Lessons in Interoperability

Posted by KatrinVerclas on Oct 23, 2009

This post is reposted from the Open Mobile Consortium, a community of mobile technologists and practitioners working to drive open source mobile solutions for more effective and efficient humanitarian relief and global social development.  It is written by Neal Lesh.

One of the Open Mobile Consortium's (OMC) primary goals is to maximize interoperability and data-sharing capabilities among the open-source mobiles technologies so that the whole of our collective effort is greater than the sum of the individual parts.

Think of the OMC technologies as a set of Lego-like building blocks that can be snapped together in different ways. It's a great goal in theory, for sure -- but in practice it is more difficult to achieve. There is redundancy, and systems that could and should share data may not.  Being open source or championing open standards is necessary -- and we certainly do – but it's not sufficient.

Building the Open Source Mobile "Lego" Pieces for Social Impact

This is a story about progress towards integrating several OMC technologies in Tanzania and some of the lessons we are learning along the way. It illustrates the potential of the OMC ‘Lego’ pieces for mobile for social impact and where we still need to do more work.

D-tree and Dimagi, two members of the Open Mobile Consortium, have been developing and implementing mobile health projects in Tanzania since 2008. We are running studies and pilot deployments with a wide range of NGO partners including Pathfinder International, the Ifakara Health Institute, PATH, Catholic Relief Services, IntraHealth, and BRAC.

Our focus is on providing phone-based tools that offer guidance and decision support to health workers while they are seeing patients. For example, our system can guide a community health worker through a series of questions to assess if a newborn needs to receive emergency care, and if so, remind the community health worker to return to make sure the family followed through and that the problem was addressed.

One of the greatest challenges in Africa is the critical shortage of human resources. Estimates suggest that there is a shortage of 4.2 million health workers worldwide but it's most severe in sub-Saharan Africa. This is where Community Health Workers (CHWs), local volunteers, come in. They have an important role to play in providing services to the poorest and most vulnerable populations.

Our work and technologies are aimed at providing as much support and advice as needed to increase the capacity of the CHWs to provide quality medical care.

All of our pilots are built on one OMC technology: JavaROSA. JavaROSA is software that runs on Java-enabled phones. It supports the collection of data on a mobile phone in a structured format and allows the data to be transmitted back to a central server.

JavaROSA is part of another consortium, the OpenROSA consortium, that also includes several OMC members. All these groups have agreed upon a single representation (called XForms) for our forms, questionnaires, and protocols—one step towards creating interchangeable Lego pieces.

Building On Each Other's Tools: Integrating RapidSMS

Over the last few months, we have incorporated two other OMC technologies, RapidSMS and Open Data Kit (ODK), into our work. Earlier this year we began thinking of more and more use cases where we needed SMS. For example, we wanted to be able to send alerts when our community health workers had not submitted any data for a period of time.

We also wanted to be able to broadcast SMS messages to help coordinate and motivate our users. The developers who were building the central server, called CommCareHQ, were familiar with RapidSMS and integrated it into our server. An administrator can now log into CommCareHQ and select some or all of our users to send an SMS message to.

For example, we can type in and send a message to all 42 of the home-based care providers in our pilot deployment with Pathfinder International, such as "Meeting changed from Friday to Thursday. Can you make it?" Any replies community health workers send on their phone, in turn, are captured and displayed on our back-end server.

We are experimenting with a wide range of other SMS-based services as well, such as alerts described above and weekly summaries showing which community health workers have reported on the most home visits this week.

But we want to go even further and use SMS with the patients directly. For example, during an assessment of an ill child, our phone-based tool will prompt the health worker to, in turn, ask the child’s care taker if she has a phone number we can use to send reminders or educational messages to help with the child’s recovery. If so, then our system will begin to send SMS messages to the phone number that is entered into our JavaROSA client.

The important point here is that we are able to leverage the work by RapidSMS to integrate these functionalities into our system. As more features are built into RapidSMS, we will be able to utilize those as well.

Testing Android and ODK: Another Building Block

The second OMC technology that we have incorporated is Open Data Kit (ODK). ODK Collect, the mobile component of the kit, is similar to and shares code with JavaROSA, but runs on Android phones. While more costly than the phones JavaROSA runs on, these phones provide a richer user experience and include functionality including GPS, barcode scanning, audio/video playback and capture.

While we favor the less expensive and more common phones for our projects involving community health workers, we want to experiment with other platforms for our projects based at health facilities. For example, we are developing a program called e-IMCI that provides step-by-step guidance for a clinician to apply the Integrated Management of Childhood Illness (IMCI) protocols for classifying and treating child illness.

A few months ago, we finished our preliminary encoding of this protocol (as an XForm) for use on JavaROSA. We soon sent it over to our colleagues at ODK, primarily to see if it would work. When it did, and we realized that it looked better than on JavaROSA, we began to make plans to test in the field. We have encountered a few technical problems and incompatibilities along the way, but with the help of the ODK team, are actively addressing them.

We are going to experiment with the ODK version of e-IMCI and expect that it will be preferred by the clinicians using it. We imagine that there will be both JavaROSA and ODK versions of e-IMCI in use, all reporting data back to the same central server which will be merged into the same reports. This data will then be used to trigger SMS alerts and reports sent and managed by RapidSMS.

Interestingly, this integration of OMC technologies happened almost without our noticing it. In this case, our systems are starting to work well together because we are working together.

We expect more and more of this to come together in the coming year. For example, when we get to the point of needing to synchronize the information on our handhelds, Mesh4X is the natural solution and it will be easy to try. We expect to use authoring tools and other services from openXdata which is hard at work in nearby Uganda.

Regardless of some of the hurdles, it is clear that some of OMCs pieces are starting to click together to improve healthcare delivery in low-income countries.

Qualcomm Creates Subsidiary for Open Source Development

Here's a related article. I didn't know Qualcomm has a near-monopoly on CDMA patents:


Post new comment

The content of this field is kept private and will not be shown publicly.
  • Allowed HTML tags: <a> <em> <strong> <cite> <code> <ul> <ol> <li> <dl> <dt> <dd><p><br> <b><i><blockquote>
  • Lines and paragraphs break automatically.

More information about formatting options