Tell us about your history in healthcare and how Jembi came about.
I’m a Biomedical Scientist by training, with a PhD in Medical Biochemistry from the University of Cape Town’s (UCT) Medical School and Postgraduate Degrees in Computer Science and Software Engineering from UNISA. I’m also an Honorary Associate Professor in the School of Mathematics, Statistics and Computer Science at the University of KwaZulu-Natal and an Honorary Research Associate in the School of Public Health and Family Medicine at UCT.
I’ve worked in the public and private sectors in Biomedical and Informatics research, as well as Information Technology Management and Software Development for over 20 years. Before I founded Jembi 10 years ago, I worked for the South African Medical Research Council (SAMRC) where my interest in digital health systems started. At the SAMRC, I started out on a number of projects working with the National and Provincial Departments of Health, as well as other local South African organisations. At one point I broadened my focus into global health informatics and started working in other countries initially Mozambique, Zimbabwe and Rwanda, mostly on research projects with other universities, collaborating on building eHealth and mHealth systems.
One of my objectives was to look at the long-term sustainability of informatics projects, and that gave me an idea for creating an entity which eventually became Jembi Health Systems. Currently, Jembi has about 75 staff members in three countries where we carry out our mission to strengthen health systems within national departments or ministries of health.
What’s Jembi’s role in developing open source solutions?
Jembi has been involved in open source software development since its inception. Almost all the software and systems that we develop are either open source or open content and we think it’s a very important paradigm for low resource settings.
Despite starting out as a niche activity often undertaken by people in universities or in the NGO sector, open source has now become mainstream. It’s an established way that software is developed and delivered internationally. Many enterprise systems are based on open source software or have open source components that make up a software stack. It terms of healthcare in low-resource settings, some of the most important systems that have reached scale are open source.
Why do you think that is?
I think it’s because open source can meet certain specific needs of healthcare in low resource settings. The free and open license often lowers the barrier to entry in terms of costs and procuring such systems, compared with commercial alternatives where the license costs and procurement processes can present a significant barrier.
Another reason that industry started adopting open source is because it has matured to the level where it is comparable in functionality to commercial solutions and able to satisfy the requirements that we have in the public health sector to a large extent.
Open source moves the cost dynamic from being cost-for-product, where you are buying a product and then often adding to that the costs of consultants or people to implement it for you, into a model where you are essentially paying for service. Often the service component associated with any kind of software is substantial. That’s why it’s not accurate to say that open source software is free, it’s only really the licence cost that is free. There are other aspects of open source that are potentially more important, such as collaboration and how you can build skills and capacity in the country, but at this stage a lot of the interest is cost-driven. The simple reality is that without open source, there wouldn’t be an affordable alternative.
How does open source actually work?
Much open source software or code is developed by a combination of paid developers and volunteers, people who have dedicated a lot of time to move this forward. Often they are funded by their employers who, in turn, use the software and derive a benefit from the contributions of others. Usually a single organisation or community curates the core code base.
There are many different open source models and different types of licences, some of which require you to keep the software open, and there are other licences which allow you to close off pieces of the code. The type of open source that Jembi has developed supports the code being kept open and community driven.
What that means is that there is a large community of developers and implementers and people with an interest in that software who basically maintain it going forward. Open source code hasn’t been around long enough to prove that it’s sustainable, so having the community aspect to it is one of the approaches to trying to achieve sustainability in the future.
How can the public sector can take advantage of open source solutions to build skills capacity and contribute to developing locally tailored national systems?
In South Africa, some of the important public sector systems in use are open source, so it’s certainly happening in the public sector. One of the benefits is that open source lends itself to collaboration and to capacity building. People with limited training in software development are able to join these communities and it’s a very fertile training ground for software developers because in return for contributing to the code base, they get mentorship. In low resource settings, this can become an effective way to build capacity within countries.
I think that a lot more could be done to leverage this kind of resource – the government or Ministry of Health often tries to do software development in-house, which has proven to be challenging. Open source software and communities can become viable alternatives to in-house software development and can really change the way that the government approaches system development. Instead of trying to build or recruit and retain engineers in-house, it can look to open source communities to provide that basic engineering skill and it can provide more of the coordination function that they are best placed to do. This approach changes the dynamic and could very well be the key to developing more appropriate solutions.
Is there a way that open source can help organisations achieve interoperability more cost effectively?
Jembi has developed a specific open source interoperability software solution that has been used effectively in several countries. In addition, open source itself definitely promotes interoperability by making it easier to interface between different systems. However, one of the main reasons for considering interoperability is to be inclusive of all kinds of software systems, including both closed or proprietary systems and open source systems, as long as there is an open interface.
Interoperability requires an open architecture, and closely related to that is the concept of open standards, which are basically used to communicate between different systems. If you have a proprietary system that has an open Application Programming Interface (API) or open Data Exchange, it can be interoperable and effectively, an open source system. Propriety systems are sometimes less inclined to be interoperable and developed more as a total solution but it’s not so much a technical issue as it is a business model.
Do you think the health system overall is mature enough to accept that there isn’t a single solution that addresses every need within a particular site and beyond?
I think this is true and core to the interoperability discussion is the fact that systems that are made up of interoperable components tend to be more robust, healthier systems than those which are built in a monolithic kind of way. I think that’s pretty well established, and a great way forward for us in South Africa and other African countries because it helps us make use of or reuse the investment that already has been made in existing systems. Leveraging those systems and getting them to be interoperable is more efficient than trying to redevelop them. In our experience, it hasn’t been that challenging to adapt multiple systems that have been developed independently of one another for specific health solutions, and make an interoperable system out of them. Interoperability holds a lot of promise to strengthen the systems by connecting up different pre-existing software applications, both propriety and open source.
Are there any regulations around using open source locally? What do you expect from regulatory developments particularly around standards in the run-up to the first phases of NHI?
There are policies and regulations that the government has developed around the use of open source that state, wherever possible, that the government should use open source solutions. However, I don’t think they were implemented that widely within government. Probably the most significant piece of legislation is the National Health Normative Standards Framework for Interoperability in eHealth (HNSF) in South Africa. That was a game changer in terms of entrenching the open architectural approach and open standards. The HNSF was promulgated as an extension to the National Health Act in 2014.
One of the challenges is how rapidly government is able to actually implement that kind of legislation. However the HNSF is the key to the way forward as it’s a very advanced and progressive piece of legislation that South Africa is fortunate to have. The HNSF sets the framework for interoperability and it goes further in actually specifying Health Information Exchange (HIE) and an enterprise architecture as central to the implementation of interoperability going forward for the public sector.
It also creates an obligation for the National Department of Health to create a National Health Standards Authority, which would set the different interoperability and content standards. The HNSF is just a framework; it’s not meant to be all things to everyone and therefore needs to be extended further in order to actually be practically useful in the context of South Africa. The HNSF could go a long way in strengthening the systems that are already in place, and setting the architecture for a new generation of health information systems that are interoperable and work together. At the same time, the NDoH could develop extensions to the HNSF and an enterprise architecture and other framework documents to guide people who are developing software.
Setting these kinds of policies and framework documents, which guide everyone and set the context within which people would go out and develop solutions, and ensuring that these all work together and become part of the public health system in general is a really sound approach for the government to take.
That kind of architecture would be very similar to what would probably be a recommended system for the private sector as well. I’m sure many of the health insurers will be looking at developing their own interoperability strategies so that they can efficiently move data between healthcare providers, facilities and administrators, and all the different players within the health insurance sector. They already have the issue of members switching providers so at a fundamental level, it seems natural that they will also be guided by strategies and the likes laid out by HIE to make themselves interoperable. Of course, extending that into the public sector would make a lot of sense, too. That is probably one of the reasons why the HNSF was promulgated, because it would serve as a good back-drop to the integration of the public and private sectors, at least at a technological level. While significant challenges may exist to doing that at other levels, at a least at a technical level it would serve as a very useful framework for making public and private health sector’s information systems interoperable. Once you have that, then that could be a parcel platform on which to start addressing some of the other integrations that are contemplated in the NHI White Paper.
Do you think as a country we’re heading in the right direction to making a more accessible and affordable healthcare system?
I support the notion of universal health coverage. We all believe in that, and we’d like to see the health system extended more equitably to people across the whole of the country than it is now. It’s really a question of how that gets done – that’s where there seems to be significant differences of opinion as to what the best model is to achieve that.
We’ve seen the NHI model work in other countries like Rwanda, which has a pretty well implemented national public health insurance system that covers 98% of the population. Of course, there are lots of issues with implementation and challenges with creating systems which make it work effectively, but it is working there.
Can we learn from that?
The challenges in South Africa are that we have a more developed and westernised kind of infrastructure; there is already a commercial market around health insurance which is going to obviously impact on the implementation of a public system. In other African countries, they don’t have such a robust private sector so it’s often easier to do. But one can look at those examples and try and understand at least from our level, where we are working from on health information systems. We can look at those countries and see how they have implemented their systems, see what’s working and what the challenges are, and learn from that as we start to do something similar in South Africa.