Sexually Transmitted Diseases:
Anonymous HIV Testing With Participant-Controlled Access to Results Using Handheld Computers: A New Model of HIV Testing Used in a Household Survey in Rural South Africa
Welz, Tanya MRCP, MSc, DTMH*†; Herbst, Kobus MBChB, MSc, FFCH*
From the *Africa Centre for Health and Population Studies, University of kwaZulu-Natal, Mtubatuba, South Africa; and †Department of HIV and Genitourinary Medicine, King’s College London School of Medicine at Guy’s, King’s College and St Thomas’ Hospitals, London, UK
The authors acknowledge Vikram Kumar of Dimagi Inc., Boston, USA, who implemented the LATPCR concept and developed the HIV Confidante and the dedicated field and laboratory staff without whom this work would not have been possible. The software developed for this study will be made available cost free on request to interested not-for-profit HIV research organizations.
This work was supported by the Wellcome Trust, UK, through grants to the Africa Centre Demographic Information System, the Africa Centre for Health and Population Studies and the Africa Centre HIV Surveillance programme (Grants 050510, 065377, 65377, 50534).
Correspondence: Dr. Tanya Welz, MRCP, MSc, DTMH, Africa Centre for Health and Population Studies, P.O. Box 198, Mtubatuba 3935, South Africa. E-mail: email@example.com.
Received for publication July 17, 2007, and accepted October 24, 2007.
ANONYMOUS HUMAN IMMUNODEFICIENCY VIRUS (HIV) surveys aim to provide unbiased estimates of HIV prevalence and are essential for monitoring the burden of HIV in a population. However, anonymous HIV surveys allow only limited linkage with risk factor data and do not allow participants to be informed of their HIV results.1 This is an important disadvantage in the face of evidence of individual and public health benefits of knowing one’s HIV status and the increasing availability of antiretroviral therapy worldwide.2–4 In accordance with ethical guidelines,5–7 anonymous HIV surveys usually refer participants to voluntary counseling and testing (VCT) services where these exist or else set up parallel VCT systems.8,9 This means that participants who wish to know their HIV status need to be bled and tested twice. Acceptance rates for such repeat VCT are not known.
The World Health Organization (WHO) has recently called for initiatives to increase access to innovative, ethical, and practical models of HIV testing and counseling.10 We devised a novel approach to anonymous HIV testing which used handheld computer and encryption technology to allow participants in an HIV sero-survey to confidentially access their results. We describe this system of ‘Linked Anonymous Testing with Participant-controlled Results’ (LATPCR) and its operational use in a very large household HIV survey in rural South Africa.
The Africa Centre Demographic Information System in the rural district of Umkhanyakude, kwaZulu-Natal, South Africa, has been collecting demographic data from approximately 12,000 households since 2000.11 Fieldworkers visit every household in a circumscribed study area of 435 km2 twice yearly to update sociodemographic and HIV risk factor data on the total resident population of 86,000 Zulu-speaking people (in June 2003). Approximately 60% of the population live in scattered homesteads in deep rural areas with a population density below 400 people per square kilometer.11 This is an area of very high HIV prevalence with rates of 38% reported amongst antenatal clinic attendees.12 Antiretroviral treatment (ART) was not available in the public sector in South Africa until October 2004. By December 2004, 56 patients were receiving ART in district health facilities (personal communication, Graham Cooke, Africa Centre for Health and Population Studies, 15 May 2007).
The HIV survey took place between June 2003 and December 2004 and has been described in detail elsewhere.13 Ethical approval was obtained from the research ethics committee of the Nelson Mandela School of Medicine, Durban. In addition to the routine collection of demographic data, fieldworkers visiting households also approached every adult male (15–54 years) and female (15–49 years) resident in the study area for HIV testing. Written, informed consent was obtained for HIV testing. Eligible participants received detailed verbal and written pretest information. This included information on what an HIV test is, the implications of a positive result, benefits and disadvantages of knowing one’s HIV status, and strategies to reduce risk. It was explained that participants who wanted to could confidentially obtain their HIV results by presenting a unique code at any of the 19 community-based counseling centers. Counseling centers were offices set up for this purpose throughout the study area in community venues such as shops. Lay counselors staffing these centers were also trained to provide health education, weight, and blood pressure checks.14
HIV tests were done on finger-prick dried blood spot samples in accordance with UNAIDS/WHO guidelines.15 HIV serology was determined by antibody testing with a broad based HIV-1/HIV-2 enzyme-linked immunosorbent assay (Vironostika, Organon Teknika, Boxtel, Netherlands) followed by a confirmatory enzyme-linked immunosorbent assay (GAC-ELISA, Abbott, Abbott Park, IL). Samples were collected from the field every 1 to 2 days and transported to the Wellcome Trust Africa Centre laboratory at the Nelson Mandela School of Medicine in Durban, 250 km away.
Each individual in the study area was assigned an alphanumeric identification number (ID), linked to his name, sex, and date of birth (Fig. 1). A separate, highly protected data table linked each individual’s ID to a pseudo-ID. Each individual’s pseudo-ID replaced his ID in the study database and could not be linked back to any identifying data except by the senior data manager. This process of ‘pseudonymization’ or ‘reversible anonymization’ always maps a given ID with the same pseudo-ID, thereby protecting participants’ identities whilst linking data from the same person irrespective of the collection time and place.16 The need to retain some linkage with individual identifiers is unavoidable in longitudinal studies where it is necessary to verify a participant’s identity at each contact.
We used public-private key encryption to protect confidential HIV results (Fig. 2). An encryption algorithm uses a set of numbers and/or characters which is openly available (a public key) to transform readable text into unreadable ‘cipher text’. A corresponding private key, which is available only to the intended recipient of the information, is needed to decrypt the message.17,18 Public-private key encryption algorithms are publicly available on the Internet and the information technology required is relatively simple. Unique, linked pairs of barcoded alphanumeric public and private keys were randomly generated and printed by an external organization. Only this so-called ‘trusted third party’16 could access the list of public-private key pairs. Three copies of the public key and one of the private key were printed as adhesive labels on a counseling slip. At HIV testing, the test sample was labeled with one copy of the public key which became the specimen number (Fig. 3). The second copy of the public key was linked to the participant’s ID on the specimen tracking form. The counseling slip containing the remaining public key and the only copy of the private key was then given to the participant. The private key was obscured by a film that had to be scratched off with a fingernail or coin.
After testing, a temporary laboratory data file containing ID’s, specimen numbers (public keys), and HIV results was imported into the study database and linked to participants’ pseudo-IDs (Fig. 4). At the same time, each participant’s public key (specimen number) encrypted his ID, name, and HIV result into a single data packet identified by the specimen number. All encrypted data packets were then downloaded to the HIV counselors’ handheld computers, which were returned to the research office at least once a week.
Survey participants who wished to know their HIV results visited a counseling center with the counseling slip containing their public and private keys. The participant or counselor entered the public key into the handheld computer (Fig. 5). This retrieved his encrypted data and produced a message ‘HIV result available. Enter private key’. Entering the private key then decrypted the ID, name, and date of birth, which then appeared on the screen. The counselor then verified the participant’s identity through his date of birth or identity documents and provided full preresult HIV counseling. If the participant chose to proceed, the HIV result was revealed on the next screen for the client and counselor to see. The counselor then provided appropriate postresult counseling and took blood for a confirmatory HIV test on all positives. HIV results could never be accessed without a private key and anyone who lost their counseling slip had to have a repeat HIV test.
Eleven thousand five hundred and fifty-one out of 19,867 eligible residents contacted (58.1%), agreed to HIV testing. In only 9 cases (0.08%) was it impossible to link an HIV result with an individual’s epidemiologic data. All 9 cases resulted from errors in manual data capture of specimen numbers during a short period when a barcode scanner was out of order. The turnaround time between testing and availability of results at counseling centers was 2 to 3 weeks. Despite the operational success of the system, the proportion of survey participants returning for their HIV results was disappointing. Only 325 of 6859 (4.7%) women and 221 of 4692 (4.7%) men tested returned for their results.
A further 146 individuals who were not eligible for the survey requested and received VCT at counseling centers using the same system. These individuals were outside the age range for the HIV survey or were not resident in the study area.
The HIV prevalence was higher among survey participants who returned for their results than among those who did not: 29.9% of women and 22.6% of men who received their results were HIV infected compared to 26.7% of women (P = 0.01) and 13.1% of men (P <0.001) who did not. This may be partly explained by the fact that participants who returned for their results were older than those who did not: 32.6 versus 28.7 years for women (P <0.0001) and 33.3 versus 26.1 years for men (P <0.0001).
Although the proportion of participants who received their results was low, the 546 individuals who learned their HIV status during the survey may not have done so if we had used conventional linked or unlinked anonymous HIV testing with referral for parallel VCT.
HIV in South Africa is highly stigmatized,19–21 and AIDS is the leading cause of death in this population resulting in 48% of adult deaths in 2000.22 Bearing in mind the extremely high HIV prevalence and the unavailability of ART at the time of the survey, the low rate of return for results is perhaps not surprising. A substudy to investigate this poor uptake of results found the main reasons to be the unavailability of ART, a perceived lack of benefit in knowing one’s HIV status, and the stigma of visiting a counseling center.23 Another possible explanation is the availability of VCT at government clinics since the start of a nevirapine-based program to prevent mother-to-child transmission of HIV in 2002. We intend to evaluate the effect of the increasing availability of ART and of home visits by HIV counselors on the uptake of results in future surveys.
Although not formally evaluated, we have no evidence that LATPCR increased survey refusal rates. The refusal rate (42%) was virtually identical to that for KwaZulu-Natal in a 2005 national linked, anonymous HIV survey (41%).24 All the national South African HIV surveys to date have had very low participation rates by international standards.24–28
Pretest HIV counseling in surveys is problematic because it requires privacy as well as substantial staff training, experience, and time7,29 – at least some of which will be spent counseling individuals who have no intention of accessing their results. Pretest counseling may also reduce survey participation as it ‘makes individual’s risks visible’.30 Conversely, anonymous HIV testing attracts more high-risk individuals.31–34 LATPCR combined the advantages of anonymous and confidential HIV testing without the need for pretest counseling in the field. It further allowed pre and postresult HIV counseling to be done by the same counselor as recommended in best practice guidelines.7 This is usually impractical in population-based HIV surveys.
Other approaches have been used to allow survey participants to access their HIV results. In ‘LAT with provision of test results’ participants receive a code which links them to their results. This is similar to the ‘nonname code’ model of anonymous VCT commonly used in testing facilities in the United States. However, this requires pretest counseling.1 Another model, ‘Voluntary Testing and Counseling’, which has been used in household HIV surveys in Uganda, also offers pre- and postresult counseling. However, survey participants have to indicate at the time of testing whether they want their results. Only if requested are named results printed and dispatched to the local counseling office.29,35,36
The capital outlay for the system was approximately $7500 for software development and $4000 for hardware (25 handheld computers and 2 barcode scanners at approximately $150 each). Equipment fared well over time with only 1 computer stolen early in the study. Routine management of samples and results was successfully performed by trained local people with high school qualifications only.
The handheld computers eliminated the need for printing and transporting results. Almost 12,000 HIV results were available at all 19 counseling centers in the study area allowing participants to receive results at any counseling venue of their choice. In settings with cellular telephone connections and Internet access, results’ turnaround time could be reduced by uploading encrypted HIV results onto handheld computers through cellular telephones.
Each handheld computer is password protected. If a handheld computer were stolen, it would be virtually impossible to decrypt even one individual’s ‘data packet’. The anonymous electronic log of all HIV results disclosed can be used to audit alleged breaches in confidentiality as well as for research into HIV testing behaviors. Examples would include identifying characteristics or events (e.g., the death of a household member in the 3 years preceding the survey) associated with participants’ accessing their HIV results or investigating how distance to counseling centers affects the uptake of results in rural areas. To date, we have not received any reports of breaches in confidentiality or unauthorized access to HIV results.
From a research point of view, LATPCR was cost-effective by optimizing the use of fieldwork, counseling and laboratory staff. It removed the need for separate blood draws and laboratory tests for parallel VCT as well as for printing and dissemination of thousands of results. VCT has been shown in many settings to be cost-effective.37–40 From a public health point of view, therefore, even the relatively small number of survey participants who received their HIV results added greatly to the overall cost-effectiveness of the survey.
LATPCR has potential for use in any research or program requiring the anonymous linkage of confidential data (which participants may choose to know) with clinical or epidemiologic information. Besides HIV, examples would include screening for sexually transmitted infections or genetic abnormalities.41 The system also has potential in VCT outreach where it would allow the paper-free, confidential dissemination of HIV results to hard-to-reach populations in nonfacility settings. In this large population-based HIV survey, LATPCR allowed us to successfully address a multitude of ethical, logistic, and scientific challenges.
2. Higgins DL, Galavotti C, O’Reilly KR, et al. Evidence for the effects of HIV antibody counseling and testing on risk behaviors. JAMA 1991; 266:2419–2429.
3. Weinhardt LS, Carey MP, Johnson BT, et al. Effects of HIV counseling and testing on sexual risk behavior: A meta-analytic review of published research, 1985–1997. Am J Public Health 1999; 89:1397–1405.
4. The Voluntary HIV-1 Counseling and Testing Efficacy Study Group. Efficacy of voluntary HIV-1 counselling and testing in individuals and couples in Kenya, Tanzania, and Trinidad: A randomised trial. Lancet 2000; 356:103–112.
7. Technical Expert Panel Review of CDC HIV Counseling, Testing, and Referral Guidelines. Revised guidelines for HIV counseling, testing, and referral. MMWR Recomm Rep 2001; 19:1–57. Available at: http://www.cdc.gov/mmwr/preview/mmwrhtml/rr5019a1.htm
. Accessed July 13, 2007.
8. Calleja JM, Marum LH, Carcamo CP, et al. Lessons learned in the conduct, validation, and interpretation of national population based HIV surveys.AIDS 2005; 19(suppl 2):S9–S17.
9. Marum L, Ngare C, Odoyo J, et al. Learning HIV status in population based prevalence surveys: Lessons from mobile VCT in the Kenya Demographic and Health Survey 2003. Paper presented at: The 13th International Conference on AIDS and STDs in Africa; September 26, 2003; Nairobi, Kenya.
11. Hosegood V, Timaeus I. Chapter 4. In: van der Walle E, ed. Household composition and dynamics in KwaZulu Natal, South Africa: Mirroring social reality in longitudinal data collection. In African Households: An Exploration of Census Data. New York: M.E. Sharpe Inc. ISBN 0-7656-1619-X; 2005.
12. Rice BD, Batzing-Feigenbaum J, Hosegood V, et al. Population and antenatal-based HIV prevalence estimates in a high contracepting female population in rural South Africa. BMC Public Health 2007;7:160.
13. Welz T, Hosegood V, Jaffar S, et al. Continued very high prevalence of HIV infection in rural KwaZulu-Natal, South Africa: A population-based longitudinal study. AIDS 2007; 21:1467–1472.
14. Welz T, Thabethe N, Radebe S, et al. Broadening access to VCT: Community-based counseling centers in a large HIV surveillance programme in rural KwaZulu-Natal, South Africa. Abstract No. TuPeD5150. Paper presented at: The XV International AIDS Conference; 2004; Bangkok, Thailand.
16. Claerhout B, DeMoor GJ. Privacy protection for clinical and genomic data. The use of privacy-enhancing techniques in medicine. Int J Med Inform 2005; 74:257–265.
17. Rivest RL, Shamir A, Adleman L. A method for obtaining digital signatures and public-key cryptosystems. Commun ACM 1978; 21:120–126.
18. Mao W. Modern Cryptography: Theory and Practice ISBN 0-13–066943-1. Upper Saddle River, NJ: Prentice Hall, 2003.
19. Simbayi LC, Kalichman S, Strebel A, et al. Internalized stigma, discrimination, and depression among men and women living with HIV/AIDS in Cape Town, South Africa. Soc Sci Med 2007; 64:1823–1831.
20. Kohi TW, Makoae L, Chirwa M, et al. HIV and AIDS stigma violates human rights in five African countries. Nurs Ethics 2006; 13:404–415.
21. Baleta A. Widespread horror over killing of AIDS activist in South Africa. Lancet 1999; 353:130.
22. Hosegood V, Vanneste AM, Timaeus IM. Levels and causes of adult mortality in rural South Africa: The impact of AIDS. AIDS 2004; 18:663–671.
23. Ndinda C, Baetzing-Feigenbaum J. An analysis of refusal to test in a community-based HIV Sero-surveillance Program in Rural KwaZulu- Natal. Paper presented at: The Demographic Association of Southern Africa Joint Population Conference; October 4–8, 2004; Durban, South Africa.
25. Shisana O, Stoker D, Simbayi LC, et al. South African national household survey of HIV/AIDS prevalence, behavioural risks and mass media impact–detailed methodology and response rate results. S Afr Med J 2004; 94:283–288.
26. Pettifor AE, Rees HV, Kleinschmidt I, et al. Young people’s sexual health in South Africa: HIV prevalence and sexual behaviors from a nationally representative household survey. AIDS 2005; 19:1525–1534.
27. Garcia-Calleja JM, Gouws E, Ghys PD. National population based HIV prevalence surveys in sub-Saharan Africa: Results and implications for HIV and AIDS estimates. Sex Transm Infect 2006;82(suppl 3):iii64–iii70.
28. Boerma JT, Ghys PD, Walker N. Estimates of HIV-1 prevalence from national population-based surveys as a new gold standard. Lancet 2003; 362:1929–1931.
29. Matovu JK, Kigozi G, Nalugoda F, et al. The Rakai Project counselling programme experience. Trop Med Int Health 2002; 7:1064–1067.
30. Kinnell AMK. So why are you here? Assessing risk in HIV prevention and test decision counselling. Sociol Health Illn 2001; 23:447–477.
31. Fehrs LJ, Fleming D, Foster LR, et al. Trial of anonymous versus confidential human immunodeficiency virus testing. Lancet 1988; 2:379–382.
32. Kegeles SM, Catania JA, Coates TJ, et al. Many people who seek anonymous HIV-antibody testing would avoid it under other circumstances. AIDS 1990; 4:585–588.
33. Levi I, Modan B, Blumstein T, et al. Characteristics of clients attending confidential versus anonymous testing clinics for human immunodeficiency virus. Isr Med Assoc J 2001; 3:184–187.
34. Sasse A, Vincent A, Galand M, et al. High HIV prevalence among patients choosing anonymous and free testing in Belgium, 1990–2002. Int J STD AIDS 2006; 17:817–820.
35. Nyblade LC, Menken J, Wawer MJ, et al. Population-based HIV testing and counseling in rural Uganda: Participation and risk characteristics. J Acquir Immun Defic Syndr 2001; 28:463–470.
36. Wolff B, Nyanzi B, Katongole G, et al. Evaluation of a home-based voluntary counselling and testing intervention in rural Uganda. Health Policy Plan 2005; 20:109–116.
37. Thielman NM, Chu HY, Ostermann J, et al. Cost-effectiveness of free HIV voluntary counseling and testing through a community-based AIDS service organization in Northern Tanzania. Am J Public Health 2006; 96:114–119.
38. Varghese B, Peterman TA. Cost-effectiveness of HIV counseling and testing in US prisons. J Urban Health 2001; 78:304–312.
39. Sweat M, Gregorich S, Sangiwa G, et al. Cost-effectiveness of voluntary HIV-1 counselling and testing in reducing sexual transmission of HIV-1 in Kenya and Tanzania. Lancet 2000; 356:113–121.
40. Liechty CA. The evolving role of HIV counseling and testing in resource-limited settings: HIV prevention and linkage to expanding HIV care access. Curr Infect Dis Rep 2005; 7:154–158.
41. Mehlman MJ, Kodish ED, Whitehouse P, et al. The need for anonymous genetic counseling and testing. Am J Hum Genet 1996; 58:393–397.
This article has been cited 2 time(s).
Effect of HIV infection on pregnancy-related mortality in sub-Saharan Africa: secondary analyses of pooled community-based data from the network for Analysing Longitudinal Population-based HIV/AIDS data on Africa (ALPHA)
AIDS Education and Prevention
Addressing Unmet Need for Hiv Testing in Emergency Care Settings: A Role for Computer-Facilitated Rapid Hiv Testing?
AIDS Education and Prevention, 25(4):
© Copyright 2008 American Sexually Transmitted Diseases Association