Do Mobile Phone ApplicationsImprove Glycemic Control (HbA1c)in the Self-management ofDiabetes? A Systematic Review,Meta-analysis, and GRADE of14 Randomized TrialsDiabetes Care 2016;39:2089–2095 | DOI: 10.2337/dc16-0346

OBJECTIVE

To investigate the effect of mobile phone applications (apps) on glycemic control(HbA1c) in the self-management of diabetes.

RESEARCH DESIGN AND METHODS

Relevant studies that were published between 1 January 1996 and 1 June 2015weresearched from five databases: Medline, CINAHL, Cochrane Library, Web of Science,and Embase. Randomized controlled trials that evaluated diabetes apps were in-cluded. We conducted a systematic review with meta-analysis and GRADE (Gradingof Recommendations Assessment, Development and Evaluation) of the evidence.

RESULTS

Participants from 14 studies (n = 1,360) were included and quality assessed.Although there may have been clinical diversity, all type 2 diabetes studiesreported a reduction in HbA1c. The mean reduction in participants using an appcompared with control was 0.49% (95% Cl 0.30, 0.68; I2 = 10%), with a moderateGRADE of evidence. Subgroup analyses indicated that younger patients weremore likely to benefit from the use of diabetes apps, and the effect size wasenhanced with health care professional feedback. There was inadequate datato describe the effectiveness of apps for type 1 diabetes.

CONCLUSIONS

Apps may be an effective component to help control HbA1c and could be consid-ered as an adjuvant intervention to the standard self-management for patientswith type 2 diabetes. Given the reported clinical effect, access, and nominal cost ofthis technology, it is likely to be effective at the population level. The functionalityand use of this technology need to be standardized, but policy and guidance areanticipated to improve diabetes self-management care.

The number of patients with diabetes globally is expected to rise to over 500 million by2030 (1). There is an urgent need for an improved self-management suite of interven-tions. For self-management tobeeffective, it needs tobe structuredand cost-effective (2)and be widely accessible across all health economies, including the developing world (2).

1Division of Population Medicine, Cardiff Univer-sity School of Medicine, College of Biomedicaland Life Sciences, Cardiff University, Heath Park,Cardiff, U.K.2Cochrane Skin Group, School of Medicine, TheUniversity of Nottingham, Nottingham, U.K.

Corresponding author: Ben Carter, ben.carter@kcl.ac.uk.

Received 17 February 2016 and accepted 3 July2016.

This article contains Supplementary Data onlineat http://care.diabetesjournals.org/lookup/suppl/doi:10.2337/dc16-0346/-/DC1.

This article is featured in a podcast available athttp://www.diabetesjournals.org/content/diabetes-core-update-podcasts.

B.C. and S.M. are joint senior authors.

B.C. is currently affiliatedwith the Department ofBiostatistics and Health Informatics, Institute ofPsychiatry, Psychology & Neuroscience, King’sCollege London, London, U.K.

© 2016 by the American Diabetes Association.Readers may use this article as long as the workis properly cited, the use is educational and notfor profit, and the work is not altered. More infor-mation is available at http://www.diabetesjournals.org/content/license.

Can Hou,1 Ben Carter,1,2 Jonathan Hewitt,1

Trevor Francisa,1 and Sharon Mayor1

Diabetes Care Volume 39, November 2016 2089

META

-ANALYSIS

As a newly emerging technology, di-abetesmobile phone applications (here-after referred to as diabetes apps) area promising tool for self-management.Wedefinediabetes apps asmobile phonesoftware that accepts data (transmittedor manual entry) and provides feedbackto patients on improved management(automated or by a health care profes-sional [HCP]). This technology combinesthe functions of the mobile phone, wire-less network for data transmission, andsometimes HCPs for providing feedback.Due to its ubiquitous, low-cost, interac-tive, and dynamic health promotion,there is potential for diabetes apps toprovide an effective intervention in dia-betes self-care.In terms of diabetes self-management,

numerous studies have proven the effec-tiveness of other telemedicine technol-ogies, such as short message service (3),computer-based interventions (4), andweb-based interventions (3,5). Com-pared with these telemedicine interven-tions, diabetes apps are advantageousin that they are global, cheaper, conve-nient, and more interactive. There is,however, current uncertainty on theclinical effectiveness of diabetes appsin diabetes self-management (6–9).

RESEARCH DESIGN AND METHODS

Data Sources and Search StrategyThe PRISMA statement and checklistwas followed. Five electronic databa-ses were searched (Medline, CINAHL,Cochrane Library, Web of Science, andEmbase) for studies published between1 January 1996 and 1 June 2015. Thereferences of the included studies werehand searched to identify any addi-tional articles. The following terms andmedical subject headings (MeSH) wereused during the search: (mobile ORmHealth OR cell phone OR MeSH “Cellu-lar Phone” OR MeSH “Smartphone” ORapp OR MeSH “Mobile Applications”)AND (MeSH “Diabetes Mellitus” OR dia-bete* OR T2DM OR T1DM OR IDDM ORNIDDM).

Inclusion and Exclusion CriteriaThe inclusion criteria were as follows:the participants were over 18 years oldand had type 1 or type 2 diabetes, thestudies were randomized controlled tri-als (RCTs), the control group in the studyreceived usual diabetes care withoutany telehealth programs, and baseline

and follow-up mean for HbA1c were re-ported (or could be calculated). Exclu-sion criteria were as follows: simulatedor self-reported HbA1c data, computeror other mobile terminal–based diabe-tes apps, diabetes apps were exclusivelydesigned for HCPs, and diabetes appswere exclusively designed for providinggeneral education or allowing commu-nication between patients and HCPs.

Two reviewers (C.H. and T.F.) searchedthe literature and assessed the studies in-dependently. Any disagreements wereresolved through discussion with a thirdreviewer (B.C.). No language restrictionswere applied.

Data ExtractionParticipant demographics, study designconsiderations, and context were ex-tracted from the included studies. Tworeviewers independently carried outthe data extraction (C.H. and T.F.). Studyauthors were contacted to provide ad-ditional data, and missing SDs were es-timated by calculation (10).

Quality AssessmentThe quality assessment was conductedby two reviewers independently (C.H.and T.F.), using the quality rating toolproposed by the U.S. Preventive Ser-vices Task Force (11). Seven criteriawere used to assess quality: baselinecomparability of the groups, the main-tenance of comparability of the groups,differential or high loss to follow-up, reli-able and valid measurement, clear defini-tion of the intervention, consideration ofimportant outcomes, and an intention-to-treat analysis. The quality of each studywasgraded as good, fair, or poor. To be rated asgood, studies needed to meet all the crite-ria. A study was rated as poor if one (ormore) domain was assessed as having a se-rious flaw. Studies that met some but notall of the criteria were rated as fair quality.

Data AnalysisChanges in HbA1c, or HbA1c at follow-up,were compared between groups using amean difference and were presentedwith an associated 95% CI.When studiesinvestigated interventions and contextsthat were both deemed clinically similarand free from statistical heterogeneity,pooling was carried out using an inversevariance random-effects model (12).Meta-analyses were conducted using theComprehensive Meta-Analysis Software(version 2.2). The level of evidence was

applied to the GRADE (Grading of Recom-mendations Assessment, Developmentand Evaluation) criteria and reported.

Heterogeneity and SubgroupAnalysesHeterogeneity was assessed and quanti-fied using the I2 statistic. When substan-tial heterogeneity was found (I2. 50%),further exploration using subgroup anal-ysis was undertaken. For type 2 diabetesstudies, subgroup analyses were as fol-lows: follow-up duration (,6 monthsvs..6 months), length of time with di-abetes (,9 years vs. .9 years), age ofparticipants (mean age ,55 years oldvs. .55 years old), number of self-monitoring tasks supported by the dia-betes apps (#3 vs. .3), and types offeedback provided. No type 1 diabetessubgroup analyses were performeddue to the small number of studies.

Sensitivity Analyses and PublicationBiasAdditional analyses were carried out onstudies with the following: good or fairquality, complete information, and abaseline HbA1c level ,9.0%. A funnelplot was used to visually inspect pub-lication bias where 10 or more studieswere pooled.

RESULTS

Identified and Included StudiesSearches identified 5,209 articles; 4,238were screened after removing duplicaterecords and 4,178 were excluded. Sixtystudies were eligible for full text reviewand 42 were excluded (Fig. 1), resultingin 14 included studies. Four studies ex-amined type 1 diabetes and 10 studiesexamined type 2 diabetes.

Characteristics of the IncludedStudies and Quality AssessmentIn the 14 studies, there were 1,360 par-ticipants: 509 and 851 with type 1 and type2 diabetes, respectively (SupplementaryTable 1). In the type 1 diabetes studies,the mean age of participants rangedfrom 34 (13) to 36 years old (14), andthe mean duration of diabetes rangedfrom 16 (13–15) to 19 years (15). Twostudies were undertaken in Europe(13,14), one in Australia (15), and onewas multinational (16). In the type 2 di-abetes studies, the mean age of the par-ticipants was much higher, ranging from51 (17) to 62 years old (18), and themean duration of diabetes ranged from

2090 Phone Apps, Diabetes, and Glycemic Control Diabetes Care Volume 39, November 2016

5 (19) to 13 years (20) from six stud-ies. Four studies were undertaken inEurope (18,20–22), three in the U.S.(17,23,24), two in Asia (19,25), and onein Africa (26).One type 1 diabetes studywas assessed

as good quality (14), two were rated asfair (13,16), and one was rated as poor(15) (for further details see Supple-mentary Table 2). For type 2 diabetes stud-ies, one was rated as good quality (21),six were rated as fair (17–19,22,24,25),and three were rated as poor (20,23,26)(Supplementary Table 2).

Apps Featured in the Included StudiesTwelve diabetes apps were identifiedand examined in this review, with six do-mains of functionality (SupplementaryTable 3); details of the feedback provided

by each can be seen in SupplementaryTable 4.

Type 1 Diabetes Apps

Three apps were used for participantswith type 1 diabetes and aimed to helppatients to calculate the most appropri-ate insulin bolus on the basis of patientblood glucose (BG) levels, food intake,and physical activity. Data for all threeapps were manually entered. One studyreported that there was little impact ofthe app on the total time spent on face-to-face or telephone follow-up and con-cluded that the software did not requiremore time for patients to manage theirdiabetes (13). A further study estimatedthe average cost to patients and educa-tors’ time was £38 per patient, attrib-uted to the app over a 9-month period

(15). HCP feedback was provided in allapps, with a frequency ranging from everyweek to every 3 weeks (SupplementaryTable 4).

Type 2 Diabetes Apps

Nine apps were used for participants withtype 2 diabetes. The apps were designedto improve patient self-managementby providing personalized feedback onself-monitoring data, such as BG, foodintake, and physical activity. In eight ofthe apps, BG was automatically trans-ferred and other data was manually en-tered, with one exception where bloodpressure, body weight, and pedometerwere also automatically transferred(25). Quinn et al. (17) reported thatthe app was associated with shorterconsultation times. Among seven appswith HCP feedback, three providedfeedback when needed (e.g., patientdata were considered abnormal). Inthe other apps, the frequency offeedback ranged from once a weekto once every 3 months (SupplementaryTable 4).

Effectiveness of the Apps

Type 1 Diabetes

There weremixed results from the type 1diabetes studies. Two studies (14,16)foundnodifferencebetween the interven-tion group and the control group andtwo studies (13,15) reported statisticallysignificant results that favored the apps.There was a statistically insignificantdifference in HbA1c between the appsand control group of 20.36% (95%Cl 20.87, 0.14; P = 0.16; I2 = 87%)(Fig. 2). No subgroup analyses werereported.

Type 2 Diabetes

All 10 studies of type 2 diabetes re-ported a reduction of HbA1c in partici-pants using an app, with a medianreduction of 0.55% (range 0.15–1.87).After pooling, the mean reduction inHbA1c was 0.49% (95% Cl 0.30, 0.68;P, 0.01; I2 = 10%) (Fig. 3). These resultsexhibited consistent findings with noheterogeneity. One study reported a re-duction larger than clinically antici-pated, which raised debate over thelegitimacy of their findings (26). Afterexcluding the subgroup of studies thatwere assessed as poor quality, we founda mean reduction of 0.41% (95% CI0.22, 0.61; P , 0.001; I2 = 0%) (Fig. 3).The level of evidence by GRADE was

Figure 1—PRISMA flowchart of included studies.

care.diabetesjournals.org Hou and Associates 2091

moderate due to the findings beingdowngraded due to quality.

Type 2 Diabetes Subgroup Analyses

The subgroup analysis by follow-up du-ration showed that five studies with ashorter follow-up duration (,6 months)displayed a larger (but nonsignificant)HbA1c reduction than those with alonger duration (.6 months), 0.62 vs.0.40% (P = 0.33), respectively. Therewas no difference in the reduction ofHbA1c in three studies with a mean di-abetes duration of ,9 years (0.53%)compared with those with a duration$9 years (0.55%; P = 0.93). Studies ofyounger participants with a mean ageof #55 years reported a larger and clini-cally significant reduction in HbA1c levelof 1.03% compared with 0.41% in thosewith an average age of .55 years, butthe result was not found to be statisti-cally significant (P = 0.10).In the subgroup analysis by number of

self-monitoring tasks, six diabetes appssupported at most three self-monitoringtasks and had results similar to the stud-ies with more than three self-monitoringtasks (mean reduction of 0.44 vs. 0.58%;

P = 0.56). Two studies of diabetes appswith only automated feedback had asmall and statistically nonsignificantchange in HbA1c of –0.26% (95% CI –0.62,0.09). When the diabetes apps that in-cluded HCP feedback were pooled,eight studies reported a reduction of0.56% (95% Cl 0.35, 0.78). There was nostatistically significant difference be-tween HCP verses automatic feedbacksubgroup (P = 0.16).

Four sensitivity analyses were under-taken to test the robustness of the re-sults. Removing three studies (20,23,26)with poor quality reported a mean re-duction of 0.41% (95% Cl 0.22, 0.61)(Fig. 3). The removal of one study (17)with incomplete statistical informationwas associated with a mean reductionof 0.48% (95% CI 0.28, 0.67), and theexclusion of one study (20) conductedon mixed participants with type 1 andtype 2 diabetes had an attendant meanreduction of 0.48% (95% Cl 0.27, 0.69).Finally, the exclusion of two studies(17,23) with baseline HbA1c levels.9.0%was associated with a mean reduction of0.47% (95% Cl 0.25, 0.69).

CONCLUSIONS

Ten studies were included for type 2 di-abetes, predominately of fair quality.The results of these indicated a con-sistent reduction in HbA1c of 0.5%. Al-though there was no indication ofheterogeneity, the study conducted byTakenga et al. (26) introduced a largeeffect that was likely to be caused bypoor study quality (high attrition rate,differential loss to follow-up, and highbaseline HbA1c level). Thus, studieswere stratified into subgroups deter-mined by their quality assessment (27).No differences were found between thesubgroups, and the studies of poorquality were included for completenessand to highlight the challenges in studydesign.

Five subgroup analyses showed thatthe effect did not differ significantly byfollow-up duration, mean diabetes du-ration of participants, mean age of par-ticipants, number of self-monitoringtasks supported by the diabetes apps,or types of feedback. Compared withstudies that investigated the effective-ness of alternative interventions such

Figure 2—Pooled type 1 diabetes studies of HbA1c comparison of apps vs. control.

2092 Phone Apps, Diabetes, and Glycemic Control Diabetes Care Volume 39, November 2016

as text messaging, mobile device use,and computer-based and conventionalself-management, we have found thatapps offer promising results and rein-force the message argued by other au-thors (3,4,28–30). The evidence for thisfinding by GRADE was moderate, afterdowngrading due to quality.The subgroup analysis by follow-up

duration suggested that the effect ofdiabetes apps on BG control may atten-uate over time. A possible rationale forthis subgroup effect is a lack in userfriendliness, a lack in perceived ad-ditional benefits, and a lack of use ofgamification elements, resulting in alack of efficacy following use (31). Thesubgroup analysis by mean age of partic-ipants indicated that younger patientswere more likely to benefit from the

use of the diabetes apps. It may be spec-ulated that younger patients are moreamenable to new technologies andmore familiar with the use of mobilephones. The subgroup analysis by per-sonalized feedback system highlightedthe gap between automated feedbackand HCP feedback. Although automatedfeedback has the advantage of beinginteractive and dynamic, there is a limitto presupposed scenarios, whereasfeedback provided by HCPs was moreindividual, especially in emergencysituations. Feedback options rangedwidely between the apps, but it is pos-tulated that it was the feedback thattriggered improved lifestyle choices,which in turn lowered HbA1c. None ofthe five sensitivity analyses changed theoverall effect size significantly, which

suggests that the findings are not sen-sitive to these scenarios. The results ofour meta-analysis lend support to theuse of diabetes apps in diabetes self-management, especially for type 2 di-abetes. However, we have highlighted anumber of limitations of current diabetesapps.

For type 1 diabetes, there was littledifference in HbA1c between interven-tion and control groups and the resultswere associated with considerable het-erogeneity. The level of evidence byGRADE was downgraded to very lowdue to study quality, inconsistency,and uncertainty, so the findings shouldbe interpreted as very uncertain andlikely to change after future research.Furthermore, none of the apps in theincluded type 1 diabetes studies had

Figure 3—Pooled type 2 diabetes studies of HbA1c comparison of apps vs. control.

care.diabetesjournals.org Hou and Associates 2093

an automatic data uploading functional-ity. In future studies for type 1 diabetes,we encourage investigators to includeapps with this functionality, not onlyfor the purpose of being user friendlybut also for safety concerns by reducingthe risk of data entry errors.Two studies reported on the cost-

effectiveness of the apps for type 1 dia-betes with inconclusive findings (15,16).Of three studies on type 2 diabetes thatdiscussed compliance, two highlightedpoor compliance, with only 35% of pa-tients being recorded as regular app users(21,24). One study (25) reported a declinein patient use over time, from 70% in the1st week to 50% in the last 2 weeks. Fourstudies tried to explore the mechanismsbehind the effects, but the conclusionswere inconsistent (15,17,21,24). Wepostulate that diabetes apps influencelifestyle choice, but how this occurs isunclear. One hypothesis is that the re-minder and feedback features of diabe-tes apps can lead to improvement inhealth beliefs, self-efficacy, and socialsupport (32).By the end of the decade, worldwide

mobile phone usage is anticipated toexceed 5 billion (33). Therefore appsmay be able to offer an affordable andwidely available adjunct to diabetes self-management. We have included stud-ies across a variety of health care systems,from both the developed and develop-ing world, so we argue that the appsare currently available and could formthe basis of improved health promo-tion on diabetes education and self-management.This study had several limitations.

Since this review was restricted to pub-lished studies, publication bias cannotbe ruled out, as highlighted by other in-vestigators (30). All included study de-signs were not blinded and so weredowngraded in the quality assessmenttool (highlighting the increased riskof ascertainment bias). Furthermore,patient-important outcomes and behav-ioral mechanisms were not consideredand highlighted as a clear gap to beaddressed in future studies. A furtherweakness is that some of the effect at-tributed to the apps could be explainedby health care providers. Finally, there isno clear definition of diabetes apps, andstudy authors defined their interven-tions in different ways as a result. In thisreview, we defined diabetes apps as

software that is designed for use on amobile phone allowing patients to enterdata into the app and receive feedback.

The implications for future researchinclude establishing a common stan-dardized platform of functionality. In-vestigators of future studies need toconsider adequately powered prag-matic RCTs with secure sequence gener-ation, concealed allocation, use of anactive control app, and comparable ac-cess to HCPs. Features such as thesemight reduce the impact of ascertain-ment bias, and effects due to HCPs.RCTs with longer duration of follow-up(.6 months) using standardized apptechnology may well demonstrate ben-eficial clinical effects in type 2 diabetes.Furthermore, there is significant scopefor research in the use of apps in otherareas of self-management, such as in-creasing physical activity, weight loss,and smoking cessation.

In a clinical context, we recommendthat HCP feedback should be central inall future app designs and supplementedwith dynamic automated feedback. Fu-ture technology should also be under-pinned by behavior change theories andgamification elements to achieve a largereffect on BG control and improve compli-ance of patients in using diabetes apps.Finally, future technology should alsoconsider the needs of older patients.

Acknowledgments. The authors acknowledgeDr. Haroon Ahmed, of the Division of PopulationMedicine, Cardiff University School ofMedicine,and the editorial base at Diabetes Care for theirclinical and methodological guidance thatstrengthened the manuscript.This article is an honest, accurate, and trans-

parent account of the study being reported; noimportant aspects of the study have been omit-ted, and any discrepancies from the study asplanned (and, if relevant, registered) have beenexplained.

Duality of Interest. No potential conflicts ofinterest relevant to this article were reported.Author Contributions. C.H. designed the pro-tocol, searched the literature, extracted thedata, carried out the analysis, and drafted themanuscript. B.C. interpreted the results anddrafted the manuscript. J.H. reviewed the man-uscript and advised on the clinical context ofthe review. T.F. searched the literature and ex-tracted the data. S.M. designed the protocol, in-terpreted the results, and contributed to themanuscript. B.C. is the guarantor of this workand, as such, had full access to all the data inthe study and takes responsibility for the in-tegrity of the data and the accuracy of the dataanalysis.

References1. Whiting DR, Guariguata L, Weil C, Shaw J. IDFdiabetes atlas: global estimates of the preva-lence of diabetes for 2011 and 2030. DiabetesRes Clin Pract 2011;94:311–3212. Polonsky WH, Fisher L. Self-monitoring ofblood glucose in noninsulin-using type 2 dia-betic patients: right answer, but wrong ques-tion: self-monitoring of blood glucose can beclinically valuable for noninsulin users. DiabetesCare 2013;36:179–1823. Liang X, Wang Q, Yang X, et al. Effect of mo-bile phone intervention for diabetes on glycae-mic control: a meta-analysis. Diabet Med 2011;28:455-4634. Pal K, Eastwood SV, Michie S, et al. Computer-based diabetes self-management interventionsfor adults with type 2 diabetesmellitus. CochraneDatabase Syst Rev 2013;3:CD0087765. Angeles RN, Howard MI, Dolovich L. The ef-fectiveness of web-based tools for improvingblood glucose control in patients with diabetesmellitus: a meta-analysis. Can J Diabetes 2011;35:344–3526. Tatara N, Arsand E, Nilsen H, Hartvigsen G. Areview of mobile terminal-based applicationsfor self-management of patients with diabetes.In Proceedings of the International Conferenceon Ehealth, Telemedicine, and Social Medicine,Cancun, Mexico, 2009, p. 166–1757. Baron J, McBain H, Newman S. The impact ofmobilemonitoring technologies on glycosylatedhemoglobin in diabetes: a systematic review.J Diabetes Sci Technol 2012;6:1185–11968. El-Gayar O, Timsina P, Nawar N, Eid W.A systematic review of IT for diabetes self-management: are we there yet? Int J Med Inform2013;82:637–6529. Klonoff DC. The current status of mHealthfor diabetes: will it be the next big thing? J Dia-betes Sci Technol 2013;7:749–75810. Ma J, LiuW, Hunter A, ZhangW. Performingmeta-analysis with incomplete statistical infor-mation in clinical trials. BMCMed ResMethodol2008;8:5611. Harris RP, Helfand M, Woolf SH, et al.;Methods Work Group, Third US Preventive Ser-vices Task Force. Current methods of the USPreventive Services Task Force: a review ofthe process. Am J Prev Med 2001;20(Suppl.):21–3512. Deeks JJ, Higgins JPT, Altman DG. Analysingdata and undertaking meta-analyses, chapter 9.In Cochrane Handbook for Systematic Reviewsof Interventions. Higgins JPT, Green S, Eds.WileyOnline Library, 200813. Charpentier G, Benhamou P-Y, Dardari D,et al.; TeleDiab Study Group. The Diabeo soft-ware enabling individualized insulin dose ad-justments combined with telemedicinesupport improves HbA1c in poorly controlledtype 1 diabetic patients: a 6-month, random-ized, open-label, parallel-group, multicentertrial (TeleDiab 1 Study). Diabetes Care 2011;34:533–53914. Rossi MC, Nicolucci A, Lucisano G, et al.; DidStudy Group. Impact of the “Diabetes Interac-tive Diary” telemedicine system on metaboliccontrol, risk of hypoglycemia, and quality oflife: a randomized clinical trial in type 1 dia-betes. Diabetes Technol Ther 2013;15:670–679

2094 Phone Apps, Diabetes, and Glycemic Control Diabetes Care Volume 39, November 2016

15. Kirwan M, Vandelanotte C, Fenning A,Duncan MJ. Diabetes self-management smart-phone application for adults with type 1 diabe-tes: randomized controlled trial. J Med InternetRes 2013;15:e23516. Rossi MC, Nicolucci A, Di Bartolo P, et al.Diabetes Interactive Diary: a new telemedicinesystem enabling flexible diet and insulin therapywhile improving quality of life: an open-label,international, multicenter, randomized study.Diabetes Care 2010;33:109–11517. Quinn CC, Clough SS, Minor JM, Lender D,Okafor MC, Gruber-Baldini A. WellDoc mobilediabetes management randomized controlledtrial: change in clinical and behavioral outcomesand patient and physician satisfaction. DiabetesTechnol Ther 2008;10:160–16818. Orsama AL, Lahteenmaki J, Harno K, et al.Active assistance technology reduces glycosy-lated hemoglobin and weight in individualswith type 2 diabetes: results of a theory-basedrandomized trial. Diabetes Technol Ther 2013;15:662–66919. Yoo HJ, Park MS, Kim TN, et al. A ubiquitouschronic disease care system using cellular phonesand the internet. Diabet Med 2009;26:628–63520. Istepanian RS, Zitouni K, Harry D, et al. Eval-uation of amobile phone telemonitoring systemfor glycaemic control in patients with diabetes.J Telemed Telecare 2009;15:125–12821. Holmen H, Torbjørnsen A, Wahl AK, et al. Amobile health intervention for self-management

and lifestyle change for persons with type 2 di-abetes, part 2: one-year results from the Norwe-gian randomized controlled trial RENEWINGHEALTH. JMIR Mhealth Uhealth 2014;2:e5722. Nagrebetsky A, Larsen M, Craven A, et al.Stepwise self-titration of oral glucose-loweringmedication using a mobile telephone-basedtelehealth platform in type 2 diabetes: a feasi-bility trial in primary care. J Diabetes Sci Technol2013;7:123–13423. Quinn CC, Shardell MD, Terrin ML, Barr EA,BallewSH,Gruber-Baldini AL. Cluster-randomizedtrial of a mobile phone personalized behavioralintervention for blood glucose control. DiabetesCare 2011;34:1934–194224. Faridi Z, Liberti L, Shuval K, Northrup V, AliA, Katz DL. Evaluating the impact of mobile tele-phone technology on type 2 diabetic patients’self-management: the NICHE pilot study. J EvalClin Pract 2008;14:465–46925. Waki K, Fujita H, Uchimura Y, et al.DialBetics: a novel smartphone-based self-management support system for type 2 diabetespatients. J Diabetes Sci Technol 2014;8:209–21526. Takenga C, Berndt R-D, Musongya O, et al.An ICT-based diabetes management sys-tem tested for health care delivery in the afri-can context. Int J Telemed Appl 2014;2014:43730727. Higgins JP, Altman DG, Gøtzsche PC, et al.;Cochrane BiasMethods Group; Cochrane Statis-ticalMethods Group. The Cochrane Collaboration’s

tool for assessing risk of bias in randomised trials.BMJ 2011;343:d592828. Steinsbekk A, Rygg LØ, Lisulo M, RiseMB, Fretheim A. Group based diabetes self-management education compared to routinetreatment for people with type 2 diabetes mel-litus. A systematic review with meta-analysis.BMC Health Serv Res 2012;12:21329. de Jongh T, Gurol-Urganci I, Vodopivec-Jamsek V, Car J, Atun R. Mobile phone messag-ing for facilitating self-management of long-termillnesses. Cochrane Database Syst Rev 2012;12:CD00745930. Su D, Zhou J, Kelley MS, et al. Does tele-medicine improve treatment outcomes for di-abetes? A meta-analysis of results from 55randomized controlled trials. Diabetes Res ClinPract 2016;116:136–14831. ScheibeM, Reichelt J, BellmannM, KirchW.Acceptance factors of mobile apps for diabetesby patients aged 50 or older: a qualitative study.Med 2 0 2015;4:e132. Nundy S, Dick JJ, Solomon MC, Peek ME.Developing a behavioral model for mobilephone-based diabetes interventions. PatientEduc Couns 2013;90:125–13233. Statista. Number of mobile phone usersworldwide from 2013 to 2019 (in billions) [In-ternet]. The Statistics Portal. Available fromhttp://www.statista.com/statistics/274774/forecast-of-mobile-phone-users-worldwide/.Accessed 1 January 2016

care.diabetesjournals.org Hou and Associates 2095