TB and HIV in Practice

This article originally appeared in HIV & AIDS treatment in practice, an email newsletter for healthcare workers and community-based organisations in resource-limited settings published by NAM between 2003 and 2014.

Successes and challenges in TB/HIV integration: reports from IAS 2011

This feature is kindly supported by the Stop TB department of the World Health Organization.

With additional reporting by Lance Sherriff.

In 2004, WHO’s STOP TB Department and the Department of HIV/AIDS jointly released the Policy on Collaborative TB/HIV activities. The goal of the policy was to direct national tuberculosis and HIV programmes to step out of their separate silos and begin to work more closely together to tackle the co-epidemic of HIV/TB.

The impetus for the policy was the realisation that TB control would never be achieved in most high-burden countries without tackling the HIV epidemic, and that unless diagnosis and treatment Improved, TB would continue to be the major cause of death in people with HIV.


retention in care

A patient’s regular and ongoing engagement with medical care at a health care facility. 

active TB

Active disease caused by Mycobacterium tuberculosis, as evidenced by a confirmatory culture, or, in the absence of culture, suggestive clinical symptoms.


The result of a statistical test which tells us whether the results of a study are likely to be due to chance and would not be confirmed if the study was repeated. All p-values are between 0 and 1; the most reliable studies have p-values very close to 0. A p-value of 0.001 means that there is a 1 in 1000 probability that the results are due to chance and do not reflect a real difference. A p-value of 0.05 means there is a 1 in 20 probability that the results are due to chance. When a p-value is 0.05 or below, the result is considered to be ‘statistically significant’. Confidence intervals give similar information to p-values but are easier to interpret. 

loss to follow up

In a research study, participants who drop out before the end of the study. In routine clinical care, patients who do not attend medical appointments and who cannot be contacted.


A healthcare professional’s recommendation that a person sees another medical specialist or service.

The policy recommended a framework for programme collaboration, and described the key interventions that each programme needed to include in order to make certain that patients received essential services to prevent, more rapidly detect and treat the other disease.

These activities included:

  • Screening TB patients for HIV, and
  • Making certain that TB patients who are HIV positive are providing with HIV care, such as cotrimoxazole and support, and timely access to antiretroviral therapy (ART);
  • Routinely screening all HIV patients for TB, provide isoniazid preventive therapy to those without symptoms of TB, and adopt good TB infection control policies to reduce TB transmission in the health facility. (These activities were subsequently rebranded as the Three I’s for HIV/TB).

Adoption of HIV and TB collaborative activities had been slow, initially restricted to pilot studies and vanguard international NGO-supported programmes. After the policy was released, Rwanda (from the HIV programme side) and Kenya (from the NTP side) reported successful implementation of some TB/HIV activities on a national scale (see HATIP 88). Over time more countries would follow suit, although the alignment of national policies with WHO guidance hasn't always translated into performance on the ground.

What are the outcomes of service integration?

But many operational questions persist for programmes regarding how to prioritise activities in a time of tight funding. Programmes are reaching different conclusions about how TB/HIV services should be structured and provided, and what support services are really needed. There are also differences in views between countries regarding the optimal time to initiate treatment for each disease in order to optimise survival while minimisjng the risk of immune reconstitution inflammatory syndrome (IRIS).

Certain details can make a big difference to how effective a service proves to be. For instance, it is easy to lose patients in the reference process, so some implementers began engaging expert patients to provide accompanied referrals between HIV and TB programmes, which has proven to be highly effective in some settings. In the absence of such mechanisms, many clinics simply gave coinfected clients referral letters sending them to access the services they couldn’t provide on-site. All too often, programmes discovered that their patients were not completing their referrals even when the TB and HIV clinics were right around the corner from each other (see HATIP 114).

The integration of TB/HIV services on-site is one alternative. Many patients prefer to receive integrated care for all their health needs at the same site — at a one-stop shop — because it is more convenient,less expensive and less time-consuming. In fact, in some countries, key HIV/TB-affected populations with individual medical needs, such as people who inject drugs, may have great difficulty accessing the care they need from rigidly disease-specific programmes, and are instead pushing for nationwide scale-up of integrated care sites, where people can get TB treatment, ART, substitution therapy and other basic health services at the same time (see HATIP 159).

But there is no guarantee that co-located services (or the patient support systems provided) will always be of the same quality or as effective compared to what could have been delivered through referral to the other programme — particularly if the effort is under-resourced.

Additionally, as ART programmes become decentralised to smaller, less well-resourced  facilities, these sites are unlikely to be able to offer the full range of services available at larger facilities. Laboratory capacity in particular will be harder to access at more remote sites.

One poster presentation at the IAS conference described a structured facility survey that was administered in 663 health facilities providing HIV care and treatment to over 900,000 patients across nine sub-Saharan African countries. Each facility was asked about the availability and types of TB diagnostic services they could provide access to, and their responses were assessed in relation to the type of facility (primary, secondary, tertiary, private), setting, years providing HIV care and their patient loads. Not surprisingly, there was better access to more sophisticated TB diagnostic capability at big public hospitals.

“As HIV care decentralizes, continued scale-up of laboratory and radiology services in lower levels of the health care system is essential,” the study’s authors wrote. But it may be difficult to achieve this before more widespread availability of low-cost, reliable point of care diagnostics, that can be operated by non-laboratory technicians.

Consequently, the models of integration may need to be adapted to the service delivery level and local resources. At virtually every step of the process, data will have to be generated that can be used to determine whether the approach is effective.

This ongoing research and evaluation process is critical, according to Dr Haileyesus Getahun, one of the chief architects of the WHO’s TB/HIV policy.

Although TB/HIV integration is increasingly being help as a model for other service intergration, "so far, it is unclear what the outcome of TB/HIV service integration has been. We know this or that model may increase the uptake of testing, for instance,. But we don’t really know how it has affected outcomes,” he told an audience attending a panel discussion at the conference on integration and financing.

He noted that despite a great push to extend the reach of PMTCT by Integrating It into other parts of the health system, a recent Cochrane review had concluded there is virtually no data to support the recommendation to integrate.1

Progress towards integration

Data do now finally show that more programmes are providing integrated TB/HIV services. For instance, data from the WHO TB database show that there is more widespread adoption of TB/HIV integration in Africa, and particularly in the PEPFAR-supported countries.2 The researchers analysed seven years of pooled country data to assess the proportion of TB patients who were tested for HIV and the proportion of estimated HIV-related TB cases receiving ART, and compared results in the WHO African Region, the rest of the world, PEPFAR (and non) supported countries in Africa.

Between 2003 and 2009, the proportion of TB patients tested for HIV increased from 4% to 58% in the African region, but from 5% to only 21% in the rest of the world. The proportion of all people living with HIV-related TB who began ART increased from 0.1% to 12.2% in Africa, compared with an increase from 3.2% to 6.8% in the rest of the world. Improvements were somewhat more pronounced in the 19 PEPFAR countries, which account for around 90% of the TB burden. 

Over the same period, HIV testing of TB patients increased from 4% to 61% in PEPFAR-supported countries, compared to an increase from 1% to 40% in non-supported countries; while the proportion of coinfected people who received ART went from 0.1% to 13.1% in PEPFAR countries, and 0 to 2.7% in non-PEPFAR countries

“A person with HIV and TB, presenting to TB services in Africa, is now more likely to have been tested than in the rest of the world, especially in PEPFAR-supported countries — this reverses the situation in 2003,” the poster’s authors wrote.

But getting reliable programmatic data on health outcomes resulting from TB/HIV integration has proven extremely challenging for the Department, according to a study by Sculier et al., co-authored by Dr Getahun. While the study suggests the push for collaborative activities has had some impact, incomplete data make it impossible to clearly determine the impact on key outcomes.

Data drawn from the World Health Organization’s Global Tuberculosis Database show that the number of PLHIV who were diagnosed and registered with TB increased from 14,188 in 2004 to 123,297 in 2008 (an 88% increase).3 Much of this was presumably due to improved case detection of TB/HIV and/or increased TB case registration. Registration among HIV-negative people with TB also increased over the same time period, but only by 20%. 

The data for other indicators seem to point in a similar direction. For instance, the reported mortality rate has gone down significantly (p<0.001) from 16% in 2004 to 12% in 2008, while it remained steady (though at a much lower rate — 5%) among the HIV-negative TB cases. The reported failure rates are low regardless of HIV status (2%). There were also reductions in treatment default and the percentage of cases that went without an evaluation — but these bad indicators were still too high — at 7% and 9% among people living with HIV/TB. Anything might have happened in these cases — they might all be dead. It is impossible to say with any certainty that the outcomes, aside from case registration, had really improved.

“The evaluation of TB treatment outcomes is still poor,” wrote Sculier et al. “Appropriate recording and reporting in both TB and HIV clinics is critical to ascertain TB treatment outcomes among PLHIV, especially mortality, and to monitor progress in TB prevention and control, particularly with increasing access to earlier antiretroviral therapy.”

But these reporting gaps may be due to the very nature of collaborative activities, according to a study from researchers at the London School of Hygiene and Tropical Medicine.4 The study looked specifically at the use of the standardised TB/HIV indicators at facilities in India, Malawi and Uganda, each operating different models of integrated TB/HIV services. Each facility provided the researchers with anonymised, aggegrated data pertaining to the TB/HIV indicators, which the researchers then assessed for strengths and weaknesses.

Problems like inconsistent data collection, and registers that have not yet been updated to include the new indicators, are common to many settings. But other problems may be intrinsic to the delivery model. Notably, facilities that relied mainly on referrals had a much more difficult time reporting outcomes for the service. This may be due to poor communication between facilities. However, Khan et al also noted that there were challenges for sites with full service integration, due to additional reporting requirements for separate national TB and HIV programmes which used entirely different reporting systems (one paper-based, one electronic).

Notably, the researchers concluded that the varying reporting challenges faced by different care models would “hinder comparisons of the effectiveness of service integration between different models of care.”

But regardless of whether WHO can get more reliable data, health departments and academic institutions also need to perform their own research to optimise TB/HIV integration and provide more effective TB/HIV services in different settings.

“There is a huge need for operational research. While we move ahead doing those things that are really no-brainer issues, at the same time we have to generate the evidence,” said Dr Getahun.

Short reports on integration

While not always conducted on a grand scale, a number of studies reported both successes and challenges for TB/HIV integration in different programmes and types of facilities.

Integrated HIV and TB screening services

Dr Sue-Ann Meehan of the Desmond Tutu TB Centre described how TB screening has been integrated into HIV counselling and testing services in Cape Town, and how HIV counselling and testing was taken out into the commun ity in order to reach groups with low testing rates, particularly men.5

Nine outreach sites were established (in shipping containers, rented property or mobile caravans) at non-medical sites in high burden communities.

After going online, these proved to reach a much larger and  notably healthier population. During 2008-2009, 34,287 clients accessed the testing services — 80% of whom were accessed through outreach, and 47% of these clients were male. 26,262 (95%) consented to be tested for HIV at the outreach sites— compared to only 6325 who consented to be tested at the main Desmond Tutu TB Centre Site.

Nine per cent of the total population tested HIV positive — 8% at the outreach sites and 15% at the main site. However, only 1093 of the 1997 who tested positive at the outreach sites accessed HIV care and treatment services, and 588 of the 922 who tested positive at the clinic site.

Every client was also given a symptom screen for TB, and 11% turned out to be symptomatic for TB (10% at the outreach sites, and 17% at the main site). About 187 ultimately received an ‘early diagnosis’ with TB, 142 of whom accessed treatment.

Among the clients who tested HIV positive, 529 were symptomatic, only 340 (64%) had a diagnostic evaluation. 37 (11%) were diagnosed with TB, and 30 (81%) accessed treatment.

Having both types of sites proved complementary since the outreach sites could “reach populations that don’t usually access HIV counselling and testing”, said Dr Meehan. “However, there are more challenges that exist getting this population into care.”

Indeed, this is a situation where a strong partnership with patient networks could improve linkage and retention in care. For example, other community members could be hired and employed as part of the health team, to provide adherence and follow-up support. As noted in a previous issue, another outreach group performing both HIV and TB screening in South Africa assigns each person who tests positive to an expert patient/community worker —which considerably improved links to care.

The 'one stop shop' model in Uganda

If diagnosed coinfected patients do make it into care, they are more likely to be retained if high quality treatment services for both illnesses can be conveniently delivered at the same site.

Accordingly, Dr Sabine Hermans, of Makerere University and the Amsterdam Institute for Global Health and Development, the Netherlands, reported on the performance of a ‘one stop shop’ model integrated TB/HIV clinic, established at the Infectious Diseases Institute — a very large urban outpatient HIV clinic serving around 15,000 patients in Kampala, Uganda.6

Previously — even though medical officers could diagnose TB during routine follow-up visits, the clinic was not performing TB screening and diagnosis systematically. If a client was diagnosed with active TB they would have to report to a TB nurse based in the pharmacy, who provided them with treatment and follow-up.

In December 2008, however, the clinic set up an outdoor integrated ‘one-stop shop’ TB/HIV clinic in the Institute’s grounds, where TB suspects and patients diagnosed with TB could access care for both diseases simultaneously.

A trained team of two medical officers, two nurses and a nurse-counsellor provided care. Once TB treatment was completed, the patient would be referred back to the general HIV clinic.

To try to assess how this one stop clinic is affecting outcomes, data were compared on ART initiation in clients with active TB in 2009 and 2007 (before the one stop clinic was established). In 2007, 161 coinfected subjects were identified, compared to 130 in 2009. Baseline characteristics (CD4 cell count, etc) were similar at TB diagnosis in both periods.

Fewer treatment-naive TB patients were initiated on ART in 2009 than 2007 (56 and 66%,p=0.03), but Dr Hermans noted these were all people with CD4 over 250 (who didn’t meet the criteria for starting ART during that period). However, in 2009, significantly more people started ART while they were still on TB treatment (94 vs 81%, p=0.001). In fact, most clients (60%) started ART during the intensive phase of TB treatment, as recommended by WHO. The median time to ART initiation fell significantly from 101 (IQR 63, 204) days in 2007 to 44 (IQR 21, 105) days in 2009 (p=0.01). Most of the early initiation of ART occurred in patients with CD4 counts below 100 at TB diagnosis.

“When looking at outcomes: overall, survival and retention improved,” said Dr Hermans. However, this wasn’t due to improvements in response of those put on earlier ART — there was no significant difference in survival between ART-treated patients in 2009 and 2007 (83% vs 80%).

In fact, the improvement in survival occurred in the group of patients who were not initiated on ART: 52% of whom survived in 2009 versus 33% in 2007 (P=0.015).

This improvement in survival certainly wasn’t explained by those in the lowest CD4 count strata (below 100) who all died or got lost to follow-up if not started on ART in both time periods. But there was a marked difference in survival and retention in people with CD4 counts between 100-249, who were not started on ART: 71% in 2009 vs. 25% in 2007.

And yet, given that this wasn’t a randomized controlled study, it is not clear how much should be made of this finding, particularly as it wasn’t seen in the next CD4 cell strata (CD4 cells above 250), even though a substantial proportion of these coinfected individuals also died or were lost to follow-up.

If anything, it underscores the need to implement the recommendations of the 2010 ART Guidelines — that ART should be started in all HIV-infected individuals with active TB, irrespective of the CD4 cell count.

A similar analysis at the same facility, only looking at TB outcomes, was presented as a poster at the meeting.7 An evaluation in 2007 had found that 30% of of the hospital‘s patients with HIV who were diagnosed with TB, defaulted after starting treatment. This was another reason why the outdoor clinic was established.  An active tracing system was established and employed whenever patients missed appointments.

It worked; in 2009, the defaulter rate had fallen to 8%, and treatment completion rates went up from 68% to 81%. Deaths increased from 2% to 11%, but the poster’s authors wrote that it was probably due to better tracking of the losses to follow-up, rather than a real increase in mortality.

The researchers also noted that contrary to what other groups have reported, most of their treatment defaults happened in the first two months on TB treatment, suggesting that ‘interventions during the first two months could drastically reduce the number of patients lost to follow-up.”

Screening but no follow-through

A less optimistic report described the challenges of trying to introduce routine intensified case finding into ART clinics in Lusaka District, Zambia — which has around 150,000 people on ART care.8

Despite having a high burden of TB and TB/HIV coinfection, in Zambia, the electronic data system wasn’t designed to collect data on TB screening. However, the low rates of TB diagnosis in Lusaka district (just 568 cases diagnosed over a year out of 99,000 people living with HIV starting ART), suggested that TB symptom screening was not being done routinely.

Since 2005, the Centres for Infectious Disease Research in Zambia (CIDRZ)has been working with the Zambian Ministry of Health trying to implement TB/HIV service integration.

To introduce TB symptom screening, CIDRZ performed baseline assessments of patient flow and registers at 17 clinics. They produced a worksheet to improve the quality and consistency of symptom screening (and link it to the diagnostic work-up). They provided nurses and clinical officers with intensive training and mentorship on TB/HIV management over three to four years.

Earlier this year, CIDRZ conducted a file review, comparing the year from late 2009-10 with the six-month period after that. They found that documentation of symptoms in the patient files increased from 87 to 93% — suggesting that staff were indeed asking the patients about their symptoms. However, the healthcare workers had refused to fill out the TB screening worksheet because “it was perceived as extra work.”

But without forms, there was no way of telling whether patients who screened positive had been sent for diagnostic work-up, and what their outcomes were. The authors suggested development of electronic data collection might improve reporting.

But it is difficult to understand why the healthcare workers’ refusal to use the worksheet wasn’t picked up and — addressed — earlier.

Primary health care clinics can integrate TB/HIV services

In stark contrast to such defeatism, a poster from Dr Kate Clouse and colleagues from the Clinical HIV Research Unit in Johannesburg, and the University of North Carolina at Chapel Hill, described rather encouraging results from a cross-sectional study to assess the extent to which HIV testing and care formed part of the normal clinical practice at a busy primary healthcare clinic in Johannesburg.

The facility, Witkoppen Health and Welfare Centre has around 8500 patient visits each month. It is supported by USAID via the NGO Right to Care, and by the Gauteng Department of Health and private donors. The clinic has dedicated HIV/TB staff five days a week (including a doctor for three days). Since 2004, it has been providing ART, with 4200 clients currently on treatment.

In addition to general HIV services, the clinic provides TB screening and treatment, pre/postnatal care, general/chronic care, paediatric care, social services and a pharmacy. Blood for CD4 cell testing is drawn at the clinic, and sent off-site. Likewise, TB diagnostic services, such as chest x-ray, microscopy, and culture have to be performed off site.

The clinic had already raised the CD4 threshold for starting ART to 350 (as the South African Ministry of Health has also recently done). In addition, it is clinic procedure to offer provider-initiated testing and counselling to all patients, especially TB suspects, with blood for CD4 cell counts drawn the same day as the positive test results. It is also clinic policy to screen all patients for TB at every visit, regardless of the reason for the visit.

Data were collected on 200 patients with suspected smear-negative TB. Participants were mostly women, 45.2% were unemployed, almost a third were born out of South Africa, 28.5% did not attend high school.

Almost half the participants did not know their status at the time they were tested for TB. All were offered HIV tests and 85% accepted (about 70% of these were tested for HIV on the same day as their TB test. 63.3% tested positive for HIV).

80% were given a CD4 test on the same day they tested positive. The median CD4 cell count in the TB suspects was 196, and 140 among those who just tested positive. Among those who already knew their status, 79.3% were on ART — 36% (n=7) of those who had not yet initiated were also eligible for ART. 86% among the newly diagnosed were eligible for ART, 70% of those had initiated ART by the end of data collection, with a median time to initiation of 26.5 days.

Overall 20% of the suspects were eventually diagnosed with TB, 90% of whom were co-infected with HIV. Conversely, 42% of the TB suspects newly diagnosed with HIV did have active TB. Of those who were eligible, 57.1% initiated ART within a median of 34 days of starting TB treatment. Note, the clinic also has tracing procedures in place to track down patients who tested TB-positive so that TB treatment is not delayed.

“The results suggest successful HIV-TB integration is possible at a high-volume clinic when all the relevant services (TB screening, TB treatment, HIV counselling and testing and ART) are available on site,” wrote Clouse et al.


[1]Tudor C L et al. Integrating prevention of mother-to-child HIV transmission (PMTCT) programmes with other health services forpreventing HIV infection and improving HIV outcomes in developing countries (Review). The Cochrane Collaboration.

[2]Gunneberg C et al. Comparison of progress in provision of HIV testing and ART for TB patients – African region, the rest of the world and PEPFAR supported countries. Sixth IAS Conference on HIV Pathogenesis, Treatment and Prevention, Rome, abstract TuPE485, 2011.

[3]Sculier D et al. Global TB treatment outcomes among people living with HIV and HIV-negative individuals from 2004 to 2008. Sixth IAS Conference on HIV Pathogenesis, Treatment and Prevention, Rome, abstract CDB 153, 2011.

[4]Khan P et al. Monitoring and evaluation of collaborative TB/HIV activities in low-income settings. Sixth IAS Conference on HIV Pathogenesis, Treatment and Prevention, Rome, abstract TuPE460, 2011.

[5] Meehan S-A, Beyers N, Naidoo P. A community HCT model for integrated HIV and TB testing in Cape Town, South Africa. Sixth IAS Conference on HIV Pathogenesis, Treatment and Prevention, Rome, abstract WEPDD0104, 2011.

[6]Hermans S et al. Integration of HIV and TB services results in earlier and more prioritised ART initiation in Uganda. Sixth IAS Conference on HIV Pathogenesis, Treatment and Prevention, Rome, abstract WePDD0103, 2011.

[7]Castelnuovo B et al. Evaluation of the impact of TB-HIV integrated care in a large urban clinic in Kampala in reducing the proportion of patients defaulting on treatment. Sixth IAS Conference on HIV Pathogenesis, Treatment and Prevention, Rome, abstract TuPE472, 2011.

[8]Kancheya N et a. Tuberculosis screening in HIV care settings in Lusaka, Zambia. Sixth IAS Conference on HIV Pathogenesis, Treatment and Prevention, Rome, abstract TuPE469, 2011.

[9] Clouse K et al. High rates of HIV testing and care among patients suspected of TB at a primary health clinic in Johannesburg, South Africa. Sixth IAS Conference on HIV Pathogenesis, Treatment and Prevention, Rome, abstract TuPE471, 2011.