, or in the case of this study, sexual risk behaviours, theory-based, tailored behavioural interventions matched to where a person may be in the process of change (stage-of-change) are effective.^29–31 The diversity in the stages of readiness for change, as well as behaviours, emphasize the need to perform an in-depth assessment during the health-care visit. While this complexity represents a potential barrier to in-depth risk assessment, it can be effectively addressed using computerized, ACASI-delivered assessments, which utilize preprogrammed algorithms to process complex historical data and will provide stage-based behavioural interventions.¹⁹ Such innovative methods that enable the provider to streamline the risk assessment process should continue to be studied.

Limitations of this study include its cross-sectional nature. In addition, the sample size was small and participants were receiving care in a single setting, factors that may impact any inferences made regarding the findings. Participants reported too few female partners to allow for reliable conclusions in this study and there are no data on solely heterosexual men and women with HIV. Future studies are warranted to better characterize sexual risk behaviours and readiness to change those behaviours among both heterosexuals and MSM who also report having female partners.

In summary, in this study risk-taking behaviours among MSM varied widely in the HIV primary care environment, and differed by race/ethnicity. Understanding the risks that patients engage in, the decision-making processes that underlie their choices and the readiness to change risk behaviours could potentially help providers to maximize an individual's health and, in the process, the public's health. Innovative methods to streamline this process in the busy primary care setting and assist the provider in prioritizing risk-reduction activities should continue to be investigated.

Footnotes

ACKNOWLEDGEMENTS

Funding was received for this study through the Department of Health and Human Services Health Resources and Services Administration Special Projects of National Significance Award No. 4H97HA01191-04-02.

References

Hall

, Song

, Rhodes

, Estimation of HIV incidence in the United States. JAMA 2008;300:520–9

CDC. Subpopulation estimates from the HIV incidence surveillance system – United States, 2006. MMWR 2008;57:985–9

McGowan

, Shah

, Ganea

, Risk behavior for transmission of human immunodeficiency virus (HIV) among HIV-seropositive individuals in an urban setting. Clin Infect Dis 2004;38:122–7

Marks

, Crepaz

, Janssen

. Estimating sexual transmission of HIV from persons aware and unaware that they are infected with the virus in the USA. AIDS 2006;20:1447–50

Fisher

, Fisher

, Cornman

, Amico

, Bryan

, Friedland

. Clinician-delivered intervention during routine clinical care reduces unprotected sexual behavior among HIV-infected patients. J Acquir Immune Defic Syndr 2006;41:44–52

Richardson

, Milam

, McCutchan

, Effect of brief safer-sex counseling by medical providers to HIV-1 seropositive patients: a multi-clinic assessment. AIDS 2004;18:1179–86

Goldstein

, Niaura

, Willey-Lessne

, Physicians counseling smokers. A population-based survey of patients' perceptions of health care provider-delivered smoking cessation interventions. Arch Intern Med 1997;157:1313–9

Goldstein

, DePue

, Kazura

, Models for provider–patient interaction: application to health behavior change. In: Shumaker

, Schron

, Ockene

, eds. The Handbook of Health Behavior Changes. New York: Springer Publishing Co, Inc., 1998

Marcus

, Goldstein

, Jette

, Training physicians to conduct physical activity counseling. Prev Med 1997;26:382–8

10.

Ockene

. Primary care-based smoking interventions. Nicotine Tob Res 1999;1:S189–93;discussion S207–10

11.

Lerman

, Rimer

, Trock

, Balshem

, Engstrom

. Factors associated with repeat adherence to breast cancer screening. Prev Med 1990;19:279–90

12.

National Cancer Institute Breast Cancer Screening Consortium. Screening mammography: a missed clinical opportunity? Results of the NCI Breast Cancer Screening and National Health Interview Survey Studies. JAMA 1990;264:54–8

13.

Fiore

, Bailey

, Cohen

, Smoking Cessation: Clinical Practice Guideline No. 18 (AHCPR Publication No. 96-0692). Bethesda, MD: Agency for Health Care Policy Research

14.

Cambell

, DeVellis

, Strecher

, Improving dietary behavior: the effectiveness of tailored messages in primary care practice. Am J Public Health 1994;84:783–7

15.

Kottke

, Battista

, DeFriese

, Brekke

. Attributes of successful smoking cessation interventions in medical practice. A meta-analysis of 39 controlled trials. JAMA 1988;259:2883–9

16.

Willey

, Goldstein

, Niaura

, Smoking Cessation Outcomes of a Physician Counseling and Home-Based Expert System Intervention. Paper Presented at the Annual Meeting of the Society of Behavioral Medicine, Washington, DC, 1996

17.

Prochaska

, DiClemente

. The Transtheoretical Approach: Crossing the Traditional Boundaries of Therapy. Homewood, IL: Dow Jones/Irwin, 1984

18.

Prochaska

, DiClemente

. Stages and processes of self-change of smoking: toward an integrative model of change. J Consult Clin Psychol 1983;51:390–5

19.

Grimley

, Bachmann

, Jenckes

, Erbelding

. Provider-delivered, theory-based, individualized prevention interventions for HIV positive adults receiving HIV comprehensive care. AIDS Behav 2007;11:S39–47

20.

Morin

, Myers

, Shade

, Koester

, Maiorana

, Rose

. Predicting HIV transmission risk among HIV-infected patients seen in clinical settings. AIDS Behav 2007;11:S6–S16

21.

Bachmann

, Grimley

, Waithaka

, Desmond

, Saag

, Hook

3rd . Sexually transmitted disease/HIV transmission risk behaviors and sexually transmitted disease prevalence among HIV-positive men receiving continuing care. Sex Transm Dis 2005;32:20–6

22.

Suarez

, Kelly

, Pinkerton

, Influence of partner's HIV serostatus, use of highly active antiretroviral therapy, and viral load on perceptions of sexual risk behavior in a community sample of men who have sex with men. J Acquir Immune Defic Syndr 2001;28:471–7

23.

Elford

, Bolding

, Maguire

, Sherr

. Sexual risk behaviour among young gay men in a relationship. AIDS 1999;13:1407–11

24.

CDC. Transmission of primary and secondary syphilis by oral sex – Chicago, Illinois, 1998—2002. MMWR 2004;53:966–8

25.

Emerson

, Lynch

, Fox

, The syphilis outbreak in Northern Ireland. Int J STD AIDS 2007;18:413–7

26.

Ruiz

, Facer

, Sun

. Risk factors for human immunodeficiency virus infection and unprotected anal intercourse among young men who have sex with men. Sex Transm Dis 1998;25:100–7

27.

Millett

, Flores

, Peterson

, Bakeman

. Explaining disparities in HIV infection among black and white men who have sex with men: a meta-analysis of HIV risk behaviors. AIDS 2007;21:2083–91

28.

Koblin

, Chesney

, Husnik

, High-risk behaviors among men who have sex with men in 6 US cities: baseline data from the EXPLORE study. Am J Public Health 2003;93:926–3

29.

Prochaska

, DiClemente

, Velicer

, Rossi

. Standardized, individualized, interactive and personalized self-help programs for smoking cessation. Health Psych 2003;12:399–405

30.

Grimley

, Hook

III . A 15-minute interactive, computerized condom use intervention with biological endpoints. Sex Transm Dis 2009;36:73–8

31.

Noar

, Benac

, Harris

. Does tailoring matter? Meta-analysis review of tailored print health behavior change interventions. Psychol Bull 2007;133:673–93