Impact of a national HIV voluntary counselling and testing (VCT) campaign on VCT in a rural hospital in Tanzania
Corresponding Author Manuel Battegay, Division of Infectious Diseases and Hospital Epidemiology, University Hospital Basel, Petersgraben 4, 4031 Basel, Switzerland. Tel.: +41 61 265 5072, Fax: +41 61 265 3198; E-mail: email@example.com
Objective To evaluate the impact of a national HIV voluntary counselling and testing (VCT) campaign on presentation to HIV care in a rural population in Tanzania.
Methods Retrospective analysis of data of the VCT and of the National AIDS Control Programme registers of the St. Francis Designated District Hospital at Ifakara for the two 6-month periods before (2007) and after (2008) the National VCT Campaign.
Results There were 4354 individuals presenting at St. Francis Hospital tested for HIV; 2065 (47.4%) before the VCT Campaign and 2289 (52.6%) afterwards. The overall HIV test positivity was 24.6% and higher in 2007 than in 2008 (26%vs. 23%, P = 0.034). This rate was much higher than the Tanzanian National HIV prevalence of 5.7%. Of 1069 individuals who tested HIV-positive, the proportion of married, divorced or widowed individuals and those who lived further than 10 km from the hospital increased from 2007 to 2008. In 356 HIV-infected persons with available data, the median CD4 cell count increased from 137 to 163 cells/mm3 (P = 0.058), while the WHO clinical stage was similar in both periods. Enroling into the National AIDS Control Programme was significantly more common in 2008 (42%vs. 30%, P < 0.001). In a multivariate analysis, the only positive predictor of testing HIV positive when presenting for care after the National VCT Campaign was being married (OR 1.61, 95%CI 1.21–2.15, P = 0.001) or divorced/widowed compared to single (OR 4.58, 95% CI 3.00–8.12, P < 0.001).
Conclusions Our results suggest that the National VCT Campaign raised awareness and readiness to test for HIV in a remote rural setting and that the HIV-positive test rate is much higher in conjunction with a specific HIV care programme.
Testing for HIV is essential for prevention, access to care and treatment. The increase of tested and counselled individuals is one of the successful characteristics of a national AIDS control programme in scaling up prevention and treatment. In 12 high-prevalence countries in sub-Saharan Africa, only 12% of men and 10% of women in the general population have ever been tested for HIV and have received their results (UNAIDS 2007). Most people test only once, mainly when they are already symptomatic and in an advanced stage of their HIV disease (Makhlouf et al. 2007). New approaches in delivering HIV testing had a positive influence (Makhlouf et al. 2007). In particular, the introduction of rapid tests provided the possibility to reach populations in rural settings and to improve convenience (Makhlouf et al. 2007). The roll-out of combination antiretroviral treatment (cART) programmes since 2003 has importantly contributed to wider HIV testing (Nuwaha et al. 2002; Warwick 2006; Roura et al. 2009). However, data on VCT facilities in conjunction with cART programmes in rural regions are still scarce (Warwick 2006; Roura et al. 2009).
The prevalence of HIV infection in Tanzania was estimated to be 5.7% in 2007 (UNAIDS 2008). With a population of 34 569 232 citizens, this accounts for about 1.9 million people living with HIV/AIDS in this country (WHO 2006). By 2007, only 37% of women and 27% of men have been tested for HIV in Tanzania (TACAIDS 2008). To promote VCT services, the government of Tanzania launched the National VCT Campaign from 14th July 2007 to 30th April 2008, which was conducted within the National HIV/AIDS Care and Treatment Plan 2003–2008. The Plan is an initiative of the Ministry of Health and Social Welfare to prevent HIV/AIDS and to provide treatment and care for patients living with HIV/AIDS (TACAIDS 2007). VCT services were offered during the campaign after educating additional health care workers at more health care facilities and public places. In the District of Kilombero and Ulanga, 30 VCT sites were added to the existing 18 facilities. Nationwide, 4 211 727 Tanzanians were tested for HIV during the National VCT Campaign, 188 237 individuals alone in the Morogoro Region from 6th October 2007 to 31st December 2007. The overall HIV prevalence was 6% and corresponded with the regional prevalence of the Morogoro Region (Ministry of Health and Social Welfare, National AIDS Control Programme 2008).
The overall aim of our study was to evaluate the impact of the National VCT Campaign on presentation to care in a rural hospital. The specific objectives were to evaluate predictors of testing HIV positive and to investigate socio-demographic and clinical differences in HIV-infected individuals presenting for care according to two calendar periods, i.e. 6 months before and after the National VCT Campaign.
Study population and setting
St. Francis Designated District Hospital (SFDDH) at Ifakara is situated in the rural Kilombero District of the Morogoro Region of Southern Tanzania and also serves the adjacent Ulanga District (Figure 1; map of the study region). The hospital is the most important health care facility in this region providing care and treatment for a population of about 600 000 inhabitants. It is estimated that about 30 000 persons in this region are living with HIV/AIDS (Ministry of Internal Affairs 2002). SFDDH was the first rural District Hospital in Tanzania accredited by the National AIDS Control Programme to become a Care and Treatment Centre at the end of 2004. Since then, uninterrupted VCT services and comprehensive care for an increasing number of persons living with HIV/AIDS have been offered at Ifakara (Stoeckle et al. 2006). By July 2008, 2394 persons living with HIV/AIDS were enroled at SFDDH of whom 1617 were followed up on a permanent basis and 825 were treated with cART.
In the Kilombero and Ulanga Districts, 18 health facilities including SFDDH were offering VCT services before the launch of the National VCT Campaign. During and after the campaign, another 30 test sites were added (Ministry of Health and Social Welfare, National AIDS Control Programme 2008). Since late 2005 and mid-2006, Ulanga District provides cART at the Mahenge District Hospital and the remote Lugala Lutheran Hospital at Malinyi.
Voluntary counselling and testing for HIV
The VCT Unit of SFDDH offers free testing according to the guidelines of the National AIDS Control Programme of Tanzania (Ministry of Health and Social Welfare, National AIDS Control Programme 2005). Individuals seeking VCT can attend the unit on a walk-in basis during working hours from Mondays to Fridays. Testing for HIV is preceded by counselling. In case of a reactive screening result with SD Bioline HIV 1/2 3.0 (Standard Diagnostics Inc., Kyonggi-do, Korea), the result was confirmed by a second test Determine™ HIV1/2 (Inverness Medical Japan Co. Ltd., Japan). In case of contradictory results, a confirmation with Uni Gold™ HIV1/2 (Trinity Biotech, Wicklow, Ireland) was performed. Test results were explained in post-test counselling. Individuals who tested reactive were motivated to enrol into the National AIDS Control Programme, which provides free treatment and care for every person living with HIV/AIDS. Those who enroled were seen by a clinician on the same day as they received the test result. Clinicians conducted a baseline clinical examination of every newly enroled individual including the CD4 T-lymphocyte count. Individuals not willing to enrol immediately were given a follow-up visit within a month after having been diagnosed with HIV infection.
The VCT and National AIDS Control Programme data of all individuals at SFDDH at Ifakara, between 21st March 2007 to 5th October 2007 (period 1) and 8th January 2008 to 12th August 2008 (period 2), i.e. two 6-month periods, were analysed. HIV testing of inpatients was excluded. Data on demographic characteristics, i.e. sex, age, living place and result of the HIV tests, were collected from VCT records. In HIV-infected individuals, additional data on marital status, CD4 cell counts and WHO stage were retrieved from the digitalized National AIDS Control Programme database. Persons with no available data on place of origin or living outside of the Morogoro Region, encompassing the Kilombero, Ulanga, Kilosa and Morogoro District, were summarized as ‘others’.
As HIV testing was strictly voluntary, no personalized data were used and no intervention was executed, there was no necessity of obtaining an informed consent form according to Council for International Organizations of Medical Science (2002) guidelines for this retrospective analysis.
Basic demographic characteristics, distance between living place and hospital, and HIV test results were compared using the Chi-square test or Fisher’s exact test for categorical variables and the Mann–Whitney test for continuous variables. Logistic regression was used to estimate the odds ratios (OR) of HIV infection when presenting for care at the SFDDH for basic socio-demographic factors and living place. In the second analysis, marital status, CD4 cell count and WHO clinical stage of HIV infection of patients who tested HIV positive were compared between the periods of testing. All analyses were performed using stata™ software version 9.2 for Windows (StataCorp, College Station, Texas).
HIV test positivity
Overall, 4354 individuals were counselled and tested for HIV as outpatients at the SFDDH at Ifakara (Table 1). Of these, 2065 (47.4%) were tested between 21st March and 5th October 2007, i.e. before the start of the National VCT Campaign, and 2289 (52.6%) were tested between 8th January and 12th August 2008, i.e. after the campaign ended. Overall, HIV test positivity was 24.6%, falling from 26% in the period before the National VCT Campaign to 23.2% thereafter (P = 0.034). During both periods, most individuals tested were women (57.2%). Children accounted for 6.6% of all tested individuals. Persons tested in the second study period tended to live more distant from our hospital than those tested during the first period (P = 0.057).
Table 1. General characteristics of all individuals tested for HIV infection before the National VCT Campaign (period 1: from 21st March to 5th October 2007) and after the National VCT Campaign (period 2: from 8th January to 12th August 2008)
| Men||1740 (42.8%)||848 (44.2%)||892 (41.6%)||0.097|
| Women||2325 (57.2%)||1072 (55.8%)||1253 (58.4%)|
|Children <15 year||289 (6.6%)||145 (7.0%)||144 (6.3%)||0.333|
|Distance to hospital|
| <10 kms||2864 (65.8%)||1367 (66.2%)||1497 (65.4%)||0.057|
| 10–49 kms||750 (17.2%)||335 (16.2%)||415 (18.1%)|
| 50–100 kms||383 (8.8%)||202 (9.8%)||181 (7.9%)|
| >100 kms and others||357 (8.2%)||161 (7.8%)||196 (8.6%)|
|HIV test positivity||1069 (24.6%)||537 (26.0%)||532 (23.2%)||0.034|
Factors associated with testing HIV positive in our study population
In a multivariable analysis (Table 2), factors associated with testing HIV positive in our adult study population (n = 4065) were female gender (OR 1.84, 95% CI 1.57-2.15, P < 0.001) and living further away from the hospital (OR 3.21, 95% CI 2.76–3.73, P < 0.001 for >10 km). In contrast, being tested in the second study period was associated with lower odds of testing HIV positive (OR 0.82, 95% CI 0.70–0.95, P = 0.008).
Table 2. Predictors for testing HIV-positive for adults (n = 4065), multivariable analysis
| Women vs. men||1.84||1.57–2.15||< 0.001|
| Period 2 vs. 1||0.82||0.70–0.95||0.008|
|Distance to hospital|
| <10 kms|| 1†||–||–|
| >10 kms||3.21||2.76–3.73||< 0.001|
Characteristics of HIV-infected individuals
Of 1069 individuals (24.6%) who tested HIV positive, 83 (7.8%) were children (Table 3). The median age and the proportion of HIV-infected women were similar in both study periods. Individuals were more frequently married or divorced (P < 0.001) and were more often living more than 10 km away from the hospital in 2008 (P = 0.005). HIV-infected individuals were more frequently willing to enrol into the treatment programme of the National AIDS Control Programme after the VCT campaign (41.9%vs. 29.6%, P < 0.001). The median CD4 cell count at presentation to the clinic tended to be higher in 2008 (163 vs. 137 cells/mm3, P = 0.058), whereas the clinical WHO stage was similar during both periods.
Table 3. Socio-demographic and clinical characteristics of all individuals tested HIV-positive (n = 1069) according to calendar period, i.e. before (period 1) and after (period 2) the National VCT Campaign
|Sex of adults|
| Men||331 (33.6%)||169 (33.9%)||162 (33.2%)||0.806|
| Women||655 (66.4%)||329 (66.1%)||326 (66.8%)|
|Children <15 year||83 (7.8%)||39 (7.3%)||44 (8.3%)||0.538|
| Median (IQR)||33 (27–40)||32 (27–40)||34 (27–42)||0.160|
| Single||378 (40.7%)||234 (49.1%)||144 (31.9%)||< 0.001|
| Married||437 (47.1%)||216 (45.2%)||221 (49.0%)|
| Divorced/widowed||113 (12.2%)||27 (5.7%)||86 (19.1%)|
| Missing data||141|| || |
| Peasant||884 (82.7%)||455 (84.7%)||429 (80.6%)||0.077|
| Other||185 (17.3%)||82 (15.3%)||103 (19.4%)|
|Distance to hospital|
| <10 kms||490 (45.8%)||269 (50.1%)||221 (41.5%)||0.016|
| 10–49 kms||302 (28.3%)||131 (40.8%)||129 (31.8%)|
| 50–100 kms||118 (11.0%)||62 (11.6%)||56 (10.5%)|
| >100 kms and others||159 (14.9%)||73 (13.6%)||86 (16.2%)|
|CD4 T-lymphocytes* c/mm3|
| Median (IQR)||152 (69–272)||137 (53–265)||163 (81–278)||0.058|
|WHO HIV Stage†|
| I||89 (30.4%)||39 (37.5%)||50 (26.5%)||0.210|
| II||77 (26.3%)||27 (26.0%)||50 (26.5%)|
| III||74 (25.3%)||23 (22.1%)||51 (27.0%)|
| IV||53 (18.1%)||15 (14.4%)||38 (20.0%)|
|NACP enrolment||382 (35.7%)||159 (29.6%)||223 (41.9%)||< 0.001|
Impact of the National Campaign on adults who tested HIV-infected-positive
In a multivariable analysis (Table 4), the marital status was the only factor associated with presenting for care after the National VCT Campaign among adults who tested HIV-positive. Persons being married (OR 1.61, 95%CI 1.21–2.15, P = 0.001) or divorced/widowed (OR 4.58, 95% CI 3.00–8.12, P < 0.001) were more likely to present after the campaign than before. In 2008, a larger proportion of persons was enroled in the National AIDS Control Programme (P = 0.002).
Table 4. Predictors for presenting for care at St. Francis Designated District Hospital for adults tested HIV-positive (n = 986) after the National VCT Campaign, multivariable analysis
|Age, per 10 years increase||1.00||0.88–1.15||0.943|
|Women vs. men||1.05||0.78–1.40||0.747|
| Married||1.61||1.21 to 2.15||< 0.001|
| Divorced/widowed||4.94||3.00–8.12||< 0.001|
|Distance to hospital >10 kms||1.22||0.93–1.59||0.156|
During our evaluation, 4354 individuals received VCT for HIV on an outpatient basis at the VCT Site of St. Francis Designated District Hospital at Ifakara. The majority of all tested (i.e. with HIV-negative and positive results) individuals were women (57.2%). This finding is consistent with other studies from Tanzania and sub-Saharan Africa, where HIV infection is more prevalent in women than men (Chu et al. 2005; Manzi et al. 2005; Ministry of Health and Social Welfare, National AIDS Control Programme 2008; UNAIDS 2008). The proportion of men and children in comparison with women tested did not vary significantly between the first and second study period. Unsurprisingly, most individuals (65.8%) attending our VCT facility lived within 10 km from the hospital. Proximity to testing facility and higher level of education (Hutchinson & Mahlalela 2006; Wringe et al. 2008), more likely in Ifakara, the rural centre, facilitate higher testing rates among both sexes. Interestingly, individuals tested in 2008 tended to live further away from the VCT site than those tested in 2007. Although statistically not significant (P = 0.057), education and information on HIV testing during the National VCT Campaign, supported by broadcasting in newspapers, on radio and television throughout Tanzania, may explain this trend (Makhlouf et al. 2007).
Difference of HIV test positivity in this VCT programme vs. the National Tanzanian prevalence
The overall HIV test positivity rate in our study was 24.6%. This is comparable to a VCT survey from Northern Tanzania (Chu et al. 2005), but considerably more than the 6% found during the National Tanzanian VCT Campaign (Ministry of Health and Social Welfare, National AIDS Control Programme 2008) and the 5.7% reported by WHO (UNAIDS 2008). A likely explanation for this difference maybe that persons seeking VCT services in relation to a hospital and an HIV care and treatment centre have a higher pre-test probability for testing HIV positive. Also, persons who attended our VCT unit may have been symptomatic for a long time, but only been motivated to be tested more recently.
Characteristics of HIV-positive tested persons
Overall, 1069 individuals (24.6%) tested HIV positive. Predictors in adults were female gender and living within 10 km of the VCT site. These findings are supported in literature (Chu et al. 2005; Manzi et al. 2005; Hutchinson & Mahlalela 2006; Stoeckle et al. 2006; Makhlouf et al. 2007; Ministry of Health and Social Welfare, National AIDS Control Programme 2008; UNAIDS 2008; Wringe et al. 2008). Importantly, the HIV test positivity rate was substantially lower after the VCT Campaign (OR 0.82 and P = 0.008), indicating that a selection of symptomatic HIV-infected persons may have decreased. Our finding on proportions of single, married, divorced or widowed HIV-infected individuals corresponded with some but not all studies (Sherr et al. 2007; Irungu et al. 2008; Wringe et al. 2008; Shorter et al. 2009), possibly reflecting local socio-cultural influences and stigma. The higher proportion of married, divorced or widowed women in 2008 might indicate the increasing awareness and readiness to test for HIV in the more vulnerable population. This is supported by data that during the National VCT Campaign 188 237 individuals throughout the Morogoro Region were tested from 6th October 2007 to 31st December 2007 (Ministry of Health and Social Welfare, National AIDS Control Programme 2008). Broad political support, consistent activity of the National VCT Campaign and continuous broadcasting throughout the country likely reduced stigma and increased test acceptance.
Importantly, individuals were more frequently willing to enrol into the cART programme of the National AIDS Control Programme after the VCT campaign (41.9%vs. 29.6%, P < 0.001). Antiretroviral treatment programmes have a positive effect on the uptake of VCT programmes (Levy et al. 2005; Stoeckle et al. 2006; Makhlouf et al. 2007), which is confirmed by our results in a rural setting.
The median CD4 cell count at presentation to our clinic tended to be higher after the VCT Campaign (163 vs. 137 cells/mm3, P = 0.058), whereas the clinical WHO stage was similar during both observation periods. Data from sub-Saharan settings usually show lower CD4 T-lymphocyte counts, i.e. of 100 CD4 T-lymphocytes/mm3, at the time of enroling into cART programmes (ART-LINC Collaboration & ART-CC 2006, 2008; Keiser et al. 2008, Nakanjako et al. 2008). In our study, the mean CD4 cell count well below 200 cells/mm3 and the high percentage (43.4%) of HIV-positive individuals in WHO stage 3 or 4 reflect that most newly diagnosed HIV-infected individuals present late and are in need of initiating cART urgently. However, the trend of a CD4 cell count increase from period 1 to 2 is noteworthy and may indicate raised awareness and readiness to test and, possibly, to start treatment for HIV in a remote setting.
Strengths and limitations
Our study is one of the largest assessing HIV test behaviour in a single VCT site in rural sub-Saharan Africa (Council for International Organizations of Medical Science 2002; Hutchinson & Mahlalela 2006; Sherr et al. 2007; Wringe et al. 2008). In particular, we analysed changes induced by a nationwide VCT campaign. A further strength of our study is the availability of the CD4 cell count, WHO stage and the assessment of readiness to enrol into a treatment programme. We excluded data from inpatients of the SFDDH, although this may have even contributed to a higher yield of positive HIV tests. However, this would have given a mixed picture.
Our evaluation has several limitations. Many other VCT services exist in the District of Kilombero and Ulanga, and data of these VCT units were not available. Hence, we could not compare our results for consistency. Information on personal data and origin was based on good faith, as official documents for verifying do not normally exist. Data on HIV clinical stage and CD4 cell count were not available for all HIV test-positive individuals included in this analysis. Hence, the CD4 count may have been higher with a complete data set for this part of the analysis, because one may speculate CD4 cell count and WHO stage were available particularly for already symptomatic individuals.
Our study suggests that the National VCT Campaign raised awareness and readiness to test for HIV in a remote rural setting and that the HIV-positive test rate is much higher in conjunction with a specific HIV care and treatment programme. VCT programmes in conjunction with care and treatment facilities seem particularly relevant as they may attract persons with higher pre-test probability for an HIV infection.