Cluster-Randomized Trial of a Mobile Phone Personalized Behavioral Intervention for Blood Glucose Control

Transcription

1 Clinial Care/Eduation/Nutrition/Psyhosoial Researh O R I G I N A L A R T I C L E Cluster-Randomized Trial of a Mobile Phone Personalized Behavioral Intervention for Blood Gluose Control CHARLENE C. QUINN, RN, PHD MICHELLE D. SHARDELL, PHD MICHAEL L. TERRIN, MD, MPH ERIK A. BARR, BA SHOSHANA H. BALLEW, BA ANN L. GRUBER-BALDINI, PHD OBJECTIVE To test whether adding mobile appliation oahing and patient/provider web portals to ommunity primary are ompared with standard diabetes management would redue glyated hemoglobin levels in patients with type 2 diabetes. RESEARCH DESIGN AND METHODS A luster-randomized linial trial, the Mobile Diabetes Intervention Study, randomly assigned 26 primary are praties to one of three stepped treatment groups or a ontrol group (usual are). A total of 163 patients were enrolled and inluded in analysis. The primary outome was hange in glyated hemoglobin levels over a 1-year treatment period. Seondary outomes were hanges in patient-reported diabetes symptoms, diabetes distress, depression, and other linial (blood pressure) and laboratory (lipid) values. Maximal treatment was a mobile and web-based self management patient oahing system and provider deision support. Patients reeived automated, real time eduational and behavioral messaging in response to individually analyzed blood gluose values, diabetes mediations, and lifestyle behaviors ommuniated by mobile phone. Providers reeived quarterly reports summarizing patient s glyemi ontrol, diabetes mediation management, lifestyle behaviors, and evidene-based treatment options. RESULTS The mean delines in glyated hemoglobin were 1.9% in the maximal treatment group and 0.7% in the usual are group, a differene of 1.2% (P, 0.001) over 12 months. Appreiable differenes were not observed between groups for patient-reported diabetes distress, depression, diabetes symptoms, or blood pressure and lipid levels (all P. 0.05). CONCLUSIONS The ombination of behavioral mobile oahing with blood gluose data, lifestyle behaviors, and patient self-management data individually analyzed and presented with evidene-based guidelines to providers substantially redued glyated hemoglobin levels over 1year. D iabetes affets 38 million people in the U.S.; 40% are undiagnosed, and another 87 million are onsidered prediabeti. Costs exeed $100 billion annually (1,2). Changes in lifestyle/self-are behaviors, omplex medial regimens, use of gluose-testing devies, and frequent data assessment by patients and providers are required to improve blood gluose and subsequent outomes. In linial trials, better self-are/lifestyle resulted in better diabetes outomes (3 5).However,these linial trials improved outomes for irumsribed patient populations (6 9). Patients with diabetes are diverse, treatment may involve multiple speialists, and are by primary are providers (PCPs) is limited to 15-min visits. Only 55% of individuals with type 2 diabetes reeive diabetes eduation (10); 16% report adhering to reommended self-management ativities (11). Conern that elevated blood gluose levels result in mirovasular omorbidity motivates behavioral hange and monitoring From the Department of Epidemiology and Publi Health, University of Maryland Shool of Mediine, Baltimore, Maryland. Corresponding author: Charlene C. Quinn, quinn@epi.umaryland.edu. Reeived 1 Marh 2011 and aepted 19 June DOI: /d Clinial trial reg. no. NCT , linialtrials.gov. This artile ontains Supplementary Data online at /d /-/DC by the Amerian Diabetes Assoiation. Readers may use this artile as long as the work is properly ited, the use is eduational and not for profit, and the work is not altered. See lienses/by-n-nd/3.0/ for details. interventions to assist patients and PCPs (12 14). The Mobile Diabetes Intervention Study, reported here, evaluated a diabetesoahing system, using mobile phones and patient/provider portals for patient-speifi treatment and ommuniation. The hypothesis tested was that mobile telephone feedbak on self-management of blood gluose results and lifestyle and linial management offered to patients with type 2 diabetes and their providers an redue glyated hemoglobin levels over 1 year. RESEARCH DESIGN AND METHODS Eligibility and study design The Mobile Diabetes Intervention Study was a luster-randomized linial trial onduted in primary are praties in four distint Maryland areas. Eligible praties inluded groups of at least three physiians without aademi affiliation who provided diabetes are to at least 10% of their patients and were identified from a list of primary are praties in the study geographi areas. A detailed desription of the study design was reported previously (13). Group assignment was onealed until a pratie agreed to partiipate in the study. Data were obtained by abstration from patients medial harts and primary olletion. As shown in Fig. 1, 26 primary are praties were randomized to one of four study groups using a stepped intervention design for groups: group 1: ontrol usual are (UC), group 2: oah-only (CO), group 3: oah PCP portal (CPP), and group 4: oah PCP portal with deisionsupport (CPDS). A total of 2,602 patients were identified by these praties for sreening; 2,103 were determined ineligible, 145 delined partiipation, 213 were enrolled, and 163 were inluded in analyses (UC, n =56;CO,n =23;CPP,n =22;and CPDS, n = 62). We aimed to identify patients treated in ommunity primary are settings who would benefit from an intensive diabetes intervention. Errors in onsent form ompletion were found on audit after study enrollment was losed. Our Institutional Review Board asked us are.diabetesjournals.org DIABETES CARE 1 Diabetes Care Publish Ahead of Print, published online July 25, 2011

2 Randomized trial of mobile diabetes oahing Figure 1 Flowhart of enrollment and patient status (n = 163). to repeat onsent proedures to assure we obtained proper signatures from all parties. We ompleted repeat onsent proedures for 163 patient partiipants and all 39 physiian partiipants. We were unable to ontat patients not reonsented; they did not signifiantly differ (P. 0.10) at baseline from inluded patients in age, sex, or baseline glyated hemoglobin. Partiipant data were analyzed aording to physiian praties original randomization treatment assignment (intention-to-treat analyses). Patients eligible for reruitment to the study met all inlusion riteria: Physiian diagnosis of type 2 diabetes for $6 months; Glyated hemoglobin $7.5% within 3 months; Age years; Patients were exluded for any of the following: Mediare or Mediaid benefiiaries; Uninsured; Insulin pump users; Not urrently managed by study physiians; Pregnant; Ative substane, alohol, or drug abuser (sober,1 year); Psyhoti or shizophreni under ative are; Severe hearing or visual impairment; or No Internet or aess. The most ommon reasons for ineligibility were as follows: glyated hemoglobin,7.5% (72%); patient s diabetes not urrently managed by study physiian (8%); not between the age of 18 and 64 years (5%); uninsured or insured by Mediaid or Mediare (3%); not type 2 diabetes (2%); no Internet or aess (2%); speified medial exlusion (2%); and psyhiatri exlusion (1%). Patients were exluded if insured by Mediare or Mediaid or were uninsured beause overage of primary are servies differs from patients ommerially insured. These patients would be exluded from planned seondary analyses of laims data provided by a ommerial insurer. Patients overed by any ommerial insurer were eligible. Patients on insulin pumps, pregnant, or not meeting other linial riteria were exluded beause their type 2 diabetes required different linial management. Of the patients identified as eligible, 42% were enrolled (213) and 77% of those enrolled ompleted the study and were inluded in the analyses. The intervention was a patient-oahing system and provider linial deision support (13). The patient-oahing system inluded a mobile diabetes management software appliation and a web portal. The mobile software allowed patients to enter diabetes self-are data (blood gluose values, arbohydrate intake, mediations, other diabetes management information) on a mobile phone and reeive automated, real-time eduational, behavioral, and motivational messaging speifi to the entered data. The patient web portal augmented the 2 DIABETES CARE are.diabetesjournals.org

3 mobile software appliation and onsisted of a seure messaging enter (for patientprovider ommuniation), personal health reord with additional diabetes information (e.g., laboratory values, eye examinations, foot sreenings), learning library, and logbook to review historial data. The provider portal had different views of patient data on the basis of study group assignment. The data-only view (group 3, CPP) allowed providers to aess unanalyzed patient data. Group 4 (CPDS) providers had aess to analyzed patient data linked to standards of are and evidene-based guidelines. Patients reeived a One Touh Ultra 2 (LifeSan, Milpitas, CA) gluose meter and supplies. Patients in the three ative treatment groups reeived idential study materials: mobile phones, 1-year unlimited data and servie plan, study mobile diabetes management software, and aess to the web-based patient portal. The mobile diabetes management software inorporated over 1,000 automated selfmanagement messages into a feedbak algorithm. The algorithm displayed eduational and motivational messages to patients after patients self-reported data into the mobile phone appliation (Supplementary Fig. 1). Diabetes eduators were virtual ase managers that intermittently reviewed patient data. Eduators ould supplement automated messages with eletroni messages sent to the patient portal. Eduator messages were based on longitudinal data trends. Patients in all three treatment groups were allowed to make telephone alls to eduators but were enouraged to ommuniate eletronially. On average,,50% of ative patients made or reeived live phone alls, with an average of one phone all per month. Lastly, patients reeived an eletroni ation plan every 2.5 months to support improved diabetes self-management and to serve as previsit summaries for physiian offie visits. Providers were not informed of the level of ommuniation to patients but knew whether patients were assigned to an intervention or to the UC group. All providers reeived the most reent Amerian Diabetes Assoiation guidelines for diabetes are and were notified when patients enrolled in the study (7,8). Providers assigned to UC were asked to are for patients as usual. Ative treatment providers were informed that their patients reeived a mobile and web-based patient oahing system. Providers in the CO group reeived data from their patients if patients hose to share it. Providers in the CPP and CPDS groups were trained on aessing the provider Internet portal on offie ompatible omputers (PCs), allowing visual aess to patients unanalyzed data. Providers in the CPDS group were trained on aessing the provider Internet portal to view patient data on offie PCs and also reeived quarterly reports (more often if needed) that summarized patients glyemi and metaboli ontrol, adherene to mediation, self-management skills, and relevant evidene-based guidelines. Reports were aessible by Internet portal or fasimile. Enrolled providers were reimbursed modestly for researh effort ($250 per patient enrolled). Primary outome The primary outome of the study was hange in glyated hemoglobin (%) omparing UC and maximal treatment (CPDS) at baseline versus 12 months. Medial hart reviews were used to asertain patient data. For patients without a glyated hemoglobin within 4 months of the desired measurement, a glyated hemoglobin test was offered at no harge at baseline to determine eligibility and at 12 months. At baseline, glyated hemoglobin was measured using one devie, the Bayer DCA 2000, by trained staff blinded to patient group assignment. At followup, if glyated hemoglobin was not asertained within 14 days of the 12-month time point, reminders were provided to patients and physiians to omplete the test. Glyated hemoglobin level at intermediate time points (3, 6, and 9 months) was olleted from patients medial harts. Seondary outomes The Patient Health Questionnaire-9 (PHQ) was administered at baseline and at follow-up interviews to assess depressive symptoms (15). We used the nineitem version of the Self-Completion Patient Outome Instrument to assess patient-reported symptoms assoiated with diabetes (16,17) and the 17-item Diabetes Distress Sale (18,19). Clinial measurement related to diabetes ompliations (blood pressure, lipid levels) was obtained from provider medial offie reords. Hypoglyemi events, hospitalization, and emergeny room visits were asertained through quarterly telephone alls to patients. Vital status was asertained through review of physiian harts if we ould not ontat patients. Study data for primary and seondary outomes were olleted Quinn and Assoiates by researh staff separately from data transmitted through the devie. A detailed desription of the study design and rationale for primary and seondary outomes has been reported previously (13). Study oversight The University of Maryland Baltimore institutional review board approved the study, and a Data and Safety Monitoring Board was appointed to review study proedures and adverse advents. Statistial analysis Praties were assigned to treatment groups aording to a 1.5:1:1:1.5 (Group 1, UC:Group 2, CO:Group 3, CPP:Group 4, CPDS) ratio using a omputer-generated list of random numbers. The ratios were higher in groups 1 and 4 for analyses of the main hypotheses. Sample size was determined on the basis of the primary outome, hange in glyated hemoglobin. The omparison of UC, whih inluded 56 patients from nine praties, to CPDS, whih inluded 62 patients from seven praties, had 80% power to detet a differene in mean glyated hemoglobin hanges of 0.65 SD, orresponding to 1.0% if SD was 1.58%, using a two-sided test with 0.05 type I error after aounting for a withinluster orrelation of 0.10, similar to a previously reported study (20,21). ComparisonsoftheUCtoCO(23patients from four praties) and CPP (22 patients from six praties) had 80% power to detet a differene in mean outome hanges of SD, %, for glyated hemoglobin. Linear mixed-effets models were used to ompare mean hanges in primary and seondary outomes between UC and eah ative intervention. The primary analysis examined 12-month hanges for glyated hemoglobin. Seondary analyses jointly ompared 3-, 6-, 9-, and 12-month hanges between groups. Random effets aounted for within-pratie lustering and within-patient orrelation. Model fixed effets were treatment group indiators, time indiators, and interations between treatment group and time. Two seondary analyses of glyated hemoglobin were performed as follows: one analysis stratified by baseline glyated hemoglobin ($9.0 vs.,9.0); the other (prespeified analysis) adjusted for baseline glyated hemoglobin as a ovariate. We performed a sensitivity analysis using weighted estimating equations (WEE) to address any residual bias from missing data (22). Statistial signifiane was defined as P, 0.05 or 95% CI are.diabetesjournals.org DIABETES CARE 3

4 Randomized trial of mobile diabetes oahing Table 1 Baseline harateristis of patients and primary and seondary study outomes Group 1: UC (n = 56) Group 2: CO (n = 23) Group 3: CPP (n = 22) Group 4: CPDS (n = 62) n % or mean 6 SD* N % or mean 6 SD* n % or mean 6 SD* n % or mean 6 SD* Baseline harateristis Glyated hemoglobin (%) $ Age (years) Sex Male Female Rae Blak (non-hispani) White (non-hispani) Other Duration of diabetes diagnosis (years) Smoking status Current smokers Former smokers Nonsmokers Eduation High shool/trade shool or less Some ollege or assoiates Bahelors degree or higher Depression (PHQ-9) Sore Minimal to mild (0 9) Moderate (10 14) Moderately severe (15 19) Severe depression (20 27) BMI BMI (kg/m 2 ) Underweight ( kg/m 2 ) Normal ( kg/m 2 ) Pre-obese ( kg/m 2 ) Obese lass 1 ( kg/m 2 ) Obese lass 2 ( kg/m 2 ) Obese lass 3 ($40 kg/m 2 ) Comorbidities Hypertension Hyperholesterolemia Coronary artery disease Mirovasular ompliations, any DIABETES CARE are.diabetesjournals.org

5 Quinn and Assoiates Table 1 Continued Group 1: UC (n = 56) Group 2: CO (n = 23) Group 3: CPP (n = 22) Group 4: CPDS (n = 62) n % or mean 6 SD* N % or mean 6 SD* n % or mean 6 SD* n % or mean 6 SD* Primary outome, glyated hemoglobin (%) Baseline Months Months Months Months Change from baseline to 12 months (mean) Change from baseline to 12 months (95% CI) 21.1 to to to to 21.5 Seondary outomes, patient-reported outomes Diabetes Distress Sale Baseline Months Change from baseline to 12 months (mean) Change from baseline to 12 months (95% CI) 20.4 to to to to 0.0 Diabetes symptom inventory Baseline Months Change from baseline to 12 months (mean) Change from baseline to 12 months (95% CI) 25.5 to to to to 1.8 Depression (PHQ-9) Baseline Months Change from baseline to 12 months (mean) Change from baseline to 12 months (95% CI) 23.2 to to to to +0.5 Seondary outomes, laboratory values Systoli blood pressure (mmhg) Baseline Months Change from baseline to 12 months (mean) Change from baseline to 12 months (95% CI) 23 to to to to +3 Diastoli blood pressure (mmhg) Baseline Months Change from baseline to 12 months (mean) Change from baseline to 12 months (95% CI) 22 to 4 22 to 7 26 to 3 24 to 2 LDL (mg/dl) Baseline Months Change from baseline to 12 months (mean) Change baseline to 12 months (95% CI) 215 to to to to 4 are.diabetesjournals.org DIABETES CARE 5

6 Randomized trial of mobile diabetes oahing Table 1 Continued Group 1: UC (n = 56) Group 2: CO (n = 23) Group 3: CPP (n = 22) Group 4: CPDS (n = 62) n % or mean 6 SD* N % or mean 6 SD* n % or mean 6 SD* n % or mean 6 SD* HDL (mg/dl) Baseline Months Change from baseline to 12 months (mean) Change from baseline to 12 months (95% CI) 21 to 3 24 to 3 23 to 4 0 to 3 Triglyerides (mg/dl) Baseline Months Change baseline to 12 months (mean) Change baseline to 12 months (95% CI) 258 to to to to 3 Total holesterol (mg/dl) Baseline Months Change baseline to 12 months (mean) Change baseline to 12 months (95% CI) 222 to to to to 2 n = 163. *Unless otherwise indiated. Primary outome, glyated hemoglobin hange over 12 months; group 4 (P, 0.001) and group 2 (P = 0.003) have signifiantly larger hanges than group 1. No other outomes are signifiant. Mean hange and CI values are from the mixed-effets model. that exludes 0. Analyses were performed using SAS version 9.1 (SAS Institute, In. Cary, NC). RESULTS The 163 study patients had a mean baseline glyated hemoglobin of 9.4% (range ) (Table 1). Mean age was 52.8 years, 50.3% were female, 39.3% were Afrian Amerian, and 31.3% were ollege-eduated. The mean duration of diabetes was 8.2 years. Most partiipants (76.1%) were obese (BMI $30 kg/m 2 ). Partiipants had a mean PHQ-9 of 5.2 (minimal to mild depression sores). Most partiipants had hypertension (63.2%) and hyperholesterolemia (58.3%). CPDS patients had higher baseline glyated hemoglobin than UC (9.9 vs. 9.2%, P = 0.04). No other baseline patient variables differed signifiantly among the four study groups. Table 1 shows primary and seondary outome measures. CPDS mean glyated hemoglobin dereased 1.9% (95% CI ) over 12 months. UC mean glyated hemoglobin dereased 0.7% ( ). Table 1 and Fig. 2 show that the mean 12-month derease in CPDS glyated hemoglobin was 1.2% more than UC (95% CI %; P, 0.001). Furthermore, the CPDS patients had a signifiantly greater derease in mean glyated hemoglobin than the UC patients when ompared at all follow-up time points (P, 0.001). CO and CPP mean glyated hemoglobin levels also dereased over 12 months. Both had greater 12-month glyated hemoglobin redutions than UC (CO, P = 0.02; CPP, P =0.45).COandCPPweresimilar to CPDS over all follow-up time points (P for both omparisons). In a stratified analysis, a greater deline was found with CPDS than UC for the stratum with baseline glyated hemoglobin,9.0% (differene in derease 0.7%, 95% CI , P = 0.003) and the stratum with baseline glyated hemoglobin at least 9.0% (differene in derease 1.3%, , P = 0.01) (shown in Fig. 2B and C). The test of interation was not signifiant (P = 0.49) for baseline glyated hemoglobin stratum and treatment group over time. We obtained the same onlusion whether or not we analyzed the baseline to 12-month hanges with intermediate glyated hemoglobin measures. Glyated hemoglobin results were unhanged after adjusting for baseline glyated hemoglobin and after performing the WEE analysis. Although there were 6 DIABETES CARE are.diabetesjournals.org

7 Quinn and Assoiates Figure 2 Primary study outome and baseline A1C stratified analyses. mean delines aross all groups in lipid values and blood pressure readings, Diabetes Distress, Diabetes Symptoms, and PHQ-9 Depression, none of the 12-month hanges omparing the UC to any of the ative interventions were signifiantly different (P. 0.05). Hypoglyemi events, hospitalizations, and emergeny-room visits were infrequent in all groups. One patient in group 4 (CPDS) was hospitalized twie for reasons not reported to the study. The DSMB determined that there were no diret studyrelated adverse events found. No patients died during the 12 months of this study. CONCLUSIONS To our knowledge, this is the first luster-randomized study of a mobile diabetes oahing intervention onduted in a ommunity setting over a 1-year treatment period. Few previous studies of eletroni or mobile ommuniation interventions for diabetes were randomized, inluded a ontrol group, or overed 1 year (13). Our study inluded minorities, found linially meaningful differenes and few adverse events, none study or treatment-related. Our study evaluated the intervention for ommerially insured patients in primary are settings, where the majority of diabetes are is provided. Enrolling and treating study partiipants aording to random assignment of physiian praties (lusters) redued the risk of bias in treatment appliation. We found that a mobile phone based treatment/behavioral oahing intervention improved glyated hemoglobin by 1.9%, ompared with 0.7% for UC, a differene of 1.2% (P, 0.001) over 12 months. This result pertained to people with poorly ontrolled glyated hemoglobin ($9.0%) and people with less severe abnormal initial glyated hemoglobin values ( %). The results stratified on baseline glyated hemoglobin (Fig. 2) demonstrate three key features. First, sine CPDS and UC had similar mean baseline glyated hemoglobin within strata of baseline A1C (,9 vs. $9.0%), and the treatment effet is similar in eah of the strata, our findings provide evidene of true 12-month treatment differenes in glyated hemoglobin, rather than regression to the mean. This stratified analysis is important, showing large hanges in A1C by adjusting for baseline A1C. Seond, the treatment effet in the higher glyated hemoglobin stratum shows this intervention to be suitable to obtain the goals of the more onservative ACCORD approah (23). Neither ACCORD nor this study olleted person-speifi data on dietary, physial ativity, and pharmaologial management adjustments made for individual patients. Beause of the personalized quality of the mobile phone tehnology, we expet to be able to make those distintions in future investigations now that its observed effets on glyated hemoglobin justify their study. Third, mobile phone are.diabetesjournals.org DIABETES CARE 7

8 Randomized trial of mobile diabetes oahing management is effiaious in patients whose glyated hemoglobin levels are learly above the desired levels as well as patients whose glyated hemoglobin levels are less egregiously elevated. Our finding is onsistent with the Cohrane Collaboration review, suggesting the benefit of individual eduation on glyemi ontrol (24). However, we did not see onvining improvements in patient-reported diabetes symptoms, diabetes distress, depression, or other linial (e.g., blood pressure) or laboratory (e.g., lipid) values. We advise aution in generalizing our findings. The interventions took plae through ommunity physiian praties and were implemented through eletroni ommuniations. Physiians in the ommunity have different experienes with and aess to resoures, inluding aess to speialists, linial pratie guidelines, and experiene or use of eletroni ommuniation. We attempted to address these differenes by enrolling multiple ommunity physiians to partiipate in the study and randomization at the pratie level. The patient population in the study may also be distintive beause private health are insurane overage and aess to the Internet (either at work or home) were required. Although not all partiipants provided data at all planned study visits, we addressed missing data in this study in two ways. First, the primary analysis used mixed-effets models, whih have the effet of impliitly imputing missing observations (25). Seond, we performed the WEE sensitivity analysis that used baseline harateristi data to upweight observations from partiipants who were most similar to partiipants with missing data (22). As a measure of long-term blood gluose ontrol, hange in glyated hemoglobin is an important, ommonly used outome. Although low glyated hemoglobin does not imply that diabetes is being well managed, well managed diabetes is haraterized by glyated hemoglobin at normal or near-normal levels (13). We sreened.2,600 patients; 72% were ineligible beause glyated hemoglobin was lower than eligibility riterion; many physiians referred patients they thought were not adequately managing their diabetes beause of poor ontrol relevant to everyday life, suh as blurred vision or pain, self-assessed ontrol of diabetes, or depression (13). In this study, we did not observe onvining hanges in these indiators. Communiations as speifi for these indiators as ours were for glyated hemoglobin may be able to make a larger differene in future studies. Future studies should also onsider how mobile ommuniation hanges behavior related to blood gluose: mediation adherene, treatment intensifiation, inreased physial ativity, and number and quality of ommuniations between providers and patients. These may be important mehanisms to explain hange in glyated hemoglobin but were not primary or seondary analyses planned for this study. Future studies of mobile health should address more speifi haraterization of patient and provider behaviors that support hange in linial health parameters. Mobile phones are ubiquitous more than 2.7 billion people own mobile phones worldwide. In the United States alone, users have inreased from 34 million in 1995 to 290 million in Mobile phone and Internet users are inreasingly diverse in age and rae. The widespread distribution of mobile phones and eletroni ommuniation, aross soioeonomi, sex, and age-groups, ombined with the ability to proess and ommuniate data in real time, make these modalities ideal platforms to reate simple, effetive, diabetes management programs (14). We found mobile phone and web portal ommuniations for diabetes to have a onsequential treatment effet when used by patients and their PCPs. Aknowledgments This researh projet is funded through ontributions by WellDo, CareFirst Blue Cross/Blue Shield of Maryland, LifeSan, and Sprint. Additional funding was provided by the Maryland Industrial Partnerships program through the University of Maryland, an initiative of the A. James Clark Shool of Engineering s Maryland Tehnology Enterprise Institute. No potential onflits of interest relevant to this artile were reported. C.C.Q., M.D.S., M.L.T. and A.L.G.-B. were responsible for the design, data analyses, writing, and review of the manusript. E.A.B. was responsible for the data analyses and manusript review. S.H.B. ontributed to the writing and review of the manusript. The funders of this study did not play a role in the design and ondut of the study; olletion, management, analysis, and interpretation of the data, or in the preparation of the manusript. WellDo did not have veto power over or have say about hanging any manusript text other than the desription of the software oahing system they provided. Dr. Ram Miller served as the DSMB.The authors thank Dr. Nanette Steinle, Assistant Professor, Division of Endorinology, Diabetes and Nutrition, University of Maryland, and Interim Chief, Endorine Setion, Baltimore Veterans Medial Center, for advie on our manusript and linial laboratory data. Referenes 1. Huang ES, Basu A, O Grady M, Capretta JC. Projeting the future diabetes population size and related osts for the U.S. Diabetes Care 2009;32: Cowie CC, Rust KF, Ford ES, et al. Full aounting of diabetes and pre-diabetes in the U.S. population in and Diabetes Care 2009;32: Gaede PH, Jepsen PV, Larsen JN, Jensen GV, Parving HH, Pedersen OB. The Steno-2 study. Intensive multifatorial intervention redues the ourrene of ardiovasular disease in patients with type 2 diabetes. Ugeskr Laeger 2003;165: Solomon CG. Reduing ardiovasular risk in type 2 diabetes. N Engl J Med 2003; 348: Knowler WC, Barrett-Connor E, Fowler SE, et al. Redution in the inidene of type 2 diabetes with lifestyle intervention or metformin. N Engl J Med 2002;346: Cleland JG, Ekman I. Enlisting the help of the largest health are workfore: patients. JAMA 2010;304: Amerian Diabetes Assoiation. Standards of medial are in diabetes: Diabetes Care 2008;31(Suppl. 1):S12 S54 8. Amerian Diabetes Assoiation. Standards of medial are in diabetes: Diabetes Care 2009;32(Suppl. 1):S13 S61 9. Saaddine JB, Cadwell B, Gregg EW, et al. Improvements in diabetes proesses of are and intermediate outomes: United States, Ann Intern Med 2006; 144: Peyrot M, Rubin RR, Lauritzen T, Snoek FJ, Matthews DR, Skovlund SE. Psyhosoial problems and barriers to improved diabetes management: results of the Cross- National Diabetes Attitudes, Wishes and Needs (DAWN) Study. Diabet Med 2005; 22: Renders CM, Valk GD, Griffin S,Wagner EH, Eijk JT, Assendelft WJ. Interventions to improve the management of diabetes mellitus in primary are, outpatient and ommunity settings. Cohrane Database Syst Rev 2001:CD Griffin S, Kinmonth AL. Diabetes are: the effetiveness of systems for routine surveillane for people with diabetes. Cohrane Database Syst Rev 2000:CD Quinn CC, Gruber-Baldini AL, Shardell M, et al. Mobile diabetes intervention study: testing a personalized treatment/behavioral ommuniationinterventionforbloodgluose ontrol. Contemp Clin Trials 2009;30: Quinn CC, Clough SS, Minor JM, Lender D, Okafor MC, Gruber-Baldini A. WellDo mobile diabetes management randomized ontrolled trial: hange in linial and behavioral outomes and patient and physiian 8 DIABETES CARE are.diabetesjournals.org

9 satisfation. Diabetes Tehnol Ther 2008;10: Kroenke K, Spitzer RL, Williams JB. The PHQ-9: validity of a brief depression severity measure. J Gen Intern Med 2001;16: Whitty P, Steen N, Eles M, et al. A new self-ompletion outome measure for diabetes: is it responsive to hange? Qual Life Res 1997;6: MColl E, Steen IN, Meadows KA, et al. Developing outome measures for ambulatory are: an appliation to asthma and diabetes. So Si Med 1995;41: Polonsky WH, Fisher L, Earles J, et al. Assessing psyhosoial distress in diabetes: development of the diabetes distress sale. Diabetes Care 2005;28: Fisher L, Glasgow RE, Mullan JT, Skaff MM, Polonsky WH. Development of a brief diabetes distress sreening instrument. Ann Fam Med 2008;6: Shea S, Weinstok RS, Starren J, et al. A randomized trial omparing telemediine ase management with usual are in older, ethnially diverse, medially underserved patients with diabetes mellitus. J Am Med Inform Asso 2006;13: Hayes RJ, Bennett S. Simple sample size alulation for luster-randomized trials. Int J Epidemiol 1999;28: Quinn and Assoiates 22. Robins J, Rotnitzky A, Zhao L. Analysis of semiparametri regression models for repeated outomes under the presene of missing data. J Am Stat Asso 1995;90: Gerstein HC, Miller ME, Byington RP, et al. Effets of intensive gluose lowering in type 2 diabetes. N Engl J Med 2008; 358: Duke SA, Colagiuri S, Colagiuri R. Individual patient eduation for people with type 2 diabetes mellitus. Cohrane Database Syst Rev 2009:CD Verbeke G, Lesaffre E. Repeated Measurements. Wiley,2007 are.diabetesjournals.org DIABETES CARE 9