In the rolling hills of the Eastern Cape, community health workers walk kilometres each day to check on tuberculosis patients who started treatment weeks earlier. They carry paper registers and symptom checklists. What they do not carry is a GeneXpert cartridge — the molecular test that could confirm a diagnosis in two hours. That machine sits in a district laboratory, often unused because cartridges are out of stock or have expired.

South Africa has one of the world's highest TB burdens, with an estimated 280,000 new cases each year. The country also has one of the largest networks of GeneXpert machines, purchased with donor and government funds. Yet a disconnect between policy design and frontline reality means that the technology often does not reach the people who need it most. Patients in remote villages wait weeks or months for a sputum smear result — a century-old method that misses roughly half of all cases — while molecular diagnostic capacity goes to waste.

This is not a story of resource scarcity alone. It is a story of how public-programme design, procurement systems, and urban bias can strand life-saving technology in rooms where no one uses it, while community workers improvise with tools from an earlier era.

GeneXpert Machines Gather Dust While TB Goes Undiagnosed

GeneXpert is a molecular diagnostic platform that detects Mycobacterium tuberculosis and rifampicin resistance directly from sputum, in about two hours. The World Health Organization has recommended it as the initial diagnostic test for all people with signs of TB since 2013. South Africa was an early adopter, deploying hundreds of modules across the country.

But in several district labs in the Eastern Cape, utilisation rates have fallen sharply. Lab managers report that cartridges — the single-use consumables required for each test — are frequently out of stock for weeks at a time. When cartridges do arrive, they sometimes expire before they can be used, because the supply is erratic and demand forecasting is poor. One lab supervisor, speaking on condition of anonymity, said that a batch of 500 cartridges expired on the shelf last year while clinicians were still sending sputum samples for smear microscopy.

The result is a diagnostic bottleneck. Patients who present at clinics with cough, fever, and weight loss give a sputum sample that is sent to the lab. Without a functioning GeneXpert, the lab falls back on smear microscopy, which requires a skilled technician and a microscope, and has low sensitivity — especially in people with HIV, who often have low bacterial loads. A smear-negative patient may be sent home without treatment, only to return months later with advanced disease.

In a recent survey of district labs in three provinces, roughly 30% reported that GeneXpert cartridges were unavailable for more than two weeks in the previous quarter. Another 15% reported that machines were non-functional due to lack of maintenance or calibration. The combination of stockouts and breakdowns means that even when a patient's sample reaches the lab, the chance of getting a molecular result is far from guaranteed.

At a clinic in the Eastern Cape, staff described a typical scenario: a patient with a persistent cough provides a sputum sample, which is sent to the district lab 40 kilometres away. The lab has no cartridges, so the sample is processed by smear microscopy. The result comes back negative. The patient is sent home with cough syrup. Three months later, the patient returns, now with haemoptysis and weight loss. This time, a chest X-ray shows cavities, and a repeat sputum sample, finally tested on GeneXpert, confirms TB. By then, the patient has infected household members and neighbours.

Community Health Workers Fill the Diagnostic Gap on Foot

In the absence of reliable molecular diagnostics, community health workers have become the backbone of TB case detection and treatment support. These workers, often women from the same villages they serve, are trained to recognise TB symptoms, collect sputum samples, and track patients who miss clinic appointments. They work for modest stipends, sometimes unpaid for months.

Their work is essential but limited. Symptom-based referral — the tool they rely on — misses a substantial proportion of smear-negative cases, particularly in people with HIV or extrapulmonary TB. A study in KwaZulu-Natal found that community health workers using symptom screening alone detected only about 40% of culture-confirmed TB cases. When GeneXpert was available at the clinic, detection rates rose to roughly 80%.

In Gauteng, a programme that placed GeneXpert machines at primary care clinics — rather than in central labs — achieved a 15–20% increase in case detection compared with clinics relying on centralised testing. The programme also reduced the time from sample collection to treatment initiation from an average of 21 days to under 3 days. But scaling such a model nationally faces barriers of cost, maintenance, and supply chain.

Community health workers also track treatment adherence, visiting patients' homes to ensure they take their daily medication. Without a confirmed diagnosis, however, some patients may be taking TB drugs unnecessarily, while others with drug-resistant TB receive standard therapy that is ineffective. The workers do their best with the tools they have, but the system asks them to compensate for a gap that technology could fill.

One community health worker in the Eastern Cape described visiting a patient who had been on first-line TB treatment for two months with no improvement. The worker suspected drug resistance but had no way to confirm it. She referred the patient back to the clinic, where a sputum sample was sent for GeneXpert — but the machine was down. The patient eventually travelled 100 kilometres to a referral hospital for testing, but by then, the disease had progressed. Stories like this are common in rural areas, where the distance to a functioning diagnostic facility can be a matter of life and death.

How a Single Molecular Test Reshapes TB Care Pathways

GeneXpert does more than diagnose TB — it detects resistance to rifampicin, one of the cornerstone drugs. This information is critical for initiating appropriate treatment. Without it, patients with multidrug-resistant TB (MDR-TB) may start standard first-line therapy, which will fail and allow resistance to amplify.

A study in KwaZulu-Natal found that each week of delay in initiating MDR-TB treatment increased the risk of death by roughly 6%. Patients whose resistance was identified by GeneXpert started appropriate therapy a median of 10 days after diagnosis, compared with 38 days for those diagnosed by conventional drug-susceptibility testing. The mortality difference was stark: 12% versus 28% at six months.

In Cape Town, a clinic that introduced on-site GeneXpert testing reduced the time from sample collection to treatment initiation from 21 days to 3 days. The clinic also saw a decline in loss to follow-up — patients who give up waiting and disappear from the system. The intervention required a dedicated technician, a stable electricity supply, and a reliable cartridge supply chain, all of which are harder to maintain in rural areas.

The test's speed is its greatest strength, but it also creates a new vulnerability: when the machine is down, the entire diagnostic pathway collapses. Unlike smear microscopy, which can be done with minimal infrastructure, GeneXpert requires a controlled temperature, a computer, and regular calibration. In many district hospitals, the machine works only when the technician is present, and the technician may be shared across multiple facilities.

There is also a trade-off between centralisation and decentralisation. Central labs can process high volumes and maintain quality control, but they introduce transport delays and loss to follow-up. Decentralised testing at clinics reduces turnaround time but requires more machines, more trained staff, and more robust supply chains. South Africa's current system attempts a hybrid approach — machines at district labs serve multiple clinics — but the supply chain failures undermine the model. Some argue that a fully decentralised system, with GeneXpert machines at every clinic, would be more effective, but the upfront cost would be substantial, and maintenance would be a challenge in remote areas.

Public-Programme Design Sends Resources to Urban Centers

The national TB programme has procured roughly 600 GeneXpert modules over the past decade. But deployment has been uneven. An analysis by the South African Medical Research Council found that about 60% of modules are located in metropolitan hospitals and reference labs, while rural clinics — where the majority of TB cases occur — have limited access.

District health managers say that budget allocations for cartridges are consistently below need. One official in the Eastern Cape estimated that the district's cartridge budget covers only about 40% of the required tests, forcing clinics to ration testing. Patients with obvious symptoms may be tested; those with ambiguous signs are often managed empirically, without a confirmed diagnosis.

Donor-funded cartridges arrive in unpredictable waves, creating feast-or-famine cycles. When a shipment arrives, labs run tests around the clock until supplies run out. Then they revert to smear microscopy for weeks or months. This stop-start pattern undermines trust in the health system and delays treatment for patients who happen to present during a drought.

The urban skew is not unique to TB diagnostics — it reflects a broader pattern in South African health spending, where major hospitals absorb a disproportionate share of resources. But TB is a disease of poverty, and poverty is concentrated in rural areas and informal settlements. The mismatch between where the machines are and where the patients are perpetuates the epidemic.

A counter-argument sometimes raised is that centralised testing is more cost-effective because it allows high-volume processing and reduces the need for multiple machines. However, this argument assumes a functional supply chain and reliable transport — conditions that do not hold in many rural districts. When centralised testing leads to long delays and high loss to follow-up, the cost per patient successfully treated may actually be higher than with decentralised testing, even if the per-test cost is lower. A modelling study from the University of Cape Town found that decentralising GeneXpert testing to primary care clinics could be cost-saving overall, once the costs of delayed treatment and secondary infections are accounted for.

Equity Gap: Wealthier Patients Bypass Public Diagnostics

Patients who can afford it bypass the public system entirely. Private laboratories in South Africa offer same-day GeneXpert testing for roughly 500 rand (about US$30). For a middle-class household, that is an inconvenience. For a family living on less than 1,000 rand per month — the reality for many TB patients — it is out of reach.

The result is a two-tier diagnostic system. Wealthier patients get a molecular result within hours and start appropriate therapy within days. Poorer patients wait weeks for a smear result that may be falsely negative, or never receive a test at all. The gap is both a health equity issue and a public health problem, because untreated TB spreads in crowded households and communities.

South Africa's proposed National Health Insurance scheme aims to cover molecular diagnostics for all, but its rollout has been repeatedly delayed. Even if NHI is implemented, the supply chain issues that plague public-sector GeneXpert testing will need to be addressed. A system that cannot keep cartridges in stock for the machines it already owns will not magically function better under a new financing model.

Some provinces have experimented with GeneXpert vans — mobile units that travel to rural clinics and offer same-day testing. These vans show promise, but they cover only a fraction of the need, and their maintenance and fuel costs are high. They are a patch, not a solution.

Another approach is the use of community-based sputum collection points, where patients can drop off samples that are then transported to a central lab. This reduces the distance patients must travel, but it does not solve the cartridge shortage or the delay inherent in centralised testing. In some areas, community health workers have been trained to collect sputum samples and transport them to the lab, but without a functioning GeneXpert at the lab, the effort is wasted.

Tracing the Cost of Idle Capacity in Lives and Currency

The human cost of diagnostic delay is difficult to quantify precisely, but the estimates are sobering. Each untreated infectious case of TB infects roughly 10 to 15 contacts per year. Delays of two to three months — common in the current system — can seed dozens of new infections, many of which will progress to active disease.

A modelling study by the South African TB Think Tank estimated that diagnostic delays contribute to roughly 1,200 preventable deaths annually in the country, simply from the failure to use existing molecular capacity effectively. The number may be higher if drug-resistant cases are missed entirely.

The economic cost is also substantial. The GeneXpert machine itself is a sunk cost — the capital has been spent. Each unused test represents a wasted investment. A cartridge costs roughly US$10 to $15, which is modest compared with the cost of treating a patient with advanced TB or MDR-TB, which can run into thousands of dollars. Yet the system continues to under-utilise the technology because of procurement and distribution failures.

Productivity losses from TB-related illness and death are estimated to exceed 2 billion rand annually. These losses fall disproportionately on working-age adults, who are the most economically active segment of the population. Reducing diagnostic delay would not only save lives but also preserve economic output.

Some critics argue that the focus on GeneXpert overlooks the need for simpler, cheaper diagnostics that could be used at the point of care. Technologies such as urine lipoarabinomannan (LAM) assays or portable molecular devices like the Truenat platform are being evaluated in South Africa. However, these alternatives have their own limitations: LAM has lower sensitivity in HIV-negative patients, and Truenat requires its own consumables and supply chain. The debate over which technology is best can distract from the immediate problem: the machines already in place are not being used effectively.

Community-Directed Delivery Models Offer a Path Forward

Several pilot projects are testing ways to bring GeneXpert closer to patients. The most promising involve portable versions of the technology, such as the GeneXpert Edge, a compact device that can run on battery power and does not require a laboratory. In Limpopo province, a pilot trained community health workers to collect sputum samples and run the test in a mobile van equipped with the Edge device.

The results were striking. Turnaround time dropped from an average of 14 days to same-day results in roughly 80% of cases. Patients who tested positive started treatment on the spot, and those with rifampicin resistance were fast-tracked to MDR-TB clinics. The pilot also reduced loss to follow-up: when patients receive a result the same day, they are far more likely to stay engaged.

Scaling this model faces real obstacles. The Edge device is cheaper than the full-size GeneXpert but still costs several thousand dollars. Training community health workers to perform molecular testing requires supervision and quality assurance. And the supply chain for cartridges remains a weak link — if the national procurement system cannot keep standard cartridges in stock, it may struggle to supply the Edge-specific cartridges as well.

But the pilots demonstrate that the technology can work in remote settings when the programme design is intentional. The question is whether the political and financial will exists to reorient the system away from urban labs and toward community-based testing. Some TB advocates argue that the biggest barrier is not technical but bureaucratic: procurement rules, laboratory regulations, and turf battles between health departments.

There is also a need for better data on where machines are located and how often they are used. Currently, the national TB programme does not publish real-time utilisation data for GeneXpert modules. Without this information, it is difficult for district managers to identify underperforming machines or redirect supplies to where they are needed most. A simple dashboard that tracks cartridge stock levels and machine uptime could help reduce idle capacity, but such systems are not yet in place.

In the meantime, community health workers continue their rounds on foot, carrying registers and hope. Their work is invaluable, but it cannot substitute for a functioning diagnostic system. South Africa has the machines. It has the workers. What it lacks is the bridge between them.

This article is for informational purposes only and does not constitute medical advice. Individuals with TB symptoms should seek evaluation at a local clinic.