Software

Developers Cubed: A Mobile Phone And A Digital Stethoscope

Gizmodo’s Developers Cubed series offers a behind the scenes look into Australia’s up and coming dev scene. This week: We chat with the recent winners of the Australian leg of Microsoft’s Imagine Cup, Team Stethocloud about how you go about developing a new and hugely less expensive stethscope as a smartphone app, what happens next for the team and why medical development might be the overlooked field that Aussie developers should target next.

Developers! Developers! Developers!
It’s not just Steve Ballmer’s favourite sweat-laden catchcry! If you’d like to be featured in Developers Cubed, shoot a note to alex.kidman at alluremedia.com.au. I’d love to hear from you.

Team Stethocloud was announced this week as the Australian winners of Microsoft’s Imagine Cup. Team member Hon Weng Chong kindly gave me some of his time to answer questions about where the team came from, and what’s next.

Giz: First of all, obviously, congratulations. Next stop is the global Imagine Cup finals in Sydney, but what was the genesis of your particular idea?

Thank you.

I was told about the Imagine Cup by a friend of mine who had competed in the Imagine Cup finals in New York last year. His solution is called LifeLens and it is a mobile application that through the attachment of a special lens can detect malaria parasites in blood smears.

At the same time, I had learnt of the proliferation of mobile phones in the developing world. That and the LifeLens project got me thinking of the potential use of mobile phones for the diagnosis of diseases in low resource settings. By then, the only question that remained was which disease we were going to tackle.

Being a medical student, I had also just started my pediatrics rotation at The Royal Children’s Hospital in Melbourne. During my pediatrics rotation, I saw a lot of cases of respiratory diseases, chief amongst them were asthma and pneumonia. We were also given a lecture on International Child Health and I was reminded of the appalling statistics presented on childhood pneumonia.

Pneumonia is called the ‘forgotten killer of children’ due to its lack of public awareness despite it being the single largest killer of children under the age of 5. Pneumonia kills nearly 2 million children per year, which is more than measles, malaria and HIV all combined. Despite such terrible statistics, pneumonia unlike malaria and HIV is neither a difficult nor expensive disease to treat as long as you catch it early enough. The problem is that diagnosis of pneumonia is hard and in developing countries you may have to walk several kilometres to the nearest clinic. Sometimes you might wait, and if you wait too long it might be too late to treat.

I had previously worked with one of our mentors, Dr Jim Black from the Gustav Nossal Institute for Global Health on several smaller projects dealing with childhood pneumonia. These projects were however done before I had any real clinical exposure to the disease and now finally seeing the effects of it in real life, I felt the need to try and solve this problem.

So I set out to build StethoCloud. I felt that if we were able to enable health workers with limited training or exposure to diagnose pneumonia through the use of a widely available device such as a mobile phone and an affordable peripheral such as our digital stethoscope we would be working to significantly address the problem of children presenting too late with pneumonia.

So, how did you come together as a team?

So the formation of the team was a very organic process. As a result of this organic process emerged one of the strongest aspects of the team; our strong cross-disciplinary make up.

Initially I was the only member of the team and I started work on the app and the cloud backend in March. As time went by, I enlisted the help of Andrew Lin who recently graduated from medicine but is now currently working as a management consultant to help me organise, run and pitch the project while I focused on the technical challenge of writing the Windows Phone app and the Windows Azure backend.

The next two members of our team, Kim Ramchen and Mahsa Salehi, joined us when I realized that I had reached the limit of my knowledge of programming and computer science when it came to writing our BioSignals Algorithms framework. So I enlisted their help to come up with an algorithm using the latest techniques in artificial intelligence and machine learning to analyse the breath sounds and to come up with a diagnosis.

You’re all studying at a tertiary level as well as being involved in the Imagine Cup. Were there particular lecturers or tutors that were instrumental in helping the Stethocloud project come to life?

We have mentors from diverse backgrounds such Public Health, Pediatric Infectious Diseases and Computer Science. All our mentors were instrumental in helping StethoCloud understand the problem at hand and the development of solutions to address them.

Dr Jim Black from the Gustav Nossal Institue was instrumental in helping us understand the problem from the developing world’s perspective. Professor Nigel Curtis helped our team understand the pathology of pneumonia and the pediatric aspect of it. Professor Chris Leckie helped Kim and Mahsa understand the signal processing and artificial intelligence aspect of the problem.

Dr Pieter Peach has been a personal mentor to me. Pieter and Professor Nigel Curtis and amongst the few individuals in medicine who truly understand the potential benefits of combining technology and medicine and it’s their optimism that keeps working on this project.

What’s your development background, in terms of both platforms and development environments?

I am a mobile and web developer. I write native Cocoa iOS, Java Android and C# Windows Phone apps. I also write backend systems in the cloud. For this project we built our backend in .Net using the Windows Azure Cloud Computing Infrastructure.

Our algorithms were prototyped in MATLAB and were subsequently translated into C# .Net to run natively in the cloud.

The video for Team Stethocloud highlights your geeky attributes, which is wonderful –  but which member of the team is the most geeky (and why?)

Kim. Surely you can’t beat SynthPop DJ and geek combo, right?

Looking beyond the Imagine Cup finals, what’s next?

We’ve so far completed three early prototypes of our StethoMic peripherals. We however are constantly striving to improve our project and we are currently working on building the next version of our peripheral that will utilize smaller, more sensitive and noise cancelling microphones to hopefully return a result with better Signal and less Noise.

The StethoCloud Windows Phone app is complete and is also localised and translated to French to cater for users in Francophone West Africa where the burden of disease for pneumonia is one of the highest in the world. We are hoping to localise the application to Spanish to cover most of Latin America. Additionally, we also realize the importance of providing a solution that does not require the availability of Internet connectivity and hence we are working on an innovative solution to transmit the audio signals via the use of a telephony stack on Windows Phone and textual data through formatted SMS.

Our cloud infrastructure on Azure is complete and we are able to store, catalogue and process audio data transmitted from the device and return a result back to the user. I believe our philosophy of constantly striving to provide the safest and most effective care for our users means that this component will never be complete. We continuously update our algorithms in the cloud to be better and safer as we accumulate more data from research clinical trials. Additionally, we hope to soon start experimenting with more sophisticated Machine Learning algorithms and data mining techniques to enhance our detection rate. We are planning to submit our research protocol very soon to The Royal Children’s hospital for ethics approval. This will enable us to begin taking audio samples from live patients.

What advice do you have for Aussie developers out there?

Find a problem that you’re passionate about and build it. Also find people in the field who are “champions of technology” as they will offer invaluable advice about the current problems that they face and how that could be solved with technology.

Medicine is particularly notorious for being an industry that shuns technology. However if you look hard enough as we did with StethoCloud, you will eventually find your champions.