Texting to Improve Testing (TextIT):

Early accurate diagnosis is one of the first crucial steps in care for infants born to HIV-infected mothers. Early initiation of antiretroviral therapy (ART) relies upon early diagnosis and results in significant reductions in infant morbidity and mortality. The investigators recently concluded a successful randomized controlled trial in Kenya entitled, "Improving uptake of early infant diagnosis of HIV for prevention of mother-to-child HIV transmission (PMTCT): a randomized trial of a text messaging intervention" (ClinicalTrials.gov # NCT01433185). In this study, text messages developed using a behavioral theoretical framework significantly improved maternal attendance at post-partum clinic appointments and rates of testing to facilitate early infant diagnosis of HIV in a selected population and controlled setting. Understanding the effectiveness of this intervention (and its limitations) in a real-world, routine-care setting represents the next step in the translational pathway to public health impact. The investigators therefore now propose a cluster randomized, stepped wedge trial in 20 clinics operated by the Kenyan Ministry of Health in the Nyanza region of Kenya and use the Reach, Effectiveness, Adoption, Implementation, and Maintenance (REAIM) framework to understand the effectiveness of the text messaging to improve testing (TextIT) intervention. Our specific aims are: 1. To determine the effect of TextIT on maternal attendance at postpartum clinic visits during the randomized stepped-wedge rollout of the intervention. Hypothesis 1: A greater proportion of women at health facilities implementing TextIT will attend clinic within eight weeks postpartum compared to women at health facilities implementing standard care. 2. To determine the effect of TextIT on virological infant HIV testing within eight weeks after birth during the randomized stepped-wedge rollout of the intervention. Hypothesis 2: Infants of women at health facilities implementing TextIT will be more likely to have virological HIV testing compared to infants of women at health facilities implementing standard care. 3. To determine the costs and cost-effectiveness of TextIT. The investigators will estimate the cost per patient and per health gain achieved (disability-adjusted life year, DALY) comparing TextIT to current standard care. Hypothesis 3: The TextIT intervention will be more cost-effective than current standard care.