Applying Diagnostic Network Optimisation analysis to inform the introduction of G6PD testing into Indonesia for improved treatment of relapsing malaria

Malaria is a disease that returns. While some forms of the illness can be cured with a single course of medicine, others hide in the liver, waiting to wake up weeks or months later and cause a new infection. To stop these relapses, doctors use a special class of drugs that clears the parasite from the liver permanently. However, these drugs carry a hidden danger for some people. A small portion of the population lacks a specific enzyme in their blood, a condition that makes their red blood cells fragile. If a person with this condition takes the anti-relapse medicine, their blood cells can burst, leading to a severe and potentially fatal reaction. To prevent this, doctors must first test patients to see if they have this enzyme deficiency. The challenge for health systems is that these tests are expensive and require equipment that many rural clinics do not have. Placing a test kit in every single small clinic is often too costly, but placing too few means patients must travel far, delaying their treatment or causing them to skip it entirely.

In Indonesia, where malaria remains a significant hurdle on the path to eliminating the disease, researchers set out to solve this puzzle of where to put the testing machines. They focused on three distinct regions: the island of Sumba, where malaria is common; North Sumatra, where it is less frequent; and Purworejo, where cases are rare. The team used a computer model designed to map out the most efficient way to build a network of testing sites. They imagined two different worlds. In the first, every public health clinic would have its own testing device, ensuring that no patient ever had to leave their local village for a test. In the second, they imagined a smarter, leaner system where devices were placed only in specific, central clinics. In this second scenario, patients from nearby villages would bring their blood samples to the central clinic, but only if the journey was short enough to be practical. The researchers calculated the costs of equipment, the price of the test kits, and the fuel needed for the staff to transport samples, all while ensuring that no patient had to travel more than fifteen kilometers one way to get their blood tested.

The results revealed that a blanket approach of equipping every clinic was not the most efficient use of resources. In Sumba, the model showed that the health system could save nearly thirty percent of its first-year budget by placing testing devices in only thirty-nine percent of the clinics, rather than all of them. This strategy meant that some patients would travel further, with the average round trip for a sample increasing from less than five kilometers to nearly ten kilometers, but the money saved on buying and maintaining dozens of extra machines was substantial. In North Sumatra, the savings were even more dramatic. By placing devices in just nineteen percent of the clinics, the total cost dropped by almost sixty percent in the first year. The model identified specific hospitals and clinics that would become the hubs for these networks, handling the extra workload of testing samples from surrounding areas. For instance, one hospital in North Sumatra would see its staff spend an additional four hundred hours a year managing these tests, a significant increase that would require careful planning to avoid overburdening the workers.

In the very low-transmission area of Purworejo, the situation was different enough that a complex network was not needed. The researchers found that placing a single device in the main district clinic would allow more than half of the patients to be tested on the spot, while the few others would have to travel a longer distance. The study did not just look at money; it also looked at the human element. The researchers noted that while the optimized networks saved money, they shifted the burden of travel and time onto the health workers who had to collect and transport the samples. They also pointed out that the model relied on data from public clinics and did not fully account for the private clinics where many people seek care, suggesting that the real-world plan might need to include these private sites to ensure everyone is covered.

Ultimately, the study suggests that Indonesia does not need to choose between expensive universal coverage and a system that leaves people behind. By using data to strategically place a smaller number of testing devices, the country can afford to test more people safely. This approach allows doctors to prescribe the stronger, more effective medicines that prevent malaria from returning, without risking the lives of those who are sensitive to the drugs. The findings offer a clear roadmap for health officials: a carefully designed network of testing sites can make the fight against relapsing malaria both safer for patients and financially sustainable for the country, moving the nation closer to its goal of eliminating the disease entirely.