Investigation into the use of short message services to expand uptake of human immunodeficiency virus testing, and whether content and dosage have impact
Objective: South Africa has one of the highest human immunodeficiency virus (HIV) prevalence rates in the world, but despite the well-established benefits of HIV counseling and testing (HCT), there is low uptake of HCT. The study aimed to investigate the effectiveness of using short message services (SMSs) to encourage HCT while interrogating the impact of altering SMS content and dosage (the number of SMSs). Materials and Methods: About 2,533 participants were recruited via an SMS sent to 24,000 mobiles randomly sampled from a pre-existing database. Recruits were randomly allocated to four intervention groups that received 3 or 10 informational (INFO) or motivational (MOTI) SMSs, and a control group. After the intervention, participants were prompted to go for HCT, and postintervention assessment was done after 3 weeks. Results: In comparison with the control, receipt of 10 MOTI messages had the most impact on uptake of HCT with a 1.7-fold increased odds of testing (confidence interval 95%; p=0.0036). The lack of efficacy of three SMSs indicates a threshold effect, that is, a minimum number of MOTI SMSs is required. INFO SMSs, whether 3 or 10 were sent, did not have a statistically significant effect. The cost can be calculated for the marginal effect of the SMSs, that is, the cost to get people to test over and above those who were likely to test without the intervention. Use of 10 MOTI SMSs yielded a cost-per-tester of $2.41. Conclusions: While there are methodological issues apparent in our study, the results demonstrate the potential of SMSs to influence the uptake of HCT, the importance of appropriate content, and the need to determine a threshold for SMS-based interventions. These results indicate a potential for SMSs to be used more generally for interventions encouraging people to take health-related actions, and the need for further research in this field. The reasonable cost-per-tester is promising for the scale-up of such an intervention. © 2012 Mary Ann Liebert, Inc.