“The potential of new technologies to re-energize the AIDS-movement is clear. We need nothing less than an HIV prevention revolution, with social media and mobile technology at its core,” – Michel Sidibé, Executive Director of UNAIDS
Part #3: Case studies: Evaluation of Ehealth projects in Uganda and Zambia, short, long term costs and benefits
Despite the rise in global mobile subscriptions there is limited evidence that health technologies are a successful and innovative tool to improve the quality of healthcare in developing countries (WHO, 2012). However, mobile technology may be particularly useful in resource constrained settings such as health care environments found in rural regions of developing countries including Uganda and Zambia. Increased affordability and wider network coverage has meant mobile subscriptions have reached 5.3 billion globally, with developing countries accounting for 73%. Uganda has seen incredible growth and now has approximately 16.7 million cell phone subscriptions or 48.38 per 100 inhabitants while Zambia has 8.1 million subscriptions or 60.59 per 100 inhabitants (International Telecommunications Union, 2011). Rural Uganda is struggling with HIV rates which are on the rise, with an average of 9.9% of prevalence in rural areas and 6.4% of prevalence as the national average, Uganda could benefit from further prevention efforts in rural areas. In addition prevention efforts need to be addressed with both sexual partners as women have a higher prevalence rate then men at 7.5% verses men at 5%. Although women however are more vulnerable to the disease due to lower levels of knowledge in comparison with men, and have less freedom to make their own sexual-related decisions (Chib et al,. 2012). Zambia differs from Uganda, in urban areas the percentage of individuals infected were approximately twice as high as the rural areas (RZMoH, 2011). However, similar to Uganda, HIV disproportionately affects more women. Women have a national average of 16.1% while approximately 12.3% (2010 statistics) of men are affected (Unicef, 2010).
This section will look at two different programs which have began to address HIV prevention strategies in Uganda and Zambia. Text to Change an HIV prevention initiative will be examined by looking at two programs, one among youth in the Arua District of Northern Uganda while the second looks at factory workers among three factories in Northern Uganda. Project Mwana will be examined in Zambia for their recent prevention efforts involving mobile phones and increasing attendance of expecting mothers to rural clinics.
Uganda: Text To Change Campaign
Text to Change (TTC) is a Dutch NGO that promotes health education specifically HIV/AIDS education via mobile phones through a short message service (SMS) quizzes. Included in their objectives is data collection, increasing awareness of HIV/AIDS, advocating testing and counseling behaviors pertaining to HIV/AIDS and testing the efficacy of incentives to participate. The TTC HIV/AIDS campaign was designed to increase knowledge and awareness about HIV/AIDS as well as to promote the regional clinic and testing centers. In conjunction with MTN a local leading telecommunication company, highly active subscribers were identified. Between January 29 and February 27 2009, text messages with HIV/AIDS related multiple choice, and true or false questions were distributed via 10 000 mobile phone numbers. When participants answered questions correctly, they received free HIV Counseling and Testing (HCT) services with the Aids information center Arua Branch, and were entered to win weekly prizes including mobile phones and airtime.
–> TTC VIDEO
TTC works in partnership with other healthcare NGOs in Africa to develop appropriate medical content for the quizzes. The quiz format was intended to provide a fun way for participants to engage with educational content. The quiz consisted of 13 questions which were sent via SMS to 10 000 subscribers falling into three knowledge areas (a) HIV/AIDS disease, (b) testing, and (c) HCT services (Chib et al., 2012). HIV knowledge questions and response choices included the following; what causes AIDS? HIV is not present in: sweat, semen, blood or breast milk? How can you tell whether one has HIV? HIV weakens an infected person’s immune system (true or false)? Do you think a healthy looking person can have HIV (true of false)? Other questions centered on testing such as, why is it important to test? Or, if you are exposed to HIV, how long should you wait to get tested? Finally HCT knowledge was tested through questions such as, where is the AIDS Information Center Uganda (AIC) in Arua? Is HIV testing at AIC accurate and confidential (Chib et al., 2012)? Once TTC sends a question to participants, participants then answer the question. Next the TTC system checks the answer and responds. If the answer is correct the participants receive more information about the topic. If the question is incorrect they’ll receive the correct answer and an explanation. This is repeated until all 13 questions in quiz are completed.
Initially, questions about demographics were asked. These demographics questions such as age and gender allowed TTC to determine who were answering the questions. The age range was 9-65 with the median age 26 and average age 28 years. Gender questions indicated that 421 were male while only 202 were female, however even though males were more likely to answer questions there was no significant difference between the number of questions. Of the 10 000 mobile numbers who were sent messages, 2363 numbers responded, of which 1954 answered quiz questions (the rest only responded to the age and gender questions). More than half of the responses replied to at least one question, while only 315 answered 7 or more. More then half (61%) answered at least one HIV testing knowledge question, while 45% answered at least one HCT service question. On average 68% of the total questions asked were answered correctly, with only 19% not answering any correctly. The questions that were answered the least also had the lowest number of correct answers. People who answered each question correctly were more likely to answer more questions than those who incorrectly responded (Chib et al., 2012).
Data was cleaned up by eliminating responses that did not use the proper formatting to respond to questions, multiple answers were eliminated from the same mobile number, and conversational responses were not accepted nor were late responses (Chib et al., 2012).
The short-term effect of this SMS campaign were an increase in the number of monthly visits for HIV testing. Chib et al., (2012) found that during the TTC campaign 677 people (376 men and 301 women) accessed HCT services at the Aura clinic, which was a 33% increase in the average monthly visitors. This included 364 people whom answered SMS text messages. Therefore, this is a good indication of the future of ehealth projects as a 33% increase in HCT services is substantial.
Limitations to TTC project
Limitations to HIV prevention via SMS TTC included cell phone usage. Although usage rates are on the rise, non-mobile phone users may have significantly benefited from prevention information. Technical limitations of this study include a 160-character limitation, and the restriction that participants must respond in a particular format. The participant must also answer the question in the language of the quiz, and therefore some individuals may know the correct answer but are unable to respond, or respond due to language and formatting barriers. Conversational answers with incorrect formatting that may have been correct did not count in the end, in order to maintain consistency in the results. The SMS campaign may be reinforcing a specific knowledge gap; such as those from high-income groups, or men benefiting more from the campaign due to better accessibility of these services and the ability act upon the information provided. However to determine this, the Quiz solicited for prior knowledge, if the responder answered a question corrected the proceeded to the next question, or if the responder answered incorrectly then the correct information was sent to them. However, the SMS campaign failed to inform those who answered any initial questions about gender or sex but failed to answer any quiz questions. Additional drawbacks of this SMS campaign included that when an individual answer a question incorrectly they were less likely to respond to the next question. This was problematic as those who answered the questions incorrectly were probably requiring additional information. All 10 000 subscriptions were to MTN, a leading telecommunications provider. By only using one telecommunication network this left out respondents from lower or higher literacy levels and or demographics to answer the question differently. Thus, including other networks might deliver different results, as would by increasing the number of identified subscribers. MTN has more than two million subscribers, yet only 10 000 were invited to participate.
A second short-term TTC project was examined using the similar quiz format as the project previously discussed, however the quizzes consisted of 18 true or false questions and involved three factories (Factory1, Factory2, Factory3) processing sugar and cobalt. The Factories were all located in factory towns, which provided housing, schooling and health facilities to their employees. There were 2 294 participants (360 Factory 1, 1294 Factory 2, and 840 Factory3) in total (Danis, et al., 2010). One of the major differences in this quiz was that multiple answers were accepted due to phone sharing among factory workers. However answers in the incorrect formant were eliminated, as were late responses. Expected results for the true or false quiz were expected to be 50%. Examples of SMS questions were; TTC Quiz Question: HIV can be prevented by using condoms correctly and consistently (1) true (2) false. Reply with “condom” and the number of your answer.
Accuracy and participation rates were above the expected values originally thought. For the first few questions accuracy rates were roughly 60% and after the first few questions increased to approximately 80%. The factory quizzes may have received higher results because they were able to answer the same question multiple times. Conversational answers were present as well. For the first question conversational answers accounted for approximately 5 % of answers, but decreased to approximately 2 % by the last question. Participation rates for the factory quiz were consistently high. One explanation for this outcome could be due to the extensive socialization among the factory workers regarding the quiz. Participation rates actually grew as the question continued to be sent out to factory workers. New participants became involved as the weeks progressed by requesting the quiz questions via the short code, which could be requested through participating individuals.
Short-term success for this project included a three-fold increase in HIV test requests from factory workers. Although we can’t assume that all of these new requests were due to the initiation of the TTC HIV prevention quiz it is reasonable to assume some of the increase was attributed to this.
Finally, we can conclude that increases in requesting HIV testing demonstrated positive outcomes in the factory TTC project. However, similar to the first TTC project in the Arua clinic area, a recommendation was that on-going evaluation must take place to determine the long term effectiveness. However as Tore Godal, special advisor to the Norwegian Prime Minister explains, e-health evidence may be lacking, however focusing on evidence rather than development is risky. Godal goes on to explain that, “imagine if Steve Wozniak and Steve Jobs had to provide evidence that the computer they made in 1976 was useful. They would not have been able to do so, but some investor believed that they had the potential for changing the world.” (WHO, 2012, pg. 331). The small short-term increases in visits to the HCT Aura clinic for testing, and increases among factor workers testing for HIV is positive reinforcement that on-going progress regarding ehealth prevention projects are making a difference. Although long term evidence assessing effectiveness may be helpful to determine how successful these projects are, it is not necessary to make a difference in HIV education, and clinic testing attendance.
Zambia: Project Mwana
Although Zambia has one of the highest cell phone usage rates per 100 inhabitants in Africa, programs dedicated to HIV prevention within Zambia are minimal. According to Donald Thea (2012) of the Boston University School of Public Health, Mwana is virtually the only project of its kind in Zambia that contains some sort of prevention effort using mobile technology. Elin Murless works for UNICEF on project Mwana and although she agrees with Thea, she has seen some new innovative ideas by partners of UNICEF but as of the August 9th interview nothing has yet to be implemented.
In Zambia it is estimated that 21% of new HIV infections are transmitted via mother-to-child. Antiretroviral’s (ARVs) are critical to help prevent the transmission from mother to child, but when this fails it is essential to ensure the infant is tested immediately for HIV. There is strong evidence to prove that if an infant is diagnosed early for HIV that morbidity and mortality can be significantly decreased. The main goal of Mwana is to reduce the time between blood sampling for the detection of infant HIV infection and notification of the test results to the relevant point-of-care facility by using an SMS based system (Seidenberg, et al., 2012). Clinics are widely dispersed and laboratories where blood samples are tested can be from 10 km to 600 km away from a district clinic, making it difficult to quickly transport samples and receive results. In April 2009, at Nameembo Rural Health Clinic an infant’s blood was collected for an HIV test. It took 66 days for the sample to travel to the lab and the results to be returned to the caregiver. In February 2011, a similar sample took 29 days to make the same trip. Mwana is the project which facilitated this decrease in travel time using ehealth technology (Republic of Zambia Ministry of Health [RZMoH], 2011). The Ministry of Health in Zambia, UNICEF, Boston University and the Zambia Center supported this project in Zambia’s Southern Province where 10 public health facilities (five in Mazabuka district and five in Monze district) were involved in this pilot project. The Zambia Center for Health and Research and Development is a local NGOs, also affiliated with Boston University also involved in this initiative (Seidenberg, et al., 2012).
Although the main purpose of this project initially was to speed up diagnosis times, preventive HIV measures were also an important part of this project during the second phase. The second phase, involved community health workers (CHW) who registered pregnancies, births, and maternal and infant deaths. Once the pregnancy was entered into the database the expected date of delivery was established. The CHW then received text messages to remind women to come in for visits to the local clinic, this encouraged mothers to come in for their first, second and third trimester visits, especially those who were travelling from long distances (Thea, 2012). During these visits a mother had access to many benefits preventing HIV transmission to her child, but at that point HIV education was also available to prevent the risk of transmission to the rest of her family. In Zambia more than 90% of women attending antenatal care (ANC) services are tested for HIV, thus the CHW could receive notifications from the local clinic and encourage expecting mothers to utilize services such as routine check ups offered by the Ministry of Health (RZHoH, 2010). Women are encouraged to visit the clinic within the first 14 weeks of her pregnancy. During the first visit to the ANC clinic, woman and their partners were encouraged to find out their HIV status if unknown (or if previously negative for HIV were then retested). Second, women could attend group health education to obtain information about child’s health and HIV status, preventing mother-to-child transmission, couples testing and disclosing results to their partner, and a wide range of other topics such as family planning. If a woman discovered that she was HIV positive she was recommended to begin ART treatment immediately. Visits were encouraged two weeks after first visit, then again within four weeks for the third visit and finally a visit in the third trimester. Follow up visits after the infant was born are critical as is HIV testing for the infant (RZHoH, 2010). CHW also kept HIV positive women accountable by reminding them of their visits to the clinic which was important especially to receive does of Co-trimoxazole which was recommend for HIV positive pregnant women after the first trimester. Comprehensive HIV care is required at minimum every four weeks during the pregnancy. If the mother lived far from the clinic she was recommend to utilize waiting homes close to their expected date to ensure a safe delivery (RZHoH, 2010). Results of this program according to Thea (2012) were impressive. During the first two months in 5 clinics, more than 800 pregnancies were registered. There are future plans to expand this project to all ten clinics as well as to include 200 additional health facilities in the next year (Murless, 2012).
Initial costs of this program were to train two individuals from each health facility, usually a nurse and a manager. The training consisted of a half-day seminar instructing individuals how to use the SMS-based system to retrieve test results. Local Zambian software developers created the SMS-based system titled ‘Results160’. The software securely and quickly delivered infant HIV results from labs to facilities using SMS messages. They were sent through a server which was centrally placed at the Ministry of Health. Ten mobile phones were distributed, one to each health clinic and used the SMS-based system between health clinics and laboratories to retrieve results. Support visits were made to each facility after two months of the program being launched to resolve any outstanding issues that might have arisen (Seidenberg, et al., 2012).
Current Role of Zambian Government
Currently the MoH is playing supportive role at both the district level as well as national level with the support of UNICEF and partners. After the pilot project was introduced, the Government of Zambia committed to a scale up of the project which would entail 587 health facilities to include HIV testing to infants. Which as both Thea (2012) and Murless (2012) acknowledged would be room for further growth, ANC services and ensure mothers had proper access to care to prevent the transmission of HIV. The National Government is currently seeking non-governmental organizations that are willing to help. The project has moved slowly Murless describes with the government’s involvement but ongoing support is being provided through UNICIEF to create what has so far been a successful initiative. UNICIEF believes the support of the government is key to ensure a sustainably run project in the near future, even if that means slow progression initially (Murless, 2012).
Both Uganda and Zambia exhibited pilot projects using ehealth technology, which directly and indirectly contributed to prevention efforts of HIV/AIDS within their countries. Both projects saw the implementation by stakeholders, Uganda made use of local health NGOs to develop appropriate medical content for the quizzes, however their was no published evidence of participants being integrated into the development process. Zambia used a collaboration of stakeholders including the government to move forward with their pilot project. Although no public evidence communicated involvement of Zambians within the developmental stage, Murless (2012) ensured that developers spent many weeks in the field including nurses and CHWs asking, designing, testing and redesigning a system that was easy to use. In the short term we can conclude that Uganda experienced a rise of 33% in monthly average testing in the HTC clinic Aura, while among factory workers there was a three-fold rise in requests made for HIV tests. While Zambia saw a large number of births recorded with the implementation of ‘Results160’ software. However, long-term analysis is necessary to determine the impact of their programs on HIV rates within the involved communities. These short term successes are evidence that moving forward and implementing district and national guidelines for would be a step forward while taking into consideration the limitations of each project and applying improvements.
Chib, A., Wilkin, H., Xue Ling, L., Hoefman, B., & Van Biejma, H. (2012). You Have an Important Message! Evaluating the Effectivness of a Text Message HIV/AIDS Campaign in Northwest Uganda. Journal of Health Communication , 17, 146-157.
Danis, C., Ellis, J., Kellogg, W., Van Beijma, H., Hoefman, B., Daniels, S., et al. (2010). Mobile Phones for Health Education in the Developing World: SMS as a User Interface. ACM Symposium on Computing for Development, London, UK.
International Telecommunications Union. (2011). World Telecommunication/ICT Development Report 2010. Retrieved 2012 10-August from http://www.itu.int/ITU-D/ict/statistics
Murless, E. (2012, August 9). Mhealth project: Mwana. (J. Burns, Interviewer)
Republic of Zambia Ministry of Health. (2011, 29-April). Project Mwana System Overview and M&E. Retrieved 2012 йил 10-August from http://www.unicefinnovation.org/sites/unicef.jjcdev2.com/files/Project_Mwana_Zambia_Evaluation_0.pdf
Republich of Zambia Ministry of Health. (2010). 2010 National Protocol Guidelines: Integrated Prevention of Mother-to-child Transmission of HIV.
Seidenberg, P., Nicholson, S., Schaefer, M., Semrau, K., Bweupe, M., Masese, N., et al. (2012, 15-March). Early infant diagnosis of HIV infection in Zambia throughmobile phone texting of blood test results. Bulletin of the World Health Organization .
Thea, D. (2012, 7-August). HIV Prevention in Zambia. (J. Burns, Interviewer.
Unicef. (2010). HIVs and AIDS. Retrieved August 10, 2012, from http://www.unicef.org/zambia/5109_8459.html
WHO. (2012). The bigger picture for e-health. Bulletin World Health Organication , 90, 330-331.