Diagnosing pneumonia and its severity requires expensive equipment such as X-ray machines and laboratory services, but in developing countries where the majority of childhood pneumonia deaths occur, the equipment is minimal or non-existing. Recently, researchers at the University of Oxford developed a more affordable tool capable of diagnosing pneumonia through the integration of easily measurable symptoms that include heart rate and temperature.
Results are described in the paper, “The power of data mining in diagnosis of childhood pneumonia,” recently published in the Journal of the Royal Society Interface.
“With the nearest hospital hours away, generalist health workers depend on a set of guidelines known as IMCI (integrated management of childhood illness),” said Elina Naydenova from Oxford’s Institute for Biomedical Engineering and the study’s first author in a press release. “These can sometimes be good at identifying cases of pneumonia but not so good at screening out cases that are not pneumonia. There is also huge variability across users.”
Globally, 1.1 million children succumb to pneumonia annually; it is a leading cause of death of children under the age of five. Studies indicate that accurate pneumonia diagnosis, including identification of severity of the disease and whether the infection is viral or bacterial, can decrease the death rate by 42 percent.
Because most of the deaths occur in poor nations that are low on the worldwide economic ladder, an urgent need for new diagnostic tools exists.
“In settings where there isn’t a clinical expert to set a conclusive diagnosis, the number of unnecessary antibiotic prescriptions has increased as a result — depleting vital drug supplies and adding to the problem of antibiotic-resistant infections,” Naydenova said. “We wanted to apply smart engineering to develop a robust automated system that was consistently more accurate.”
To develop the system, the team examined data from a clinical study in Gambia. They focused on four measurements that could be assessed with two basic equipment items and then used a computer to integrate the measures into an algorithm to diagnose pneumonia.
“Heart rate, respiratory rate and oxygen saturation can all be measured using a pulse oximeter. Temperature requires a thermometer. These are things that can be made available to a health worker with basic training,” Naydenova said.
With the four measures, the team was able to achieve 98.2% sensitivity and 97.5% specificity — the tool identified 982 out of 1,000 pneumonia cases, and only gave 25 false positive pneumonia cases among 1,000 people without the disease. Using the IMCI strategy developed by WHO and UNICEF to identify common childhood conditions, the performance was found to have 94% sensitivity and 69% specificity.
The team was also able to identify the severity of the disease with a 72.4% sensitivity and 82.2%specificity, through assessment of two lung sounds that only required a stethoscope. IMCI, conversely, was found to have a 79.3% sensitivity and 67.7% specificity in determining pneumonia severity. Such diagnostic measure could be improved to 89.1% sensitivity and 81.3% specificity by adding a test for the biomarker C Reactive Protein, but the team says that the addition would have involved additional costs.
The automated tool was also found to be much more specific in discriminating bacterial from viral infections than IMCI. In fact, by assessing heart rate, respiratory rate, oxygen saturation, and a biomarker called lipocalin-2, the investigators distinguished bacterial from viral pneumonia with 81.8% sensitivity and 90.6% specificity, whereas IMCI had a 100% sensitivity and 0% specificity — meaning all cases were diagnosed as bacterial infections.
Although no low-cost tests for the biomarkers required for the new automated tool are currently available, several research teams are working on ways to perform those tests in settings of limited resources.
“We have identified a set of features that could offer an alternative to the combination of X-rays and blood cultures only available in a well-equipped hospital,” Naydenova said. “These will be used in a mobile application linked to a set low-cost diagnostic equipment, which we will be trialling in the next couple of years.”