Bulletin of the World Health Organization

Acceptability of routine HIV counselling and testing, and HIV seroprevalence in Ugandan hospitals

Rhoda K Wanyenze a, Cecilia Nawavvu a, Alice S Namale b, Bernard Mayanja a, Rebecca Bunnell b, Betty Abang b, Gideon Amanyire a, Nelson K Sewankambo a, Moses R Kamya a


Both counselling and testing are key components of HIV prevention and care programmes.1 Through HIV counselling and testing (HCT), uninfected individuals can take steps to avoid becoming infected, while infected individuals can avoid transmission to sexual partners or children.25 Moreover, HCT is the first step in referral to care and support services.6,7 It has also become important for preventing mother-to-child transmission and increasing access to HIV/AIDS care, including antiretroviral therapy (ART).8,9 Even with limited availability of ART, early diagnosis of HIV and access to basic preventive care, including co-trimoxazole, can slow progression to AIDS.10,11 Providing HCT for family members of HIV-positive patients can identify other HIV-infected individuals in their households, facilitate partner disclosure and testing, identify HIV-discordant couples, and support care and medication adherence in HIV-infected individuals.12 Worldwide, it is estimated that over 90% of HIV-infected individuals are still unaware of their status.1315 In Uganda, 15% of the general population has received HCT, while more than 70% would like to be tested.16

Hospitals in high-prevalence settings are crowded with HIV/AIDS patients,1719 though the majority only learn about their infection late in the disease course, if ever.2024 A survey at Mulago hospital in Uganda found that half of medical inpatients with HIV-related diagnoses left hospital without HCT.25

It has been proposed that offering HCT routinely in health-care settings will increase access to care.6,7,13,17 Routine HCT is initiated by health-care providers and offers testing to all patients irrespective of their presenting illness. This approach differs from voluntary counselling and testing, which is client-initiated. The guidelines on provider-initiated HCT at health-care facilities, released by WHO in May 2007, recommend that testing should be part of standard medical care for all patients during widespread HIV epidemics.26 The United States Centers for Disease Control and Prevention (CDC) also revised HCT guidelines to recommend routine screening for HIV infection in health-care settings for individuals aged 13–64 years.27 When Botswana introduced routine opt-out HCT, HIV testing and interventions to prevent mother-to-child transmission both increased.28 The revised Ugandan HCT policy recommends routine HCT in health-care facilities, including antenatal clinics.29

We implemented routine HCT at Mulago and Mbarara hospitals in Uganda, assessed testing uptake, and estimated HIV seroprevalence among patients and their family members.



Mulago and Mbarara hospitals are the largest, tertiary, public, university teaching hospitals in Uganda. Mulago hospital has more than 600 000 patient-visits per year while Mbarara hospital serves over 100 000 patients annually. Although outpatient HIV clinics have been operational since 1987 in Mulago and since 1998 in Mbarara, the services offered were limited. Until recently, HCT was provided for patients only on request and when they were able to pay the fee of 5000 Ugandan shillings (US$ 3). A survey in July 2003 revealed that 67% of medical inpatients at Mulago had never been tested for HIV and only 20% had been tested during hospitalization, although 64% wanted to be tested.24

HCT programme

The Mulago and Mbarara routine HCT programme was established in November 2004. Coverage had expanded from four to 25 wards or clinics (16 in Mulago and nine in Mbarara) by February 2006; this represents 31% of the wards and clinics in Mulago and 50% of those in Mbarara. Units that were thought to have patient populations in which there was a high HIV prevalence were prioritized. The units offering routine HCT included four medical wards, four obstetrics and gynaecology wards, two sexually transmitted disease (STD) outpatient clinics, two dermatology outpatient clinics, five paediatric inpatient wards, two cancer inpatient wards, one tuberculosis ward and two medical emergency outpatient units. Two surgical wards and one dental clinic started offering routine HCT in October 2005 to assess the need among surgical patients.

In the selected wards and clinics, health-care providers offered HCT on an opt-out basis to all patients whose HIV status was unknown. Patients who reported a previous negative test result more than three months before the current hospital contact were also offered HCT. Individuals whose medical records showed that they were HIV-positive were not re-tested. In practice, HCT was offered alongside other clinical investigations and consent was provided for all tests. Patients were informed that they were free to decline testing and that those who opted-out would still receive the medical care required. Testing was provided to all patients unless they declined. Pretest information was provided to groups of patients, in outpatient waiting areas and general wards, and emphasized the benefits of HCT, such as the care available for HIV-infected individuals. Information on risk reduction was repeated during one-on-one post-test sessions.

Rapid HIV testing was used. The sequential rapid testing algorithm included the Determine HIV-1/2 assay (Abbott Laboratories, Illinois, United States of America) for screening, the HIV-1/2 STAT-PAK Dipstick assay (Chembio Diagnostic Systems Inc., New York, USA) for confirmatory testing, and the Uni-Gold test (Trinity Biotech, Wicklow, Ireland) as the tie-breaker. An HIV-negative result with the Determine assay was reported as negative. An HIV-positive result with the Determine assay was confirmed using the STAT-PAK assay and was reported as positive if both tests gave positive results. If the Determine and STAT-PAK assay results were discordant, the sample was subjected to a third test, the Uni-Gold test. The result was reported as positive if the Uni-Gold test result was positive and negative if both STAT-PAK assay and Uni-Gold test results were negative. Children aged less than 18 months who were HIV-positive had their HIV status confirmed by a deoxyribonucleic acid (DNA) polymerase chain reaction (PCR) test. Patients received pretest and post-test counselling on the same day. Test results were also provided to the medical team. Health-care providers and counsellors encouraged patients to disclose their HIV status to sexual partners and family members when they felt ready to do so. The counsellor assisted with disclosure if requested by the patient. Health-care providers who had undergone appropriate training also participated in disclosing results to patients and their family members.

Any of the patients’ spouses, children, parents or other household members who were present in hospital was also offered HCT. Couples testing was encouraged whenever the patients’ sexual partners were present. In paediatric wards, HIV testing was offered to children more than 3 months of age and their parents or carers. Some parents with documented HIV-positive results had children whose HIV status was unknown. In these cases, only the children were tested. Parents or guardians provided consent for children aged less than 12 years. Children aged 12 to 17 years had to agree to testing, in addition to receiving consent from their parents or guardians.

All patients, whether HIV-negative or HIV-positive, were counselled on risk reduction after the test. In addition, HIV-positive patients were also given information on the HIV/AIDS care available. Co-trimoxazole prophylaxis and tuberculosis screening were initiated on diagnosis. On discharge, HIV-positive patients were given referrals to HIV/AIDS clinics for follow-up care.

Data collection

Data collected from individuals who were offered HCT included age, sex, educational level, marital status, history of HIV testing, and last sexual partner’s HIV status (for those who reported sexual contact within 12 months). Reasons for declining current testing and hospital category (i.e. medical inpatient, medical outpatient, cancer inpatient, surgical inpatient, or paediatric, obstetrics and gynaecology, STD clinic or dermatology clinic patient) were also documented. Family members who were tested were recorded as a father or mother (for paediatric patients), sexual partner or other family member.


Data from all units that offered routine HCT between 1 November 2004 and 28 February 2006 were included. The characteristics of patients who were tested and those who declined testing were compared. In addition, HIV seroprevalence was analysed by age, sex and hospital ward or clinic. The overall HIV burden in each unit (which included patients admitted with a documented HIV-positive test result and HIV-positive patients identified by routine HCT during the current visit) was also calculated. The accessibility of routine HCT in the wards and clinics was quantified as the proportion of patients who were offered HCT relative to those eligible for the test. The uptake of routine HCT was quantified as the proportion of patients who accepted HCT relative to those who were offered it. In addition, bivariate and multivariate logistic regression analyses were carried out for patients aged 15 years or older to identify associations between an HIV-positive status and sociodemographic characteristics and the type of ward or clinic. All the variables included in the bivariate analysis were also included in the multivariate model. We excluded children aged less than 15 years from this part of the analysis because the model included variables related to sexual behaviour.


Overall, 51 642 patients were offered HCT between 1 November 2004 and 28 February 2006. Of these, 50 649 (uptake rate 98%) accepted. The uptake rate ranged from 94–99% between the hospital units, except that the STD clinic had a low rate of 84%. The accessibility of HIV testing was high in all wards and clinics, with 90–100% of eligible patients being offered HCT. Among the 993 patients who declined testing, 247 (25%) said they did not wish to be tested, 207 (21%) said they would get tested after their current illness improved, 198 (20%) said they had previously tested HIV-positive, 65 (7%) said they had tested HIV-negative numerous times, 54 (6%) said that testing offered no benefits, 43 (4%) said they had given blood samples too many times, and 30 (5%) wanted to consult their spouses before testing. Other reasons for declining HCT included not being emotionally prepared, fear of an HIV-positive result, being confident they were not HIV-infected, and lack of time. There were no significant demographic differences between patients who declined and those who accepted testing.

During post-test counselling sessions, 2194 patients said they had previously tested HIV-positive but there was no documentation and they were not receiving care. Overall, 48 454 patients had never tested HIV-positive. The majority (61%) were female, 47% were married, and their median age was 27 years (range 3 months to 106 years). Of these 48 454, 12 107 (25%) were HIV-positive and 39 037 (81%) had never been tested for HIV. Among the 39 037 being tested for the first time, 11 108 (28%) tested HIV-positive. Of the 9102 who reported that they had previously tested HIV-negative, 850 (9%) were HIV-positive. Among 7156 children less than 15 years of age, 1024 (14%) were HIV-positive and 6961 (96%), which included all those with an HIV infection, had never been tested before. The prevalence of HIV among the patients tested varied between the wards and clinics and was highest among medical inpatients (Table 1). When patients who were previously known to have an HIV infection were included, the overall HIV burden in all inpatient wards was 38%. The greatest burdens were observed in cancer wards (54%), medical wards (46%) and emergency medical wards (45%); the lowest burdens were seen in paediatric wards (18%), obstetrics and gynaecology wards (16%), and surgical wards (13%).

Of the 10 439 family members who were offered HCT, 9720 (93%) accepted. Family members tested included 1510 spouses, 4122 mothers and 683 fathers of paediatric patients, 905 children belonging to patients, and 2500 other household members. Of these, 7100 (73%) had never been tested. The prevalence of HIV among family members was 20% (n = 1983). Among 1510 spouses of patients, 447 (30%) were HIV-positive, as were 862 (21%) mothers of paediatric patients and 127 (14%) children belonging to patients. Of 1213 couples (i.e. 2426 individuals), one member of which was a patient, 224 (19%) were HIV-discordant; in 69% of discordant couples, the patient was seropositive while the spouse tested negative.

The majority of patients (97%) and family members (99%) who were tested received their results in hospital. Of the 1514 patients who did not receive their results in hospital, 39% were discharged before receiving them, 29% died before receiving them, 22% had post-test counselling and results disclosure deferred because of the severity of their illness, and 10% refused to receive their results although they had initially agreed to testing.

Bivariate analysis showed that HIV infection in patients was associated with age (highest in the 31–40 age group), being married, divorced or widowed, having an HIV-positive partner, or having a spouse of unknown HIV status (Table 2). Patients who reported no sexual contact in the previous 12 months also had an increased risk of HIV infection. Independent predictors of HIV infection included age, hospital ward, partner’s HIV status and marital status. Widowed patients were more than four times as likely to have an HIV infection as those who had never been married, while separated patients were more than three times as likely and married patients were 1.7 times as likely. In addition, patients who said they did not know the HIV status of their sexual partners were three times as likely to have an HIV infection. Medical inpatients were more than three times as likely to have an HIV infection as surgical inpatients (Table 2).


This study demonstrates that providing free HCT was feasible and highly acceptable at two large public hospitals in Uganda. Routine HCT identified a large number of undiagnosed HIV infections and HIV-discordant partnerships among patients and their families. Similarly, other pilot programmes have shown that more HIV infections can be identified by routine HCT than by testing based on risk assessment and clinical diagnosis.22,23,28,30,31 In our programme, 81% of patients had never been tested before. Access to HIV testing for children was even more limited.

The high uptake of HCT may have been because it was free. In an assessment carried out at Mulago hospital before implementation of this programme, the most common reason for not undergoing testing during hospitalization, cited by 42% of patients, was a lack of money.25 High uptake may also be associated with improved access to HIV care and treatment in Uganda. Free home-based HCT programmes in the country have a similarly high testing uptake.12,16,32 A population-based survey in Botswana also showed that most people were in favour of routine HIV testing,33 and the introduction of testing in women receiving antenatal care in Botswana resulted in a high uptake.28

In addition, hospitals also provided a good setting for effective HIV diagnosis in patients’ family members. Although married people are not traditionally highlighted as a risk group, they had an increased risk of HIV infection. Moreover, the risk was higher in patients who did not know the HIV status of their sexual partners. Among couples who were tested, 100% of the HIV-negative members of discordant couples were previously unaware of their discordant status. These findings highlight the need to promote HIV testing for couples. Offering HCT to patients attending hospital saved them both time and transport expenses. Routine HCT can also reduce delays in diagnosis and improve the clinical management of HIV-infected patients.

The HIV prevalence rates observed here may not be representative of the overall HIV prevalence in the two hospitals because the HCT programme started with wards and clinics in which the prevalence was thought to be high.

Quick and efficient referral to care is necessary to reduce morbidity and mortality in patients who are HIV-positive. Because testing was carried out in a hospital, it was possible to start providing care immediately, with referrals for follow-up care after discharge. However, with the planned expansion of routine HCT in Uganda, thousands of HIV-infected individuals will be identified and existing HIV clinics may be overwhelmed. This increase in routine HCT should be coupled to an expansion of HIV/AIDS follow-up care and treatment services. Increasing routine HCT also means that HIV testing must become a core component of hospital services. Health-care workers will have to be trained and involved in areas such as pretest and post-test counselling and specimen collection and testing, as well as carrying out the usual diagnostic work-ups and routine patient care. The use of trained lay and peer health-care providers, including other HIV-infected individuals, to assist with HCT should be evaluated because human resources at Ugandan health-care units are limited. Confining counselling to the provision of key prevention and care messages would decrease the time, personnel and space required. However, research will be needed to establish whether this will have a negative impact on risk reduction and linkage with care, or will threaten the patient’s right to confidentiality and privacy and to decline testing.

In circumstances in which it is not possible to provide routine HCT for all wards and clinics, HCT could, at a minimum, be offered to all patients in wards and clinics where the HIV burden is high; in our case, these were medical inpatient and outpatient units, emergency wards and STD clinics. Although paediatric, obstetrics and gynaecology, and surgical units are not usually regarded as being at risk of HIV infection, our findings show that the prevalence was 15%, 13% and 12% in these three units, respectively. This is not surprising as the prevalence of HIV in Uganda is 6.4%.16 Moreover, it highlights the need to include all hospital patients in areas where the HIV prevalence is high. Following an HCT policy revision and successful pilot programmes in Uganda, routine HIV testing in health-care settings has been incorporated into the five-year HIV strategic plan, both advocacy of and resources for an increase in health-care provider-initiated HCT have grown, and HCT provision is currently being expanded.

Our findings show that offering HCT routinely is feasible and may increase linkage to HIV care and treatment for many individuals with HIV infections. Routine HCT should be an integral component of any expansion in HIV prevention, care and treatment services in Africa and other areas where the prevalence of HIV is high. ■


The authors acknowledge the contributions made by the management of Mulago and Mbarara teaching hospitals, by the routine testing and counselling (RTC) technical committee and by staff of the two hospitals in implementing the RTC programme, and the advice on writing the paper provided by Wolfgang Hladik from the United States Centers for Disease Control and Prevention (CDC)–Uganda.


United States Centers for Disease Control and Prevention (CDC), Department of Health and Human Services, and Presidential Emergency Plan for AIDS Relief (PEPFAR). This project was supported by cooperative agreement number U62/CCU024297-02 from the CDC.

Competing interests

None declared.



  • Mulago–Mbarara Teaching Hospitals’ Joint AIDS Program (MJAP), Kampala, Uganda.
  • Centers for Disease Control and Prevention–Uganda, Entebbe, Uganda.