Training community health workers to perform intensified TB case finding associated with a surge in TB diagnoses in a rural Malawi study

Intensified tuberculosis case finding (TB ICF) by community health workers was associated with a dramatic (20-fold) increase in TB case detection at a very busy antiretroviral therapy (ART) clinic in rural Malawi, according to findings presented at the 21st International AIDS Conference (AIDS 2016) in Durban, South Africa, last month.

The study findings were presented by Dr Katie Simon of the Baylor College of Medicine Children’s Foundation Malawi who also noted that the intervention led to the detection of more paediatric TB, which previously was not being diagnosed at the facility.

HIV and TB in Malawi

TB is the most common infectious cause of death among people living with HIV in Malawi (and globally). TB is undiagnosed in at least half of the HIV-positive people who have died of TB – and some autopsy studies report that TB is identified in up to two-thirds of the people who die in a hospital in sub-Saharan Africa.

TB diagnosis can be particularly difficult in settings such as Malawi, which has to manage a high prevalence of HIV (9.6% among adults) despite severe resource constraints – including a profound healthcare worker shortage. Consequently, “Malawi has task shifted to the extreme,” said Simon, “with the bulk of the care provided by nurses and clinical officers to nearly 980,000 people living with HIV.” The demands placed upon these health care workers is only set to increase, as the country is now transitioning to offering ART to everyone diagnosed with HIV.

An unintended consequence of these overburdened health workers and facilities may be that numerous cases of TB are being missed. The TB prevalence in the country has recently been revised upwards from 160 cases per 100,000 to 280 cases per 100,000 – and almost 70% of the TB cases in Malawi are in HIV-positive individuals. The national guidelines recommend routine ICF whenever a person with HIV attends a health facility; and the Ministry of Health’s data suggest that ICF coverage stands at 98%.

“But if you look at the yield of ICF, it’s only 0.2 percent – so something doesn’t quite add up there,” said Simon. The World Health Organization’s recommended 4-symptom TB screen is highly sensitive for TB in people with HIV. The screen looks for a positive response to any of the following questions: have you had a cough in the last 24 hours, any fever, night sweats or unexplained weight loss? However, in practice, some over-burdened health providers may not perform the screen at every clinic visit. Simon said there is relatively little evidence of how to improve case detection in such routine settings.

One potential solution tried in other settings has been more task shifting, so Simon and her colleagues considered whether Tingathe, a US-funded community healthworker programme, could assist with TB ICF. This programme is in operation at Salima District Hospital, a large rural public hospital in Lilongwe, Malawi, and a number of other facilities near the central and southeastern districts of Malawi.

The Tingathe community health workers who are based at facilities but have a reach into the community focus primarily on the prevention of mother-to-child transmission of HIV (PMTCT) and early infant diagnosis (EID) of HIV. They are trained to provide case management for pregnant women and their children but also conduct active HIV case finding and assist with tracing anyone who was lost to follow-up.

The intervention

An intervention was implemented at Salima District Hospital, which has about 4800 people on ART. The community health workers who were stationed at the facility received one-day training on TB screening. This involved an introduction to the four-question TB screen as well as an orientation to the process or flow that individuals who screened positive would have to go through in order to be diagnosed with TB.

Once trained, these community health workers provided screening to people attending the ART clinic. Anyone who responded affirmatively to any of the four questions was moved to the front of the queue and the clinician would then evaluate per protocol for TB, using smear microscopy and Xpert MTB/RIF tests in parallel (as opposed to serially) with a chest x-ray if necessary, according to Malawian guidelines.

Critically, these individuals were then assigned a community health worker case manager who would be responsible for helping them through the cascade until their final diagnosis. Those who were diagnosed with TB were then started on TB treatment, and provided adherence support by the community health worker at the facility, or via home visits if a problem with adherence was identified. The community healthworker case managers assisted with community tracing for anyone who was lost to follow-up, and linked with other community health workers for combined home-based TB and HIV contact tracing.

The evaluation

To evaluate the effectiveness of the intervention, Simon and colleagues looked at 16 months of pre- and post-intervention data, which were abstracted from registers and tools used by the community health workers. A single-group interrupted time series analysis (a method developed to assess the impact of a policy or practice when randomisation is deemed unethical or practical) was used to assess impact.

Results

"During the sixteen months prior to the interventions a case of TB was diagnosed approximately every other month, and following the intervention there were about ten diagnoses made monthly," said Simon. This represented a 20-fold increase, p < 0.0001. The increase in case detection began immediately in the first month of the intervention, with an immediate increase of 6.7 monthly diagnoses (p < 0.0001). The rate of increase of monthly TB diagnoses improved by 0.78 diagnoses per month in the post-intervention period compared to the pre-intervention period (p = 0.026).

Engaging community health workers involved in maternal and child health may have also helped prioritise diagnoses in children. “Particularly of note to me as a paediatrician is that paediatric TB was not diagnosed at all in this cohort during the sixteen months prior to the intervention, but in the six months following, the nine cases of childhood TB were diagnosed and started on treatment,” said Simon.

Simon noted there are a number of limitations to the evaluation. For instance, the study is limited to one site, and there may be some bias in how data were collected pre- and post-intervention. Because of the uncontrolled nature of the study, “we have to interpret causation with care,” she said. She added that it is possible that the healthworkers may have ‘cherry picked’ people, screening and triaging those they thought were most likely to have TB, but she wasn’t convinced that it was a weakness in a clinic that sees over 200 people per day.

“Obviously the intent is that every patient with HIV should be screened for TB at every clinic visit. But practically speaking, this doesn’t happen,” she said.

Conclusions and looking forward

Future work will be needed to assess whether the impact of the intervention persists and what impact it has on the treatment outcomes of these patients. The evaluation has some patient-level analyses still pending to take a look at what the impact of this intervention is on ‘time-to-treatment’ initiation and on mortality.

Regardless, they have now expanded this intervention to other facilities covered by the community health worker programme. “We now have twenty facilities implementing this model so it will be interesting to find out if this expansion sees similar results,” she concluded. She added that it will also be important to see whether the programme and contact tracing will be sustainable in this setting.

She closed her talk with a Malawian proverb so applicable to the context and need for task shifting to accomplish HIV and TB targets in these settings “Mutu umodzi suzenga denga” – one person alone cannot raise a roof.