Type 2 Diabetes in the Real World:The Elusive Nature of GlycemicControlDiabetes Care 2017;40:1425–1432 | https://doi.org/10.2337/dc16-1974

Despite U.S. Food and Drug Administration (FDA) approval of over 40 new treatmentoptions for type 2 diabetes since 2005, the latest data from the National Health andNutrition Examination Survey show that the proportion of patients achieving glycatedhemoglobin (HbA1c) <7.0% (<53 mmol/mol) remains around 50%, with a negligibledecline between the periods 2003–2006 and 2011–2014. The Healthcare Effective-ness Data and Information Set reports even more alarming rates, with only about40%and 30%of patients achievingHbA1c<7.0% (<53mmol/mol) in the commerciallyinsured (HMO) andMedicaid populations, respectively, again with virtually no changeover the past decade. A recent retrospective cohort study using a large U.S. claimsdatabase explored why clinical outcomes are not keeping pace with the availability ofnew treatment options. The study found that HbA1c reductions fell far short of thosereported in randomized clinical trials (RCTs),with poormedication adherence emerg-ing as the key driver behind the disconnect. In this Perspective, we examine theimplications of these findings in conjunction with other data to highlight the discrep-ancy between RCT findings and the real world, all pointing toward the underrealizedpromise of FDA-approved therapies and the critical importance of medication adher-ence. While poor medication adherence is not a new issue, it has yet to be effectivelyaddressed in clinical practicedoften, we suspect, because it goes unrecognized. Tosupport the busy health care professional, innovative approaches are sorely needed.

ACHIEVEMENT OF HbA1c GOALS IN THE REAL WORLD

Large randomized controlled trials (RCTs) have clearly demonstrated that achieving andsustaining optimal glycemic control prevents or delays the development of microvas-cular and macrovascular disease (1–3). Although the risk of developing diabetes-related complications rises steadily when glycated hemoglobin (HbA1c) values are inexcess of 6.5% (48 mmol/mol) (4,5), an HbA1c of ,7% (,53 mmol/mol) is generallyconsidered a target goal for diabetes management (6). In truth, it is now widelyrecognized that individualizing the HbA1c goal for each patient is critically important,especially when the presence of comorbidities necessitates less stringent targets (6).However, most government and private insurer reports continue to use 7% as a pointof reference; therefore, we focus herein on this marker of glycemic control.Despite our growing understanding of diabetes and the availability of new medica-

tions and technologies, a substantial number of individuals are not at their glycemicgoal. Of note, recent data indicate that 85.6% of adults with diagnosed diabetes aretreated with diabetes medication (7). Results from the National Health and NutritionExamination Survey (NHANES) indicate that only about 50% of American adults withdiabetes are achieving HbA1c,7.0% (,53 mmol/mol) (8), and it is estimated that only64% are reaching individualized glycemic goals (9). These findings are noteworthy

1Division of Endocrinology and Metabolism,School of Medicine, University of California,San Diego, San Diego, CA2Veterans Affairs Medical Center, San Diego, CA3Taking Control of Your Diabetes, Del Mar, CA4Department of Psychiatry, University of Califor-nia, San Diego, San Diego, CA5Behavioral Diabetes Institute, San Diego, CA

Corresponding author: Steven V. Edelman,svedelman@vapop.ucsd.edu.

Received 13 September 2016 and accepted 25March 2017.

© 2017 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.

See accompanying articles, pp. 1469and 1588.

Steven V. Edelman1,2,3 and

William H. Polonsky4,5

Diabetes Care Volume 40, November 2017 1425

PER

SPEC

TIVES

INCARE

because they provide nationally repre-sentative snapshots with which to assesscare outcomes realized over time in real-world settings. After an initial increase inthe percentage of patients attainingHbA1c ,7.0% (,53 mmol/mol) betweenthe survey periods 1988–1994 (44%) and2003–2006 (57%), performance stag-nated, with a new wave of data showingpercentages for the 2007–2010 and2011–2014 periods at 52% and 51%, re-spectively (9) (Fig. 1A).The Healthcare Effectiveness Data and

Information Set (HEDIS) results, whichinclude data from over 1,000 health planscoveringover171millionpeople in2014, areevenmore troubling. In 2014, approximately

40% of commercially insured HMO patientsand 30% of government-insured patientsachieved HbA1c ,7.0 (,53 mmol/mol),again with no change over the past de-cade (10) (Fig. 1B).

While these worrisome findings arecertainly skewed, at least to some extent,by the unknown number of individualswhose individualized HbA1c targetsare .7%, it remains evident that a con-siderable percentage of U.S. adults withdiabetes are in persistent poor glycemiccontrol. The question of why these pop-ulation trends stand in such sharp con-trast to the generally favorable resultsemerging from RCTs has not been fullyexplored. This Perspective calls out the

distressing lack of improved glycemiccontrol in this country over the past de-cade, as well as the notable gap betweenclinical trial and real-world results, andprovides an urgent challenge to considernew approaches to achieving meaningfuland sustained outcomes.

WHY IS OUR LARGE AND EVER-GROWING DIABETES TOOLBOXNOT LEADING TO IMPROVED ANDSUSTAINED GLYCEMIC CONTROL?

The number of diagnosed cases of diabe-tes has increased fourfold from 1980through 2014 (from 5.5 million to 22.0million), with type 2 diabetes making upthe vast majority of cases (11). While agrowing number of options are availableto treat type 2 diabetesdover 40 newdrugs, including combination products,for type 2 diabetes have been approvedsince2005 (12)dthemost recent analysisof NHANES data currently being preparedfor publication suggests that this explo-sion in available options has not contrib-uted to a meaningful improvement inglycemic control (8,9). This has profoundimplications, considering that poorly con-trolled diabetes is a recognized cause ofsevere microvascular and macrovascularcomplications (7). If the present trends inincidence and prevalence continue with-out change, it is estimated that one-thirdof Americans will have diabetes by 2050,portending evenhigher rates ofmorbidityand mortality as a result of poor glycemiccontrol (13).

Efficacy Unrealized: The DisconnectBetween Clinical Trial and Real-WorldResultsRCTs have been the “gold standard” ofstudy design because they tightly controlthe setting and delivery of interventions,minimize the effect of external factors onoutcomes, and lead to a random distribu-tion of unmeasured confounders. Yet thedegree towhich results obtainedunderRCTconditions can be extrapolated to real liferemains an open question (14,15). Indeed,it is important to be aware of the ways inwhich clinical trial results might inflateexpectations of treatment efficacy. First,therapy interventions are often morefocused in the unusual setting of the clin-ical trial, where patientsmay benefit frommore frequent face-to-face visits, conve-nient access to therapy, closer monitor-ing, and wider availability of educationalresources and support services (14,16–18).Second, trial participants are often more

Figure 1—Percentage of patients with diabetes achieving HbA1c,7% over the last decade has notimproved. A: Nationally representative real-world snapshot of the percentage of patients withtype 1 or type 2 diabetes achieving HbA1c levels ,7% from 2003 to 2006 (N = 999) (8), 2007 to2010 (N = 1,444) (7), and 2011 to 2014 (N = 2,677) (9) using data derived from theNHANESdatabase.B: Only approximately 40% of patients with either type 1 or type 2 diabetes in the HMO populationor 30% of patients in the Medicaid population consistently achieved HbA1c levels,7% over a timeperiod spanning 2007 to 2014 accordingly to data obtained from the HEDIS database (10).

1426 Elusive Glycemic Control in the Real World Diabetes Care Volume 40, November 2017

concerned with their health and treat-ment and thusmoremotivated to activelyparticipate in their own care. Third, com-mitting to a protocol over a defined pe-riod of time, sometimes with the benefitof financial incentives, maymake it easierfor patients to follow medication instruc-tions appropriately (19,20).We now have the opportunity to use

administrative claims data in conjunctionwith clinical trial data to guide clinicaldecision-making. In the case of type 2 di-abetes, evidence from a new claims datastudy (21) may help to explain the dis-couraging NHANES and HEDIS findingsmentioned above and, in so doing, vali-date what we intuitively know to betruedthat clinical trial outcomes and ac-tual clinical care often tell different stories.Using a large electronic medical record–administrative claims database (OptumHumedica SmartFile, 2007–2014), thisretrospective analysis was undertakento assess HbA1c reductions in patientsinitiating a glucagon-like peptide 1 recep-tor agonist (GLP-1 RA) or a dipeptidyl pep-tidase 4 (DPP-4) inhibitor, to quantify thegap between real-world (i.e., usual care)and RCT efficacy, and to calculate the rel-ative contribution of various biobehav-ioral factors in explaining this gap (21).Specifically, investigators identified

adult patients with type 2 diabetes whoinitiated a GLP-1 RA (n = 221) or DPP-4inhibitor (n = 652) and then comparedtheir real-world HbA1c change over a12-month periodwith that of participantsin 11 pivotal RCTs, seven for the GLP-1 RAclass and four for theDPP-4 inhibitor class(22–32). Of note is that although HbA1cchange was measured over a 12-monthperiod for the real-world samples, onlythree of the 11 comparator RCTs were ofsimilar length (27,29,32). This may some-what reduce the gap between real-worldHbA1c change and the changes seen in thecomparator RCTs. Additionally, the dataset does not provide data regarding dos-ages in the real-world samples, whichmay differ from the dosages used in thecomparator RCTs.Nevertheless, the study(21) revealed that in patients with similarinclusion criteria characteristics, HbA1c re-ductions in the usual care group fell farshort of those reported in the RCTs: HbA1creductions in the real-world group were20.52% for the GLP-1 RAs and 20.51%for the DPP-4 inhibitors, whereas in theRCTs the changes ranged from 20.84%to 21.60% for the GLP-1 RA groups and

from 20.47% to 20.90% for the DPP-4inhibitor groups (Fig. 2). The finding thatHbA1c reductions were similar betweenthe real-world treatment groups forGLP-1 RAs and DPP-4 inhibitors was sur-prising considering that clinical trial com-parisons of incretin-based therapiesconsistently show greater HbA1c reduc-tions from baseline with GLP-1 RAs thanwith DPP-4 inhibitors (33).

To better understand the differences be-tween usual care and clinical trial HbA1cresults, multivariate regression analysisassessed the relative contributions of

key biobehavioral factors, including base-line patient characteristics, drug therapy,and medication adherence (21). Signifi-cantly, the key driver was poor medica-tion adherence, accounting for 75%of thegap (Fig. 3). Adherence was defined, fol-lowing the Centers for Medicare & Med-icaid Services recommendation, as thefilling of one’s diabetes prescription oftenenough to cover $80% of the time onewas recommended to be taking the med-ication (34). By this metric, proportion ofdays covered (PDC) $80%, only 29% ofpatients were adherent to GLP-1 RA

Figure 2—Reductions in tightly controlled clinical trials are not being translated into real-worldHbA1c outcomes. A: Conceptually, there is an efficacy gap between clinical trial results and real-world outcomes. Patients with diabetes in the real world are experiencing less meaningful and lesssustained improvements resulting in an efficacy gap. B: A retrospective study identified 11 pivotalRCTs with patients who initiated GLP-1 RAs (seven studies, n = 2,600) or DPP-4 inhibitors (fourstudies, n = 1,889) that included measurements of HbA1c at both drug initiation and after 1 year oftreatment. Data from the 2007–2014 Optum Humedica database served as a resource for the real-world data, and a cohort of patients with characteristics similar to the pivotal clinical trials wasidentified (21). DPP-4i, DPP-4 inhibitor.

care.diabetesjournals.org Edelman and Polonsky 1427

treatment and 37% to DPP-4 inhibitortreatment.

A CLOSER LOOK AT THE PROBLEM

Thesedataareconsistentwithprevious real-world studies, which have demonstratedthat poor medication adherence to bothoral and injectable antidiabetes agents isvery common (35–37). For example, a ret-rospective analysis of the MarketScanCommercial and Medicare Supplementaldatabases targeting adults initiating oralagents in the DPP-4 inhibitor (n = 61,399),sulfonylurea (n = 134,961), and thiazolidi-nedione (n = 42,012) classes found thatadherence rates, as measured by PDC$80% at the 1-year mark after the initialprescription,were below50% for all threeclasses, at 47.3%, 41.2%, and 36.7%, re-spectively (36). Rates dropped even lowerat the 2-year follow-up (36) (Fig. 4A).

Another claimsdatabase study of 1,321patients with type 2 diabetes who weretreated with once-daily liraglutide under-scores the particular challenge posed bythe injectable GLP-1 RA class, as only 34%(n = 454) were classified as adherent asmeasured by PDC $80% (37). Further-more, as one might expect, patientsin the adherent versus poorly adherentgroup were more likely to achieve HbA1cgoals of,7.0% (,53mmol/mol) (50% vs.39%, P , 0.001) and to have at least a1.0% reduction in their HbA1c (38% vs.32%, P = 0.022).

Data exploring persistence, defined asthe continuation of treatment for the pre-scribed duration but allowing for a strictlydefined permissible gap (e.g., days ormonths) after the last expected refilldate, show even more alarming trends. A2013 utilization study conducted in theTruven Health Analytics MarketScan com-mercial health insurance database re-vealed that only 18% of DPP-4 inhibitorusers (n = 208,683) and 11% of GLP-1 RAusers (n = 124,925) were persistent withtherapy (nonpersistence in this study wasdefined as a gap of over 30 days after thelast expected refill date) during the firstyear of treatment (38) (Fig. 4B).

The Adherence and PersistenceChallenge May Be UnderestimatedOur current ability to assess adherenceand persistence is based primarily on re-view of pharmacy records, whichmay un-derestimate the extent of the problem.For example, using the definition of ad-herence of the Centers for Medicare &

Figure 3—75% of the efficacy gap is due to poor adherence (21). Amultivariate regressionmodelwas generated to estimate the change in HbA1c 1 year after initiating GLP-1 RAs (A) or DPP-4inhibitors (B). Adherence to the index drugwasmeasured using a single variable based on PDCwitha nonoverlapping supply of the index drug (GLP-1 RA or DPP-4) during the follow-up period. Patientswere classified as adherent if PDCwas$80%.Dosing of the indexdrug during the follow-up periodwas also captured based on the last fill of the index drug in the follow-up period. The estimatedcoefficients were used to calculate the predicted change in HbA1c levels in trial and real-worldsettings, estimate the gap, and describe factors contributing to the gap in outcomes. Thepredicted change in HbA1c in trial settings was calculated bymultiplying each estimated coefficientby the value of that covariate in trials and summing the products. The predicted change in HbA1clevels in real-world settings was estimated in the same way and was found to be mathematicallyidentical to the actual real-world change in HbA1c levels. The predictionmodel controlled for variousparameters, including baseline characteristics (such as age, diabetes complications, and prior drugtherapy), addition of diabetes medications, and differences in adherence. The contribution of thefactors to the gap between real-world (GLP-1 RAs, 221 patients; DPP-4 inhibitors, 652 patients)outcomes and clinical trial (predicted) results was calculated; it was found that for both GLP-1 RAsand DPP-4 inhibitors, 75% of the gap was due to poor adherence. Adherence rate in the real worldwas 29% for GLP-1 RAs and 37% for DPP-4 inhibitors. Adherencewas defined as PDC by drug$80%.It should be noted that only three of the comparator RCTs were 52 weeks in length (27,29,32);among the remaining eight studies, two were 24 weeks (25,29), fivewere 26 weeks (22–25,28), andone was 30 weeks in length (31).

1428 Elusive Glycemic Control in the Real World Diabetes Care Volume 40, November 2017

Medicaid ServicesdPDC $80%da pa-tient could miss up to 20% of days cov-ered and still be considered adherent. In

retrospective studies of persistence, thepermissible gap after the last expectedrefill date often extends up to 90 days

(39,40). Thus, a patient may have a gapof up to 90 days and still be consideredpersistent.

Additionally, one must also considerthe issue of primary nonadherence;adherence and persistence studies typ-ically only include patients who havecompleted a first refill. A recent study ofe-prescription data among 75,589 insuredpatients found that nearly one-third ofnew e-prescriptions for diabetes medica-tions were never filled (41). Finally, noneof these measures take into account ifthe patient is actually ingesting or inject-ing themedication after acquiring his or herrefills.

The Ultimate Price of Poor Adherenceand PersistenceAcknowledging and addressing the prob-lem of poor medication adherence is piv-otal because of thewell-documented direconsequences: a greater likelihood oflong-term complications, more frequenthospitalizations, higher health care costs,and elevated mortality rates (42–45). Inpatients younger than 65, hospitalizationrisk in one study (n = 137,277) was foundto be 30% at the lowest level of adher-ence to antidiabetes medications (1–19%)versus 13% at the highest adherencequintile (80–100%); hospitalization riskwas defined as the probability of one ormore all-cause hospitalizations during a12-month period (43). In patients over65, a separate study (n = 123,235) foundthat all-cause hospitalization risk was37.4% in adherent cohorts (PDC $80%)versus 56.2% in poorly adherent cohorts(PDC ,20%) (45). This latter study alsofound better adherence to be associatedwith significantly lower all-cause outpa-tient costs, acute care costs, and totalmedical costswhen comparedwith pooreradherence thresholds (all P, 0.001) over3 years. Better adherence was also linkedto fewer emergency room visits, shorterhospital length of stay, and lower risk ofacute complications (all P , 0.001) (45).Furthermore, for every 1,000 patientswho increased adherence to their antidia-betes medications by just 1%, the totalmedical cost savings was estimated tobe $65,464 over 3 years (45).

The cumulative impact of poor adher-ence is also supported by an earlier retro-spective cohort study of 11,532 patientswith diabetes in the Kaiser Permanente ofColorado registry. There was a 39% in-creased risk for all-causemortalityassociated

Figure 4—Patientswith type2diabetes demonstrate adherence rates less than50%andpersistence ratesless than 25%over 1 year.A: A real-world study of 238,372 patientswith type 2 diabetes taking oral diabetesmedications found that adherence rates, definedas aPDC$80%, were consistently less than 50% acrossthree oral drug classes at the 1-year time point and dropped to approximately 40% at the 2-yearfollow-up (35). Using administrative claims from a U.S. health plan affiliated with Optum, a retrospectivestudy of 1,321 patients with type 2 diabetes treated with liraglutide once daily found that adherence rateswere34%for injectableGLP-1RAs(36).B: A retrospective cohort study, conductedusingdata extractedfromTruvenHealthAnalyticsMarketScan commercial health insurancedatabase, found that less than 25%of patients on DPP-4 inhibitors were persistent with therapy at 12 months; the rate for GLP-1 RAs evenworse,with only approximately 15%of themore than134,000patientswho initiatedGLP-1RAs consideredpersistent with therapy (38). Themedian time todiscontinuationwas90days for theGLP-1RAcohort and120days for theDPP-4 inhibitor cohort. Apatientwasdefinedaspersistentwith therapy if heor shehad lessthan a 30-day gap in therapy. DPP-4i, DPP-4 inhibitor; SU, sulfonylurea; TZD, thiazolidinedione.

care.diabetesjournals.org Edelman and Polonsky 1429

with poor adherence to oral hypoglycemicagents (42). Additionally, an overlookedresult of the recently completed Liraglu-tide Effect and Action in Diabetes: Evalu-ation of Cardiovascular Outcome Results(LEADER) trial provides further insight re-garding the influence of adherence.Whilethe primary analysis showed that liraglu-tide lowered cardiovascular deaths by22% compared with placebo in the globalfindings, this reduction in risk was notobserved in the North American sample;according to the study sponsor, “for rea-sons that are still unclear, the N.A. [NorthAmerican] patient groups tend to havelower compliance and adherence com-pared to global rates during large cardio-vascular studies” (46,47).

Poor Adherence Is Difficult to AddressIn total, these new data oblige those of uswho treat people with diabetes to beeven more mindful of the potential forpoor adherence. There are many poten-tial contributors to poor medication ad-herence, including depressive affect,negative treatment perceptions, lack ofpatient-physician trust, complexity ofthe medication regimen, tolerability, andcost (48). Although these factors have notbeenwell studied, it is believed thatmanyof them can be addressed with appropri-ate education, thorough instruction inmedication administration, ongoing andrespectful patient-physician interactions,and shared decision-making betweenpatients and health care professionals(49–51). However, these results are oftennot easy to achieve, especially given thelimited time available for a meaningfuldiscussion during typically short clinic vis-its. Medication adherence is not a singlebehavior but rather a dynamic constella-tion of behaviors influenced by social, en-vironmental, and individual circumstancesthat defy “one-size-fits-all” solutions (52).A recent review of interventions address-ing problematic medication adherence intype 2 diabetes found that few strategieshave been shown consistently to have amarked positive impact, particularly withrespect to HbA1c lowering, and no singleintervention was identified that could beapplied successfully to all patients withtype 2 diabetes (53). Additional evidenceindicates that improvements resultingfrom the few effective interventions,such as pharmacy-based counseling ornurse-managed home telemonitoring,oftenwaneonce theprogramsend (54,55).

We suspect that the efficacy of behav-ioral interventions to address medicationadherence will continue to be limited untilthere are more focused efforts to addressthree common and often unappreciatedpatient obstacles. First, taking diabetesmedications is a burdensome and oftendifficult activity for many of our patients.Rather than just encouraging patients todo a better job of tolerating this burden,more work is needed to make the processeasier and more convenient. For example,once-weekly medications, combinationpills, and new innovations in deliverymethods may be beneficial.

Second, poor medication adherenceoften represents underlying attitudinalproblems that may not be a strictly be-havioral issue. Specifically, negative be-liefs about prescribed medications arepervasive among patients, and behavioralinterventions cannot be effective unlessthese beliefs are addressed directly (35).Therefore, improved communication be-tween patients and clinicians regardingthe benefits and risks of specific treat-ment options, as well as wider adoptionof shared decision-making strategies, iswarranted (35).

Third, the issue of access tomedicationsremains a primary concern. A study byKurlander et al. (51) found that patientsselectively forgo medications becauseof cost; however, noncost factors, suchas beliefs, satisfaction with medication-related information, and depression, arealso influential.

SUMMARY

Only about half of patients with type 2diabetes are meeting glycemic goals(8–10), and there has been negligiblechange in the percentage of individualsachieving their target goals over the lastdecade. Despite the approval of 40 newtreatment options for type 2 diabetessince 2005 (12), these therapies and ap-proaches have not had a meaningful im-pact on the degree of glycemic controlin a large subset of the population withdiabetes (9). Although these treatmentoptions have shown notable efficacy inRCTs, their impact on glycemic controlin real-world clinical practice has beenminimal.

Poormedication adherence and persis-tence continue at alarming rates. Theyare demonstrably key contributors tothe disconnect between RCT and real-world results in lowering HbA1c and are

implicated in higher morbidity, mortality,and health care costs (43,44). Achievingmeaningful and sustained glycemic con-trol requires innovative approaches forthe real world. In order to help our pa-tients achieve meaningful and sustainedHbA1c reductionsdand move the dialpositively on the next NHANES and HEDISHbA1c analysesdwe appeal to the diabe-tes community to drive even harder forinnovative approaches designed with thereal world in mind.

Acknowledgments. The authors thank Carol A.Verderese, The Diabetes Education Group,Lakeville, CT, and Christopher G. Parkin, CGParkinCommunications, Inc., Boulder City, NV, for theireditorial support.Funding. Intarcia Therapeutics, Inc. providedfunding for editorial support in thewritingof thismanuscript.Duality of Interest. S.V.E. is a consultant for In-tarcia, Sanofi, AstraZeneca, Eli Lilly, Dexcom,Bayer, Abbott, and Johnson & Johnson. W.H.P.is a consultant for Intarcia, Sanofi, Eli Lilly, NovoNordisk,Dexcom,Roche,Abbott, andAstraZeneca.No other potential conflicts of interest relevant tothis article were reported.

References1. UK Prospective Diabetes Study (UKPDS)Group. Intensive blood-glucose control withsulphonylureas or insulin compared with conven-tional treatment and risk of complications inpatients with type 2 diabetes (UKPDS 33). Lancet1998;352:837–8532. The Diabetes Control and Complications TrialResearchGroup. The effect of intensive treatmentof diabetes on the development and progressionof long-term complications in insulin-dependentdiabetes mellitus. N Engl J Med 1993;329:977–9863. Epidemiology of Diabetes Interventions andComplications (EDIC) Research Group. Epidemiol-ogy of Diabetes Interventions and Complications(EDIC). Design and implementation of a long-termfollow-up of the Diabetes Control and Complica-tions Trial cohort. Diabetes Care 1999;22:99–1114. Selvin E, Steffes MW, Zhu H, et al. Glycatedhemoglobin, diabetes, and cardiovascular risk innondiabetic adults. N Engl J Med 2010;362:800–8115. Selvin E, Ning Y, Steffes MW, et al. Glycatedhemoglobin and the risk of kidney disease andretinopathy in adults with and without diabetes.Diabetes 2011;60:298–3056. Inzucchi SE, Bergenstal RM,Buse JB, et al.Man-agement of hyperglycemia in type 2 diabetes,2015: a patient-centered approach. Update to aposition statement of the American Diabetes As-sociation and the European Association for theStudy of Diabetes. Diabetes Care 2015;38:140–1497. Centers for Disease Control and Prevention.National diabetes statistics report: estimates ofdiabetes and its burden in the United States,2014 [Internet], 2014. Atlanta, GA, U.S. Departmentof Health and Human Services. Available from

1430 Elusive Glycemic Control in the Real World Diabetes Care Volume 40, November 2017

http://www.cdc.gov/diabetes/pubs/statsreport14/national-diabetes-report-web.pdf. Accessed 10January 20168. Ali MK, Bullard KM, Saaddine JB, Cowie CC,Imperatore G, Gregg EW. Achievement of goalsin U.S. diabetes care, 1999–2010. N Engl J Med2013;368:1613–16249. Carls GS, Huynh J, Tuttle E, Edelman SV.Achievement of glycated hemoglobin goals intheU.S. remains unchanged through 2014. Poster(1515-P) presented at the 76th Scientific Sessionsof the American Diabetes Association, 10–14 June2016, New Orleans, LA10. National Committee for Quality Assurance.The State of Health Care Quality Report 2015 [In-ternet], 2015. Washington, DC, National Committeefor Quality Assurance. Available for downloadfrom http://store.ncqa.org/index.php/catalog/product/view/id/2341/s/2015-state-of-health-care-quality-report/. Accessed 5 July 201611. Centers for Disease Control and Prevention.PowerPoint slides on diabetes: Long-term trendsin diabetes [Internet], 2017. Available fromhttps://www.cdc.gov/diabetes/data/center/slides.html. Accessed 8 September 201712. U.S. FoodandDrugAdministration.Drugs@FDA:FDA approved drug products [Internet]. Availablefrom https://www.accessdata.fda.gov/scripts/cder/drugsatfda/. Accessed 12 July 201613. Centers for Disease Control and Prevention.Chronic disease prevention and health promo-tion: at a glance fact sheets: diabetes [Internet],2016. Available from https://www.cdc.gov/chronicdisease/resources/publications/aag/diabetes.htm. Accessed 22 November 201614. van Onzenoort HA, Menger FE, Neef C, et al.Participation in a clinical trial enhances adherenceand persistence to treatment: a retrospective co-hort study. Hypertension 2011;58:573–57815. Rothwell PM. External validity of randomisedcontrolled trials: “to whom do the results of thistrial apply?” Lancet 2005;365:82–9316. National Heart, Lung, and Blood Institute.What are the possible benefits and risks of clinicaltrials? [Internet], 2017. Available from https://www.nhlbi.nih.gov/studies/clinicaltrials/benefitsrisks.Accessed 8 September 201717. Harvard Women’s Health Watch. What clini-cal studies can do for you [Internet], July 2012.Available from http://www.health.harvard.edu/staying-healthy/what-clinical-studies-can-do-for-you. Accessed 23 July 201618. Burgess LJ, Sulzer NU, Hoosain F, Leverton N,Bliganut S, Emanuel S. Patients’ motivations forparticipating in cardiovascular clinical trials: alocal perspective. Cardiovasc J Afr 2009;20:220–22319. van Dulmen S, Sluijs E, van Dijk L, de Ridder D,Heerdink R, Bensing J. Patient adherence to med-ical treatment: a review of reviews. BMC HealthServ Res 2007;7:55–6720. Jin J, Sklar GE, Min Sen Oh V, Chuen Li S.Factors affecting therapeutic compliance: a re-view from the patient’s perspective. Ther ClinRisk Manag 2008;4:269–28621. Carls GS, Tuttle E, Tan R-D, et al. Understand-ing the gap between efficacy in randomized con-trolled trials and effectiveness in real-world use ofGLP-1 RA and DPP-4 therapies in patients withtype 2 diabetes. Diabetes Care 2017;40:1469–1478

22. Pratley RE, Nauck M, Bailey T, et al.; 1860-LIRA-DPP-4 Study Group. Liraglutide versus sita-gliptin for patients with type 2 diabetes who didnot have adequate glycaemic control with met-formin: a 26-week, randomised, parallel-group,open-label trial. Lancet 2010;375:1447–145623. Marre M, Shaw J, Brandle M, et al.; LEAD-1SU study group. Liraglutide, a once-daily humanGLP-1 analogue, added to a sulphonylurea over26 weeks produces greater improvements in gly-caemic and weight control comparedwith addingrosiglitazone or placebo in subjects with type 2diabetes (LEAD-1 SU). Diabet Med 2009;26:268–27824. Nauck M, Frid A, Hermansen K, et al.; LEAD-2Study Group. Efficacy and safety comparison ofliraglutide, glimepiride, and placebo, all in combi-nation with metformin, in type 2 diabetes: theLEAD (Liraglutide Effect and Action in Diabetes)-2study. Diabetes Care 2009;32:84–9025. Diamant M, Van Gaal L, Stranks S, et al. Onceweekly exenatide compared with insulin glarginetitrated to target in patients with type 2 diabetes(DURATION-3): an open-label randomised trial.Lancet 2010;375:2234–224326. Blevins T, Pullman J, Malloy J, et al.DURATION-5: exenatide once weekly resulted ingreater improvements in glycemic control com-pared with exenatide twice daily in patientswith type 2 diabetes. J Clin Endocrinol Metab2011;96:1301–131027. Garber A, Henry R, Ratner R, et al.; LEAD-3(Mono) Study Group. Liraglutide versus glimepir-ide monotherapy for type 2 diabetes (LEAD-3Mono): a randomised, 52-week, phase III, double-blind, parallel-treatment trial. Lancet 2009;373:473–48128. Zinman B, Gerich J, Buse JB, et al.; LEAD-4Study Investigators. Efficacy and safety of the hu-man glucagon-like peptide-1 analog liraglutide incombination with metformin and thiazolidine-dione in patients with type 2 diabetes (LEAD-4Met+TZD). Diabetes Care 2009;32:1224–123029. Nauck MA, Meininger G, Sheng D, TerranellaL, Stein PP; Sitagliptin Study 024 Group. Efficacyand safety of the dipeptidyl peptidase-4 inhibitor,sitagliptin, compared with the sulfonylurea, glipi-zide, in patientswith type 2 diabetes inadequatelycontrolled on metformin alone: a randomized,double-blind, non-inferiority trial. Diabetes ObesMetab 2007;9:194–20530. DeFronzo RA, Hissa MN, Garber AJ, et al.;Saxagliptin 014 Study Group. The efficacy andsafety of saxagliptin when added to metformintherapy in patients with inadequately controlledtype 2 diabetes with metformin alone. DiabetesCare 2009;32:1649–165531. Arechavaleta R, Seck T, Chen Y, et al. Efficacyand safety of treatment with sitagliptin or glime-piride in patients with type 2 diabetes inade-quately controlled on metformin monotherapy:a randomized, double-blind, non-inferiority trial.Diabetes Obes Metab 2011;13:160–16832. Goke B, Gallwitz B, Eriksson J, Hellqvist A,Gause-Nilsson I; D1680C00001 Investigators. Sax-agliptin is non-inferior to glipizide in patients withtype 2 diabetes mellitus inadequately controlledon metformin alone: a 52-week randomised con-trolled trial. Int J Clin Pract 2010;64:1619–163133. Aroda VR, Henry RR, Han J, et al. Efficacy ofGLP-1 receptor agonists and DPP-4 inhibitors:

meta-analysis and systematic review. Clin Ther2012;34:1247–1258.e2234. Centers for Medicare & Medicaid Services.Medicare 2016 Part C & D star rating technicalnotes (draft) [Internet], 2015. Available fromhttps://www.cms.gov/Medicare/Prescription-Drug-Coverage/PrescriptionDrugCovGenIn/Downloads/2016-Technical-Notes-Preview-1-v2015_08_05.pdf. Accessed 22 November 201635. Aikens JE, Piette JD. Diabetic patients’ medi-cation underuse, illness outcomes, and beliefsabout antihyperglycemic and antihypertensivetreatments. Diabetes Care 2009;32:19–2436. Farr AM, Sheehan JJ, Curkendall SM, SmithDM, Johnston SS, Kalsekar I. Retrospective analysisof long-term adherence to and persistence withDPP-4 inhibitors in US adults with type 2 diabetesmellitus. Adv Ther 2014;31:1287–130537. Buysman EK, Liu F, Hammer M, Langer J. Im-pact of medication adherence and persistenceon clinical and economic outcomes in patientswith type 2 diabetes treated with liraglutide: aretrospective cohort study. Adv Ther 2015;32:341–35538. Koro CE, Chhabra P, Stender M, Spain CV,Allen JK, KrzywyHJ. Treatment utilization patternsof GLP-1 agonists and DPP-4 inhibitors amongtype 2 diabetics in a U.S. commercially insuredpopulation: 2005–2011. Poster (1589-P) pre-sented at the 73rd Scientific Sessions of the Amer-ican Diabetes Association, 21–25 June 2013,Chicago, IL39. Cramer JA, Roy A, Burrell A, et al. Medicationcompliance and persistence: terminology anddefinitions. Value Health 2008;11:44–4740. Cai J, Wang Y, Baser O, et al. Comparativepersistence with antihyperglycemic agents usedto treat type 2 diabetesmellitus in the real world.Poster (1155-P) presented at the 76th ScientificSessions of the American Diabetes Association,10–14 June 2016, New Orleans, LA41. Fischer MA, Stedman MR, Lii J, et al. Primarymedication non-adherence: analysis of 195,930electronic prescriptions. J Gen Intern Med 2010;25:284–29042. Ho PM, Rumsfeld JS, Masoudi FA, et al.Effect of medication nonadherence on hospitali-zation and mortality among patients with di-abetes mellitus. Arch Intern Med 2006;166:1836–184143. Sokol MC, McGuigan KA, Verbrugge RR,Epstein RS. Impact of medication adherence onhospitalization risk and healthcare cost.Med Care2005;43:521–53044. Egede LE, Gebregziabher M, Echols C, LynchCP. Longitudinal effects of medication nonadher-ence on glycemic control. Ann Pharmacother2014;48:562–57045. Boye KS, Curtis SE, Lage MJ, Garcia-Perez LE.Associations between adherence and outcomesamong older, type 2 diabetes patients: evidencefrom a Medicare Supplemental database. PatientPrefer Adherence 2016;10:1573–158146. Marso SP, Daniels GH, Brown-Frandsen K,et al.; LEADER Steering Committee; LEADER TrialInvestigators. Liraglutide and cardiovascular out-comes in type 2 diabetes. N Engl J Med 2016;375:311–32247. Reeves A. Novo Nordisk drug cuts diabeticdeaths, but not enough for Wall Street [articleonline]. Investor’s Business Daily, 14 June 2016.

care.diabetesjournals.org Edelman and Polonsky 1431

Available from http://www.investors.com/news/technology/novo-nordisk-drug-cuts-diabetic-deaths-but-not-enough-for-street/. Accessed18 July 201648. Garcıa-Perez LE, AlvarezM, Dilla T, Gil-GuillenV, Orozco-Beltran D. Adherence to therapies inpatients with type 2 diabetes. Diabetes Ther2013;4:175–19449. Beck RS, Daughtridge R, Sloane PD. Physician-patient communication in the primary care office:a systematic review. J Am Board Fam Pract 2002;15:25–38

50. Zolnierek KB, Dimatteo MR. Physician com-munication and patient adherence to treatment:a meta-analysis. Med Care 2009;47:826–83451. Kurlander JE, Kerr EA, Krein S, Heisler M,Piette JD. Cost-related nonadherence to medica-tions among patients with diabetes and chronicpain: factors beyond finances. Diabetes Care2009;32:2143–214852. Steiner JF. Rethinking adherence. Ann InternMed 2012;157:580–58553. Sapkota S, Brien JA, Greenfield J, Aslani P. Asystematic review of interventions addressing

adherence to anti-diabetic medications in pa-tients with type 2 diabetesdimpact on adher-ence. PLoS One 2015;10:e011829654. Brennan TA, Dollear TJ, Hu M, et al. An in-tegrated pharmacy-based program improvedmedication prescription and adherence rates indiabetes patients. Health Aff (Millwood) 2012;31:120–12955. Wakefield BJ, Holman JE, Ray A, et al. Effec-tiveness of home telehealth in comorbid diabe-tes and hypertension: a randomized, controlledtrial. Telemed J E Health 2011;17:254–261

1432 Elusive Glycemic Control in the Real World Diabetes Care Volume 40, November 2017