Last week I attended the Idea Jam Night for Random Hacks of Kindness (RHoK) here in Toronto. The event was held in preparation of our upcoming hackathon on June 2nd and 3rd. In a hackathon, subject matter experts submit problems they are facing in their field of expertise where there is potential for technological solutions. Over the course of a weekend, they are grouped with interested developers to work on building a working prototype of the proposed solution. I am happy to be a part of the organizing committee for this hackathon and even happier that one of the problems presented was healthcare-related. The idea was submitted by Leanne Tran who conducts radiology research in Toronto.

The problem:

Approximately 80% of maternal deaths globally are due to obstrectic complications such as haemorrhage, sepsis, etc. 99% of those deaths are in developing countries, and most of them are highly preventable with the right obstetric care. (Chuni, 2008). Taking Nepal as an example, obstructed labour is the second highest leading cause of maternal deaths. Often these soon-to-be mothers do not have access to specialized care and don’t discover potential complications such as feto-pelvic disproportion (when the size of the fetus’ head is larger than the pelvic) until it’s too late. The problem as outlined by Leanne was two-fold 1) lack of specialized training in operating ultrasound devices, as the accuracy and quality of ultrasound scanning is highly operator-dependent and 2) lack of access to radiologists and OBGYNs to interpret those scans.

The Analysis:

Several process-related questions must be asked prior to discussing potential technological solutions, none of which I have the answer for at the moment but hope to know before the hackathon:

  1. What type of ultrasound devices are used on the field in remote regions?
  2. Do these devices have the interface to export the ultrasound images? If so, in what format?
  3. Are there multiple points from which an optimal ultrasound image can be taken? If so, does it depend on the age of the fetus?
  4. What are the most common perinatal complications in each trimester?
  5. Can a clear diagnosis of an emergency case be made solely based on an ultrasound image? If not, what additional data elements need to be attached to that image?
  6. If the solution is to be implemented online, what measures must be taken to ensure the privacy and confidentiality of Personal Health Information (PHI)?

Answering the questions above will significantly help in shaping how the solution is to be designed.

The Proposed Solution:

Leanne would like an mHealth solution that will help guide and train the ultrasound technicians on the best practices and techniques for scanning, with a focus on pregnancy. Assuming our solution will be a smart phone application, the first can be accomplished by providing instructional videos. If the approach/angle differ based on the age of the fetus, or any other factors, perhaps the app can start by presenting a couple of questions to determine which training video to display. Of course a more advanced approach would be the use of image processing technologies (much like the ones used in cameras to detect faces) to detect the pregnant belly. Clear instructions can be given to the technicians before starting the scan about the optimal location from where the picture should be taken for the app. Once that picture is taken and the app identifies the location and proportions of the belly, a new layer is added to the actual image containing the ultrasound scanner and where to place it to get the optimal image(s).

If this is captured before labour, it could help save the mother’s life by getting her to the hospital or the specialized care she needs before it is too late. The technician would simply upload it to a website and email it to a specific doctor/hospital participating in the program where they can review it and send back further instructions to the technician.

Personally, I see potential value for using Ushahidi and Crowdmap in streamlining the process of communication if we take out the diagnosis automation component.

The Call for Help:

The hackathon is taking place on June 2nd & 3rd  and we need developers to work with us on a prototype for this system. We need experienced Mobile App developers (iPhone, Android, or Blackberry) who can design a phone app that allows for bi-directional data transmission as we need to be able to send information from the phone to an online database, then send instructions back from the database to the same phone that transmitted the data. We also need web developers, preferably with experience in Ushahidi integration with mobile apps. If you are interested, please sign up here or contact me over Twitter @YaserAlyounes

Like this post? Subscribe to my RSS feed and get loads more!