A blood test can quickly diagnose malaria, but similar tests are lacking for other diseases that cause fever.

Jack Kurtz/Zuma Press, Inc./Alamy Stock Photo

Malaria-free but still sick: What’s giving millions of kids fevers?

The toddler on her mother's lap is listless, her eyes dull. She has a fever, little appetite, and a cough. Her journey to the health clinic took an hour by bush taxi, and she had to wait two more hours to be examined. When it's finally her turn, the nurse practitioner pricks her finger and blots a drop of blood onto a rapid diagnostic test (RDT) for malaria. In 15 minutes the answer is clear: The child has malaria. She receives antimalarial drugs, which will most likely vanquish the parasites from her bloodstream within days, and she is sent home to recover.

If the test is negative, however, things get complicated. If malaria isn't making her sick, what is? Is it pneumonia, typhoid, or Lassa fever? Meningitis? Or more than one infection at the same time? If she has bacterial meningitis, the right antibiotic could save her life. If she has Lassa fever, antibiotics won't help.

Until recently, nearly every child with a temperature above 38.5°C was treated for malaria in regions where the disease is endemic. It was one of the most common and deadliest causes of fever, and there was no easy way to rule it out: A definitive diagnosis required a microscope and a skilled technician—unavailable in many places. To be safe, health workers were trained to treat most fevers with a dose of antimalarial medicine. Public health campaigns helped spread the word: If your child has a fever, get them treated for malaria!

In the past decade, malaria RDTs—which use antibodies to detect the parasite's proteins—have transformed the landscape. The tests help reduce unnecessary prescriptions for malaria medicines, but they have exposed a new problem: the previously hidden prevalence of "negative syndrome"—feverish kids who don't have malaria. Even in places with the highest rates of malaria, only about half of fevers are actually due to the disease. In many places, that figure is 10% or less. In 2014, the World Health Organization (WHO) estimated that 142 million suspected malaria cases tested negative worldwide.

Negative test results pose a dilemma for health care workers, who in remote areas may be community volunteers with minimal training. When their one diagnostic test comes up negative, they are left empty-handed, with nothing to offer except some advice: Return if the child gets sicker. But often the family lives hours from the nearest clinic and even farther from a hospital. And patients, or their parents, expect to receive some sort of treatment. So health workers "usually give all the medicine they have," says Didier Ménard, a malaria expert at the Pasteur Institute in Paris. That approach often means antibiotics.

A health worker examines a feverish child using a diagnostic app.

Magali Rochat

Unintended effects

Several research teams have documented dramatic increases in antibiotic prescriptions in places where malaria testing has been introduced, with some clinics giving the drugs to almost all patients who test negative for malaria. At the same time, studies of the actual causes of fever in places such as Tanzania and Thailand suggest that only 5% to 10% of patients have a bacterial infection that antibiotics could help treat. Those unnecessary prescriptions can cause side effects in patients and increase the risk that antibiotic-resistant strains of bacteria will emerge and spread.

Meanwhile, because there are few reliable rapid tests for infections other than malaria, HIV/AIDS, and tuberculosis, most nonmalarial fevers go undiagnosed and untreated. Patients who have a fungal, parasitic, or viral infection and incorrectly receive antibiotics don't get the drugs or supportive treatments that might actually help them recover.

Public health advocates want to change that. "People are moving from malaria to fever diagnosis and management" as a primary challenge for public health in malaria risk areas, says Heidi Hopkins, an infectious disease researcher at the London School of Hygiene & Tropical Medicine (LSHTM).

To tackle the problem, researchers are starting to gather data on the myriad causes of fever—no small challenge in regions where diagnostic labs are scarce. Several teams are working on robust RDTs that can be used without refrigeration or electricity and can withstand many months of shipping and storage in tropical heat and humidity. Others are developing electronic tools—diagnostic apps—that can help identify the patients most in need of additional treatment.

Countless pathogens can trigger fever, one of the body's most basic immune reactions. "Based on symptoms alone, there is no way to determine what is causing a fever," Hopkins says. Common illnesses vary widely from place to place and from season to season. "What is a prominent cause of fever in Laos is not the same as in Tanzania," says Sabine Dittrich, who heads the program on malaria and fever at the Foundation for Innovative New Diagnostics (FIND), a nonprofit organization in Geneva, Switzerland. In one region, herpesvirus, typhoid fever, and leptospirosis might be the top culprits, whereas dengue, scrub typhus, and Japanese encephalitis might be more common in another.

When fever isn’t malaria

Researchers have estimated the percentage of fevers that are—and are not—due to malaria in parts of Africa where the disease is endemic. But data are sparse, especially in west and central Africa.

Malaria atlas project adapted by J. You/ Science

"What we are really missing—desperately, desperately missing—is surveillance data" on the main pathogens in different regions that make people sick with fever, says Teri Roberts, a diagnostics expert at the Doctors Without Borders Access Campaign in Geneva. "If we were to ask a developer tomorrow to make a test for the top 20 pathogens you'd want to measure if someone has a febrile illness, we would not be able to tell them what to test for."

Without a clear diagnosis, simply knowing whether the patient's infection is bacterial or viral can help. In wealthy countries, tests to distinguish between the two are fairly common. They detect biomarkers—proteins or other molecules that the body produces in response to various infections. For example, one key sign of inflammation or infection is an increased level of C-reactive protein (CRP) in the blood, and bacterial infections typically trigger higher levels of CRP than viral ones do.

However, such tests can be misleading in poorer regions. Malnutrition, for example, can suppress CRP levels, whereas parasitic infections can raise them. The tests "are promising in Europe, but we don't know enough about how they would work in areas where there are malaria and malnutrition," Dittrich says. To better understand the advantages and pitfalls, FIND and its partners are conducting clinical trials using CRP and other biomarker tests to help diagnose fever in Thailand, Myanmar, and Malawi.

A search for causes

Identifying the sickest patients also can aid treatment because they are the most likely to need either drugs or more complex care at a hospital. Valérie D'Acremont and her colleagues at the Swiss Tropical and Public Health Institute in Basel have found a way to help do that: They developed an app for a smartphone or tablet that builds on WHO guidelines for treating sick children, adding information from half a dozen simple diagnostic tests. A monitor that fits over a child's finger measures pulse and oxygen saturation in the blood, helping determine whether their breathing is seriously impaired and whether they are dehydrated. A glucose test can flag acute hypoglycemia, which is potentially deadly and easy to treat. A hemoglobin test diagnoses severe anemia, and an arm circumference test identifies malnourished patients. Both conditions can make infections more dangerous. If a child is severely ill, the algorithm recommends both immediate treatment and referral to a hospital or better-equipped clinic.

For children who are not severely ill, the algorithm uses the rate of breaths per minute to distinguish between an upper respiratory tract infection, which usually needs no further treatment, and a lower respiratory tract infection, which can lead to pneumonia. Patients at risk of pneumonia have their CRP levels tested to find out whether a bacterial cause is likely, in which case they receive antibiotics. If a virus is more likely, the child receives inhaled salbutamol, a common asthma medicine that widens the airways. Feverish children without cough, vomiting, or other symptoms also have their CRP levels tested. Kids with levels above the cutoff receive antibiotics; those below probably have a virus and don't need medicine.

In the end, it’s not just the tool that’s going to improve patient care and save lives. You need to reorganize the system.

Lee Schroeder, University of Michigan

In a trial involving 3600 children in Dar es Salaam, Tanzania, the algorithm cut in half the rate of "clinical failures"—children who got significantly sicker or were still seriously ill a week after their initial visit. At the same time, the app dramatically reduced the use of antibiotics: Only 11% of children treated according to the algorithm received antibiotics, compared with 95% of children treated according to standard guidelines. Dittrich calls the work extremely promising.

Other teams are developing tests that could search for multiple causes of fever in a single blood sample. Chembio Diagnostic Systems in Medford, New York, is working with FIND to develop a test that would detect some of the most common fever-causing pathogens in the Asia-Pacific region, based on antibodies or marker proteins. The test will cover four types of malaria, dengue virus, Zika virus, chikungunya virus, and the potentially deadly bacterial diseases leptospirosis, murine typhus, melioidosis, and scrub typhus—which are treatable but require the correct antibiotics.

No simple fix

But even if such a test proves both reliable and durable enough to withstand tough field conditions, it would not be a quick fix for the problem of diagnosing fever. Often, says Clare Chandler, a medical anthropologist at LSHTM, tests that seem mobile and simple "aren't really simple at all." Even the rapid malaria tests, for example, have several unexpected pitfalls, she and her colleagues found: They take 15 minutes to get results, a strain for clinic workers who might have dozens of people waiting to see them. The lancets used to prick fingers pose another problem. After one use, they need to be disposed of. "You need a sharps bin, which has been a huge issue," she says. "We've seen people just throwing them on the floor, or trying to put them down latrines, not really knowing how to destroy them."

"We need to learn from our experience with malaria tests" and consider the potential downsides of new diagnostics, Hopkins agrees. Another problem: Because viruses are the most common cause of fever, the tests will often end up steering health workers away from giving improper treatment—rather than toward something that can help. "The tests are perceived as gatekeepers, to keep people from things we don't want them to have," Chandler says. That situation can frustrate both the patients and the people caring for them. In one study of RDTs, Chandler notes, patients avoided the clinic that offered the tests because they were less likely to receive medicine.

Chandler and others say the only lasting solution to the fever dilemma is to build stronger health systems with highly qualified health workers, a reliable supply of essential medicines, and well-equipped hospitals to treat severe disease. "In the end, it's not just the tool that's going to improve patient care and save lives. You need to reorganize the system," says Lee Schroeder, a pathologist and diagnostics expert at the University of Michigan in Ann Arbor. That's an effort that will take decades. But in the meantime, knowing what to offer a feverish child would make an immeasurable difference.