Diabetes Care 1 Addie L. Fortmann,1 Linda C. Gallo,2 Dulce Digital: An mHealth SMS- Maria Isabel Garcia,1 Mariam Taleb,1 Johanna A. Euyoque,1 Taylor Clark,1 Based Intervention Improves Jessica Skidmore,1 Monica Ruiz,1 Sapna Dharkar-Surber,3 James Schultz,4 Glycemic Control in Hispanics and Athena Philis-Tsimikas1 With Type 2 Diabetes https://doi.org/10.2337/dc17-0230 OBJECTIVE Type 2 diabetes is growing in epidemic proportions and disproportionately affects lower income, diverse communities. Text messaging may provide one of the most rapid methods to overcome the “digital divide” to improve care. EMERGING TECHNOLOGIES AND THERAPEUTICS RESEARCH DESIGN AND METHODS A randomized, nonblinded, parallel-groups clinical trial design allocated N = 126 low- income, Hispanic participants with poorly controlled type 2 diabetes to receive the Dulce Digital intervention or usual care (UC). Dulce Digital participants received up to three motivational, educational, and/or call-to-action text messages per day over 6 months. The primary outcome was HbA1c; lipids, blood pressure, and BMI were secondary outcomes. Satisfaction and acceptability were evaluated via focus groups and self-report survey items. RESULTS The majority of patients were middle aged (mean age 48.43 years, SD 9.80 years), female (75%), born in Mexico (91%), uninsured (75%), and reported less than a ninth- 1Scripps Whittier Diabetes Institute, Scripps grade education level (73%); mean baseline HbA1c 9.5% (80 mmol/mol), SD 1.3, and Health, San Diego, CA fasting plasma glucose 187.17 mg/dL, SD 64.75. A statistically significant time-by- 2Department of Psychology, San Diego State fi University, San Diego, CA group interaction effect indicated that the Dulce Digital group achieved a signi cantly 3 P University of California, San Diego, San Diego, greater reduction in HbA1c over time compared with UC ( = 0.03). No statistically CA significant effects were observed for secondary clinical indicators. The number of 4Neighborhood Healthcare, San Diego, CA blood glucose values texted in by participants was a statistically significant predictor Corresponding authors: Athena Philis-Tsimikas, of month 6 HbA1c (P < 0.05). Satisfaction and acceptability ratings for the Dulce [email protected], Linda C. Digital intervention were high. Gallo, [email protected], and James Schultz, [email protected]. CONCLUSIONS Received 30 January 2017 and accepted 11 May Use of a simple, low-cost text-messaging program was found to be highly acceptable 2017. in this sample of high-risk, Hispanic individuals with type 2 diabetes and resulted in Clinical trial reg. no. NCT01749176, clinicaltrials .gov. greater improvement in glycemic control compared with UC. This article contains Supplementary Data online at http://care.diabetesjournals.org/lookup/ Type 2 diabetes is growing in epidemic proportions in the U.S. and worldwide. The suppl/doi:10.2337/dc17-0230/-/DC1. International Diabetes Federation estimates that by 2040 there will be 642 million © 2017 by the American Diabetes Association. people living with diabetes worldwide, an increase of .50% compared with the pre- Readers may use this article as long as the work is properly cited, the use is educational and not sent day (1). The U.S. has the highest prevalence of diabetes among developed nations for profit, and the work is not altered. More infor- (i.e., 11% of the population that was 20–79 years of age) (1), and individuals of ethnic/ mation is available at http://www.diabetesjournals minority and low socioeconomic status are disproportionately affected (2,3). A recent .org/content/license. Diabetes Care Publish Ahead of Print, published online June 9, 2017 2 Texting Intervention in Hispanics With Diabetes Diabetes Care study (4) in the 26 states and District of practical barriers inherent to traditional the feasibility and acceptability of such Columbia that expanded Medicaid under (e.g., face-to-face) visits. approaches in underserved populations the Affordable Care Act found that diabe- Short messaging service (SMS), or text (19,24). The current study addresses tes diagnoses increased by 23% in 2014 messaging, is among the most frequently these gaps by investigating the glycemic compared with the previous year. Hispanic used mobile communication methods and benefit and acceptability of a culturally individuals in the U.S. experience higher has been adopted by an estimated three- tailored, SMS-based DSME and support rates of type 2 diabetes and, once diag- quarters of mobile users worldwide (14). intervention (Dulce Digital) among under- nosed, exhibit poorer glycemic control Text messaging is simple to implement served Hispanics with poor control in fed- than non-Hispanic white individuals (3,5). and may provide one of the most rapid erally qualified health centers in Southern Diabetes self-management education methods to overcome the recently impli- California. (DSME) and support is an effective method cated limitation of the “digital divide” (15) to improve clinical and cost outcomes (6,7) to improve care. In the U.S., texting among RESEARCH DESIGN AND METHODS and can be successfully tailored for ethni- adults in 2011 was higher among Hispanics Study Sample and Setting cally diverse populations (e.g., Philis- (83%) and African Americans (76%) than Between October 2012 and February Tsimikas et al. [8]). However, many at-risk among non-Hispanic whites (70%) (16). 2014, 126 individuals were consented individuals are unable to access DSME Ninety-nine percent of received text mes- and enrolled into the Dulce Digital study and support because of practical (e.g., sages are opened, and 90% are read within (Fig. 1). The sample included Spanish- work, transportation, caregiving) and 3 min of receipt (17). Thus, text messaging speaking and English-speaking His- health care access barriers (9–12). In represents an opportunity to provide fre- panic men and women, 18–75 years of fact, in 2012 only 4.7% of the 21 million quent, daily, low-cost, and interactive age, who were uninsured or underinsured people with diagnosed diabetes accessed communication that could prove ben- (Medicaid), and had type 2 diabetes and any accredited DSME program (9). To im- eficial for population-level diabetes poor glycemic control (as indicated by an prove patient and practice performance interventions. HbA1c level of $7.5% [58 mmol/mol]). outcomes, alternative methods must effi- Recent research syntheses have shown Individuals with plans to move outside ciently and effectively extend the reach of that mHealth interventions improve ad- the region and those with a severe phys- the care team to those in need of addi- herence and clinical control in patients ical or mental condition that would inter- tional support to reach clinical targets. with type 2 diabetes (18–20). However, fere with participation were excluded. The widespread adoption of mobile most studies were small, nonrandomized, Participants were recruited from clinic phone technologies, including among and resulted in limited clinical improve- sites within Neighborhood Healthcare, a low-income and older adults (13), high- ments (20–23). Further, few studies have network of federally qualified health cen- lights the potential for mobile health examined the implications of integrating ters in San Diego and Riverside counties (mHealth) technology to circumvent the these mHealth technologies into care or that serves predominantly low-income Figure 1—CONSORT flow diagram. care.diabetesjournals.org Fortmann and Associates 3 individuals of an ethnic/racial minority. “Tick, tock. Take your medication at the Hierarchical Linear and Nonlinear Model- All procedures were approved by the same time every day!”), and blood glu- ing software (HLM7; Scientific Software Scripps Health Institutional Review Board. cose monitoring prompts (e.g., “Time to International, Lincolnwood, IL) by A.L.F. check your blood sugar. Please text back Descriptive statistics were obtained, and Study Design and Intervention your results.”). All content was converted distributions were examined for normal- The intervention was tested using a par- into 119 brief, #160 character, text mes- ity. The triglycerides variable was signifi- allel-groups, nonblinded, randomized de- sage–friendly format and sent out via a cantly skewed as was the natural log sign. Blocked random assignment with contracted patient health management transformed to normalize the distribu- equal allocation was used to assign par- technology platform (Rip Road LLC, New tion; however, because no appreciable ticipants to Dulce Digital or usual care (UC), York, NY). Two to three messages a day differences between analyses using trans- using a randomly generated numbers were sent at study start, with frequency formed versus untransformed variables sequence. Participants were informed tapering over 6 months. Message timing were observed, results are presented for of group assignment after the baseline was standardized across all participants untransformed data only. assessment. and correlated with traditional meal or Mixed models were used to examine At the baseline visit, all participants testing times. Blood glucose–monitoring whether the two groups evidenced differ- viewed a 15-min diabetes educational prompts encouraged participants to text ential rates of change over time for HbA1c video developed by Scripps. All partici- message in their next observed value; one and secondary outcomes (i.e., time-by- pants received a blood glucose meter value $250 or #70 mg/dL or three values group interactions). To evaluate a possi- (OneTouch Verio Meter; LifeScan, Inc., Mil- between 181 and 250 mg/dL prompted ble dosage effect on the primary outcome pitas, CA), testing strips, and instructions a bilingual study coordinator to call the in Dulce Digital, the number of text mes- on use. A physical assessment with fasting participant to assess possible reasons sages and the number and duration of venous blood draw and study question- for hyperglycemia/hypoglycemia and to study coordinator phone calls were exam- naires were completed at baseline, month encourage as-needed follow-up with ined as predictors of month 6 HbA1c level 3, and month 6. Assessments were per- providers.