University of California, San Francisco Logo

University of California, San Francisco | About UCSF | Search UCSF | UCSF Medical Center

Home > Global Health Literature Digest > A Comparison of HIV
A comparison of HIV detection rates using routine opt-out provider-initiated HIV testing and counseling versus a standard of care approach in a rural African setting.
Global Health Sciences Literature Digest
Published May 12, 2011
Journal Article

Silvestri DM, Modjarrad K, Blevins ML, Halale E, Vermund SH, McKinzie JP. A comparison of HIV detection rates using routine opt-out provider-initiated HIV testing and counseling versus a standard of care approach in a rural African setting . J Acquir Immune Defic Syndr. 2011 Jan 1;56(1):e9-32.


To determine whether the HIV prevalence of those being tested by Provider Initiated Testing and Counseling (PITC) is different from those whose HIV tests are based on standard counseling and testing; and to assess client attitudes towards PITC.


Mobile clinic providing general medical services in four rural sites in southern Zambia.


Clinical trial. Convenience sampling of clients for questionnaire administration.


Adult patients aged 15-49 years seeking general medical care.

Main Outcome Measures

HIV serostatus, history of prior HIV test, and attitude toward PITC associated with HIV referral method; multivariable logistic regression modeling was used to assess the association of referral method and HIV status, adjusting for age, sex, and clinic site; multivariable analysis was also used to evaluate the association of age, sex, HIV stigma, HIV knowledge and education with history of HIV testing.


Of 22 health care workers (HCW), 16 were assigned to identify patients who voluntarily requested HIV testing (VCT), exhibited signs or symptoms suggestive of HIV/AIDS, or reported HIV risk factors (all considered standard counseling and testing); the remaining six providers (who were all physicians) referred every medical patient they saw for HIV counseling and testing regardless of presenting complaint (PITC). Patients in both arms underwent counseling and testing (CT) by the same staff. A separate group of clinic assistants interviewed a convenience sample of PITC and non-PITC patients, using a questionnaire that included items evaluating client and perceived partner attitude toward PITC.


A total of 1,221 adults presenting to the mobile clinic were tested, of whom 681 (56%) were in the PITC arm; 317 (26%) of the total sample completed the questionnaire interview. In the PITC group, 222 (33%) would have been referred for testing in the absence of PITC because of client request for VCT (142, 64%), clinical suspicion of HIV infection (63, 28%), or both (17, 8%). The standard non-PITC group included 540 patients of whom 454 (84%) requested HIV testing, 59 (11%) were referred because of clinical suspicion of HIV infection, and 27 (5%) because of both. In the PITC group, 114 clients per practitioner underwent HIV testing; in the standard CT arm, 34 clients per practitioner were tested. The client groups were similar in baseline characteristics, except that women comprised 60% of the standard group, and 81% of the PITC group (p<0.01); there were also some significant differences in client distribution by site between arms. Approximately 30% of male and female clients were in polygamous marriages; 36% had little or no knowledge of HIV testing; only 44% did not express stigma towards people living with HIV (PLHIV); 96% were favorable to PITC, and 86% perceived their partners as favorable to PITC. Overall, 53% had had a previous HIV test. Being a woman (OR 2.4, CI 1.1-4.0) and having fewer stigmatizing attitudes towards PLHIV (OR 1.4, CI 1.1-1.9) were independently associated with prior HIV testing. The authors reported that marital status was significantly associated with prior HIV testing, but provided no data. Of the 1,221 adults who had HIV testing, 130 (10.6%) were HIV positive; HIV prevalence did not differ by testing arm; 11.2% were seropositive in the PITC group, and 10.0% were HIV positive in the standard CT group. Among all patients from both groups, the HIV prevalence was 10.4% among those who requested HIV testing, 16.3% among those with a clinical presentation prompting referral, and 9.3% among those voluntarily requesting HIV testing. In the PITC group, 12.7 HIV infected persons were identified per provider, compared to 3.4 HIV infected persons identified per provider. In multivariable analysis, there was no difference in HIV prevalence among clients in either arm (OR 1.01, CI 0.7-1.5). No data were given on the persons who opted-out in the PITC arm.


The HIV prevalence among clients who underwent PITC was similar to that among clients who underwent standard CT referral. This means that in Zambia, the HIV prevalence among the general medical patient population may equal that in traditionally-defined volunteering or high risk groups; 44 patients who otherwise would not have been referred for HIV testing were found to be HIV positive in the PITC group.

Quality Rating

This was a fair quality trial. Although the authors describe the trial as randomized, the patients were not randomized, but assigned to either arm, dependent on the provider they saw in the clinic. Only the providers in one arm (PITC) were described as being chosen randomly. However, there were 22 providers total, and 16 were assigned to the standard arm, leaving only 6 for the PITC arm, so it is not clear how randomization took place. Although staff administering questionnaires were described as ‘masked to HIV status’, clients were interviewed prior to HIV testing. Intention-to-treat analysis was used to evaluate some outcomes. It was not possible given the study design to blind the providers. As a result, it is possible that providers in either arm may have been more likely to inquire about risk factors for HIV or suspect HIV-related symptoms in identifying clients to be tested. This might be more likely in the PITC arm in which all providers were physicians. The authors did not indicate why only physicians were chosen for the PITC arm, whereas a range of providers were part of the standard arm. Clients who were interviewed were selected based on client and provider availability (convenience sampling), which could have introduced bias. The results and interpretation are somewhat marred by poor analytic procedures and potential biases that were not explored: 1) The number of clients who either did not consent for HIV testing or defaulted before taking the test was not tracked. Thus the authors did not address opting-out, which is part of the PITC procedure, as indicated in their title. 2) Covariates included in the multivariable model of patient factors with prior HIV testing were chosen a priori, rather than based on associations with outcome in bivariate analyses. 3) Sample size calculations and assumptions to determine number of persons required in each arm were not described. 4) Data were not shown for differences in prevalence among patients who had testing for different reasons (self referral, clinical suspicion) between the arms. The authors stated there was no difference, although the p-value (0.08) was close to significance. Data could have suggested possible provider bias in patient referral.

In Context

Previous studies of opt-out HIV testing have focused on specific populations, such as infants,(1) pediatric clients,(2) antenatal women,(3,4) and high risk patients.(5,6) WHO recommends PITC for all adolescents and adults in areas with generalized epidemics(7) or overall HIV prevalence >1%. However, PITC remains underutilized because of the expense of test kits and human resources.(8)

Programmatic Implications

This study indicates that in a setting with a generalized HIV epidemic and high HIV rates (approximately 10%), PITC may identify as many HIV-positive persons as standard VCT and referral based on clinical symptoms. These results need to be replicated elsewhere, and cost analyses performed to determine cost per HIV positive person identified. However, at least in this area of rural Zambia, PITC may have significant benefit in identifying persons who can then obtain HIV care and treatment.


  1. Rollins N, Mzolo S, Moodley T, Esterhuizen T, van Rooyen H. Universal HIV testing of infants at immunization clinics: an acceptable and feasible approach for early infant diagnosis in high HIV prevalence settings. AIDS. 2009 Sep 10;23(14):1851-7.
  2. Kankasa C, Carter RJ, Briggs N, et al. Routine offering of HIV testing to hospitalized pediatric patients at university teaching hospital, Lusaka, Zambia: acceptability and feasibility. J Acquir Immune Defic Syndr. 2009; 51:202-208.
  3. Perez F, Zvandaziva C, Engelsmann B, et al. Acceptability of routine HIV testing (“opt-out”) in antenatal services in two rural districts of Zimbabwe. J Acquir Immune Defic Syndr. 2006;41:514–520
  4. Creek TL, Ntumy R, Seipone K, et al. Successful introduction of routine opt-out HIV testing in antenatal care in Botswana. J Acquir Immune Defic Syndr. 2007;45:102–107.
  5. Leon N, Naidoo P, Mathews C, et al. The impact of provider-initiated (optout) HIV testing and counseling of patients with sexually transmitted infection in Cape Town, South Africa: a controlled trial. Implement Sci. 2010;5:8.
  6. Pope DS, Deluca AN, Kali P, Hausler H, Sheard C, Hoosain E, Chaudhary MA, Celentano DD, Chaisson RE. A cluster-randomized trial of provider-initiated (opt-out) HIV counseling and testing of tuberculosis patients in South Africa. J Acquir Immune Defic Syndr. 2008;48:190–195.
  7. World Health Organization (WHO). Guidance on Provider-Initiated HIV Testing and Counselling in Health Facilities. Geneva, Switzerland: World Health Organization; 2007. Accessed. December 13, 2009
  8. Monjok E, Smesny A, Mgbere O, et al. Routine HIV testing in health care settings: the deterrent factors to maximal implementation in sub-Saharan Africa. J Int Assoc Physicians AIDS Care. 2010;9:23–29