With the development of more effective therapies, which significantly delay the onset of AIDS for many people, public health experts believe the numbers of AIDS cases and deaths are no longer reliable measures for monitoring the epidemic.1 Guidelines issued by the Centers for Disease Control (CDC)2 have recommended that all states extend their reporting requirements from late-stage HIV infection (AIDS cases) to initial HIV infection diagnosis. California has adopted regulations to implement a statewide HIV surveillance system using a non-name code as opposed to the use of an individual's name. California's HIV surveillance activity began on July 1, 2002. This required laboratories, health care providers, and counseling and testing providers to report all cases of HIV infection to their local health department.
Some states, such as New York, Texas, and Florida, have adopted name-based reporting while other states, such as Maryland, Massachusetts, and California, have chosen a non-name code system. There were no recommendations from the CDC regarding components of the non-name code, and different states have devised their own non-name reporting systems. By May of 2003, 37 states had adopted or were implementing name-based HIV case surveillance; 8 states and Puerto Rico had adopted non-name coded systems; 4 states had adopted a name-to-code system; and the system remained unresolved in only 1 state.2
At the national level and in the literature there has been substantial debate over the relative merits of name-based vs. non-name HIV infection reporting.1,3-11 Proponents of HIV name reporting argue that HIV should be treated like any other sexually transmitted disease for which names are reported. They also argue that having an individual's name assists with referral to medical care, risk reduction counseling, and partner notification. Opponents argue that the public health benefit of names reporting is outweighed by the deterrent effect it has on HIV testing and early identification. They also contend that name reporting is not necessary to achieve referral to medical care, risk reduction counseling, or partner notification.3,12 Whether the benefits to individual and public health outweigh the risks is still the subject of debate.9,13
In the California system, unlike AIDS case reporting, HIV case reports and laboratory notifications do not include personal-identifying information. Health care providers and laboratories instead use a non-name code that consists of 4 components: Soundex (an algorithm to convert a surname to an alphanumeric code14), an 8-digit date of birth, a 1-digit code for gender, and the last 4 digits of the social security number. Health care providers are responsible for providing the client's surname, date of birth, and gender when submitting laboratory requisitions for any test used to identify HIV, a component of HIV, or antibodies or antigens to HIV.15 An initial study in San Francisco assessing the feasibility of a non-name based HIV surveillance system indicated that the level of completeness in reporting using this system exceeded the minimum level required by the CDC.16
The goal of this study was to evaluate the acceptability of California's HIV Surveillance Program by non-name code through interviews with HIV test takers and to assess potential changes in testing behavior under coded, name-to-code, and name-based HIV infection reporting systems. Thus we assessed preferences among the 3 major types of HIV infection reporting and potential testing behavior changes through interviewer-administered surveys among HIV test takers exiting from confidential and anonymous testing visits in 4 diverse California counties.
This study was conducted as part of a contract between the California State Office of AIDS and the University of California, San Francisco (UCSF), AIDS Policy Research Center to conduct an independent external evaluation of the acceptability of California's HIV non-name code reporting system to key stakeholders in the state, including HIV test takers. The evaluations were conducted immediately prior to California's implementation of its HIV non-name code reporting system and approximately 1 year after implementation. Four counties-Los Angeles, Riverside, Fresno, and Santa Clara-were chosen in collaboration with the State Office of AIDS, to reflect the diversity of HIV risk populations in California. To evaluate the acceptability of policy options for HIV infection reporting, we conducted exit interviews with HIV test takers immediately following appointments for pretest counseling and blood collection at HIV testing sites. Local county health department surveillance officers chose from existing publicly funded HIV testing sites a sample of testing centers to represent a mixture of confidential and anonymous testing sites and populations served. Three testing sites were selected in Los Angeles, 5 each in Riverside (anonymous) and Fresno, and 1 site in Santa Clara (both anonymous and confidential). All 14 testing sites contacted by the study team agreed to participate. Data reported here were collected during May and June of 2002 immediately prior to implementation of regulations requiring statewide reporting of HIV infection, with the exception of Santa Clara County, where data was collected in July of 2002 due to delays in obtaining local institutional review board approval. Postimplementation data were collected from the same sites, using the same methodology, between May and June of 2003.
Participant Eligibility and Recruitment
Eligible participants were individuals who had just taken an HIV test, were aged 18 years or older, who spoke either English or Spanish, and who were able to provide verbal consent to the interviewer-administered questionnaire. Interviews were conducted in private settings out of hearing distance of other individuals. Potential study participants were identified by HIV test counselors and clinic staff and informed of the opportunity to participate in the exit interview. If willing, they were directed to study field staff who obtained verbal consent of the participants. Two field study staff visited the clinics in each city during operating hours for a minimum of 5 days. Participants were reimbursed $10 for participating in the exit survey. The UCSF Committee on Human Research approved the study procedures for approaching and interviewing participants; 2 testing sites required a second local institutional review board approval.
The exit survey questionnaire and materials were developed in consultation with the California State Office of AIDS. The exit questionnaire obtained information on demographic characteristics of the respondent, sexual activity, injection drug use, and prior HIV testing history. Respondents were asked about the type of HIV test they had just taken, the likelihood of currently testing HIV positive on a scale of 1-10, and the likelihood of HIV testing in the next 12 months.
Because of the complexity of the policy issues, spoken scripts in both English and Spanish with matching printed materials outlining the concepts being assessed were developed. Figure 1 shows the definitions used to distinguish anonymous vs. confidential HIV testing sites. Figure 2 lists the definitions used to distinguish the 3 types of HIV reporting systems.
Following a description of the 3 HIV reporting systems, respondents were asked about the likelihood of HIV testing in the next 12 months under each of the 3 reporting systems as well as if there were no reporting system at all. The exact wording of the question with regard to name reporting for a confidential HIV test was “If California were to start using a name reporting system, how likely is it that you would get a confidential HIV test in the next 12 months?” The questions for anonymous and confidential HIV testing and testing under name-to-code and code-based reporting system used the same wording structure. Likelihood of testing was assessed using a 4-point scale consisting of 1) very likely, 2) somewhat likely, 3) somewhat unlikely, and 4) very unlikely. Likelihood of testing was asked first for confidential testing and then again for anonymous testing. Respondents were also asked which of the 3 reporting systems would be most acceptable to them, if they knew which HIV reporting systems was adopted in California, and how they had heard about the new reporting system.
Data from the interview-administered questionnaire were computer entered with a 10% double-entry validation check. All statistical analyses were conducted using the SAS statistical system version 8.1 (SAS Institute, Cary, NC). Response proportions and 95% CIs were calculated using standard equations.17 Contingency table analyses were evaluated using Fisher exact test and multivariate logistic regression was used to determine independent predictors of binary outcome measures. The effect of clustering by county or testing site was evaluated using generalized estimating equations. For analyses of continuous outcome measures, analysis of variance was used unless the underlying distributions did not meet statistical assumptions, in which case nonparametric tests were used (Wilcoxon 2-sample test or Kruskal-Wallis test). The McNemar χ2 test of related proportions was used to determine whether the proportion of individuals reporting likelihood of confidential testing under HIV reporting is significantly different from the current system (no reporting) or different from the likelihood of testing at an anonymous testing site under the same type of HIV reporting system.
Preimplementation Data: Participant Demographics
HIV exit interviews were obtained from persons attending 14 HIV testing sites covering the 4 California counties. A total of 208 individuals completed the exit interviews. The demographics of the study respondents are presented in Table 1. As planned in the study design, the 4 counties exhibited significant diversity in respondent demographics. The proportion of individuals self-reporting as African American differed significantly between the 4 counties (P = 0.049), with Los Angeles county sites reporting the highest proportion of African Americans (23%) and Santa Clara county sites reporting the lowest proportion of African Americans (3%). The proportion of individuals self-reporting as gay or bisexual also varied significantly (P = 0.011) by county, with the highest proportion of men who have sex with men (MSM) respondents sampled from Los Angeles County (40%) and the least proportion of gay or bisexual respondents coming from Fresno County (14%). Other significant differences in respondent demographics between the counties were gender (P < 0.0001), age <40 years (P < 0.009), highest educational grade completed (P < 0.001), household monthly income (P < 0.001), history of injection drug use (P < 0.001), and type of HIV test just taken (anonymous vs. confidential, P < 0.002).
Anonymous vs. Confidential Testing
After a concise explanation of the differences between anonymous and confidential HIV testing, 43% of respondents reporting having just taken a confidential HIV test, 55% reported an anonymous HIV test, and 1.4% (n = 3) were unsure as to the type of HIV test they had just taken. When asked whether a confidential HIV test result would be reported to the local health department, 51% reported that a positive HIV test would be reported to the health department, 29% reported that a positive HIV test would not be reported to the health department, and 20% were unsure whether a positive HIV test would be reported. Among those with an answer (n = 164), there was no statistically significant difference in the proportion of the individuals stating that a positive HIV test would be reported to the health department by the type of HIV test (anonymous vs. confidential) taken by the respondent.
Knowledge of HIV Infection Reporting System to Be Used in California
When respondents were asked if they knew which of the 3 types of HIV infection reporting systems was about to start being used in California, only 20% reported knowing which of the 3 systems was about to start. Among those stating they knew which system was about to start (n = 38), 58% correctly identified the non-name code system, whereas 21% misidentified the California system as name and 21% as name-to-code HIV reporting.
Preferences Among HIV Infection Reporting Systems
Following a concise description of the name, name-to-code, and code-based HIV infection reporting systems, respondents were asked which of the 3 systems would be most acceptable to them. The distribution of respondents' preferences of HIV infection reporting system is presented in Figure 3. Overall with respect to which would be the most acceptable system, 67% reported non-name code (95% CI: 59-75%), 19% reported name-to-code (95% CI: 7-31%), and 12% reported name-based HIV reporting (95% CI: 0-25%). Only 2% of respondents reported having no preference or did not know which of the 3 HIV infection reporting systems would be most acceptable to them. Across the 4 counties, the non-name code-based HIV infection reporting system was reported as the most acceptable system by the majority of the respondents in each of the 4 counties.
Predictors of Preference for Coded HIV Infection Reporting
Univariate and multivariate predictors of preference for coded vs. name-based HIV infection reporting systems are given in Table 2. For the purposes of analysis, preferences for name-to-code or code-based reporting systems were collapsed into a single group. This outcome measure was chosen to best represent the major dichotomy between name HIV infection reporting and the 2 similar coded systems. Significant univariate predictors of a preference for coded HIV reporting were female gender, age <40 years, some college or more education, monthly income ≥$2000, and having just taken an anonymous HIV test. Injection drug use was the only significant univariate predictor of a preference for name HIV reporting. Ethnicity, not completing high school, number of prior HIV tests, and self-report of a >50% chance of being HIV infected did not predict a preference for coded or name-based reporting.
Significant independent predictors of a preference for coded HIV reporting in multivariate analysis were female gender (OR = 6.2, 95% CI: 1.7-22, P = 0.006), MSM (OR = 5.7, 95% CI: 1.2-26), and having just taken an anonymous HIV test (OR = 3.6, 95% CI: 1.4-9.3, P = 0.009). Clustering by county or testing site had negligible effects on identified risk factors and standard errors.
Likelihood of Confidential and Anonymous HIV Testing Under the 3 HIV Infection Reporting Systems
Table 3 summarizes responses to a series of questions about the likelihood of confidential and anonymous HIV testing in the next 12 months under no HIV reporting and under the 3 HIV reporting system options. We found no significant difference in the proportion of respondents reporting that they were likely (very likely or somewhat likely) to have a confidential HIV test between the no-reporting system (76% likely) and the code reporting system (72% likely, P ≤ 0.0001). However, there were significant differences reported in the likelihood of confidential HIV testing between the no-reporting or code-based systems and the name-based and name-to-code systems. Under the no-reporting system, 76% of respondents reported being likely to test in the next 12 months as opposed to only 56% under name reporting (P < 0.0001) and 62% under name-to-code reporting (P ≤ 0.0002). Similar significant differences were seen for the proportion of individuals reporting that they were likely to have a confidential HIV test between the code-based system (78% likely) and the name-based and name-to-code based systems (P < 0.0001 and P = 0.0003, respectively).
When we compared the likelihood of confidential vs. anonymous HIV testing in the next 12 months among individuals, we found that there was no significant difference in the proportion of individuals who reported being likely (very likely or somewhat likely) to test in the next 12 months within the no-reporting systems and within the code-based reporting system. Significant differences in the likelihood of testing were seen within the name-based and name-to-code reporting systems. Significantly fewer individuals reported being likely to have a confidential HIV test than reported being likely to have an anonymous HIV test for name-based HIV reporting (51% likely confidential vs. 76% likely anonymous, P < 0.0001) and for name-to-code reporting (63% likely confidential vs. 75% likely anonymous, P = 0.0061).
Postimplementation Data: Participant Demographics
HIV exit interviews were obtained from a total of 226 individuals completing the exit interviews (95 Los Angeles, 48 Fresno, 50 Riverside, 33 Santa Clara) approximately 1 year after the implementation of HIV infection reporting. The distribution of men and women was similar to that of respondents in the preimplementation period (69% men and 31% women) as was the distribution of those reporting homosexual or bisexual orientation (25% homosexual and 7% bisexual). The postimplementation sample was also similar in its distribution by race (32% white, 31% Hispanic, 29% African American, 2% Asian, and 6% other), with a slightly higher proportion of African Americans.
Postimplementation Knowledge of California's HIV Infection Reporting System
After implementation of California's HIV infection reporting system, only 14 persons (6%) reported knowing which of the 3 systems was in operation within the state. Approximately 45% of respondents reported that they had just taken a confidential HIV test, with 49% reporting having taken an anonymous HIV test, and 6% unsure of which type of HIV test they had just taken. Forty-nine percent of respondents correctly reported that a positive confidential HIV test would be reported to the local health department whereas 16% believed that a positive confidential HIV test would be not be reported. Thirty-five percent of respondents did not know whether a positive confidential HIV test would be reported.
Postimplementation Preferences Among HIV Infection Reporting Systems and Predictors of Preference
After implementation of California's HIV infection reporting system, the non-name code system was still reported as the most acceptable reporting system by HIV test takers. Overall with respect to which would be the most acceptable system, 54% reported non-name code, 19% reported name-to-code, and 17% reported name-based HIV reporting (see Figure 4). Only 4% of respondents reported having no preference or did not know which of the 3 HIV infection reporting systems would be most acceptable to them. Across the 4 counties, the non-name code-based HIV infection reporting system was reported as the most acceptable system by the majority of the respondents in each of the 4 counties (52% of Fresno respondents, 47% Los Angeles, 58% Riverside, and 74% Riverside).
Postimplementation independent multivariate predictors of a preference for a coded HIV infection reporting system included 2 of the predictors for the preimplementation period; MSM (adjusted OR = 3.4, 95% CI: 1.2-9.5, P = 0.022) and having just taken an anonymous HIV test (adjusted OR = 2.6, 95% CI: 1.2-5.5, P = 0.016) and a new predictor of not being of Hispanic origins (adjusted OR = 2.7, 95% CI: 1.2-5.7, P = 0.012). Female sex was no longer significantly associated with a preference for a coded HIV infection reporting system in the postimplementation period. Again, clustering by county or testing site had negligible effects on identified risk factors and standard errors.
The findings from this study document strong support among HIV test takers in California for a non-name coded reporting system. This is important because the goal of the state in adopting this approach was to increase acceptability to potential HIV test takers and minimize any deterrent effects that might discourage awareness of HIV serostatus. It is clear that a coded reporting system is more difficult to develop and implement in terms of requirements from laboratories, health care providers, and local health departments than a name-based system. Thus, it is important that this added effort is worth it in terms of acceptability to those who seek publicly supported testing. The magnitude of the preference for a non-name code-preferred by >5 to 1 over name reporting preimplementation and 3 to 1 postimplementation-provides a justification for the added effort.
In addition to documenting acceptability to test takers, this study also estimates the potential deterrent effects of a name-based reporting system. With regard to confidential HIV testing, which is the basis of the HIV surveillance system, under a no-reporting scenario 52% of respondents report being very likely to test in the next 12 months. Under a name reporting system, this drops by more than one-third to only 32% being very likely to have a confidential HIV test, suggesting a significant potential deterrent effect. Under a coded system, the likelihood of testing is nearly identical to that under a no-reporting system. This finding is consistent with prior studies.18
This study also estimates the probability of a shift away from confidential testing, which is the basis of the HIV surveillance system, to anonymous testing under the scenario of name-based reporting. Under a name reporting system, 32% of individuals report being very likely to have a confidential test whereas 57% reported being very likely to have an anonymous HIV test. This reflects a significant potential shift from confidential to anonymous testing.
The goal of adopting HIV case surveillance is to improve our understanding of size and trends of HIV infection in California's HIV epidemic. Data from this study suggest that the likelihood of accomplishing these goals is best achieved under a non-name coded system. Specifically, we found no difference in the likelihood of confidential testing between a non-name coded system and no reporting at all. Conversely, we found an increased likelihood of anonymous testing and decreased likelihood of confidential testing if the state adopted a name-based system. Both of these effects would result in capturing fewer HIV cases in the surveillance system.
Further, these findings suggest that those lost to the surveillance system may introduce a systematic bias to the surveillance system. In particular we found that in 2 independent samples, MSM and individuals who currently opt for anonymous testing may be disproportionately less likely to be included in a name-based system. What we know about the current epidemiology of the HIV epidemic in California suggests that these are key groups that we would want to capture in any surveillance system. For example, of the cumulative number of AIDS cases reported through December 2001 with reported risk factors, 82% were among MSM (including MSM who use injection drugs) in California, compared with 58% in the United States as a whole.
In the year following implementation of California's non-name code HIV infection reporting system, HIV test takers' preferences among the 3 potential HIV reporting options remained remarkably similar to those prior to its implementation. Unfortunately, knowledge of the new HIV infection reporting system in place in California also remained quite low. This low level of knowledge of HIV reporting practices points to a failure of the HIV test informed-consent process currently in place in California and in other parts of the United States. Integral to the concept of informed consent is the communication of information about what may happen as a result of participation in the HIV testing and counseling procedure. Certainly, reporting of personal risk information and an HIV-positive test result to a local health department constitutes knowledge that a potential test taker should have in order to make a fully informed decision. This is regardless of whether a non-name or name-based HIV infection reporting system is in place. We urge California's publicly funded HIV test sites and local health departments to examine their current HIV informed consent guidelines and procedures and modify them to include information on HIV reporting practices.
The results of this study are congruent with several studies from the literature. These studies have found a potential deterrent effect among MSM7,8,19-21 and among Latinos and African Americans.22 The deterrent effect of name reporting is the greatest in groups at highest risk of HIV infection, particularly for high-risk repeat testers.23 Potentially at odds with the finding from our study in California that a significant potential deterrent effect is possible with a switch to a name-based HIV infection reporting systems is a report from studies funded by the CDC. In this study conducted in 6 states, knowledge of state policies was low among all individuals and the deterrent effect was minimal on heterosexuals at sexually transmitted disease clinics, moderate among injection drug users, and highest among MSM.24 Overall, the level of testing went up in 3 states, was level in 2 states, and declined in 1 state. The mixed impact of the switch to name-based HIV infection reporting of this CDC study (a deterrent in some states and for some groups, but not in others) is difficult to interpret, as the impact of temporal trends is not easily controlled for in statistical analyses.
This study has a number of limitations. To have data from a real-world setting, ie, current HIV test takers, we sampled at publicly funded HIV testing sites. One limitation introduced by this choice is that our findings may not be generalizable to all HIV testing situations such as persons testing in private medical settings or hospitals. However, publicly funded HIV testing sites provide a significant proportion of all of California's HIV tests and these sites have been an important part of the HIV reporting policy debate. In general, publicly funded testing sites provide free HIV testing in targeted communities at higher risk of HIV infection. Purposeful sampling of these sites has enabled us to estimate the impact of HIV reporting policies on this important segment of HIV testing in California.
This study examined HIV test takers from only 4 California counties. It may be that results from these counties are not completely generalizable to the remaining California counties; however, the lack of differences in the major findings of the study across the 4 counties suggests that these major results will hold throughout California. Because participating test sites were more comfortable with local staff making the initial contact with potential participants, we are not able to report a specific interview refusal rate or estimate the extent to which selection bias may have modified the reported results.
This study also has a number of strengths over previous research into the issues surrounding HIV infection reporting. By sampling individuals exiting an actual HIV test, we were able to target individuals who are much more invested in HIV testing and reporting policies than prior studies that sampled persons who were recruited from convenience samples in community settings, a significant number of whom have never been tested for HIV. Additionally, this study was designed to estimate the potential shift from confidential testing to anonymous testing under the different reporting options. While it is possible that name-based HIV infection reporting might not deter individuals from HIV testing, a shift to anonymous testing would significantly undermine HIV surveillance efforts. Unlike prior research, this study undertook specific procedures to present the complex policy options both orally and in written form, making clear distinctions between the policy options and concepts under study, rather than assuming prior and accurate knowledge of how the reporting systems would work. This study was specially designed to address the acceptability of the 3 potential HIV reporting policy options and to estimate the potential deterrent effect on HIV testing associated with each option. Previous published literature into these questions often involved less direct inference from studies designed to answer other research questions or failed to address the potential shift to anonymous testing.
Views on HIV name reporting are often strongly held on both sides of the debate. These views have been reflected in California ballot initiatives, where HIV name reporting was defeated in 1986 by 71% of the voters and again in 1998 by 66% of the voters. Legislative initiatives for name reporting were also defeated in the California legislature each year between 1988 and 1999 before the legislature agreed on a coded system. This political context is quite different from most other states that have adopted name-based reporting. The findings on acceptability to test takers respond to important questions for policy makers who are in the process of implementing and evaluating the California HIV system.
The authors thank Juan Ruiz, MD, DrPH, Director, Surveillance Branch, and Raphael Hess, MA, who served as project officer. We also thank other members of the UCSF research team-Maricarmen Arjona, Steve Riffe, T. Anne Richards, Kim Koester, Karen Vernon, Tim Lane, Jay Newberry, Sandra Schwarcz, Maree Kay Parisi, and Heather Leatherwood-for their input and assistance. We also thank the HIV test sites that assisted us with recruitment and interviews and the departments of public health in each of the participating counties. Most of all, the authors thank the participants for their willingness to provide their views on this important issue.
1. Gostin LO, Ward JW, Baker AC. National HIV case reporting for the United States: a defining moment in the history of the epidemic. N Engl J Med
2. Centers for Disease Control. Guidelines for National Human Immunodeficiency Virus Case Surveillance, Including Monitoring for Human Immunodeficiency Virus Infection and Acquired Immunodeficiency Syndrome
. Atlanta, GA: Centers for Disease Control and Prevention; 1999.
3. Woods WJ, Dilley JW, Lihatsh T, et al. Name-based reporting of HIV-positive test results as a deterrent to testing. Am J Public Health
4. Woods WJ, Binson D, Morin S, et al. HIV testing after implementation of name-based reporting. JAMA
. 1999;281:1378. Author reply 1379-1380.
5. Nakashima AK, Horsley R, Frey RL, et al. Effect of HIV reporting by name on use of HIV testing in publicly funded counseling and testing programs. JAMA
6. Lansky A, Lehman JS, Gatwood J, et al. Changes in HIV testing after implementation of name-based HIV case surveillance
in New Mexico. Am J Public Health
7. Kegeles SM, Catania JA, Coates TJ, et al. Many people who seek anonymous HIV-antibody testing would avoid it under other circumstances. AIDS
8. Kegeles SM, Coates TJ, Lo B, et al. Mandatory reporting of HIV testing would deter men from being tested. JAMA
9. Colfax GN, Bindman AB. Health benefits and risks of reporting HIV-infected individuals by name. Am J Public Health
10. Solomon L, Benjamin G, Wasserman M. HIV testing after implementation of name-based reporting. JAMA
. 1999;281:1378-1379. Author reply 1379-1380.
11. Burris S. Surveillance
, social risk, and symbolism: framing the analysis for research and policy
. J Acquir Immune Defic Syndr
. 2000;25(Suppl 2):S120-S127.
12. Bindman AB, Osmond D, Hecht FM, et al. Multistate evaluation of anonymous HIV testing and access to medical care. JAMA
13. Centers for Disease Control. Public Health Service guidelines for the management of health-care worker exposures to HIV and recommendations for postexposure prophylaxis. Centers for Disease Control and Prevention. MMWR Morb Mortal Wkly Rep
14. Office of AIDS. Soundex generation. California Department of Health Services. Available at: http://www.dhs.cahwnet.gov/org/ps/ooa/HIVReporting/SOUNDEX/SOUNDEX.htm
. Accessed May 8, 2003.
15. Los Angeles Department of Health Services. HIV Reporting with non-name Code. Available at: http://www.lapublichealth.org/hiv/hivreporting.htm
. Accessed May 8, 2003.
16. Schwarcz S, Hsu L, Chu PL, et al. Evaluation of a non-named based HIV reporting system in San Francisco. J Acquir Immune Defic Syndr
17. Fleiss JL. Statistical Methods for Rates and Proportions
. 2nd ed. New York: Wiley; 1981.
18. Hecht FM, Chesney MA, Lehman JS, et al. Does HIV reporting by name deter testing? MESH Study Group. AIDS
19. Hecht F. HIV testing among populations at risk for HIV infection
: nine states, November 1995-December 1996. MMWR Morb Mortal Wkly Rep
20. Johnson WD. The impact of mandatory reporting on HIV seropositive persons in South Carolina. Paper presented at: Fourth International Conference on AIDS; 1988; Stockholm, Sweden.
21. Hirano D. Anonymous HIV testing: the impact of availability on demand in Arizona. Am J Public Health
22. Reed G. The impact of mandatory name reporting on HIV testing and treatment. Poster presentation at: XI International Conference on AIDS; July 7-12, 1996; Vancouver, BC, Canada.
23. Fordyce EJ, Sambula S, Stoneburner R. Mandatory reporting of human immunodeficiency virus testing would deter blacks and Hispanics from being tested. JAMA
. 1989;262:349. Letter; see comments.
24. Osmond DH, Bindman AB, Vranizan K, et al. Name-based surveillance
and public health interventions for persons with HIV infection
. Multistate Evaluation of Surveillance
for HIV Study Group. Ann Intern Med