Factors Driving Diabetes Care Improvement in a Large Medical Group: Ten Years of Progress

P

P

P

P

The purpose of this study was to documenttrends in diabetes quality of care and coincidingstrategies for quality improvement over 10 years ina large medical group. Adults with diagnosed diabetesmellitus were identified each year from 1994(N = 5610) to 2003 (N = 7650), and internal medicalgroup data quantified improvement trends.Multivariate analysis was used to identify factorsthat did and did not contribute to improvementtrends. Median glycosylated hemoglobin A_1C (A_1C)levels improved from 8.3% in 1994 to 6.9% in2003 (<.001). Mean low-density lipoprotein(LDL) cholesterol measurements improved from132 mg/dL in 1995 to 97 mg/dL in 2003 (<.001).Both A_1C (<.01) and LDL improvement (<.0001)were driven by drug intensification, leadershipcommitment to diabetes improvement, greater continuityof primary care, participation in local andnational diabetes care improvement initiatives, andallocation of multidisciplinary resources at the cliniclevel to improve diabetes care. Resources werespent on nurse and dietitian educators, active outreachto high-risk patients facilitated by registries,physician opinion leader activities including clinicbasedtraining programs, and financial incentives toprimary care clinics. Use of endocrinology referralswas stable throughout the period at about 10% ofpatients per year, and there were no disease managementcontracts to outside vendors over the studyperiod. Electronic medical records did not favorablyaffect glycemic control or lipid control in this setting.This primary care-based system achieved A_1Cand LDL reductions sufficient to reduce macrovascularand microvascular risk by about 50% accordingto landmark studies; further risk reductionshould be attainable through better blood pressurecontrol. Strategies for diabetes improvement needto be customized to address documented gaps inquality of care, provider prescribing behaviors, andpatient characteristics.

(Am J Manag Care. 2005;11:S177-

S185)

The benefits of good glycemic and lipidcontrol in type 2 diabetes mellitus arewell known.^1-4 Although wide gapsbetween recommended and actual levels ofcare have been reported, recent nationalreports document little improvement inglycemic control in US adults with diabetes.^5,6 For example, Saydah et al usednational data to show little glycosylatedhemoglobin A_1C (A_1C) improvement andmoderate low-density lipoprotein (LDL)cholesterol improvement among adults withdiabetes from 1994 to 2000.⁷ One of the fewlong-term analyses of glycemic control intype 2 diabetes recently found that controlrates dropped from 44.5% in 1994 to 35.8%in 2000, a period of time during which moreeffective treatment regimens became widelyavailable.⁸

Despite generally discouraging nationaldata, there have been some reports of significantimprovement in diabetes care.^9-12However, most of these reports cover shorttime periods, and few reports systematicallyexplore the wide range of factors that mayaffect diabetes care improvement.

In this article, we present 10-year trendsin glycemic control (A_1C levels) and lipidcontrol (LDL levels) in a well-defined populationof adults with type 2 diabetes andaddress the following questions: (1) Whatlevels of improvement in diabetes care canbe achieved and sustained in primary carepractice? (2) To what degree is intensificationof pharmacotherapy driving improvementin diabetes care quality? and (3)What strategies may be used by medicalgroups or health plans to accelerate diabetescare improvement? The answers tothese questions may guide medical groupsand health plans currently contemplatingefforts to improve diabetes care. As at thelevel of the individual patient with diabetes,success and sustainability in A_1C and lipidcontrol at the institutional level mayrequire a renewed commitment of resourcesto meet changing needs as well as frequentevaluation of results to ensure progresstoward goals.

Methods

We evaluated diabetes care deliveredfrom January 1, 1994, to December 31,2003, at HealthPartners Medical Group(HPMG), a Minnesota multispecialty medicalgroup that at the time of the study providedcomprehensive healthcare services to adefined population of 175 000 adult membersat 17 clinics. Internists and familyphysicians delivered most of the adult diabetescare but easily referred patients tosubspecialists as needed. From 9% to 12% ofadult patients with diabetes saw an endocrinologisteach year, most for a single visit todevelop a treatment plan that was thenjointly implemented with their primary carephysician. Diabetes education nurses ineach clinic worked closely with primary carephysicians to provide patient education andself-management training. A drug formularyfacilitated use of sulfonylureas, metformin,insulin, fibrates, and HMG-CoA [hydroxymethylglutaryl coenzyme A] reductaseinhibitors (statins). The use of alpha-glucosidaseinhibitors, meglitinides, and thiazolidinedionesrequired special authorizationor prior use of alternative agents.

HPMG Diabetes Programs

International Statistical

Classification of Diseases, 9th Revision

ICD-9

In 1995, HPMG leaders identified diabetesas 1 of 8 priority areas for clinical qualityimprovement. This led to a multifaceted diabetesimprovement strategy that began withthe development of a registry for patientswith diabetes and evolved throughout theperiod of interest to include a variety of initiatives.In each calendar year from 1994 to2003, HPMG members aged 19 years andolder were classified as having diagnoseddiabetes if they met 1 or both of the followingcriteria: (a) 2 () 250.xx diagnostic codes at outpatientor inpatient visits that calendar year,or (b) a filled prescription for a diabetes-specificmedication in that calendar year. Thisvalidated method of diabetes identificationhad an estimated sensitivity of 0.91 and anestimated specificity of 0.99; the positivepredictive value was 0.94 in 1994¹³ and 0.95when revalidated in 2001. Those plan membersidentified as having diabetes were listedin diabetes registries first provided to physician/nurse teams in 1997; these registrieswere progressively expanded to include A_1Cand LDL data and identification of comorbidcoronary heart disease. Nurses used the registriesto guide "active outreach" to high-riskpatients not in metabolic control or missingrecommended tests.

simultaneously

Beginning in 1997, the medical groupleadership maintained a commitment toimplement the locally accepted diabetesguideline, called Institute for ClinicalSystems Improvement (ICSI) DiabetesGuideline (www.icsi.org), and participatedin a medical group shared learning initiativeto improve diabetes care called the ICSIDiabetes Action Group. HPMG also achievedrecognition in the National Committee forQuality Assurance (NCQA)/American DiabetesAssociation (ADA) Diabetes PhysicianRecognition Program (DPRP) in 1999, andmaintains that standing to date. All basicmeasures of diabetes care were continuouslytracked over the 10-year period, and A_1Cvalues were provided as a feedback andimprovement tool to clinics and providersstarting in 1997. LDL values were added tothe feedback around 1999. Starting in 2001,feedback to providers was changed so thatthe percentage of patients with diabeteswho were at A_1C and LDLgoal was reported, in an effort to focus diabetescare on both lipid and A_1C control.Although a diabetes registry was alwaysavailable to the clinics, the format evolvedas resources committed to this projectwaxed and waned. In the final year of thestudy period, financial incentives weremade available to clinics for good performanceon diabetes measures.

Tracking Trends in Diabetes Care

The proportion of patients with diabeteshaving 1 or more A_1C or LDL tests in eachcalendar year and the mean and median testvalues were measured for 1994 through2003. When more than 1 A_1C or LDL test wasdone within a calendar year, the valueobtained latest in the year was selected foranalysis. A single accredited clinical chemistrylaboratory did all tests. A_1C was measuredby a liquid chromatographic assay,with a normal range of 4.5% to 6.1% and acoefficient of variation of 0.58% at an A_1Cvalue of 8.8%.¹⁴ LDL was calculated usingstandard equations only when blood sampleswere drawn after a minimum 12-hour fastand when triglycerides were less than 400mg/dL. Patient age and gender wereobtained from medical group administrativedata. Age was reported as age in years as ofJanuary 1 of each year.

All study subjects had basic health insurancebenefits through HPMG. More than91% of study subjects younger than 65years of age and 80% to 82% of those 65years and older had pharmacy coverage ineach year. For these patients, filled prescriptionswere enumerated in each calendaryear for the following drug classes:insulins, sulfonylureas, biguanides, thiazolidinediones,alpha-glucosidase inhibitors,meglitinides, and statins. Less than 5% of allHPMG members reported using any allopathicmedical care services from sourcesoutside HealthPartners.¹⁵

Results

P

P

Demographic and clinical informationfor adults classified as having diabetes ineach calendar year from 1994 to 2003 aregiven in Table 1. The number of patientswith diabetes in each yearly cross-sectionalsample rose from 5610 in 1994 to 7650 in2003, despite the absence of sustainedmembership growth. The increased numberof patients in the cross-sections overtime is consistent with changes in diabetesdiagnostic criteria and rising prevalence ofdiabetes.^16,17 Median A_1C fell from 8.3% in1994 to 6.9% in 2003 (<.001) and meanLDL fell from 132 mg/dL in 1995 to 97mg/dL in 2003 (<.001) across the cross-sectionalsamples (Figures 1 and 2).

Previous analysis showed that A_1C in thispatient population was not a predictor ofsubsequent death or disenrollment (althoughbetter glycemic control was correlatedwith lower medical care costs¹⁸).Analysis of the 1994 patients with diabetesas a cohort through 1999 showed similarimprovement trends in both A_1C and LDL(data not shown), indicating that death orselective disenrollment did not account forthe noted improvements.

P

P

P

P

P

P

In unadjusted bivariate comparisons, A_1Cand LDL improved significantly from 1994to 1999 in nearly all defined patient subgroups.Multivariate models with change inA_1C from 1994 to 1999 as the dependentvariable showed that older age (<.0001),higher baseline A_1C (<.0001), addition ofsulfonylurea (= .006), and addition ofmetformin (= .01) were related to greaterA_1C improvement. Baseline sulfonylurea use(= .002) and evidence of depression duringthe study (<.03) were related to lessA_1C improvement. Also, younger adults (aged18-44 years), patients on insulin treatment,and patients with no pharmacy coverage(about 18.3% of patients) had changes inA that were less desirable than thoseachieved by other patients. Patient gender,patient comorbidity, primary care physicianvariables (age, gender, and specialty), diabeteseducator visits, and a limited set ofinteraction terms were not significantlyrelated to change in A_1C.

Patients who were older, had very highbaseline A_1C scores, or had major comorbidityhad the greatest improvement inLDL levels. As expected, those with higherbaseline A_1C were more likely to receiveintensified treatment, and A_1C improvementwas greatest in those with initiationof insulin or combination therapy. Thecombination of insulin and metformin wasparticularly potent, for all patients as wellas for patients with baseline A_1C greaterthan 9%, but other combinations (insulinand sulfonylurea, metformin and sulfonylurea)were also effective. Combinationtherapy using metformin and/or insulinincreased substantially during the studyperiod.

Thiazolidinediones, alpha-glucosidaseinhibitors, and meglitinides were used inthe aggregate by less than 2% of all patientsthrough 1999. Thiazolidinedione useincreased from about 2% in 1999 to about9.5% of patients with diabetes by 2003.

P

P

P

P

P

Statin use accounted for most of theobserved improvement in LDL; statin usewas less than 20% in the mid-1990s, butreached 36% in 2001, and further increasedto 50% in 2003. From 1995 to 1999, thosepatients taking statins had LDL decreasesaveraging about 30 mg/dL, whereas thosenot taking statins saw LDL drops of onlyabout 7 mg/dL. In multivariate models,LDL improvement from 1995 to 1999 wassignificantly related to higher baseline LDL(<.0001), older age (= .03), highercomorbidity score (= .05), taking a statinat baseline (<.0001), or initiating statinsduring the study (<.0001).

Table 2 provides an overview of diabetescare measurement trends coupled with careimprovement strategies implemented from1994 to 2003. The table summarizes learningsfrom 6 federally funded research grantsthat collected data within the HPMG patientpopulation during these years, as well asqualitative data from interviews with HPMGand HealthPartners leaders, providers, clinicstaff, and quality improvement staff. The relativecontributions of specific improvementstrategies to observed trends in diabetescare quality are discussed next along withgeneral observations about the ongoingHPMG efforts.

Additional data on attitudes of patientstoward diabetes and diabetes care providerswere obtained in surveys sent in 2001 to1900 randomly selected HPMG adults withdiabetes. Analyses of these data have beenpreviously published, and reference to thesesurvey results and trends is made in theDiscussion that follows.^19-21

Discussion

1. Primary care clinics can successfully

improve diabetes care in the absence of

carve-out disease management. Primary

care physician continuity of care is significantly

related to better diabetes care.

This medical group implemented manycommon disease management tools, such asregistries with ongoing monitoring ofpatients and active outreach to high-riskpatients and those missing necessary tests.Previsit planning and case managementoften occurred in a multidisciplinary fashionamong the diabetes nurses, dietitians, andphysicians at each clinic. These activitieswere done within the clinic and medicalgroup, rather than being contracted out to acommercial disease management vendor. Inseeking to improve access to primary carevisits, we discovered a strong relationshipbetween higher primary care continuity ofcare and quality of diabetes care. The HPMGdiabetes care strategy is to invest availableresources to develop chronic disease careinfrastructure within the medical group,rather than siphoning off resources to outsidevendors.

2. The final common pathway to A

_1C

and LDL improvement is intensification

of pharmacotherapy.

In this medicalgroup, drug therapy with combinations ofinsulin, metformin, and sulfonylureas ledto the greatest improvement in A_1C. Theimprovements noted through 1999 do notreflect benefits from widespread use ofthiazolidinediones, alpha-glucosidase inhibitors,or nonsulfonylurea secretogogues,because use rates of these classes of drugswere low prior to year 2000. Clinical inertiais also a major barrier to better diabetescare. Clinical inertia is defined as failure tointensify therapy at a visit when A_1C, LDL,or blood pressure are not at evidence-basedgoals, and is reported to occur atabout 60% of all diabetes visits.^22-24 However,our data suggest that HPMG physicianswho participated in clinic-basedStaged Diabetes Management training sessionswere more likely to intensify medicationsthan physicians who did not receivethis training. Intensification of pharmacotherapyappears to be the "final commonpathway" to successful control of both A_1Cand LDL over long periods of time.

3. Certain groups of patients have had

less improvement in A_1C and LDL than

other groups. Those with the most difficulty

included younger adults and those with

a current or former diagnosis of depression.

In addition to the willingness of physiciansand patients to increase doses or usecombination therapy, there were also anumber of additional variables that emergedas independent predictors of A_1C or LDL control.For example, LDL change was robust inolder, sicker patients—perhaps a reflection ofcurrent lipid control guidelines²⁵ that recommendtargeting this population. Alternatively,the lesser degree of A_1C improvement inyounger adults may indicate problems withaccess to care for younger patients who aremore often employed. Age-based attitudinaldifferences toward diabetes may also accountfor the observed relative weakness inglycemic control in younger patients.²⁶ Datafrom the 2001 HPMG survey of adults withdiabetes indicated that appreciation of theserious risk of diabetes (which patients withasymptomatic diabetes often fail to perceive)also independently predicts improvement inA_1C.^17,27 Tailoring of diabetes care improvementstrategies to the needs of particular subgroupsof patients may well be the engineneeded to drive future improvement in diabetescare.²⁷

4. As overall care improved, the

"recidivism

vector" became an increasingly

important drag on further improvement.

Each year, the net small incrementalimprovement in A_1C was the sum of 2 vectors.The improvement vector (roughly 35% of diabetespatients had better A_1C values than theyear before) was largely offset by the recidivismvector (roughly 30% of diabetes patientshad worse A_1C values than the year before).As overall A_1C values improve, the recidivismvector increases and acts as an increasinglyprominent brake on further population-levelimprovements in A_1C levels. Thus, in additionto focused efforts to target high-A_1C patientswith education and support, more aggressive"proactive" or "feed-forward" care also needsto be sustained across the entire populationto continue the observed trends in A_1Cimprovement.²⁷ For example, results fromdiabetes prevention trials provide clinicaljustification of aggressive management ofinsulin resistance among patients with prediabetes.This supports the proposition thatthose with near-normal A_1C values shouldreceive continued lifestyle support andongoing aggressive pharmacotherapy asneeded to preempt deterioration in A_1C.^28,29

5. The benefits of patient education

could be enhanced by more careful targeting

of this resource.

The presence ofdiabetes educators in clinics has been associatedwith improved A_1C levels.³⁰ Paymentfor such services by Medicare has recentlyimproved; however, significant barriers stillremain to self-management support. Forexample, nurse and dietitian encounterscannot be reimbursed if services are providedon the same day, even if this is more convenientfor the patient. In addition, theamount of time per year that will be reimbursedis limited and bears no relationshipto the actual needs of the patient. The valueof diabetes educators to medical groups maybe greatly enhanced if educators collaboratewith physicians to improve blood pressureand lipid control, as well as glycemic control.Additional efficiencies may be obtainedby allocating educator time in a way that isconsistent with a given patient's readinessto change.¹⁹

6. Participation in national and local

long-term care improvement activities

facilitated improvements, but competing

priorities were sometimes a barrier.

Themedical group participated in both local andnational quality improvement initiatives,including initiatives sponsored by the ICSI,the Institute for Healthcare Improvement(www.ihi.org), and the Minnesota Departmentof Health. However, the impact ofthese and other activities (such as participationin a national physician recognitionprogram—see Sidebar, "The Diabetes PhysicianRecognition Program [DPRP] atHealthPartners Medical Group [HPMG]") onquality of care may be attenuated if thereare many improvement initiatives competingfor the physician's attention and priority.A primary example in the HPMG setting wasan improvement initiative funded by a privatefoundation designed to address clinicsystems globally, without any disease-specificfocus.

This program, which had strong buy-infrom health plan and medical group leaders,led to temporary discontinuation of ongoingdisease-specific efforts to improve diabetescare, and likely accounted for the temporaryworsening of A_1C levels in the period from2001 to 2002. A second example was implementationof the electronic medical record(EMR), which many thought would solve theproblems of poor quality care through bestpractice reminders. However, the process ofEMR implementation diverted time andattention from clinical care and disruptedestablished chronic disease care routines forabout 6 to 12 months. In clinics where theEMR was first implemented, EMR-relatedimprovements in process measures, such asincreased A_1C or cholesterol testing, did nottranslate to better levels of A_1C or LDL relativeto similar clinics without EMRs.^31-33

7. Financial accountability and performance

incentives for diabetes performance

may facilitate improvement.

It wasnot until 2003 that HPMG establishedexplicit financial incentives for better diabetescare. The impact of these incentiveson diabetes care quality began to be feltalmost immediately and continued through2004. Incentives were initially directed toclinics and to practicing physicians in leadershippositions rather than to individualphysicians.

Thus, the improvements noted in diabetescare from 1994 to 2003 cannot beattributed to positive financial incentives.However, HPMG leaders plan to amplify therole of positive financial incentives to clinicsas a strategy to further improve diabetescare and other care in the future. The EMRfacilitates detailed tracking of diabetes careperformance at the patient, provider, andclinic levels. A comprehensive diabetes"optimal care measure" is used to rewardclinics in the system for the number ofpatients with diabetes who simultaneouslymeet all of the following standards:

• A_1C tested and result <7%;
• LDL tested and result <100 mg/dL;
• Systolic blood pressure <130 mm Hg;
• Aspirin use (for ages >40 years); and

Nonsmoker status.

Another financial incentive program wasoffered by HealthPartners health plan tocontracted medical groups other than HPMGfor many years. Data suggest that this financialincentive program, referred to as theOutcomes Recognition Program, did have apositive impact not only on glycemic controlin patients with diabetes, but also on lipidcontrol in patients with heart disease and inoverall rates of appropriate preventive care.These incentives were condition-specificand were provided to medical groups ratherthan to individual clinics or physicians.

Summary and Future Directions

P

The 10-year trends presented here indicatethat widespread and sustained improvementin diabetes care can be achievedwithin a primary care-oriented deliverysystem. The magnitude of improvement indiabetes care from 1994 to 2003 was sufficientto reduce cardiovascular risk by about50% in adults with diabetes. In addition tothe noted improvements in A_1C and LDL levels,the rate of major cardiovascular eventsin adults with diabetes has decreased overtime, and rates of patient-reported blindnessamong those with diabetes decreased from1995 to 2001 (= .052).

These results question conventional wisdomthat EMRs, disease management contractswith outside vendors, and widespreaduse of expensive new classes of pharmacologicagents are necessary to improve diabetescare. An alternative strategy—toinvest resources in enhanced primary caredelivery systems and to increase primarycare physician continuity of care—appearedquite effective in this report. It is clear thateven the major 10-year improvement trendobserved in this group leaves much to bedesired. In settings such as this, with medianA_1C levels already below 7%, higher priorityshould be given to better blood pressurecontrol while maintaining gains achieved inA_1C and LDL levels.

Important future directions include theneed to reduce clinical inertia and increasepatient activation.³⁴ There is strong evidencethat improved primary care physiciancontinuity of care is a driver of diabetes careimprovement, and efforts to maintain continuityof care will likely continue to be a priority.HPMG will also give more attention toexternal accountability and to physician orclinic financial incentives, which havedemonstrated potential to induce careimprovements in several clinical domains. Amajor need is to develop the potential of ourrecently implemented EMR system to providemore sophisticated decision support tophysicians (and perhaps ultimately topatients as well) in a customized and timelyfashion.

Acknowledgment

This study was supported in part bygrant HS 09946 from the US Agency forHealthcare Research and Quality.

Lancet.

1. UK Prospective Diabetes Study (UKPDS) Group.Intensive blood-glucose control with sulphonylureas orinsulin compared with conventional treatment and riskof complications in patients with type 2 diabetes(UKPDS 33). 1998;352:837-853.

Lancet.

2. UK Prospective Diabetes Study (UKPDS) Group.Effect of intensive blood-glucose control with metforminon complications in overweight patients with type 2 diabetes(UKPDS 34). 1998;352:854-865.

Diabetes

Care.

3. Pyorala K, Pedersen TR, Kjekshus J, Faergeman O,Olsson AG, Thorgeirsson G. Cholesterol lowering withsimvastatin improves prognosis of diabetic patients withcoronary heart disease. A subgroup analysis of theScandinavian Simvastatin Survival Study (4S). 1997;20:614-620.

4. Nathan DM. Late breaking clinical trials: Impact ofDCCT intervention on subsequent cardiovascular disease.Presented at: 65th Scientific Sessions of theAmerican Diabetes Association; June 10-14, 2005; SanDiego, Calif.

Diabetes.

5. Ford E, Mokdad A. Trends in glycosylated hemoglobinconcentrations among United States adults. 2003;52(suppl 1):A219.

6. Saaddine JB, et al. Diabetes quality of care in the US:improvement in the last decade 1990-2000. Presentedat: 64th Scientific Sessions of the American DiabetesAssociation; June 4-8, 2004; Orlando, Fla.

JAMA.

7. Saydah SH, Fradkin J, Cowie CC. Poor control of riskfactors for vascular disease among adults with previouslydiagnosed diabetes. 2004;291:335-342.

Diabetes Care.

8. Koro CE, Bowlin SJ, Bourgeois N, Fedder DO.Glycemic control from 1988 to 2000 among U.S. adultsdiagnosed with type 2 diabetes: a preliminary report.2004;27:17-20.

Diabetes Care.

9. Sidorov J, Shull R, Tomcavage J, Girolami S, LawtonN, Harris R. Does diabetes disease management savemoney and improve outcomes? A report of simultaneousshort-term savings and quality improvement associatedwith a health maintenance organization-sponsored diseasemanagement program among patients fulfilling healthemployer data and information set criteria. 2002;25:684-689.

Jt Comm J Qual

Improv.

10. Sperl-Hillen J, O'Connor PJ, Carlson RR, et al.Improving diabetes care in a large health care system: anenhanced primary care approach. 2000;26:615-622.

Ann Intern

Med.

11. Kerr EA, Gerzoff RB, Krein SL, et al. Diabetes carequality in the Veterans Affairs Health Care System andcommercial managed care: the TRIAD study. 2004;141:272-281.

Am J Manag Care.

12. McBean AM, Jung K, Virnig BA. Improved care andoutcomes among elderly Medicare managed care beneficiarieswith diabetes. 2005;11:213-222.

Am J Manag Care.

13. O'Connor PJ, Rush WA, Pronk NP, Cherney LM.Identifying diabetes mellitus or heart disease amonghealth maintenance organization members: sensitivity,specificity, predictive value, and cost of survey and databasemethods. 1998;4:335-342.

J Lab Clin Med.

14. Huisman TH, Henson JB, Wilson JB. A new high-performanceliquid chromatographic procedure to quantitatehemoglobin A_1C and other minor hemoglobins inblood of normal, diabetic, and alcoholic individuals.1983;102:163-173.

HMO Pract.

15. O'Connor PJ, Rush WA, Rardin KA, Isham G. AreHMO members willing to engage in two-way communicationto improve health? 1996;10:17-19.

Diabetes Care.

16. Report of the Expert Committee on the Diagnosisand Classification of Diabetes Mellitus. 1997;20:1183-1197.

Eff Clin Pract.

17. O'Connor PJ, Rush WA, Cherney LM, Pronk NP.Screening for diabetes mellitus in high-risk patients: cost,yield, and acceptability. 2001;4:271-277.

Diabetes Care.

18. Gilmer TP, O'Connor PJ, Manning WG, Rush WA.The cost to health plans of poor glycemic control.1997;20:1847-1853.

Diabetes Care.

19. O'Connor PJ, Asche SE, Crain AL, et al. Is patientreadiness to change a predictor of improved glycemiccontrol? 2004;27:2325-2329.

Jt

Comm J Qual Saf.

20. Solberg LI, O'Connor PJ, Christianson JB,Whitebird RR, Rush WA, Amundson GU. The QUESTfor quality: what are medical groups doing about it? 2005;31:211-219.

Diabetes

Care.

21. Gilmer TP, O'Connor PJ, Rush WA, et al. Predictorsof health care costs in adults with diabetes. 2005;28:59-64.

Ann Intern Med.

22. Phillips LS, Branch WT, Cook CB, et al. Clinicalinertia. 2001;135:825-834.

Am J

Manag Care.

23. Karter AJ, Moffet HH, Liu J, et al. Achieving goodglycemic control: initiation of new antihyperglycemictherapies in patients with type 2 diabetes from the KaiserPermanente Northern California Diabetes Registry. 2005;11:262-270.

Advances in Patient Safety: From Research to

Implementation.

24. O'Connor PJ, Sperl-Hillen JM, Johnson PE, RushWA, Blitz G. Clinical inertia and outpatient medicalerrors. In: Volume 2, Concepts and Methodologies.AHRQ Publication No. 050021(2). February 2005.Agency for Healthcare Research and Quality; Rockville,Md. Available online at: http://www.ahrq.gov/qual/advances/. Accessed June 23, 2005.

Circulation.

25. Grundy SM, Cleeman JI, Merz CN, et al.Implications of recent clinical trials for the NationalCholesterol Education Program Adult Treatment Panel IIIguidelines. 2004;110:227-239.

BMC Fam Pract.

26. O'Connor PJ, Desai JR, Solberg LI, Rush WA,Bishop DB. Variation in diabetes care by age: opportunitiesfor customization of care. 2003;4:16.

Health

Care Manag Sci.

27. Johnson PE, Veazie PJ, Kochevar L, et al. Understandingvariation in chronic disease outcomes. 2002;5:175-189.

N Engl J Med.

28. Tuomilehto J, Lindstrom J, Eriksson JG, et al;Finnish Diabetes Prevention Study Group. Prevention oftype 2 diabetes mellitus by changes in lifestyle amongsubjects with impaired glucose tolerance. 2001;344:1343-1350.

N Engl J Med.

29. Knowler WC, Barrett-Connor E, Fowler SE, et al;Diabetes Prevention Research Group. Reduction in theincidence of type 2 diabetes with lifestyle intervention ormetformin. 2002;346:393-403.

Diabetes.

30. Gilmer TP. Relative impact of glycemic control andcardiovascular comorbidity on subsequent resource useamong adults with diabetes. 2004;53(suppl2):A50.

Ann Fam Med.

31. O'Connor PJ, et al. Impact of an electronic medicalrecord on diabetes quality of care. 2005.In press.

Diabetes Care.

32. Meigs JB, Cagliero E, Dubey A, et al. A controlledtrial of web-based diabetes disease management: theMGH diabetes primary care improvement project.2003;26:750-757.

Diabetes

Care.

33. Montori VM, Dinneen SF, Gorman CA, et al;Translation Project Investigator Group. The impact ofplanned care and a diabetes electronic management systemon community-based diabetes care: the MayoHealth System Diabetes Translation Project. 2002;25:1952-1957.

J Gen Intern Med.

34. Greenfield S, Kaplan SH, Ware JE Jr, Yano EM,Frank HJ. Patients' participation in medical care: effectson blood sugar control and quality of life in diabetes.1988;3:448-457.

Download PDF: Factors Driving Diabetes Care Improvement in a Large Medical Group: Ten Years of Progress