Secondary Logo

Journal Logo


Factors Associated With Testing for Hepatitis C in an Internet-Recruited Sample of Men Who Have Sex With Men


Author Information
  • Free


HEPATITIS C VIRUS (HCV) infection is the most common chronic blood-borne infection in the United States. 1,2 Nearly 4 million U.S. residents (1.8%) have been infected with HCV, 3,4 yet less than 50% are aware of their infection. 1,5 While serving as a source of HCV transmission to others, infected individuals themselves are at risk for significant morbidity and mortality from such complications as chronic liver disease, cirrhosis, and hepatocellular carcinoma. 6–10 In the United States, chronic liver disease is the 10th leading cause of death among adults, accounting for up to 30,000 deaths per year, or 1% of all deaths. 11 Because 30- to 49-year-olds have the highest prevalence rates and 20- to 39-year-olds have the highest incidence rates of HCV infection, the number of deaths attributable to HCV-related chronic liver disease could increase substantially over the next 10 to 20 years as the infected individuals reach ages when complications from chronic liver disease typically occur. 1,12

Having its highest concentration in the blood, HCV is transmitted primarily through permucosal and percutaneous routes. 2 Identified risk factors for HCV infection include injection drug use, 11 HIV infection, 8,13,14 intranasal cocaine use, 15 multiple sexual partners, 1–3,15,16 blood transfusions, 2 body and ear piercing, 5,15 employment in clinical laboratory work or inpatient care, 1 exposure to a sex partner or household member with a history of hepatitis, 1 and perinatal exposure. 5 Studies also have implicated tattooing 5,17 and shared razors in HCV infection. 18

Testing for HCV became available in 1989, yet research suggests that most individuals infected are unaware of their status. 1,5 Approximately 36,000 of the 6 million individuals who donate blood each year in the United States repeatedly show anti-HCV positive test results. 15

Testing for HCV offers several potential benefits for individuals infected with HCV. First, behavior changes can be made to decrease the risk for transmission of HCV to others. Individuals also can make lifestyle modifications, such as avoiding alcohol consumption, to promote physical health. Finally, treatment can be considered to reduce morbidity and possible mortality, and perhaps to cure the individual. The Internet was used as a method of data collection in an effort to identify correlates of HCV testing among men who have sex with men (MSM), self-identified as at risk for HCV.


After approval from the Institutional Review Board of the University of Alabama at Birmingham, an online Internet data collection site was established, which was accessible online for one complete month, May 1 to 31, 1999. Results from Internet searches for gay- and bisexually oriented Web sites were used to identify Web masters solicited to establish linkages from their Web sites to the authors’ data collection Web site. Notices also were posted in Web-site “guest books” of gay- and bisexually oriented Web sites that included the uniform resource locator (URL) address of the authors’ data collection Web site. Active recruitment of respondents was accomplished by announcing the anonymous study via electronic mailing lists, targeting lists maintained for specific MSM interest groups.

The questionnaire was translated into HTML, the de facto language of the Internet. With a point-and-click interface, respondents completed a questionnaire that was visually and functionally similar to traditional questionnaires: The background was white and the text black to resemble a paper-and-pencil questionnaire. Form elements included check boxes, radio buttons, text entry boxes, and selection lists to facilitate data entry and minimize the chance of error. A common gateway interface (CGI) script was used for automatic compilation and exportation of data into SAS software (Cary, NC) for data analysis.


The questionnaire contained 31 questions designed to survey hepatitis-related knowledge, attitudes, and behaviors. The questionnaire assessed demographic information, knowledge about hepatitis transmission and prevention, risk behavior, and resources used to obtain information about hepatitis. Anticipating the difficulty respondents may have had distinguishing between the various types of viral hepatitis, the questionnaire included specific questions about the more commonly known hepatitis B virus (HBV) in addition to HCV.

To assess respondents’ knowledge of hepatitis, a knowledge scale was created from a total of all the correct responses to knowledge items, with scores ranging from 0 to 4. The standardized alpha coefficient for the knowledge scale, a measure of internal reliability, was 0.76.

Respondents reported numbers of sexual partners during the preceding 30 days and over their lifetime. Inconsistencies exist among studies on the efficiency of sexual transmission of HCV. 16 However, research suggests that a greater number of sexual partners and a failure to use condoms are associated with HCV transmission among both heterosexual and MSM populations. 6 Thus, respondents who reported more than 10 partners in their lifetime or more than 1 partner in the past 30 days and less consistent condom use were considered at risk for infection. The following nonsexual risk behaviors were assessed over the lifetime and the preceding year: needle use for drugs or steroids, receiving a tattoo, receiving an ear or body piercing, sharing a razor, and sharing straws to snort cocaine. Respondents reported blood or blood products received before 1990.

Prior testing and treatment for hepatitis C, HIV status, health insurance coverage, and resources used for information about hepatitis were assessed. Sources of information about hepatitis included news magazines, the popular press, men’s magazines, the gay media, the health department, healthcare providers, friends, television, and professional training. Scores also were created to examine whether the number of information sources affected HCV testing.

To allow for categorical analyses, continuous variables such as age were collapsed into the following categories: younger than 20, 20 to 29 years, 30 to 39 years, 40 to 49 years, 50 to 59 years, and 60 years and older.

To ensure respondent anonymity, no electronic individual user tracking data (e.g., Internet protocol [IP] address) were examined. However, respondents who indicated that they lived in the United States were asked to supply their U.S. postal (zip) code. Respondent zip code, gender, and date of birth (month/day/year) were examined to identify possibly duplicate responses.

Statistical Analysis

The statistical analysis comprised several steps. Frequencies of risk factors were computed. Correlations between sexual risk behaviors and nonsexual risk factors were examined using the χ2 statistic. The associations between potential correlates and hepatitis C testing were assessed using contingency table analyses. The χ2 statistic and Cramer’s V, the corresponding P levels, and confidence limits were computed to assess the magnitude of these bivariate associations. 19

Next, factors identified as associated with HCV testing were entered into a multivariate logistic regression model to identify the independent contribution of each factor while adjustments were made for the effects of other factors in the model. 20 The χ2 statistic was used to assess the overall statistical significance of the logistic model. An adjusted odds ratio was calculated for each factor to assess the magnitude of association between statistically significant predictors and HCV testing.


Of the 628 respondents who completed the survey, 90.3% reported that they currently lived in the United States. Of these US residents, 21.7% were women and 11.1% were heterosexual men. Of the gay or bisexual male respondents who reported living in the United States, 328 (86.1%) identified themselves as “gay” or as “having sex with members of the same sex,” and 53 (13.9%) identified themselves as “bisexual” or as “having sex with both men and women.” Of the 381 male respondents reporting sex with men, 94.75% (n = 361) reported one or more HCV risk factors. These 361 respondents comprise the sample used for analysis.

The average age of the 361 respondents was 37.7 years (range, 18–78 years). The sample was predominately white (85%). Among these respondents, 38.8% reported that they had been tested for HCV, and 24 reported a history of positive test results for HCV. The frequency of risk factors reported by respondents is presented in Table 1. Engagement in sexual risk behavior was correlated with engagement in nonsexual risk behaviors (P = 0.001). Of the respondents at risk because of sexual behavior, 88.7% also were at risk because of other HCV-related risk factors,

Table 1:
Risk Factor Frequencies

On the basis of bivariate associations, 14 characteristics were associated with HCV testing (Table 2). Any variable whose bivariate test had a P value less than 0.25 was considered a candidate for modeling, according to recommended model-building techniques. 20 These factors were entered into a logistic regression analysis. The three factors that retained significance after adjustments for all the other factors in the model were a lifetime history of nonsexual risk behaviors (tattooing and body or ear piercing), increased knowledge of HCV, and healthcare provider communication about hepatitis (Table 3). The predictive power of the overall model (χ2 = 79.59;P = 0.0001) was high, correctly classifying 74.6% of the respondents into their self-reported HCV testing status categories (P = 0.0001). No interactions among these factors were observed.

Table 2:
Correlates of Hepatitis C Testing
Table 3:
Results From Logistic Regression of Correlates on Hepatitis C Virus Testing


There is an urgent need to enhance HCV awareness and facilitate HCV testing among populations at high risk for infection. Although it has been suggested that MSM are not at increased risk for HCV over their heterosexual counterparts, the current sample of MSM were at risk for infection through both sexual and nonsexual means. Nearly 95% of this MSM sample reported one or more factors that put them at risk for HCV, and most of the respondents who reported multiple sexual partners were also at risk for HCV through other risk factors.

The findings from this study demonstrate that although 38.8% of the respondents reported being tested for HCV, 26.3% reported having no information about hepatitis. Respondents who reportedly underwent tattooing or body piercing were more likely to have been tested for HCV. This may be partly the result of education efforts aimed at individuals and businesses participating in these activities. Not only does it seem imperative to develop innovative interventions that encourage increased knowledge of HCV among MSM specifically, but as this study also suggests, strategies must be developed to facilitate accurate risk assessment of patients, and to improve provider communication with patients.

Healthcare providers have an impact on the testing rates of individuals at increased risk for HCV infection. Respondents who reported provider communication about HCV were three times more likely (95% CI, 1.85–5.06) to have been HCV tested than respondents who reported no provider communication. The quality and content of communication with a healthcare provider may be important. Unfortunately, the perceived quality of patient–provider interactions was not assessed. Another issue that needs further investigation is the importance of patients’ openness to their healthcare providers about risk behavior. Because HCV testing was not associated with having health insurance, at-risk patients may need to perceive a level of comfort with their providers before they will discuss stigmatized risk behaviors. Providers may need to create environments conducive to risk disclosure.

Using the Internet to Collect Data

The use of the Internet as a data collection method deserves discussion. First, and most importantly, the results of this study may not generalize to other MSM populations. Use of the Internet does not ensure true random sampling. 21 Self-selection of survey respondents is likely to produce bias. 22,23 However, with the rapid growth of the Internet and the multiplication of users, the Internet provides a vehicle to reach an increasing number of individuals for both data collection and intervention;21 Internet-based survey research represents a trade-off between easy, low-cost access to large numbers of respondents and sample representativeness of the population studied. However, the degree of fit between a survey sample and the target population about which generalizations can be made is a common challenge for many types of survey methods. 24,25 An advantage of Internet data collection is that the physical locations of the server and respondent are inconsequential, affording a sample unconfined by geographic boundaries. This cannot be accomplished easily with some paper-and-pencil methods, for example. 22

Second, Internet respondents may differ from other populations because a computer with Internet access is required. Also, the Internet may reach a different segment of the target population, which may differ from other MSM in level of openness about sexual behavior. 26,27 Individuals who are uncomfortable about disclosing stigmatized behavior may be more accessible, but they may have greater hesitation about participating, despite assurances of anonymity or confidentiality. A link to a free service that guarantees anonymous access to the Internet, such as, may alleviate respondent distrust. However, the effect of such assurance is largely unstudied and unknown. Thus, more research is needed to understand who is reached through the Internet and the characteristics of the respondents.

Third, this study used a cross-sectional research design. Additional studies using a prospective cohort design will be necessary to evaluate the significance and stability of the factors identified in this study on HCV testing rates over time. The Internet may offer a quick, low-labor, inexpensive method for collecting data prospectively. 22,25,28–31

Fourth, the sample was not racially or ethnically diverse, but predominantly a white sample (85.1%). However, the demographics of Internet users are expanding rapidly. 32 To illustrate, in 1996 only 27% of computer users were going online, as compared with 80% in 1999.

Fifth, multiple submissions, whether by error or intent, of a completed electronic survey can be difficult to prevent. However, unless the individual respondent creates an entirely new demographic and health profile for each submission, duplicate responses can be identified, 33 and a well-written common gateway interface program may detect and delete multiple submissions. 25 Furthermore, this problem is not unique to electronic media. Multiple submissions are a risk also with mail-in questionnaires.

Sixth, because respondents interact with the questionnaire in a structured way, erroneous or unacceptable data may be avoided. Respondents can be prohibited from entering multiple responses to questions requesting a single answer. By limiting the ability to make mistakes, the online survey may yield more usable data than traditional method.

Seventh, response rates were incalculable because it was not known how many possible respondents may have read electronic advertisements, such as solicitation e-mails, or examined the Web site itself but chose not to complete the questionnaire.

Eighth, although a self-administered format was used to minimize response bias, the results remain based on self-report data, which has well-established limitations. 34 However, the number of respondents in this study who reported a positive HIV status may confirm the validity of these data. Although the impact of using the Internet on self-report is not well documented, research in other disciplines suggests considerable similarity of responses between Internet and paper-and-pencil data collection methods. 35 The most notable exception is the significance of increased socially undesirable responses in data collected via the Internet.

Finally, the self-administered questionnaire was developed as an epidemiologic tool to survey relevant factors related to vaccination. Therefore, it is limited in scope. More detailed information could be gathered through other data-collection method (e.g., in-depth interviews). Researchers have begun to explore data collection using open-ended questions via the Internet, and preliminary results suggest that respondents to open-ended Internet questionnaires tend to be better educated than other Internet users. 33


In the United States, a large group of individuals with HIV infection exists who may transmit the infection to others, and who are at risk for HCV-related chronic diseases. To prevent new infections, programs must focus on preventing the initiation of HCV-risk behaviors, and on providing risk-reduction counseling and services to those who engage in risk behaviors. The MSM in the current sample were engaging in multiple behaviors that put them at risk for HCV. However, they were unaware for their status and had little information about HCV. These findings suggest that whereas MSM as a group may not be at increased risk for HCV infection, 8 there may be subgroups or pockets of MSM that are at increased risk.

The lack of an effective vaccine, the high rates of chronic infection, and the limited effectiveness of treatment mandate that at-risk individuals be identified, counseled, tested, and treated. Increased public awareness, education, and perhaps screening are needed, particularly in communities whose members are at increased risk for transmission and infection.

Furthermore, the role of the healthcare provider cannot be underestimated. Because recommendation by a healthcare provider is a strong predictor of preventive behavior in other health-related areas with other populations, 36 healthcare providers must inform patients about HCV. With no vaccine available for HCV, effective preventive measures are key to reducing the burden of disease. Identifying individuals who could benefit from behavioral and medical intervention is vital in any effort to reduce HCV infection, morbidity, and mortality.

Advances in computer technology and the exponential growth of the Internet not only provide innovative strategies for collecting data, but also may be incorporated into provider practice. Providers are beginning to experiment with computer-based delivery of patient education. 37–39 For example, an interactive software program on computers placed in a clinic or made available on the Internet may allow for low-labor risk assessment and theoretically based risk reduction intervention. 25,39 Individuals who are either unaware of their risks or embarrassed about their behaviors may progress along a continuum to act on their information and vaccination needs. Computer technology may provide “cues to action,” encouraging individuals either to discuss their risks with or to seek testing from their healthcare providers. This medium may facilitate dialogue about risks between at-risk individuals and their providers.


1. Centers for Disease Control and Prevention. Recommendations for prevention and control of hepatitis C virus (HCV) infection and HCV-related chronic disease. Centers for Disease Control and Prevention. MMWR Morb Mortal Wkly Rep 1998; 47: 1–39.
2. Williams I. Epidemiology of hepatitis C in the United States. Am J Med 1999; 107: 2S–9S.
3. Alter MJ, Kruszon-Moran D, Nainan OV, et al. The prevalence of hepatitis C virus infection in the United States, 1988 through 1994. N Engl J Med 1999; 341: 556–562.
4. Shakil AO, Conry-Cantilena C, Alter HJ, et al. Volunteer blood donors with antibody to hepatitis C virus: clinical, biochemical, virologic, and histologic features. The Hepatitis C Study Group [see comments]. Ann Intern Med 1995; 123: 330–337.
5. Gordon FD. Cost effectiveness of screening patients for hepatitis C. Am J Med 1999; 107: 36S–40S.
6. Alter MJ, Mast EE. The epidemiology of viral hepatitis in the United States. Gastroenterol Clin North Am 1994; 23: 437–455.
7. Koff RS. Advances in the treatment of chronic hepatitis. JAMA 1999; 282: 511–512.
8. Osella AR, Massa MA, Joekes S, et al. Hepatitis B and C virus sexual transmission among homosexual men. Am J Gastroenterol 1998; 93: 49–52.
9. Schafer DF, Sorrell MF. Hepatocellular carcinoma [see comments]. Lancet 1999; 353: 1253–1257.
10. van Leeuwen DJ, Dadrat A. Viral hepatitis and its imitators. In: Blackwell RE, ed. Women’s Medical Text. Boston: Blackwell Science, 1996: 324–338.
11. Gross JB. Hepatitis C: infection, transmission, recognition, and treatment. Minn Med 1998; 81: 28–32.
12. McQuillan GM, Alter MJ, Moyer LA, Lambert SB, Margolis HS. A population-based serologic study of hepatitis C virus infection in the United States. In: Rizzetto M, Purcell RH, Gerin JL, Verme G, eds. Viral Hepatitis and Liver Disease. Turin: Edizioni Minerva Medica, 1997: 267–270.
13. Becherer PR. Viral hepatitis: what have we learned about risk factors and transmission? Postgrad Med 1995; 98:65–68, 71–74.
14. Brook MG. Sexual transmission and prevention of the hepatitis viruses A–E and G. Sex Transm Infect 1998; 74: 395–398.
15. Conry-Cantilena C, VanRaden M, Gibble J, et al. Routes of infection, viremia, and liver disease in blood donors found to have hepatitis C virus infection [see comments]. N Engl J Med 1996; 334: 1691–1696.
16. Feldman JG, Minkoff H, Landesman S, Dehovitz J. Heterosexual transmission of hepatitis C, hepatitis B, and HIV-1 in a sample of inner-city women. Sex Transm Dis 2000; 27: 338–342.
17. Mele A, Sagliocca L, Manzillo G, et al. Risk factors for acute non-A, non-B hepatitis and their relationship to antibodies for hepatitis C virus: a case–control study. Am J Public Health 1994; 84: 1640–1643.
18. Tumminelli F, Marcellin P, Rizzo S, et al. Shaving as potential source of hepatitis C virus infection [letter]. Lancet 1995; 345: 658.
19. Fleiss JL. Statistical Methods for Rates and Proportions. New York: Wiley, 1973.
20. Hosmer DW, Lemeshow S. Applied Logistic Regression. New York: Wiley, 1989.
21. Baehring TU, Schulze H, Bornstein SR, Scherbaum WA. Using the World Wide Web: a new approach to risk identification of diabetes mellitus. Med Inform 1997; 46: 31–39.
22. Houston JD, Fiore DC. Online medical surveys: using the Internet as a research tool. MD Comput 1998; 15: 116–120.
23. Eysenbach G, Diepgen TL. Epidemiological data can be gathered with the World Wide Web. BMJ 1998; 316: 72.
24. Schmidt WC. World Wide Web survey research: benefits, potential problems, and solutions. Behav Res Methods Instrum Comput 1997; 29: 274–279.
25. Pealer LN, Weiler RM. Web-based health survey research: a primer. Am J Health Behav 2000; 24: 69–72.
26. Hays RB, Kegeles SM, Coates TJ. Unprotected sex and HIV risk-taking among young gay men within boyfriend relationships. AIDS Educ Prev 1997; 9: 314–329.
27. Prieur A. Norwegian gay men: reasons for continued practice of unsafe sex. AIDS Educ Prev 1990; 2: 109–115.
28. Kiuchi T, Kaihara S. Automated generation of a World Wide Web–based data entry and check program for medical applications. Comput Methods Programs Biomed 1996; 52: 129–138.
29. Mrad R lenert LA. Epidemiology and the Internet. Epidemiology 1998; 9: 364.
30. Rothman KJ, Cann CI. Epidemiology and the Internet. Epidemiology 1997; 8: 123–125.
31. Suchard MA, Adamson S, Kennedy S. Netpoints: piloting patient attitudinal surveys on the Web. BMJ 1997; 315: 529.
32. Budman SH. Behavioral health care dot-com and beyond: computer-mediated communications in mental health and substance abuse treatment. Am Psychol 2000; 55: 1290–1300.
33. Fawsett J, Buhle EL. Using the Internet for data collection: an innovative electronic strategy. Comput Nurs 1995; 13: 273–279.
34. Giuliani M, Rezza G, Lepri AC. Stigmatized behaviors, cultural context, and accuracy of self-reported data in the STD field. Sex Transm Dis 1998; 25: 221.
35. Kiesler S, Sproull LS. Response effects in the electronic survey. Public Opinion Q 1986; 50: 402–413.
36. Goldstein MG, Depue J, Kazura A, Niaura R. Models of provider–patient interaction: applications to health behavior change. In: Shumaker SA, Schron EB, Ockene JK, McBee WL, eds. The Handbook of Health Behavior Change. New York: Springer, 1998: 85–113.
37. Marcus BH, Bock BC, Pinto BM, Forsyth LH, Roberts MB, Traficante RM. Efficacy of an individualized, motivationally tailored physical activity intervention. Ann Behav Med 1998; 20: 174–180.
38. Paskett ED, McMahon K, Tatum C, et al. Clinic-based interventions to promote breast and cervical cancer screening. Prev Med 1998; 27: 120–128.
39. Krishna S, Balas EA, Spencer DC, Griffin JZ, Boren SA. Clinical trials of interactive computerized patient education: implications for family practice. J Fam Pract 1997; 45: 25–33.
© Copyright 2001 American Sexually Transmitted Diseases Association