mHealth, the use of mobile phones to support health care, is often seen to have great potential, as mobile technology has a high prevalence, with 6.8 billion subscriptions worldwide [1]. This extends to the developing world, with Africa having the fastest growing market for cell phone technology, and over 75% of the low-income South African population already owning, or having access to, a cell phone number [2]. mHealth can be implemented for a variety of health-related functions: to provide information, act as reminders of clinic visits, encourage behavior changes, and be actively used in symptom diagnoses and drug adherence checking [3][4][5]. However, there are few evaluations that thoroughly investigate the viability of the mHealth intervention, typically drawing on either quantitative or qualitative data, but not both. Furthermore, cost-effectiveness, context, and acceptability are still contentious issues, and therefore, even though strong potential exists, the viability of mHealth interventions in all situations is not clear and further investigation into such matters is still merited [6][7].

The use of mHealth interventions in the control and management of non-communicable diseases is relatively well established; however, there is a paucity of research on its viability from lower-middle-income countries [8]. Hypertension is one of the many non-communicable diseases on the rise in the developing world, including South Africa [9], where the prevalence is approximately 21% [10]. This can in part be attributed to increasing rates of rural-urban migration, as hypertension is relatively uncommon in rural South Africa [11]. It is further exacerbated by socioeconomic factors that encourage poor diet and sedentary lifestyles and contribute to low knowledge levels (both of risk factors and treatment) amongst the poorer urban population, which ultimately increases complications in hypertension [12]. Low knowledge levels in hypertensive patients are one of the key barriers to hypertensive care [13]. Furthermore, there is a particular demand for lifestyle modification information in South Africa [14], and treatment and control of non-communicable diseases are part of South Africa’s National Development Plan [15].

mHealth interventions represent an opportunity to overcome the significant access barriers to health care seen in developing countries [16][17], due to the traditional socioeconomic and rural and urban divides [6], as well as an opportunity to address low knowledge levels and contribute to positive behavior change. According to social cognitive theory, short message service (SMS) interventions may be particularly good at increasing knowledge, as they allow for repeated learning, grant a high level of autonomous learning, and allow the learner to learn at their own pace and in a low-stress environment [18]. This paper therefore seeks to test whether the dissemination of health information via SMS leads to improvements in health knowledge and self-reported health-related behaviors amongst chronic hypertensive patients in a resource-poor setting.

An SMS campaign consisting of both knowledge and lifestyle information was designed. A mixed methods approach that employed a controlled clinical trial and a focus group evaluated the SMS campaign. The project was approved by the University of Cape Town’s Health Sciences Faculty Human Research Ethics Committee (HREC REF 043/2011).

The details of this mixed methods approach are described below.

The results demonstrated almost no differences in knowledge between the control and intervention arms at exit, with the exception of 1 question. Self-reported behavior change in the intervention group showed significant trends toward healthier lifestyles. This observation was confirmed by focus groups, which emphasized the utility of the SMSes as a reminder, and a source of motivation to change, rather than a source of new knowledge.

There are a few possible explanations for no observed difference in knowledge results between the control and intervention groups. Firstly, the survey questions may not have been suitable in assessing participants’ knowledge; they were not sufficiently difficult for measuring a change. Census data on baseline education levels for our population demographic indicate the majority have not finished secondary schooling [17], and this informed the design of our survey questions. Furthermore, baseline questions were formulated in conjunction with health care professionals around what information was felt was important for the patient to know. While low levels of education were confirmed by our baseline assessment demographic data, it appears participants’ specific health knowledge behavior was relatively high. Indeed, they may have been so high (participants knew on average 54.8% of the information based on their baseline questionnaires) that there may not have been enough leeway for any statistically significant gains to be observed. In many ways, this is a positive, as it suggests that the health care staff are providing good health information, among other possible sources. This may also be due to the fact that our patient population was sampled from a hypertensive club, which may not be reflective of general patients with hypertension. Subgroup analysis stratified by some of the baseline demographics, such as education level, may have accounted for such factors but could not be performed due to limited sample size. Secondly, methodological issues were encountered with the SMSes service delivery. On average, 29.7% of SMSes failed to deliver to the intervention group. This was due to the service provider and casts doubt on the validity of the results. Lastly, loss to follow-up was also a significant problem, which reduced the statistical power of the results.

The findings do nonetheless suggest that SMSes are effective in motivating behavior change. However, there is a high potential for bias, as only self-reported behavior changes were reported. The health behavior change results could not be confirmed objectively with blood pressure and weight changes before and after the intervention, due to poor medical record keeping in patient clinic folders. While the focus group findings supported the self-reported behavior change findings from the questionnaire, the same self-reporting biases exist, and there may also be selection bias at play in the focus group, as participants self-selected into the focus group. The focus group further suggested that the SMSes were functioning as “cues to action”—see, for example, the Health Belief Model [19]. The Health Belief Model theorizes that health seeking behavior is influenced by a range of beliefs, such as perceived severity, perceived benefits, self-efficacy (confidence in one’s ability to take action), as well as reminders to action that activate the readiness of the patient to seek good health. The SMSes may act as one such reminder, as they provide information as well as remind the patient of this information on a daily basis. Furthermore, the Stages of Change theory [20] postulates that behavior change is a staged process starting at precontemplation and working through 5 stages toward maintenance or termination of the behavior. The SMS campaign appears to be particularly good at moving from the precontemplation to the contemplation or preparation stage. It is uncertain whether the campaign was long enough or particularly well suited to moving participants into the next stage of “action,” although self-reported behavior change suggests it might have been. Combined, these results suggest an expressed interest in changing behavior following the SMS intervention, even if it is uncertain whether this resulted in behavior change.

In conclusion, the results suggest that more research is required before implementing SMS interventions for chronic disease in developing countries, and that any intervention should be entered with caution, realizing some of the limitations of the developing world. Firstly, the infrastructure for their successful implementation may not be available [21], a problem experienced in this intervention by a high SMS delivery failure rate. Secondly, they may not be a necessary, or good, vehicle for knowledge dissemination. However, SMSes do appear to act as a reminder for the patient, and participants reported they felt they were being better looked after by receiving the SMSes, and that the SMSes were cueing positive health-related behavior changes. This could potentially make them a very useful tool for assisting in the management of chronic conditions. Future studies should incorporate objective measures of behavior change, such as weight and blood pressure changes. Recent trials have also suggested that personalization of the SMSes and tailoring toward the individual receiving the SMS can have a marked improvement on outcomes, a theme that could also be explored in future work [22].