Publication
Article
The American Journal of Managed Care
Continuity of Care and Cardiovascular Risk Factor Management: Does Care by a Single Clinician Add to Informational Continuity Provided by Electronic Medical Records?
Author(s):
Background: Electronic medical records allow informationsharing among multiple clinicians treating the same patient,enabling informational continuity between visits.
Objective: To assess the contribution of continuity of care(COC) with a single clinician to short-term outcomes in a setting inwhich electronic medical records are used.
Study Design: Retrospective cohort study.
Methods: Between January 1, 2003, and October 1, 2004, weidentified 3718 patients assessed for lipid and blood pressure controland a subgroup of 1448 patients with diabetes mellitusassessed for glycemic control in the primary care clinics of a largeDepartment of Veterans Affairs healthcare facility. Continuity ofcare was defined as having been seen by the same clinician(physician or nurse practitioner) in the year before testing.Analytic techniques accounting for clustering of patients byproviders yielded robust estimators for the association betweencontinuity with a single clinician and control of these cardiovasculardisease risk factors.
P
Results: Patients with complete COC were more likely to bemen with few medical problems and visits during the study period.Controlling for these differences, we detected no associationbetween COC and patient attainment of recommended goals forcardiovascular disease risk factor control (< .05 for all).
Conclusion: Continuity of care with a single clinician contributeslittle to cardiovascular risk factor management in a settingin which electronic medical records provide enhanced informationalcontinuity, although its value may be greater in the managementand outcomes of established diseases that requirecoordination of care and ongoing collaboration between clinicianand patient.
(Am J Manag Care. 2005;11:689-696)
One of the central characteristics of primarycare is the coordinated and comprehensivemanagement of acute and chronic disease, aswell as risks for future disease occurrence. Achievingthis objective, however, has become increasingly challenginggiven the presence of competing demands thatarise from several sources and have the potential toaffect patient care delivery.1 Having a continuous relationshipwith a single provider has been proposed as ameans of balancing these effects, yet a study2 examiningthis issue found mixed results, related to differencesin the methods used to define and measure continuityof care.
Previous conceptual work has identified several componentsor levels of continuity of care.2 The first aspect,informational continuity, is present when a knowledgebase containing information on a patient's past medicalexperiences, the services provided, and the results ofthese encounters is readily accessible to all cliniciansproviding care. A second component, present when thelocation of care and the composition of practice membersencountering and providing treatment to thepatient remain constant, is referred to as longitudinal orteam-based continuity. A third and more abstractaspect of continuity has been labeled interpersonal continuity,which represents the extent to which a relationshipbetween a clinician and a healthcare consumerhas developed and is characterized by trust, respect, anunderstanding of preferences in communication anddecision making, and a clinician's sense of responsibilityor commitment to providing care to a patient.Although each of these components may exist alone,continuity of care has been described as a hierarchy,with the levels comprising a progressively more complexstructure that includes previous elements.
Establishing systems of care that ensure continuityof care in each of these dimensions will likely requireconsiderable investment in infrastructure and personnel.The incremental value of each element, however,has not yet been established, and it may be that somecomponents of continuity of care are more important inensuring favorable outcomes than others. Healthcareinformation management systems or electronic medicalrecords (EMRs) offer a mechanism through which informationabout a patient may be shared by multiple clinicianstreating a patient. Although functionalities differ,many EMRs can be used to provide a record of currentand past therapies, procedure results, and consultants'notes, leading to greater treatment coordination whenmultiple clinicians participate in a patient's care. SomeEMRs, including the system used in the Department ofVeterans Affairs (VA), offer enhanced capabilities thatfacilitate delivery of care through the use of clinicalreminders or that monitor quality of care and generatereports for physician feedback on their performance. Itis unclear whether receiving care over time from thesame clinician at the same site contributes to bettertreatment outcomes among systems using EMRs. Theobjective of this study was to assess the incrementalvalue of longitudinal continuity to the short-termresults of cardiovascular risk factor management, whichis often delayed or omitted when a usual source of careis not available.3,4 We hypothesized that continuity ofcare with a single provider would add little to the controlof cardiovascular risk factors in a setting in whichinformational continuity and other treatment-enablingsystems were already present.
METHODS
Cardiovascular diseases represent the most commoncause of adult mortality in the United States,5 and thescience base supporting the value of risk factor controlfor this group of conditions is substantial.6-10 In this retrospectivecohort study, we focused on the extent towhich short-term outcomes related to cardiovascularrisk factor management were consistent with recommendedgoals, rather than on the process of screening ortreatment initiation. Methods for the effective control ofprimary risk factors for cardiovascular diseases, includinglipid, blood pressure, and glycemic control, are widelyavailable in the form of clinical guidelines andconsensus recommendations.11-15 This study was conductedin 2 adjacent primary care clinics of equal sizewith comparable staffing at the Louis Stokes ClevelandDepartment of Veterans Affairs Medical Center, a tertiarycare facility affiliated with a major teaching hospitaland a physician residency training program.
Study Subjects
The laboratory database at our facility was searchedto identify all patients who had undergone assessmentof low-density lipoprotein cholesterol (LDL-C) betweenJanuary 1, 2003, and October 1, 2004. Patients withfewer than 2 visits in the preceding 12 months to theprimary care clinics at the study site were excluded.The 12 months preceding the date of LDL-C testing (theindex date) constituted the exposure period for eachpatient. Outpatient medical records were obtained forthis period for each individual to identify all encountersat the Louis Stokes Cleveland Department of VeteransAffairs Medical Center, the diagnoses or proceduresassociated with each date of service, and the identificationnumber for the clinician (physician or nurse practitioner)responsible for each visit.
Study Measures
Informational and Longitudinal Continuity.
All cliniciansproviding care during the study period did sousing the VA EMR, an information management systemthat stores and organizes data on patients' procedureresults, radiographic reports, specialist consultations,immunizations, and medications, as well as their medical,surgical, and social histories.16 This system offeredinformational continuity for each patient in our sample,while providing other advanced functions thatenable delivery of high-quality care. The EMR developedby the VA is an information system that is readilyavailable on a continuous basis, simultaneously orsequentially, to multiple health professionals providingcare within a clinic or at remote locations. In addition,it has features that facilitate delivery of manyaspects of care through targeted clinical reminders, aswell as features that provide feedback to healthcareproviders on their ability to meet performance goals inthese areas.
We used provider identification numbers in the EMRand dates of service for each visit during the study periodto assess visit patterns to the primary care clinics.Although we intended to base our analyses on indices ofcontinuity of care used in other studies,17-19 we foundthat the distribution was significantly skewed and wouldhave violated assumptions of linear modeling. Therefore,we created a straightforward measure reflecting completecontinuity with a single clinician vs treatment bymultiple clinicians and used this dichotomous indicatoras the independent variable in all analyses.
Because clinicians (physicians, nurse practitioners,and physicians' assistants) at the study site often outlineplans of care that involve follow-up visits with nursesor clinical pharmacists to assess adherence, toeducate patients on self-management strategies, todetermine response to therapy, and to titrate up medicationdosages, we created models that included andexcluded these team-related visits.
Indicators of Cardiovascular Risk Factor Control.
The EMR provided data for the end points examined inthis study. In addition to LDL-C levels, we obtained systolicand diastolic blood pressures for each patient at themost recent date of service during the preceding 12months, as well as the most recent hemoglobin A1C (A1C)level for patients with diabetes mellitus. From these data,we created a series of dichotomous measures indicatingwhether goals for cardiovascular risk factor managementhad been attained according to the recommendations ofthe Third Report of the National Cholesterol EducationProgram (NCEP) Expert Panel on Detection, Evaluation,and Treatment of High Blood Cholesterol in Adults11 ATPIII; the Seventh Report of the Joint National Committeeon Prevention, Detection, Evaluation, and Treatment ofHigh Blood Pressure (JNV 7)20; and the AmericanDiabetes Association (ADA).21
To follow the recommendations of the Third Reportof the National Cholesterol Education Program ExpertPanel on Detection, Evaluation, and Treatment of HighBlood Cholesterol in Adults11 as closely as possible, weused clinical, laboratory, and demographic data fromthe EMR to determine the presence of major risk factorsthat modified goals for lipid control. These included adiagnosis of hypertension, a high-density lipoproteincholesterol level less than 40 mg/dL (< 1.03 mmol/L) orgreater than 60 mg/dL (> 1.55 mmol/L), age older than45 years for men or older than 55 years for women, andthe presence of coronary heart disease.11 Conditionssuch as diabetes mellitus or peripheral vascular diseaseare closely linked with development of atheroscleroticdisease and have therefore been considered by NCEP tobe coronary heart disease equivalents for the purpose ofdetermining LDL-C treatment goals.11 The presence ofany of these conditions was determined by searchingthe VA inpatient and outpatient clinical databases usinga strategy based on diagnosis and procedure codesdescribed in previous work.22 Tobacco use, another factorused in determining lipid control goals, was assessedusing data from physician responses to system-generatedelectronic clinical reminders in the EMR thatprompted discussion of smoking cessation withpatients. We used threshold blood pressures of 140 mmHg systolic and 90 mm Hg diastolic, or pressures lessthan 130/80 mm Hg for patients with diabetes mellitus,20 to indicate attainment of desired blood pressurecontrol. Although we considered a variety of thresholdsfor acceptable long-term glycemic control, we followedthe current recommendation of ADA,21 which definesdesirable long-term glycemic control by A1C levels lessthan 7.0%.
Covariates.
International Classification of
Diseases, Ninth Revision, Clinical Modification
To account for the effect of otherpatient-related factors on the probability of achievingcardiovascular risk factor control, we consideredpatient gender, age at the end of the study period, numberof primary care visits during the 12 months beforeLDL-C testing, and number of health conditions, identifiedusing a set of diagnosiscodes described previously.23 We were unable toinclude patient race or ethnicity as a covariate in ouranalysis because it was missing in up to 30% of patientdata in our sample and, when available, had been determinedby means other than patient self-report, the preferredmethod of assessment.24
Statistical Analysis
t
A variety of analytic strategies tested the associationbetween our measure of longitudinal continuity andcardiovascular risk factor management. We used the χ2test, test, or Mann-Whitney test as appropriate to thedata type to assess unadjusted differences in selectedfactors according to receipt of care from a single clinicianvs multiple clinicians. Separate multivariate logisticregression models assessed patients' likelihood ofbeing at nationally recommended goals for cardiovascularrisk factor control, adjusting for the covariatesdescribed in the previous paragraph and for clusteringby clinician, to yield robust standard errors for modelparameters. Similar linear regression models weredeveloped for continuous outcomes such as LDL-Clevel, systolic blood pressure, and A1C level. Becausewe anticipated finding no difference in outcomes relatedto the presence or absence of longitudinal continuity,we did not reduce the level of significance to account formultiple comparisons, as this would have made it moredifficult to reject the study hypothesis. Stata version 7.0(StataCorp LP, College Station, Tex) was used for allanalyses. The institutional review board of the LouisStokes Cleveland Department of Veterans AffairsMedical Center approved the study.
RESULTS
P
P
P
P
Of the 3718 patients in our study, most were menwith few chronic medical conditions other than coronaryheart disease or equivalent conditions (Table 1).Most patients in our sample had been seen in primarycare outpatient clinics 2 to 4 times in the 12 monthspreceding LDL-C testing, and 2712 (73%) patients hadcomplete continuity of care with a single clinician orclinician team. Compared with patients who had beenseen by more than 1 clinician, these patients weremore likely to be men, to have fewer medical illnesses,to have less frequent visits to the primary careclinic, and to have been referredless often in the past to a VA nutritionclinic (< .05 for all). No significantdifferences betweenpatients seen by a single clinicianvs multiple clinicians were noted inunadjusted comparisons of continuousoutcomes for LDL-C level (98.7mg/dL [2.56 mmol/L] vs 100.8 mg/dL[2.61 mmol/L], = .08), systolicblood pressure (139.9 mm Hg vs139.7 mm Hg, = .81), or A1C level(7.01% vs 6.99%, = .80).
Overall, more than 60% of patientsachieved acceptable LDL-C levelcontrol and approximately 50%attained blood pressure control withvalues less than 140/90 mm Hg.Among patients with diabetes mellitus,more than 45% achieved the recommendedlevel of glycemic control,although fewer than one thirdattained the level of blood pressurecontrol (< 130/80 mm Hg) recommendedby the JNC 7.20
The level of cardiovascular riskfactor control did not differ significantlybetween those who receivedall of their primary care from a singleclinician vs multiple clinicians(Figure). After accounting forpatients' demographic characteristics,number of outpatient visits,number of chronic medical illnessesthey had been treated for, and referralfor nutritional advice in the past,the association with continuity ofcare did not achieve statistical significancefor any aspect of cardiovascularrisk factor control studied (Table2). We were similarly unable todemonstrate an association withcontinuity of care in linear regressionmodels that used LDL-C level,systolic blood pressure, or A1C levelas a continuous outcome. These findingswere unchanged using alternativecontinuity measures in modelsbased on visits with a single clinicianonly or in models stratified by thepresence of multiple (≥ 2) chronicillnesses or frequent (≥ 5) clinic visits.
DISCUSSION
In this study examiningcardiovascular riskfactor management inthe primary clinics of alarge VA facility, continuouscare with a singleclinician was not associatedwith better cardiovascularrisk factorcontrol. Although weobserved that discontinuityin the provision ofcare was uncommon—more than 70% of thestudy subjects saw thesame primary care clinicianor clinician teamduring the 12 monthsbefore LDL-C testing—some argue that evengreater continuity ofcare would further optimizethe potential forbetter outcomes.25 Our results from a setting in whichenhanced informational continuity is present do notsupport this assertion and suggest instead that longitudinalcontinuity with a single clinician adds little to theshort-term outcomes of cardiovascular risk factor management.
Healthcare professionals are expected to advocate fortheir patients and to maintain patient welfare as the primaryconsideration in providing care, although manyfactors affect the type and quantity of services theydeliver.26 Recent findings, for example, indicate thatvariation in treatment processes and outcomes occur toa greater extent at the facility level than at the providerlevel.27 This observation supports a theoretical frameworkthat highlights the need to consider the organizationalcontext in which care is delivered and thestructure of the practice when interpreting outcomesgenerated in these settings.28 We used this approach tobetter understand the results reported herein.
In the mid 1990s, the VA undertook a systemwideredesign process that emphasized the importance ofdocumenting the processes and outcomes of healthcarefor acute and chronic conditions, as well as identifyingopportunities for improving care, for increasing accessto preventive care for veterans,29 and for developingoverlapping systems of care that better support healthcareproviders in achieving these goals.30,31 Followingthis, the VA established a network of research centers togenerate data on and evaluate interventions to improvethe quality of healthcare for veterans. The results ofthese efforts have led to the evolution of a culture ofquality within the organization and to improvement inthe processes and outcomes of many forms of chronicand preventive care delivery, matching those experiencedby patients with other forms of public or privatehealth insurance.32-34
Redesign of the VA during the past decade has alsoshaped practice structure through the introduction ofcare management protocols and systems.31,33,35 Examplesinclude computerized network-based programsthat search for diagnosis codes or demographic andclinical data in EMRs to identify patients eligible for specifichealthcare services and that prompt clinicians ateach encounter to take action until the service has beendelivered or obviating circumstances are documented.A new generation of information management modules,now in use, is likely to assist clinicians in rapidly recognizingtrends in patient responses to therapy and to providedecision support for recommended therapeuticalternatives. In addition, automated mechanisms forauditing and providing feedback assist clinicians ingauging their performance against achievable benchmarksand desired targets. Continued work to ensurethe accuracy and reliability of data on which auditsand feedback are based and careful consideration ofthe treatment processes and outcomes that constituteoptimal quality are essential in achieving thepotential of the EMR as a tool for evaluation andimprovement in healthcare. Finally, future VA initiativessimilar to those being considered by the Centersfor Medicare & Medicaid Services may motivateproviders to achieve higher levels of cardiovascularrisk factor control through financial incentives that"pay for performance."
Our findings should not be interpreted to mean thatcontinuity of care with a single provider is an irrelevantor anachronistic aspect of primary care. Indeed, severalfactors may have explained our failure to detect anincremental benefit associated with longitudinal continuity.First, our analysis failed to consider patients' contributionsto outcomes in implementing the care oradvice they received. Because the health beliefs andpreferences of healthcare consumers contribute to theoutcomes they attain,36 it is possible that individualsreceiving continuous care from a single clinician wereprovided more preventive services but simply failed tofollow through. Our study was unable to determine theextent to which this occurred. Second, care obtainedoutside the VA may have affected the preventive careoutcomes that we observed. Although low copaymentsand a prescription benefit provide a strong financialincentive for veterans to receive care within the VA system,older veterans may have used their Medicare benefitsto obtain care from physicians outside of the VA.Previous studies,32,34 however, demonstrate that preventivecare is typically provided at higher frequencieswithin the VA than in non-VA settings, limiting the contributionof non-VA care to the outcomes we observed.Third, our finding of no association between continuitywith a single clinician and control of cardiovascular diseaserisk factors may have resulted from measuring longitudinalcontinuity during a short (12 month) period.Although prior studies37,38 examining continuity useintervals of similar duration, a longer time frame mightnonetheless have allowed important differences toemerge, leading to a stronger association with cardiovascularrisk factor control. Confirmation of our findingsat other VA sites during longer periods would beuseful in assessing their generalizability and sensitivityto the length of time during which continuity is measured.Fourth, the outcomes we chose to measure mayhave been unlikely to be affected by longitudinal continuity.Applying a guideline in patient care is a fundamentallydifferent task than managing a complexestablished illness, in which coordination of care andphysician-patient collaboration are often required.
Several limitations should be considered in interpretingthese findings. First, unmeasured factors, especiallythose reflecting other characteristics of ourfacility or the systems in use, may have affected outcomesor moderated associations with longitudinalcontinuity. Prospective studies in our setting are underway to identify organizational features and dynamicsthat affect the processes and outcomes of care.
Second, our results may not be generalizablebecause of unique characteristics of our study sample.Although the VA serves older adults from low-incomebackgrounds, a group potentially more vulnerable tothe negative effects of interruptions in continuity ofcare,39 future studies using different samples in otherorganizational contexts would be nonetheless useful inexploring further the effect of continuity of care ontreatment outcomes.
Third, the negative finding in our study may haveresulted from a type II error. However, by setting the alevel at .05, taking into account the number of clinicians,and considering the effect of clustering ofpatients within clinician practices, we had sufficientpower to detect an odds ratio as low as 1.14 in the multivariateanalyses for our summary measures of LDL-Clevel and blood pressure control. For models ofglycemic and blood pressure control in patients withdiabetes mellitus, we had sufficient power to detect anodds ratio as low as 1.25.
Fourth, our study did not include measures of thequality of interactions between clinicians and patientsbut focused instead on other aspects of continuity ofcare. Although some evidence suggests the importanceof interpersonal continuity in many areas of care,40,41the retrospective study design and limitations of thedata available to us did not permit a test of the relationshipbetween clinician-patient interaction and cardiovascularrisk factor management.
CONCLUSIONS
Although longitudinal continuity with a singleprovider is often associated with greater patient andprovider satisfaction, systems that enhance informationalcontinuity may be more important and mayserve to counterbalance the effect of interruptions incontinuity of care in some circumstances. Designingand implementing information management systemssimilar to those in the VA may offer clinicians in otherorganizational settings an approach that is less dependenton longitudinal continuity in managing cardiovascularrisk factors. Understanding whether thisapproach will translate similarly to better longer-termcardiovascular outcomes or improvements in otherareas of disease management will require additionalstudy.
Acknowledgements
We are grateful to practice personnel at the study site for their interest,participation, and support in conducting this study. We also wish to thankSusan Flocke, PhD, and Kurt Stange, MD, PhD, for their helpful commentson early versions of the manuscript.
From the Department of Medicine, Louis Stokes Cleveland Department of VeteransAffairs Medical Center (DL, SO, DA); Department of Epidemiology and Biostatistics, CaseWestern Reserve University (DL, DA); and University Hospitals of Cleveland (CR);Cleveland, Ohio.
This study was supported by grants RCD-03028-1 (DL) and REA 01-100 (DA) from theDepartment of Veterans Affairs, Veterans Health Administration, Health Services Researchand Development Service, Washington, DC. The views expressed in this article are those ofthe authors and do not necessarily represent the views of the Department of VeteransAffairs.
Address correspondence to: David Litaker, MD, PhD, Department of Medicine, LouisStokes Cleveland Department of Veterans Affairs Medical Center, 10701 East Boulevard(151W), Cleveland, OH 44106. E-mail: david.litaker@med.va.gov.
J Fam Pract.
1. Jaén CR, Stange KC, Nutting PA. The competing demands of primary care: amodel for the delivery of clinical preventive services. 1994;38:166-171.
Ann Fam
Med.
2. Saultz JW. Defining and measuring interpersonal continuity of care. 2003;1:134-143.
Am J Public Health.
3. DeVoe JE, Fryer GE, Phillips R, Green L. Receipt of preventive care amongadults: insurance status and usual source of care. 2003;93:786-791.
Fam
Pract.
4. Devroey D, Coigniez P, Vandevoorde J, Kartounian J, Betz W. Prevention andfollow-up of cardiovascular disease among patients without a personal GP. 2003;20:420-424.
Health, United States, 2003:
Chartbook on Trends in the Health of Americans.
5. Fried VM, Prager K, MacKay AP, Zia H. Hyattsville, Md: NationalCenter for Health Statistics; 2003.
Lancet.
6. Scandinavian Simvastatin Survival Study (4S) Group. Randomised trial of cholesterollowering in 4444 patients with coronary heart disease: the ScandinavianSimvastatin Survival Study (4S). 1994;344:1383-1389.
Lancet.
7. Health Protection Study Collaborative Group. MRC/BHF Heart ProtectionStudy of cholesterol lowering with simvastatin in 20,536 high-risk individuals: arandomised placebo-controlled trial. 2002;360:7-22.
N Engl J Med.
8. Shepherd J, Cobbe SM, Ford I, et al. Prevention of coronary heart disease withpravastatin in men with hypercholesterolemia. 1995;333:1301-1307.
Lancet.
9. Shepherd J, Blauw GJ, Murphy MB, et al. Pravastatin in Elderly Individuals atRisk of Vascular Disease (PROSPER): a randomised controlled trial. 2002;360:1623-1630.
Arch Intern Med.
10. Pearson TA, Laurora I, Chu H, Kafonek S. The Lipid Treatment AssessmentProject (L-TAP): a multicenter survey to evaluate the percentages of dyslipidemicpatients receiving lipid-lowering therapy and achieving low-density lipoproteincholesterol goals. 2000;160:459-467.
JAMA.
11. Expert Panel on Detection, Evaluation, and Treatment of High BloodCholesterol in Adults. Executive summary of the Third Report of the NationalCholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, andTreatment of High Blood Cholesterol in Adults (Adult Treatment Panel III). 2001;285:2486-2497.
Treating Tobacco Use and
Dependence: Clinical Practice Guideline.
12. Fiore MC, Bailey WC, Cohen SJ, et al. Rockville, Md: US Dept of Health andHuman Services; June 2000.
Circulation.
13. Pearson TA, Blair SN, Daniels SR, et al; American Heart Association ScienceAdvisory and Coordinating Committee. AHA guidelines for primary preventionof cardiovascular disease and stroke: 2002 update: consensus panel guide tocomprehensive risk reduction for adult patients without coronary or other atheroscleroticvascular diseases. 2002;106:388-391.
Guide to Clinical Preventive
Services.
14. United States Preventive Services Task Force. Washington, DC: Office of Disease Prevention and Health Promotion,US Dept of Health and Human Services; 1996.
15. Management of Dyslipidemia in Primary Care Web site. VHA/DoDguideline for diagnosis and management of lipid disorders. Available at:http://www.oqp.med.va.gov/cpg/DL/G/DL_about.htm. Accessed December 6,2004.
Cancer.
16. Fletcher RD, Dayhoff RE, Wu CM, Graves A, Jones RE. Computerized medicalrecords in the Department of Veterans Affairs. 2001;91:1603-1606.
J Fam Pract.
17. Flocke SA, Stange KC, Zyzanski SJ. The impact of insurance type and forceddiscontinuity on the delivery of primary care. 1997;45:129-135.
Med
Care.
18. Bice TW, Boxerman SB. A quantitative measure of continuity of care. 1977;15:347-349.
J Fam Pract.
19. Magill M, Senf J. A new method for measuring continuity of care in familypractice residencies. 1987;24:165-168.
Hypertension.
20. Chobanian AV, Bakris GL, Black HR, et al; Joint National Committee onPrevention, Detection, Evaluation, and Treatment of High Blood Pressure,National Heart, Lung, and Blood Institute; National High Blood PressureEducation Program Coordinating Committee. Seventh Report of the JointNational Committee on Prevention, Detection, Evaluation, and Treatment of HighBlood Pressure. 2003;42:1206-1252.
Diabetes Care.
21. American Diabetes Association. Standards of medical care for patients withdiabetes mellitus. 2003;26(suppl 1):533-550.
Med Care.
22. Litaker D, Koroukian SM. Racial differences in lipid-lowering agent use inMedicaid patients with cardiovascular disease. 2004;42:1009-1018.
ICD-9-CM
J Clin Epidemiol.
23. Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index for usewith administrative databases. 1992;45:613-619.
Federal Register
24. Office of Management and Budget. Revisions to the standards for the classificationof federal data on race and ethnicity. notice, October 30,1997. Available at: http://www.whitehouse.gov/omb/fedreg/ombdir15.html.Accessed September 30, 2005.
J Fam Pract.
25. Stevens GD, Shi L. Racial and ethnic disparities in the quality of primary carefor children. 2002;51:e573.
Ann Intern Med.
26. ABIM Foundation, American Board of Internal Medicine; ACP-ASIMFoundation, American College of Physicians-American Society of InternalMedicine; European Federation of Internal Medicine. Medical professionalism inthe new millennium: a physician charter. 2002;136:243-246.
Health Serv Res.
27. Krein SL, Hofer TP, Kerr EA, Hayward RA. Whom should we profile? examiningdiabetes care practice variation among primary care providers, provider groups,and healthcare facilities. 2002;37:1159-1180.
Milbank Mem Fund Q.
28. Donabedian A. Evaluating the quality of care. 1966;44(suppl):166-206.
Med Care.
29. Chapko MK, Borowsky SJ, Fortney JC, et al. Evaluation of the Department ofVeterans Affairs community-based outpatient clinics. 2002;40:555-560.
JAMA.
30. Kizer KW. From the Veterans Health Administration. 1996;275:1069.
Am J Med Qual.
31. Kizer KW. The "new VA": a national laboratory for health care quality management.1999;14:3-20.
Prev Med.
32. Rabiner DJ, Branch LG, Sullivan RJ. The receipt of prevention services by veteransusing VA versus non-VA facilities. 1998;27:690-696.
N Engl J Med.
33. Jha AK, Perlin JB, Kizer KW, Dudley RA. Effect of the transformation of theVeterans Affairs Health Care System on the quality of care. 2003;348:2218-2227.
Ann Intern
Med.
34. Asch S, McGlynn E, Hogan M, et al. Comparison of quality of care for patientsin the Veterans Health Administration and patients in a national sample. 2004;141:938-945.
Ann
Intern Med.
35. Kerr EA, Gerzoff RB, Krein SL, et al. Diabetes care quality in the VeteransAffairs Health Care System and commercial managed care: the TRIAD study. 2004;141:272-281.
J Health Soc Behav.
36. Andersen R. Revisiting the behavioral model and access to medical care: doesit matter? 1995;3:1-10.
Arch Fam Med.
37. Gill JM, Mainous AG. The role of provider continuity in preventing hospitalizations.1998;7:352-357.
Fam Med.
38. Parkerton PH, Smith DG, Straley HL. Primary care practice coordination versusphysician continuity. 2004;36:15-21.
Ann Fam Med.
39. Nutting PA, Goodwin MA, Flocke SA, Zyzanski SJ, Stange KC. Continuityof primary care: to whom does it matter and when? 2003;1:149-155.
Ann Fam Med.
40. Saultz JW, Albedaiwi W. Interpersonal continuity of care and patient satisfaction:a critical review. 2004;2:445-451.
Ann Fam Med.
41. Saultz JW, Lochner J. Interpersonal continuity of care and care outcomes: acritical review. 2005;3:159-166.
