Development and evaluation of DiabeText, a personalized mHealth intervention to support medication adherence and lifestyle change behaviour in patients with type 2 diabetes in Spain: A mixed-methods phase II pragmatic randomized controlled clinical trial

•DiabeText is a new text messaging intervention for diabetes self-management.•It is integrated with patient-generated and with routinely collected clinical data.•A 3-month phase II trial showed that it was feasible to deliver the intervention.•DiabeText produced high levels of patient acceptability and perceived utility.•More robust trials are needed to determine its effectiveness and cost-efficacy. Despite the increasing interest in text-messaging interventions to support healthcare delivery, the available evidence about their effectiveness is still limited. 1) to develop DiabeText, an intervention delivering automated, tailored brief text messages to support diabetes self-management; 2) to explore the potential impact of DiabeText on self-management behavior and glycaemic control, and; 3) to examine the feasibility of conducting a future phase III randomized clinical trial to evaluate the effectiveness of DiabeText. 3-month, two-arm, randomized feasibility trial (ClinicalTrials.gov: NCT04738591) with patients with type 2 diabetes (HbA1c > 8%). Participants were allocated to the control (usual care) or DiabeText group (usual care + five text messages per week). Outcomes were: recruitment rate; follow-up rate, missing data; medication adherence; adherence to Mediterranean diet; physical activity; and HbA1c. In addition, after delivering the intervention, we conducted a qualitative study involving 14 semi-structured interviews with participants allocated to the DiabeText group, to understand their views about the intervention. From 444 screened people, we recruited 207 participants (recruitment rate = 47%), of which 179 completed the post-intervention interview (follow-up rate = 86%). We sent 7,355 SMS during the intervention period, of which 99% successfully reached the participants. At post-intervention, DiabeText was associated with non-statistically significant (p > 0.05) improvements in adherence to medication (OR = 2.0; 95%CI 1.0 to 4.2), Mediterranean diet (1.7; 0.9 to 3.2), and physical activity (1.7; 0.9 to 3.1). No between-group differences were observed in mean HbA1c (p = 0.670). The qualitative study indicated that participants perceived DiabeText as a helpful resource because it increased their awareness about the importance of adequate self-management and the sense of being cared for. DiabeText is the first system in Spain to integrate patient-generated and routinely collected clinical data to deliver tailored text messages to support diabetes