The bridges.org/IICD Case Study Series on ICT-Enabled Development sets out to illustrate how ICT contributes to development in Africa. The aim of this series is to help ground level initiatives imagine the possibilities of what can happen if they use ICT successfully to overcome development obstacles, and to contribute to the existing body of knowledge on the digital divide.

I. Overview

Initiative: The Tsilitwa telehealth project uses a system in which a nurse can send live pictures of a patient over a wireless network to a remote doctor, and at the same time they can speak to one another using a Voice over IP (VoIP) phone. The project involves researchers at the University of Cape Town (UCT) developing software applications to meet the needs of local health workers.

Implemented by: The wireless network, including VoIP was set up by the Centre for Scientific and Industrial Research (CSIR). The telehealth project is being run by researchers from the computer science department, UCT.

Funding or financial model: Funding originally came from the Department of Trade and Industry and now comes from the Department of Health. The telehealth project is also funded by Telkom and the South African - Netherlands Programme on Alternatives in Development (SANPAD), with more funding in the pipe-line from the World Bank and the International Development Research Council (IDRC).

Time frame: The wireless LAN was set up in 1999 and the telehealth project in March 2003.

Local context: Tsilitwa lies approximately 20km along dirt track from Qumba, which is a small town 30km north of Umtata in the Eastern Cape. The local population is poor because there are few jobs and little subsistence farming. For many families the only source of income is often a disability grant given to one family member. This is a high crime area, blamed largely on bored, unemployed youth.

The development problem/obstacle addressed: There is no doctor at the clinic in Tsilitwa. If a patient has symptoms a nurse cannot recognise, she has no choice but to send him to the hospital in Sulenkama some 20km away. As transport is expensive and does not run directly between the two villages, the patient may be forced to walk. Further, there are no phone lines in Tsilitwa, and only one in Sulenkama, making communication between the hospital and clinic extremely difficult. In addition, electricity blackouts are almost a daily occurrence in this area.

How ICT is used to overcome the problem: The CSIR set up a wireless local area network (LAN) in 1999, which connects the hospital in Sulenkama with the clinic in Tsilitwa, the local school, a community centre and police station. It uses wireless fidelity (WiFi, otherwise known as 802.11b) signals with a booster between Tsilitwa and Sulenkama, as Sulenkama lies in a slight valley. At present just the clinic and hospital use the system in a telehealth project being pioneered by researchers at UCT. The system works as follows: a nurse at Tsilitwa focuses a web camera on a patient, the picture being displayed on a computer running Microsoft Internet Explorer. In Sulenkama the doctor opens up an Internet Explorer window which is configured to load a web page associated with the IP address of the web camera and hence the image. At the same time they are able to speak to one another over the wireless network using VoIP. The doctor is then able to advise the nurse as to a course of treatment or possible referral of the patient to a specialist elsewhere, thus saving time and money. Sending the image over a wireless network, as opposed to a fibre or wired network provides a cost effective means of communication that is appropriate to the rural setting. In addition, the use of VoIP means the clinic is only charged for a local call.

Next steps: Researchers Bill Tucker and Marshini Chetty have visited the site twice and spoken to all involved to find out what enhancements can be made to the system. Based on these needs, they will design and write a software application for the telehealth project. Thus the outputs of this project will be a working application the success of which may allow UCT to influence telecommunications policy; an MSc; a Phd and a methodology that is replicable.

Geographical area targeted: Some twenty square kilometers of rural land around Tsilitwa in the Eastern Cape. The clinic servers around 10,000 people and the hospital is the only one in the area.

Contact information:
Contact person: Bill Tucker
Email: btucker@cs.uct.ac.za
Computer Science Department, UCT

II. Gauging Real Impact

This section considers whether and how the initiative has made a Real Impact at the ground level by looking through the lens of basic best practice guidelines for successful initiatives. The bridges.org 7 Habits of Highly Effective ICT-Enabled Development Initiatives are used here as a framework to highlight what the initiative has done well.

The 7 Habits of Highly Effective ICT-for-Development Initiatives

1. Implement and disseminate best practice
The telehealth project was first designed to use critical action research: a system of successive rounds of consultation with local stakeholders as part of the development of an ICT solution. As the project got under way, it was realised that this methodology resulted in the project being delayed from interference by local community leaders with their own agendas and it did not address the human-computer interface sufficiently. The critical action research element was subsequently modified in favour of establishing better rapport with a larger number of local stakeholders.

2. Ensure ownership, get local buy-in, find a champion
The CSIR trained two local people to look after the wireless LAN and provide technical support to health workers for the telehealth project. When interviewed, nurses and and the doctor at Sulenkama could clearly see the benefits of the system and were keen to use it.

3. Do a needs assessment
Local health issues have been discussed with nurses and the clinic, the doctor at Sulenkama and the technical support team. The area has been visited for a few days at a time so that researchers could fully appreciate the conditions under which people in this area live. In addition, ideas garnered from an IDRC workshop entitled "first mile modalities" were used to asses local needs.

4. Set concrete goals and take small achievable steps
There is a project plan with clear mile-stones. A record is kept of all visits to Tsilitwa, in addition to all conversations held with health workers and other stakeholders. The first goal was to resolve issues with sending images over the wireless LAN and sound quality of the VoIP telephone. Once these were resolved the research team made a number of suggestions to health workers based on their observations of seeing the system in use, with the aim of developing a software application that would help them.

5. Critically evaluate efforts, report back to clients and supporters, and adapt as needed
The project has an ethos of trying to fully understand local issues before suggesting solutions. To do this researchers have met with a number of local stakeholders, such as community leaders, teachers and health workers. The clinic has been visited twice to find out how well the system is working and the research team report back to nurses on the progress they have made in writing a telehealth application. It is intended for this two way process to continue for the duration of the project.

6. Address key external challenges
Power outages occur on an almost daily basis, such that the computers had to be placed on uninterrupted power supplies (UPS). Unfortunately outages can last more than the life time of a UPS battery, so the research team are exploring ways in which to implement a store and forward means of accessing data. This would also be useful for when the wireless LAN is not available, either due to a power outage or an antenna being blown astray in high winds. The nurse could take pictures and notes of a patient and store them on the hard drive of the computer, to be sent to the doctor, along with a time stamp, when services are restored. Theft is a huge problem in the area: Telkom regularly replaces stolen or vandalised solar panels that power the batteries of public phones, only to have them stolen again within days. The wireless network has avoided theft so far, the antenna being powered by mains electricity. The remote location of Tsilitwa has made it difficult for researchers to visit and there are some concerns as to their safety. To address this, accommodation for researchers has been specially built and a local watch man can be obtained. Getting the buy-in of the local community leader -- who is coincidentally the headmaster of the local school -- has been difficult for reasons unknown, and had some impact on the wireless LAN not being used as much as it could be. Researchers for the telehealth project have been careful to avoid upsetting this person. Finally, local politics is potentially an issue: if a development project is seen to be associated with a particular political party then supporters of an opposing party will try to undermine it.

7. Make it sustainable
The telehealth project has attracted an impressive amount of funding so far, but long term sustainability needs to be addressed. Researchers hope to get the support of the Department of Health in that they will recognise the cost savings from this system and want to put it into their budget.

III. Lessons Learned

We invited Bill Tucker, co-researcher and supervisor of Marshini Chetty for the Tsilitwa telehealth project, to share his views on the greatest success, the challenges faced, key constraints and dependencies that affect the project, opportunities for future improvement, and other lessons learned. This is what he said:

So far, most lessons have come from initiating a project with a rural community. We learned that things take a long time! In the beginning, we were a bit naive thinking we could directly engage a
community in order to come with ICT requirements. We found ourselves involved with our particular communities by going indirectly, via a research effort that was already in place. Thus, we were able to come into a community with a bit of momentum. That momentum helped us define the technological requirements for our work by building on what was already there. More importantly, this "piggy back" helped to lay the groundwork for a trusting relationship with the community. In particular, I have learned that it's just as important to spend "quality" time with the community-based ICT activists as it is to discuss the research with them. I can't over emphasize enough how important it is just to become friends with the people with whom you are conducting the research. The trust that comes along with friendship is perhaps the key ingredient to having them put an actual investment into the research. We were also fortunate to be able to meet members from other communities that have similar needs and experiences, as well as the researchers working in those other communities. This came about due to a funding initiative, a workshop. This was instrumental in putting together a bigger picture of what we are doing. Taken together, we learned that a much wider approach to defining "stakeholders" in the project meant a much wider appreciation of the kinds of problems we need to deal with.

IV. The Story

This section presents a narrative description of the initiative that highlights why this use of ICT for development is particularly interesting.

Tsilitwa is a sprawling community in the grassy hills of the Eastern Cape, approximately 50km north of Umtata. Unemployment is rife and crime is high in this area: there are few phone lines because the solar panels and batteries that power them are stolen each time they are installed. The local health clinic servers around 10,000 people and relies on a doctor some 20km away at the Nessie Knight hospital in Sulenkama to treat patients nurses at the clinic cannot help.

When we visited the hospital, there was one Cuban doctor to see 200 patients a day. There was no electricity at the time; no water supply as the water tank had not been refilled; and no phone lines as the generator which powers the one line supplied to them (a wireless DECT phone) had run out of diesel. The doctor was clearly stressed and had to use his personal cell phone, at his own expense, to arrange appointments for patients to see specialists in Umtata. Against this backdrop, we were amazed that we could always pick up a cellular signal and Bill Tucker was able to connect to the Internet using GPRS through his cell phone.

The telehealth project piggy backs on a wireless LAN installed by the CSIR in 1999, which connects the hospital to the clinic in Tsilitwa, the local school, a community centre and the police station. The nurse at Tsilitwa focuses a web camera on a patient, the picture being displayed on a web page set up for this purpose. In Sulenkama the doctor loads this web page, data being transmitted over WiFi, and can ask the nurse to refocus the web camera on the patient if need be. At the same time, they speak using a phone which uses VoIP. At present, only the telehealth project is using the LAN, which partially explains why the speed at which the image loaded at the hospital was so good. The line quality of the VoIP phone was also checked and proved to be excellent.

Bill Tucker and Marshini Chetty consulted with health workers and local technical support to find out how the system could be improved and found out that nurses at the clinic would prefer a two way visual communication if possible; that the image should be made sharper and that a means of addressing the effects of frequent power outages would be useful. They plan to develop a software application that will address some of these issues and are exploring the use of laptops with long battery lives and that come with inbuilt cameras, as a means of circumventing the power outages. Despite the doctor having little time to sit behind a computer, he nevertheless appreciated the value of the system and would be keen to use it providing it was quick and easy to use.