Bibliographic database and search engines were PubMed (incorporating Medline), EMBASE, CINAHL, PsycINFO, ScienceDirect, SpringerLink, EBSCOhost, ABI/INFORM, Google Scholar, and Google. Synthesis-producing organizations comprised the Cochrane Collaboration, the Centre for Reviews and Dissemination, and the Campbell Collaboration.

The search terms were health promotion, public health, preventive health services, disease prevention, population surveillance, patient compliance, patient care management, disease management, self care, HIV, acquired immunodeficiency syndrome, tuberculosis, communicable diseases, communicable disease control, chronic disease, noncommunicable disease, cellular phone, mobile phone, cell phone, SMS, MMS, text messag*, picture messag*, developing countries, developing world, Asia, Africa, South America, Latin America, and Europe, Eastern. We used the medical subject headings in Medline and Emtree tools in EMBASE. When necessary, these words were translated into French and German to minimize language bias.

The gray literature search strategy was inspired by a guideline for finding gray literature in public health [18]. Requests for cases and papers were posted on email lists, through personal connections, and directly to health organizations and SMS providers. Key gray literature sources were the Digital Repository Infrastructure Vision for European Research, the Electronic Theses Online System, the Networked Digital Library of Theses and Dissertations, the Directory of Open Access Repositories, DissOnline.de, SUDOC Catalogue, Social Science Research Network, Scirus, and Catholic Media Council Library. Other sources of gray literature comprised proceedings from related conferences, funding and not-for-profit organizations, listservs, blogs, SMS, and mobile phone providers. For identified projects, we contacted the corresponding agency to request documents and white papers that described the application and evaluation results. References of included documents and citation tracking of articles were also searched.

Following a detailed protocol, an initial reviewer identified and retrieved all eligible documents. The titles and abstracts were first examined to remove obviously irrelevant reports. A second reviewer confirmed eligibility and relevance. The full text of the potentially relevant ones was reviewed to finally select eligible papers.

English-, French-, and German-language documents were included when they were related to interventions using SMS in developing countries for disease prevention. For the purpose of this review, SMS referred to all applications that used SMS or texting for disease prevention purposes. These included applications that automatically shifted between SMS and general packet radio service channels, had live-person components, and connected to the Internet. Developing countries were defined as developing and emerging economies, a World Economic Outlook classification system based on (1) per capita income level, (2) export diversification, and (3) degree of integration into the global financial system [20]. For this study, we considered countries classified as developing and emerging economies in the International Monetary Fund’s World Economic Outlook report in 2008 to be developing countries (eg, Algeria, Brazil, Cambodia, India, Mexico, Poland, and Zambia) [20].

Documents were excluded if the applications did not use SMS, if they were not used in developing countries, or when they did not focus on the prevention of communicable and noncommunicable diseases. Documents published in languages other than English, French, or German were also excluded.

Eligible articles were reviewed and data extraction was inspired by a template provided in the Cochrane Handbook for Systematic Reviews of Interventions [21]. Applications were given a unique identification number. Disease focus, country or countries of implementation, project objectives, target audience, status (planned, ongoing, or completed), funding source, and contact details were recorded. When an evaluation was reported, the methods used and outcomes measured (health outcome, process of care, relevance, acceptability, and cost effectiveness) were recorded. Each personal communication attempt was documented along with contact information and the outcome of the effort. Finally, SMS-based interventions used for health promotion and disease prevention were extracted from the larger dataset and used for this review.

The 34 applications, listed in Table 1 [22-67], were described using varying levels of detail about the purpose of the initiative and features. Most provided information about the method used for communication (one-way or two-way), the use of games and contests, and incentives to increase the adoption and use of the application.

Some projects generated participation with a rewards system. Text To Change in Uganda designed an SMS multiple choice quiz, three times per week, to raise HIV/AIDS awareness and increase VCT [45]. Parallel to radio and newspaper advertisements, background information was provided in a special column in a magazine. Users submitting correct answers to the quiz were offered free VCT services and the possibility to win mobile phones or airtime. The program also monitored knowledge and beliefs through responses. The project Beat It in South Africa used free SMS to enter a prize draw where the results were announced during a television show designed to promote positive living and HIV/AIDS prevention [38,39]. This drew viewers and provided a contactable database. Campaigns such as Learning about Living in Nigeria (Figure 3) [35,36,68] and eQuest in Kenya [28] often targeted youth, encouraging them to correctly answer questions on HIV/AIDS issues for free airtime, T-shirts, mobile phones, computers, or DVD players.

Most of the reported applications did not provide information about an evaluation. Five interventions reported outcomes, with varying levels of detail (Table 2 [26,27,36,41,49-51,63]): Learning About Living [36], Cell for Life [41], Text Me! Flash Me! [27], Text to Change [49-51], and Breast Cancer Awareness project [63]. Each of these is described below including available information about the aim and method used to evaluate it. Findings are described, and stated benefits and limitations are then noted.

The project Learning about Living, an HIV and SRH education campaign, was evaluated after 14 months of implementation [36]. The project reached 9000 young people (median age of 24 years) who submitted 60,000 HIV/AIDS-related questions by SMS, and 4500 people (50% of users) accessed the service more than once. Most users (93%) were from urban or suburban setting and were male (79%). The majority (76%) were satisfied with the service, mostly because it was free (24%), quick (12%), or easily available (7%). Among dissatisfied users (24%), half did not receive a response to their SMS due to mobile network fluctuation, and 50% complained about the timing of answers. Despite the lack of information on methodology and variation in methods of evaluation between different sites, this initiative may have increased access to information for young people. The strengths of the project were regular formal and informal communication between all stakeholders. Noted challenges included network fluctuations, spam issues, and low reach of women and rural populations.

The pilot service of Cell-Life was evaluated using quantitative information from 210 respondents as well as qualitative inputs from 10 in-depth interviews and 5 focus groups [41]. Among the 210 participants, 66% owned a mobile phone, 80% felt comfortable using a mobile phone, and 64% were comfortable using SMS. Awareness of Cell-Life was poor, and its primary role was perceived as information dissemination at multiple levels. Users ranked the two-way communication and 24-hour availability of the messages for consultation on the mobile phone as important aspects, with the principle advantage that Cell-Life would save them time and money. Many requests were received for information about HIV and health in general, medication reminders, clinic locators, and access to reference sources of experts. Barriers included concerns about maintenance and SMS costs, limited access for poor people, high mobile phone turnover, and potential misuse or private use of SMS.

The Text Me! Flash Me! Service was an HIV/AIDS education initiative in Ghana that strived to strengthen HIV/AIDS knowledge among the most-at-risk population. After 6 months of implementation, the initiative was evaluated [27]. The aim was to understand the main reasons for using the service. Participants reported sharing information they received with others (87%) and 40% of them did so by SMS. The most popular service of the system was referral to VCT services (71% of the calls), which resulted in a substantial increase in VCT uptakes. Traditionally, counseling occurred at a rate of 50 clients per week per counselor. This program resulted in each counselor providing advice for approximately 88 clients per week. Reported challenges included a lack of financial incentives for counselors.

The project Text to Change aimed to increase the use of VCT centers. The system was evaluated in several papers, including 1 published in peer-reviewed literature [49-51]. In 2 papers, the project showed a 40% increase of VCT uptake after using an SMS quiz [49,50]. Identified challenges included language barriers due to English-language messages, a majority of male participants, confidentiality concerns (33.8%), and missing data (no information collected at VCT centers). The low number of respondents highlighted the importance of marketing at the beginning of the campaign and reassuring anonymity. In the quantitative study published in 2011 [51], the reported response rate was higher (86.7%) and still characterized by a majority of male participants. Among the SMS users, 19% mentioned they used it to get health information, and slightly more than half (51%) said they would be extremely likely to access health education programs through SMS, which was correlated with owning a mobile phone. High-risk and low-risk adolescents were equally likely to be interested in receiving HIV prevention programs via SMS. This study suggested that not all adolescents would have access to a mobile phone intervention and that nearly half would not necessarily choose to receive HIV/AIDS health information via this channel. It shows that adolescents may engage with education in different ways, thus emphasizing the importance of making HIV/AIDS prevention messages available in different modes and environments.

In India, the community-based study to evaluate the effect of SMS on breast cancer prevention targeting working women in the private sector in Delhi reported a significant increase of women performing breast self-examination after 2 months of SMS reminders [63]. The study sample was small and targeted only highly educated women working in the private sector who own mobile phones.

There are many SMS health initiatives for disease prevention in developing countries, yet few are being evaluated and reported. Those that did conduct evaluations reported process evaluation and uptake, providing limited data about behavior change. Moreover, with a low number of documents found in the peer-reviewed literature, it appears that, to date, little is being done to advance our understanding of what works and what outcomes could be achieved in using SMS for disease prevention in a developing country contexts. Major opportunities are perceived, evident by the number and wide variety of projects, the recent creation of the United Nations Foundation’s mHealth Alliance, and papers published describing SMS applications in the developed world [6-14,75]. However, the need remains for evidence-based dissemination of information about using mobile phones and SMS for improving health in the developing world. The limited evidence found in this systematic review highlights the need for research that assesses behavioral, social, economic, and health outcomes of mobile phone interventions aimed at promoting health in developing country contexts. Sharing best practices and providing strategic directions will allow the building of evidence for the use of mobile health technologies that address the needs of health systems, communities, and populations. Consequently, the promises of using mobile phones and SMS may translate into an equitable improvement in health.