Pneumonia is an infection of the lungs, filling them with fluid. It causes cough and fever and can make breathing difficult. Pneumonia can be caused by viruses, fungi or bacteria. Bacterial organisms such as Haemophilus influenzae and Streptoccocus pneumoniae cause more than 50% of pneumonia deaths in children under 5 years of age. Severe pneumonia can be deadly.
Pneumonia kills more children under the age of five than any other disease. Approximately 1.5 million children die from pneumonia every year; 98% of those occur in developing nations. One child dies from pneumonia every 20 seconds; that’s 4,300 young lives lost every day.
The World Health Organization and UNICEF produce a set of guidelines to train health workers to diagnose and manage diseases such as pneumonia, diarrhoea, and malaria called the Integrated Management of Childhood Illnesses (IMCI).
The IMCI guidelines train health workers to diagnose pneumonia by counting the respiratory rate and recognizing signs of respiratory distress such as subcostal recessions, chest indrawing and tracheal tug. Respiratory rate counting is only moderately accurate and many health workers count rates badly or not at all. Improved diagnosis is not the whole solution but could save many thousands of children’s lives every year.
For more information on pneumonia in the developing world, please visit our the World Pneumonia Day website and help them raise the awareness of this problem.
The StethoMic is a low-cost digital stethoscope that we've prototyped from off-the-shelf components. By utilising the powerful digital signal processing chips on modern smartphones, the StethoMic minimizes the need for custom circuitry and electronic components. This allows us to develop the device at a much lower-cost than regular digital stethoscopes out in the market.
A benefit of simplifying the design of the digital stethoscope attachment is that it is now possible for people in the developing world to make their own devices based on our design. A local technician with access to sufficient training and the right materials can produce our digital stethoscopes thereby empowering communities to be self-reliant.
By virtue of of the StethoMic's connection to smartphones, our digital stethoscope unlike other digital stethoscopes is able to directly connect to the Internet. By connecting to the Internet, we harness our cloud database of breath sounds and sophisticated algorithms to provide diagnostic clinical decision support to the health worker at the point of care.
StethoCloud currently runs on Windows Phone, Android and legacy J2ME compatible phones. An iOS version of the app is currently in development
Our apps have been developed for clinical use by health workers in developing countries. Thus, we base the design, flow and content of the apps on the IMCI Guidelines. The apps guide the health worker in the assessment of a child with questions based off the IMCI Guidelines. We augment the current guidelines with content such as photos and videos explaining the different signs of the disease.
By connecting the StethoMic to a smartphone running our app, a health worker is able to record a child's breath sounds. The StethoCloud app is able to perform basic offline analaysis on the recorded breath sounds thanks to our prorprietary BioSignals algorithm suite.
The BioSignals algorithms are optimized digital signal processing algorithms that enable even low-powered feature phones to analyse breath sounds and calculate the respiratory rate of a child. Based off the IMCI guidelines, obtaining an objective measure of respiratory rate will help with the diagnosis and triage of the disease.
Although obtaining a respiratory rate is a highly sensitive measure of a child's risk of pneumonia, it is unfortunately also highly non-specific. A child's increase in respiratory rate may be caused by many other different diseases such as anemia or anxiety. This leads to many false positive results.
By connecting our apps to the cloud, we are able to upload the recorded breath sounds for further processing by sophisticated machine learning algorithms based on Big Data data mining techniques to increase the specificty of the diagnosis and reduce the false positive results.
StethoCloud is unique from other digital stethoscopes because we harness the power of cloud computing to create the artificial intelligence that makes our stethoscopes "smart".
By using scalable cloud storage facilities provided by Microsoft's Windows Azure platform, we are building the world's largest database on breath and heart sounds. Through this database, we will be able to train our machine learning algorithms to extract features and perform knowledge discovery of breath sound patterns and their related pathologies. This training will enable the machine to accurately diagnose pneumonia, asthma and other respiratory diseases with at least the same level of sensitivty and specificity as human diagnosticians.
© StethoCloud Pty Ltd 2013
