“We want to use mobile health (mHealth) technologies to make a difference to global health – to tackle the burden of disease in low-resource settings,” said Guy Dumont of the Department of Electrical and Computer Engineering of the University of British Columbia. Prof. Dumont was presenting an update on his group’s work on a Phone Oximeter – a smartphone-based pulse oximeter developed to make pulse oximetry universally available for disease diagnostics and management in low and middle-income countries. He was presenting this work at STIAS where he is currently a fellow.
“We are looking at technologies and applications that might only make a small difference in more developed settings where there is greater access to sophisticated equipment but can make a huge difference in settings where this expensive equipment is not available,” he said.
“Our goal is to develop an easy‐to‐use and robust mobile health platform for early screening and diagnosis in low-resource settings,” he added. “We want to use smart phones and technology to measure physical parameters and use this data to build risk models with the aim of reducing mortality and improving health outcomes.”
The multidisciplinary group’s current focus is child and maternal mortality.
“Child and maternal mortality is still too high in many settings – the Millennium Development Goals 4 and 5 were not reached in this regard,” said Dumont. “For example, in South Africa the maternal mortality rate is still 140 per 100 000 live births – which is much too high.”
Standard oximeters are typically used in operating rooms to make sure patients don’t become hypoxic, thus increasing the safety of anaesthesia – however, they can also be used to diagnose disease.
“A standard oximeter is expensive – up to US $2000, and, even if available, cannot always be used by frontline health care workers who may not have the necessary training,” said Dumont. “Our aim is to produce a low-cost product that can do an early assessment and predict risk in patients developing various conditions.”
Oxygen saturation as disease predictor
The Phone Oximeter, developed by the researchers at the University of British Columbia, measures the blood’s oxygen saturation (SpO2), which is a strong predictor of critical illness. The oximeter integrates pulse oximetry with the technology of a standard smart phone, by connecting a sensor via the phone’s audio jack. It uses the phone battery and processing power to power the sensor and to function as a low-cost standalone device, and has an intuitive display to ensure ease of use, regardless of literacy or language. The Phone Oximeter is able to measure oxygen saturation, heart and respiratory rate and it is hoped it will be made available at a cost of about US $25.
“This technology platform has currently been tested in three studies,” said Dumont. “all focused on maternal and child health – namely pre-eclampsia which is often diagnosed late and is still a major killer of women and babies; pneumonia and sepsis in children under five; and, the problem of post-discharge mortality in children with acute infectious disease.”
“Using predictive analytics, risk predictors can be developed,” said Dumont. “With appropriate software, we can then provide frontline health care workers expert diagnostic, referral, and treatment advice based on clinically established guidelines such as the WHO Integrated Management of Childhood Illness.”
“For example, early data collected in over 2000 women in three countries (South Africa, India and Pakistan) has allowed us to develop a predictor of risk for pre-eclampsia.”
“High blood pressure is the major although not the only symptom in pre-eclampsia. Therefore oxygen saturation is a significant predictor.”
“This risk assessment then comes with recommendations for treatment for individuals,” he added.
This initial work has led to a larger randomised controlled trial – PRE-EMPT (PRE-eclampsia, Eclampsia Monitoring, Prevention and Treatment), a Bill & Melinda Gates Foundation-funded initiative designed to develop and test new tools to reduce the global impact of pre-eclampsia. This trial has already enrolled about 52 000 patients out of a projected 80 000 and hopes to validate the technology with results expected in about a year.
Feedback from the frontline
Two usability sub-studies were also being conducted in South Africa – at Tygerberg Hospital in Cape Town and Frere Maternal Hospital in East London. Information from these studies was used to change the app based on local conditions and health worker feedback.
“You cannot just rely on engineers and clinicians to develop such products,” said Dumont. “The frontline user – the health care worker – needs to be part of the design process.”
The second study, conducted in Bangladesh, used the same technology but focused on identifying children under five requiring hospitalisation due to pneumonia or sepsis.
“Pneumonia is still a major killer of children under five,” said Dumont. “It is very treatable if caught early, however, in many settings there is no or too late diagnosis.”
“A recent study showed that heart rate, respiratory rate, oxygen saturation and temperature can diagnose pneumonia as successfully as X-rays,” said Dumont. “And you can measure all of these using a phone.”
Finally, the last study was conducted in Uganda to look at the problem of post-discharge mortality in children with acute infectious diseases.
“The idea is to use simple parameters – including upper-arm circumference (as an indicator of malnutrition), oxygen saturation, HIV and malaria status, and previous admissions, to identify children at risk,” said Dumont. “This predictive information can then be used to inform the hospital and to educate parents and try to convince them not to discharge their children early.”
Dumont pointed out though that this is a complex issue because the reasons parents discharge their children would include non-medical ones such as transport, logistics, cost and access to carers.
The overall aim is to make mobile apps that are locally relevant incorporating culturally appropriate visuals and language. The clinical trials are also using locally sourced – and, in some cases, locally produced – Android phones to ensure accessibility. If the trials prove successful the next challenge will be to get health authorities to take on these technologies.
“We want to develop reliable tool apps that are as simple as possible to fit the capabilities of the health personnel who will use them to screen patients,” said Dumont.
“We also want to take advantage of the tremendous amount of computing power people carry around in their pockets,” he added.
“However,” Dumont concluded. “Technology is just one aspect. The entire healthcare system must have the capability to properly manage and treat those screened patients.”
Prof. Dumont’s group will hold a conference on mHealth in maternal and child health in February 2017. This will immediately follow the STIAS Roundtable on mHealth for informed decision making at point-of-care in resource-poor settings to be held on 20 and 21 February 2017.
Michelle Galloway: Part-time media officer at STIAS
Photo: Christoff Pauw