COACH Trial: A Randomized Controlled Trial of NursePractitioner/Community Health Worker Cardiovascular DiseaseRisk Reduction in Urban Community Health CentersJerilyn K. Allen, RN, ScD, FAAN1,2,3, Cheryl R. Dennison Himmelfarb, RN, ANP, PhD,FAAN1,2, Sarah L. Szanton, PhD, CRNP1, Lee Bone, MPH, RN3, Martha N. Hill, PhD, RN,FAAN1,2,3, David M. Levine, MD, MPH, ScD1,2,3, Murray West, MD4, Amy Barlow, CRNP,CDE1, LaPricia Lewis-Boyer, BA1, Mary Donnelly-Strozzo, MS, MPH, ACNP-BC, ANP-BC1,Carol Curtis, BS, CCRP1, and Katherine Anderson, MS11Johns Hopkins University School of Nursing, Baltimore, MD2Johns Hopkins University School of Medicine, Baltimore, MD3Johns Hopkins University School of Public Health, Baltimore, MD4Baltimore Medical Systems, Baltimore, MDAbstractBackground—Despite well-publicized guidelines on the appropriate management ofcardiovascular disease (CVD) and type 2 diabetes, the implementation of risk-reducing practicesremains poor. This paper describes the results of a randomized controlled clinical trial evaluatingthe effectiveness of a comprehensive program of cardiovascular disease risk reduction deliveredby nurse practitioner/community health worker (NP/CHW) teams versus enhanced usual care(EUC) to improve lipids, blood pressure, glycated hemoglobin (HbA1c), and patients’ perceptionsof the quality of their chronic illness care in patients in urban community health centers.Methods and Results—A total of 525 patients with documented cardiovascular disease, type 2diabetes, hypercholesterolemia, or hypertension and levels of LDL-cholesterol, blood pressure orHbA1c that exceeded goals established by national guidelines were randomized to NP/CHW(n=261) or EUC (n=264) groups. The NP/CHW intervention included aggressive pharmacologicmanagement and tailored educational and behavioral counseling for lifestyle modification andproblem solving to address barriers to adherence and control. As compared to EUC, patients in theNP/CHW group had significantly greater 12 month improvement in total cholesterol (difference,19.7mg/dL), LDL cholesterol (difference,15.9 mg/dL), triglycerides (difference, 16.3 mg/dL),systolic blood pressure (difference, 6.2 mm Hg), diastolic blood pressure (difference, 3.1 mm Hg),HbA1c (difference, 0.5%), and perceptions of the quality of their chronic illness care (difference,1.2 points).Conclusions—An intervention delivered by a NP/CHW team using individualized treatmentregimens based on treat-to-target algorithms can be an effective approach to improve risk factorstatus and perceptions of chronic illnes care in high risk patients.Corresponding author: Jerilyn K. Allen, RN, ScD, FAAN, M Adelaide Nutting Professor and Associate Dean for Research, JohnsHopkins University School of Nursing, 525 North Wolfe Street, Baltimore, MD 21205, P: 410-614-4882, F: 410-614-1446,jallen@son.jhmi.edu.Conflict of Interest Disclosures: None.Clinical Trial Registration: www.clinicaltrials.gov Identifier NCT00241904NIH Public AccessAuthor ManuscriptCirc Cardiovasc Qual Outcomes. Author manuscript; available in PMC 2012 November 1.Published in final edited form as:Circ Cardiovasc Qual Outcomes. 2011 November 1; 4(6): 595–602. doi:10.1161/CIRCOUTCOMES.111.961573.NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

KeywordsRandomized trial; Cardiovascular disease; Diabetes; PreventionINTRODUCTIONApproximately 831,000 Americans die annually from cardiovascular disease (CVD), withlower income, prior coronary heart disease (CHD), and diabetes populations differentiallyrepresented in these deaths.1 Despite well-publicized guidelines on the appropriatemanagement of cardiovascular disease (CVD) and type 2 diabetes, implementation of risk-reducing practices remains poor. Several different models of chronic disease casemanagement have emerged to respond to growing concerns about the quality and increasingcosts of health care; however, evaluation of their impact on patient outcomes or cost islimited.Case management by a specially trained nurse-led team, including community healthworkers, has been shown to be among the most efficacious strategies to improvemanagement of CVD risk factors in many studies.2–6 Several studies have shown that nursemanagement clinics are at least as beneficial in achieving goals as are other clinics managedby physicians, and in many cases actually result in marked improvement in the outcomesincluding patient satisfaction and utilization of healthcare services, compared with usualcare. For example, trained nurses providing care have demonstrated successful strategies forimproving lipid levels in patients with elevated low-density lipoprotein cholesterol (LDL-C)and blood pressure (BP).4–9 In the nurse management models, factors such as patienteducation and counseling and even regular telephone follow-up by a nurse showed markedsustained improvement in medication adherence and goal achievement. Nurse casemanagers have been shown to improve adherence to guidelines in part by serving as a bridgeto physician care and by adhering more strictly to management algorithms, including manycounseling features that are not within the time frame of a busy physician in practice.Several studies suggest that a nurse-led team management program is the most effectivestrategy to date for reducing LDL-C.6, 10Nurses managing patients with diabetes also have a more favorable impact on chronicdisease parameters, including adherence to recommendations for diet and for renal testing.11Nurse case management improves control of diabetes in clinical settings, with significantreductions in fasting blood glucose, body weight, glycosylated hemoglobin, and LDL-C.12Telephone management of diabetic patients by a nurse has been shown to markedly improveCHD risk factors, including lipids. Diabetic patients were more likely to be appropriatelytreated with a lipid lowering therapy when managed by the nurse over the phone thanpatients managed solely with usual care.13In low income and minority populations, community health workers (CHWs) or lay healthadvisors have often participated in team-based care for the management of CVD risk factors,particularly hypertension and diabetes.14, 15 While there are too few randomized clinicaltrials of the role of these individuals, there is sufficient collective experience to suggest thatthis role can be an important one in improving adherence in high risk subsets of thepopulation.9, 16–18 Trained CHWs, front line health and human service care providers, mostoften share the same ethnicity, geographic community, and socioeconomic background ofthe patients they serve. The theoretical rationale for using CHWs is a shared perspective andexperience that enhances trust between CHW and the patient, and enables the CHW toeffectively link underserved populations to healthcare resources where traditional healtheducation and outreach efforts have failed.19, 20 CHWs also bridge the communicationAllen et al. Page 2Circ Cardiovasc Qual Outcomes. Author manuscript; available in PMC 2012 November 1.NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

barriers between patients and health care providers which can catalyze provider and healthsystem changes.21, 22 CHWs have been shown to improve quality of care, satisfaction withcare, increase access to care, reduce healthcare costs, strengthen local economies andfamilies, and foster community capacity building.15, 20, 23–25 CHWs also have been shownto be effective in research as interviewers and interventionists.9, 17, 21, 26The aim of this study was to evaluate the effectiveness of a comprehensive program of CVDrisk reduction delivered by nurse practitioner (NP)/CHW teams versus enhanced usual careto improve lipids, blood pressure, and HbA1c levels in patients in urban community healthcenters. This effectiveness research is one of the first studies testing a model of NP/CHWteam care in urban federally-qualified community health centers. We used community basedparticipatory research establishing a true academic health center-community practicepartnership to enhance the applicability and sustainability of the intervention.METHODSStudy designThe complete methods of the Community Outreach and Cardiovascular Health (COACH)study have been detailed elsewhere.27 Briefly, we used community-based participatoryresearch (CBPR) as a theoretical framework for this study. CBPR is a methodology thatpromotes active community involvement in the processes that shape research andintervention strategies, as well as the conduct of research studies.28. This research utilized aCommunity-Provider Advisory Committee to guide all aspects of the study.COACH was a randomized controlled trial in which 525 patients were randomly assigned toone of two groups: comprehensive intensive management of cardiovascular disease (CVD)risk factors by a NP/CHW team or an enhanced usual care (EUC) control group. Individualsin the control group received usual care from their primary provider which was enhanced byfeedback regarding CVD risk factors provided to the patient and their provider. Those in theintensive intervention group received enhanced usual care plus management by the NP/CHW team. The program included aggressive pharmacologic management, tailorededucational and behavioral counseling for lifestyle modification, identification of barriers toadherence and control, phone follow-ups between visits and pre-appointment reminders.Participants Patients were recruited between July 2006 and July 2009 from two community healthcenters which are part of the federally - qualified community health center entitledBaltimore Medical Systems Incorporated (BMS). The focus of these clinics is on primarycare in communities designated as medically underserved areas.Patients identified from clinic-based computerized ICD 9 codes were eligible if they wereAfrican American or Caucasian and had diagnosed CVD defined as a prior MI,revascularization procedure for coronary disease, ischemic heart disease, stroke, peripheralvascular disease, or hypercholesterolemia, hypertension, or had diagnosed type 2 diabetesreceiving any therapy. They had to be ≥ 21 years of age, and be able to speak andunderstand English. Patients were enrolled in the trial if they had at least one of thefollowing criteria within the past six months at the time of the medical record reviews: (1)an LDL-C ≥ 100 mg/dl or LDL-C ≥ 130 mg/dl if no diagnosed CVD or diabetes, (2) a bloodpressure > BP 140/90 mm Hg or > 130/80 mm Hg if diabetic or renal insufficiency, or (3) ifdiabetic, a HbA1c 7% or greater or glucose ≥ 125 mg. Patients were excluded if they had aserious life-threatening non-cardiac co-morbidity with a life expectancy of less than 5 years(AIDS or cancer for example), had a serious physician-recorded psychiatric morbidity thatAllen et al. Page 3Circ Cardiovasc Qual Outcomes. Author manuscript; available in PMC 2012 November 1.NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

would preclude participating in their own care, or were sufficiently neurologically impairedto preclude participation in their own care.Of the 3899 screened for eligibility, 525 were enrolled in the trial (Figure 1). Theparticipants were randomly assigned, stratified by race and sex, to receive the NP/CHWintervention or EUC. All participants provided written informed consent. The protocol wasapproved by the Johns Hopkins University Institutional Review Board.Intervention The NP/CHW intervention focused on behavioral interventions to effect therapeutic lifestylechanges (TLC) and adherence to medications and appointments as well as the prescriptionand titration of medications. Patients were followed for one year. The NP and CHW workedas a team. The NP functioned as the case coordinator for each study participant. Sheoversaw the initial assessment and tailored the intervention plan, conducted the interventionincluding lifestyle modification, counseling and medication titration and prescription,consulted with the physician, and supervised the CHW. Specific algorithms for drugtreatment of hyperlipidemia, hypertension (HBP), hyperglycemia, as well for ACE, and β-blocker therapy were developed for this study based on current guidelines and standards ofcare. (Algorithms can be found in the Appendix)In addition to meeting with the NP, patients and their support person met with the CHWwho spent additional time problem solving anticipated barriers to treatment adherence,including issues important to the patient’s life which might not be directly related tocardiovascular health. The CHW also reinforced instructions by the NP related to integrationof lifestyle modifications and medication therapies and assisted patients in designing a set ofreminders, prompts, logs, pill organizers, alarm clocks, or whatever the individual believedwould work for them to assist in following complex regimens. The intensity of the nurse/CHW intervention was greater among those who had not yet achieved goals. Follow-upalgorithms guided the frequency and type of follow-up. Those patients not making progresstowards their goal levels received more frequent telephone follow-up from the CHW.A low-literacy Wellness Guide was developed specifically for the study as a behavioral toolfor the NP, CHW and patient team to promote TLC. The patient received the Guide at thefirst encounter, took it home as a tool for making changes and was asked to bring it to eachvisit. The Wellness Guide had sections focusing on the patient’s laboratory results andtherapeutic goals for weight, blood pressure, lipids, and HbA1c (for patients with diabetes);medication reconciliation and customized tips for taking medicine; healthy eating, includingstrategies for portion control; increasing physical activity and a customized walkingprogram; smoking cessation; and a place to record questions for future visits. Each sectionhad a place for recording the patient’s goals, potential barriers, strategies to deal withdifficult situations, ways to reward oneself; and identification of support people to helpfacilitate meeting goals. This section was completed during the counseling sessions with theNP and CHW.The lifestyle behaviors of healthy low-fat, low sodium diet, regular moderate-intensityphysical activity, and smoking cessation were the focus of TLC counseling interventions.The nurse initiated recommendations for healthy low fat, low sodium eating recommendedin the TLC diet, adapted for diabetics according to standards of the American DiabetesAssociation. 29 The importance of dietary adherence was emphasized as an adjunct topharmacotherapy. Recognition of food preferences were important along with how tochoose low-fat, low sodium foods, modify recipes, self-monitor fat and sodium intake, anddevelop individualized low-fat, low-sodium eating plans. Some areas of focus includedreducing portion size, reducing fast food intake, avoiding processed foods high in sodiumAllen et al. Page 4Circ Cardiovasc Qual Outcomes. Author manuscript; available in PMC 2012 November 1.NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

and carbohydrates. Progress review of dietary patterns, strategies for dietary change, andguides for managing difficult situations were addressed with patients by the CHW.Patients were instructed to participate in a moderate-intensity home-based exercise program.The patient selected the mode of moderate-intensity physical activity and set realistic goals.Telephone contact was initiated by the CHW two weeks after inception of the program andonce a month until the sixth month to monitor progress, answer questions, and provideindividualized feedback and positive reinforcement.The intervention teams included NPs who were certified adult nurse practitioners withexperience in the delivery of primary care and CHWs with experience working withunderserved minority populations. The NPs completed additional continuing education inthe management of hypertension, hypercholesterolemia, and diabetes and the CHWs weretrained in the disease pathophysiology of CHD and diabetes and therapeutic lifestylemanagement approaches of nutrition and physical activity. Both NPs and CHWs weretrained in motivational interviewing behavior change techniques.Documenting the team’s adherence to protocols was important to assure interventionfidelity. Encounter forms for the NP and the CHW tracked the number, length, and contentof the encounters (such as counseling on diet, exercise, medications, smoking cessation, andadherence) to determine the delivered dose of the intervention. In addition there was aCOACH Program Intervention Quality Assurance (QA) Plan to assure adherence to studyintervention protocols and treatment algorithms to promote intervention integrity throughoutthe study. QA assessments were conducted on a quarterly basis. The QA assessmentincluded analysis of audio-tape recorded intervention sessions and interventiondocumentation in medical records. QA assessments were independently conducted by twoCOACH study investigators. The two independent reviewers discussed assessments andprovided feedback to interventionists to provide positive reinforcement and/or a plan foradditional training in a timely basisPatients and their providers in the EUC group received the results of baseline lipids, BP, andHbA1c along with the recommended goal levels and a pamphlet on controlling risk factorspublished by the American Heart Association. In addition, providers received copies of theAHA/ACC Guidelines for Secondary Prevention.30Outcome MeasuresThe primary outcomes were changes from baseline to one year in lipids, BP, HbA1c andpatients’ perceptions of the quality of their chronic illness care. The primary outcomes alsowere operationalized as meeting the goals for secondary prevention or experiencing aclinically significant change as follows: HbA1c < 7% or clinically significant decrease of ≥0.5%; systolic BP < 140 mm Hg or < 130 mm Hg if patient had diabetes or kidney diseaseor clinically significant decrease of ≥ 10 mm Hg; and LDL cholesterol level, < 100 mg/dLor < 130 if no CVD or diabetes or a clinically significant decrease of ≥ 20%. The chemistrylaboratory at Johns Hopkins performed all biochemical measures. Total cholesterol,triglycerides, and high-density lipoprotein cholesterol (HDL-C) were measured directly aftera 12 hour fast. Low-density lipoprotein cholesterol (LDL-C) was estimated using theFriedewald equation.31 In the event of triglyceride levels greater than 400 mg/dL, directmeasurement of LDL-C through ultracentrifugation methods was performed. In participantswith diabetes, HbA1c was measured using high pressure liquid chromatography. Bloodpressure was measured using the Omron Digital Blood Pressure Monitor HEM-907XLautomatic blood pressure device according to JNC VII guidelines, following five minutes ofquiet rest, in the right arm with the person seated in a chair with arm supported at heartlevel. The average of three blood pressures was recorded.Allen et al. Page 5Circ Cardiovasc Qual Outcomes. Author manuscript; available in PMC 2012 November 1.NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

The patient’s ratings of care received from their health care team was measured by thePatient Assessment of Chronic Illness Care (PACIC) Survey, a 20-item patient reportinstrument that assesses patient’s perceptions of the receipt of clinical services and actionsconsistent with quality care defined by the Chronic Care Model.32 The five subscales arePatient Activation; Delivery System/Decision Support; Goal Setting; Problem-solving/Contextual Counseling; and Follow-up/Coordination.Secondary outcomes included the lifestyle behaviors of dietary intake measured by theHabits and History Food Frequency Questionnaire, Block 2005.1,33, 34 and physical activityevaluated with the Stanford 7-Day Physical Activity Recall.35, 36 Quality of life wasmeasured by the 5 item EuroQol questionnaire37 and resource utilization and patients’ healthcare utilization data were collected to conduct a cost effectiveness analysis which will bereported separately.Statistical AnalysisThe sample size for this study was determined considering the effect sizes observed in theinvestigators’ preliminary work. Based on the calculations, a minimum of 450 participants(225 per group) were needed to detect clinically significant differences in changes in theprimary outcomes of BP, LDL-C, and HbA1c at one year to ensure 80% power at a 0.05significance level. This sample size accounted for an expected 25% attrition at the one-yearfollow-up yielding 180 participants per intervention.The data analysis for this paper was generated using SAS® version 9.2 for Windows.Statistical tests were used to study differences in baseline demographic, clinical, and riskfactor characteristics, with a t-test used for continuous variables and a chi-square test forcategorical variables. Similar statistical tests were used to compare baseline characteristicsfor subjects completing the study to those lost to follow-up for any reason.The primary outcomes are changes from baseline to one year in lipids, BP, HgA1c, and thepatient’s perception of chronic illness care. Analysis followed the intention- to- treat modelincluding all randomized participants in the analyses according to their original assignment.Participants who withdrew or did not complete the one year assessment were included in theanalysis. Missing data were imputed with multiple imputation. Multiple imputation is anadvanced statistical method for handling missing data.38 This computationally intensiveapproach uses multiple linear regression to predict missing values with observed data. Theprocedure is repeated with five iterations. Repeatedly imputing missing values allows forquantifying the uncertainty resulting from sampling error.Generalized linear mixed models (GLMM) using a random patient-level intercept modelwere used to build multilevel models comparing the effectiveness of the NP/CHWintervention with EUC on each outcome, controlling for the covariates of age, sex, race,body mass index, insurance status which were determined by univariate analyses to bepredictive of outcomes. Mixed models are the optimal statistical method to use with pre- andpost-intervention repeated measures data, as this modeling approach accounts for thecorrelated data structure.RESULTS The sample was predominantly female (71%) and Black (79%). A majority had at least ahigh school education; however, a majority had annual incomes less than $20,000 and fewerthan half had private health insurance. There were no significant differences insociodemographic and baseline measures between the two groups except for higher totalcholesterol and HbA1c levels in the NP/CHW intervention group compared to the EUCAllen et al. Page 6Circ Cardiovasc Qual Outcomes. Author manuscript; available in PMC 2012 November 1.NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

group (Table 1). We did not find statistically significant differential attrition between thetwo groups. Ninety four percent (n=467) completed the 1-year assessment with nodifferences between completers and noncompleters in baseline lipids, HbA1c, BP, age,education, race, or sex.A total of 84 percent of patients randomized to the intervention group completed an initialvisit, and 70 percent had at least four in-person visits with the nurse. Patients in theintervention group had a mean of 7 ± 3 in-person visits and 6 ± 5 telephone visits with theNP/CHW team. A comprehensive cost-effectiveness analysis will be published separately.At 12 months, patients in the intervention group had significantly greater overallimprovement in total cholesterol, LDL cholesterol, triglycerides, systolic and diastolic BP,HbA1c, and perceptions of the quality of their chronic illness care compared to patientsreceiving EUC (Table 2). The analyses using general linear mixed models controlled forage, sex, education, race, body mass index, insurance status, and an indicator of controlstatus at baseline. The estimated between group differences were clinically significant. Atthe 12 month follow-up, a significantly higher percentage of patients in the interventiongroup compared to the EUC group had values that reached guideline goals or showedclinically significant improvements in LDL cholesterol (EUC=58%; I=75%, p<0.001),systolic BP (EUC=74%; I=82%, p=0.018), and HbA1c (EUC=47%; I=60%, p=0.016).Patients’ assessments of their chronic illness care improved significantly from baseline toone year in the intervention group (Table 3). This increase was significantly greater than themodest increase in the EUC group. These significantly different changes were present forthe total score on the PACIC instrument as well as the five subscales of Patient Activation,Delivery System Design/Decision Support, Goal Setting, Problem Solving/ContextualCounseling, and Follow-up Coordination.Although there were greater changes in the recommended direction in the intervention groupcompared to EUC, there were no statistically significant differences in changes betweengroups in the level of physical activity, body mass index, total energy intake, saturated fat orsodium intake from baseline to one year of follow-up.DISCUSSIONThis study demonstrated that vulnerable patients with uncontrolled CVD risk factorsmanaged by a NP/CHW intervention team achieved significant improvement in their CVDrisk profiles. The improvements in the primary outcomes in this study compare favorablywith changes in other studies targeting improvement in clinical outcomes and quality of carein patients with CVD and/or diabetes. In a meta-analysis of 66 trials to improve theoutcomes of diabetes care, the HbA1c level decreased by a mean of 0.42% (95% CI, 0.29 to0.54)39 versus a mean of 0.50% (95% CI, 0.2 to 0.9) in this trial. In a recent review of 11studies of nurse-led interventions used to improve control of high blood pressure in peoplewith diabetes, meta-analysis showed greater reductions in blood pressure in favor of nurse-led interventions (systolic weighted mean difference −5.8 mmHg, 95% CI, −9.6 to −2.0;diastolic weighted mean difference −4.2 mmHg, 95% CI, −7.6 to −0.7).40 In a systematicreview of 44 trials, systolic BP decreased by a mean of 4.5 mm Hg (95% CI, 1.8 to 6.6)41versus a mean of 6.2 (95% CI, 2.1 to 10.2) in this trial. The changes in HbA1c, BP, total andLDL cholesterol and triglycerides in this study are clinically meaningful. On a populationlevel, they should lead to a meaningful decreases in macrovascular and microvasculardisease in people with diabetes42, 43 and decreases in events in people with CVD.1Patients in the intervention group rated the quality of care that they received for themanagement of their chronic conditions as increasing significantly more than patients whoAllen et al. Page 7Circ Cardiovasc Qual Outcomes. Author manuscript; available in PMC 2012 November 1.NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

received EUC. Whether this translates into greater satisfaction with care is unclear;however, we know that satisfaction with care predicts better self-care and more favorableoutcomes.44 The NP/CHW team enhanced patient self-care by encouraging self-monitoring,mutual goal-setting and decision-making, addressing barriers to improve adherence tomedications and appointment keeping, and making proactive contact with patients to assessprogress. These types of interventions are consistent with the strategies described by theChronic Care Model to improve the performance of health care systems.45The results of this trial support the potential for nurse-led patient-centered medical homes(PCMH) to improve the quality of care in high risk underserved populations. The concept ofa PCMH is receiving increased attention as a means to improve care and potentially reducecosts. The PCMH has its origins in care for children with chronic conditions.46 In March2007, the American Academy of Pediatrics, American Academy of Family Practice,American College of Physicians, and American Osteopathic Association published a jointstatement of principles calling for accessible, continuous, team-based care that focuses onthe whole person, with the PCMH taking responsibility for care coordination.47 In 2009, theAmerican College of Physicians endorsed the inclusion of NP-led practices to test differentPCMH models within the Medical Home Demonstration Project.48 As the costs of healthcare for chronic diseases continues to increase, NPs are in pivotal positions to address theneed for safe, effective, patient-centered, efficient, and equitable health care.49This study also provides evidence that a nurse-led team which includes CHWs is aneffective model of care. However, adoption and sustainability of this model of care willrequire financing mechanisms for CHWs. Funding, reimbursement and payment policies forCHWs must be established to ensure that CHW models are adopted in mainstream healthcare.14, 50The limitations of the COACH Trial include the fact that it was conducted in one federally-qualified community health system and used highly trained NPs and CHWs, which maylimit generalizability. Second, the recruitment and screening process resulted in theinclusion of a sample of predominately Black women. However this represents the majorityof patients seen in these and other similar community health clinics which increasesconfidence in the generalizability of findings to similar settings. Third, physicians hadpatients in both the intervention and EUC groups. This may have resulted in a change in thelevel of care provided to their patients in the EUC group as they received laboratory reportsat baseline and tended to become more vigilant with the assessment, treatment and follow-up for cardiovascular risk factor management. This may explain the improvements inclinical measures in the EUC group. Nevertheless, improvements in clinical outcomes andperceptions of the quality of care were significantly greater among patients in theintervention group compared to the EUC group. Finally, there was a higher attrition rate inthe intervention group (13%) as compared to the EUC group, 9%. However, the study waspowered to account for a dropout rate of 25%. The slightly differential dropout rate in theintervention group may be due to the increased commitment to participate in the interventiongroup, including more visits to the clinic resulting in more costs to the participant.CONCLUSIONSIn summary, an intervention delivered by a NP/CHW team using individualized treatmentregimens based on treat-to-target algorithms can be an effective approach to improve riskfactor status and perceptions of chronic illness care in high risk patients. The translation ofnew knowledge and efficacious interventions into the care of populations, particularly thoseat highest risk of multiple chronic diseases, disability and mortality, remains a nationalproblem. This study developed a partnership with Baltimore Medical Systems, specificallyAllen et al. Page 8Circ Cardiovasc Qual Outcomes. Author manuscript; available in PMC 2012 November 1.NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

to implement a program to reach this high risk population. Moreover, the design andintervention were developed in collaboration with the Community Health Centers within thisSystem, so that the program, if found efficacious, would be sustained for long termeffectiveness. Further analyses will evaluate the cost effectiveness of NP/CHW model.Further study is needed to determine if this translates into improved morbidity and mortalityfrom CVD.Supplementary MaterialRefer to Web version on PubMed Central for supplementary material.AcknowledgmentsWe gratefully thank the patients who participated in the program and the administration and staff of BaltimoreMedical Systems for their collaboration in the design and implementation of the program. We also thank MargaretDenny for her assistance in the preparation of this manuscript.Funding Sources: This study was supported by the National Heart Lung and Blood Institute, National Institutes ofHealth grant # R01HL082638.References1. Roger VL, Go AS, Lloyd-Jones DM, Adams RJ, Berry JD, Brown TM, Carnethon MR, Dai S, DeSimone G, Ford ES, Fox CS, Fullerton HJ, Gillespie C, Greenlund KJ, Hailpern SM, Heit JA, HoPM, Howard VJ, Kissela BM, Kittner SJ, Lackland DT, Lichtman JH, Lisabeth LD, Makuc DM,Marcus GM, Marelli A, Matchar DB, McDermott MM, Meigs JB, Moy CS, Mozaffarian D,Mussolino ME, Nichol G, Paynter NP, Rosamond WD, Sorlie PD, Stafford RS, Turan TN, TurnerMB, Wong ND, Wylie-Rosett J. Heart disease and stroke statistics--2011 update: a report from theAmerican Heart Association. Circulation. 2011; 123:e18–e209. [PubMed: 21160056]2. Allen JK. Cholesterol management: an opportunity for nurse case managers. J Cardiovasc Nurs.2000; 14:50–58. [PubMed: 10653276]3. Allen JK, Dennison CR. Randomized trials of nursing interventions for secondary prevention inpatients with coronary artery disease and heart failure: systematic review. J Cardiovasc Nurs. 2010;25:207–220. [PubMed: 20386243]4. Becker DM, Raqueno JV, Yook RM, Kral BG, Blumenthal RS, Moy TF, Bezirdjian PJ, Becker LC.Nurse-mediated cholesterol management compared with enhanced primary care in siblings ofindividuals with premature coronary disease. Arch Intern Med. 1998; 158:1533–1539. [PubMed:9679794]5. DeBusk RF, Miller NH, Superko HR, Dennis CA, Thomas RJ, Lew HT, Berger WE III, Heller RS,Rompf J, Gee D. A case-management system for coronary risk factor modification after acutemyocardial infarction. Ann Intern Med. 1994; 120:721–729. [PubMed: 8147544]6. Allen JK, Blumenthal RS, Margolis S, Young DR, Miller ER III, Kelly K. Nurse case managementof hypercholesterolemia in patients with coronary heart disease: results of a randomized clinicaltrial. Am Heart J. 2002; 144:678–686. [PubMed: 12360165]7. Allison TG, Squires RW, Johnson BD, Gau GT. Achieving National Cholesterol Education Programgoals for low-density lipoprotein cholesterol in cardiac patients: importance of diet, exercise, weightcontrol, and drug therapy. Mayo Clin Proc. 1999; 74:466–473. [PubMed: 10319076]8. Brown AS, Cofer LA. Lipid management in a private cardiology practice (the Midwest Heartexperience). Am J Cardiol. 2000; 85:18A–22A.9. Hill MN, Han HR, Dennison CR, Kim MT, Roary MC, Blumenthal RS, Bone LR, Levine DM, PostWS. Hypertension care and control in underserved urban African American men: behavioral andphysiologic outcomes at 36 months. Am J Hypertens. 2003; 16:906–913. [PubMed: 14573327]10. Shaffer J, Wexler LF. Reducing low-density lipoprotein cholesterol levels in an ambulatory caresystem. Results of a multidisciplinary collaborative practice lipid clinic compared with traditionalphysician-based care. Arch Intern Med. 1995; 155:2330–2335. [PubMed: 7487258]Allen et al. Page 9Circ Cardiovasc Qual Outcomes. Author manuscript; available in PMC 2012 November 1.NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

11. Sikka R, Waters J, Moore W, Sutton DR, Herman WH, Aubert RE. Renal assessment practices andthe effect of nurse case management of health maintenance organization patients with diabetes.Diabetes Care. 1999; 22:1–6. [PubMed: 10333895]12. Ridgeway NA, Harvill DR, Harvill LM, Falin TM, Forester GM, Gose OD. Improved control oftype 2 diabetes mellitus: a practical education/behavior modification program in a primary careclinic. South Med J. 1999; 92:667–672. [PubMed: 10414474]13. Kirkman MS, Weinberger M, Landsman PB, Samsa GP, Shortliffe EA, Simel DL, Feussner JR. Atelephone-delivered intervention for patients with NIDDM. Effect on coronary risk factors.Diabetes Care. 1994; 17:840–846. [PubMed: 7956628]14. Brownstein JN, Bone LR, Dennison CR, Hill MN, Levine DM. Community health workers asinterventionists in research and practice for the prevention of heart disease and stroke. AmericanJournal of Preventative Medicine. 2005; 29:128–133.15. Fedder DO, Chang RJ, Curry S, Nichols G. The effectiveness of a community health workeroutreach program on healthcare utilization of west Baltimore City Medicaid patients with diabetes,with or without hypertension. Ethn Dis. 2003; 13:22–27. [PubMed: 12723008]16. Batts ML, Gary TL, Huss K, Hill MN, Bone L, Brancati FL. Patient priorities and needs fordiabetes care among urban African American adults. Diabetes Educ. 2001; 27:405–412. [PubMed:11912801]17. Hill MN, Becker DM. Roles of nurses and health workers in cardiovascular health promotion. AmJ Med Sci. 1995; 310(suppl 1):S123–S126. [PubMed: 7503115]18. Hill MN, Bone LR, Hilton SC, Roary MC, Kelen GD, Levine DM. A clinical trial to improve highblood pressure care in young urban black men: recruitment, follow-up, and outcomes. Am JHypertens. 1999; 12:548–554. [PubMed: 10371363]19. Bird JA, Otero-Sabogal R, Ngoc-The H, McPhee SJ. Tailoring lay health worker interventions fordiverse cultures: Lessons learned from Vietnamese and Latina communities. Health Educ &Behav. 1996; 23:S105–S122.20. Brownstein, JN.; Rosenthal, EL. The final report of the National Advisor Study: Weaving theFuture. Tucson, AZ: University of Arizona; 1998. The challenge of evaluating CHA services; p.50-74.21. Love, MB.; Gardner, K. The emerging role of the community health worker in California: resultsof a statewide university survey and San Francisco Bay Area focus group on the CommunityHealth Workers in California’s public health system. San Francisco, CA: California Department ofHealth Services, Health Promotion Section; 1992.22. Stone, RI.; Wiener, J. Who will care for us? Addressing the long-term care workforce crisis.Washington, DC: The Urban Institute and the American Association of Homes and Services forthe Aging; 2001.23. Koch, E.; Thompson, A.; Keegan, P. Community health workers: a leadership brief on preventivehealth programs. Washington, DC: The CivicHealth Institute at Codman Square Health Center andthe Harrison Institute for Public Law at Georgetown University Law Center and the Center forPolicy Alternatives; 1998.24. Levine DM, Bone LR, Hill MN, Stallings R, Gelber AC, Barker A, Harris EC, Zeger SL, Felix-Aaron KL, Clark JM. The effectiveness of a community/academic health center partnership indecreasing the level of blood pressure in an urban African-American population. Ethn Dis. 2003;13:354–361. [PubMed: 12894960]25. Witmer A, Seifer SD, Finocchio L, Leslie J, O’Neil EH. Community health workers: integralmembers of the health care work force. Am J Public Health. 1995; 85:1055–1058. [PubMed:7625495]26. Hill MN, Bone LR, Butz AM. Enhancing the role of community-health workers in research. ImageJ Nurs Sch. 1996; 28:221–226. [PubMed: 8854543]27. Allen JK, Himmelfarb CR, Szanton SL, Bone L, Hill MN, Levine DM. COACH trial: Arandomized controlled trial of nurse practitioner/community health worker cardiovascular diseaserisk reduction in urban community health centers: Rationale and design. Contemp Clin Trials.2011; 32:403–411. [PubMed: 21241828]Allen et al. Page 10Circ Cardiovasc Qual Outcomes. Author manuscript; available in PMC 2012 November 1.NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

28. Green L, Daniel M, Novack L. Partnerships and coalitions for community-based research. PublicHealth Reports. 2001; 116(suppl 1):20–31. [PubMed: 11889272]29. American Diabetes Association. Nutrition recommendations and principles for people withdiabetes mellitus. Tenn Med. 2000; 93:430–433. [PubMed: 11077891]30. Smith SC Jr, Allen J, Blair SN, Bonow RO, Brass LM, Fonarow GC, Grundy SM, Hiratzka L,Jones D, Krumholz HM, Mosca L, Pearson T, Pfeffer MA, Taubert KA. AHA/ACC guidelines forsecondary prevention for patients with coronary and other atherosclerotic vascular disease: 2006update endorsed by the National Heart, Lung, and Blood Institute. J Am Coll Cardiol. 2006;47:2130–2139. [PubMed: 16697342]31. Friedewald WT, Levy RI, Fredrickson DS. Estimation of the concentration of low-densitylipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin Chem. 1972;18:499–502. [PubMed: 4337382]32. Glasgow RE, Wagner EH, Schaefer J, Mahoney LD, Reid RJ, Greene SM. Development andvalidation of the Patient Assessment of Chronic Illness Care (PACIC). Med Care. 2005; 43:436–444. [PubMed: 15838407]33. Block G, Hartman AM, Dresser CM, Carroll MD, Gannon J, Gardner L. A data-based approach todiet questionnaire design and testing. Am J Epidemiol. 1986; 124:453–469. [PubMed: 3740045]34. Block G, Thompson FE, Hartman AM, Larkin FA, Guire KE. Comparison of two dietaryquestionnaires validated against multiple dietary records collected during a 1-year period. J AmDiet Assoc. 1992; 92:686–693. [PubMed: 1607564]35. Blair SN, Haskell WL, Ho P, Paffenbarger RS Jr, Vranizan KM, Farquhar JW, Wood PD.Assessment of habitual physical activity by a seven-day recall in a community survey andcontrolled experiments. Am J Epidemiol. 1985; 122:794–804. [PubMed: 3876763]36. Sallis JF, Haskell WL, Wood PD, Fortmann SP, Rogers T, Blair SN, Paffenbarger RS Jr. Physicalactivity assessment methodology in the Five-City Project. Am J Epidemiol. 1985; 121:91–106.[PubMed: 3964995]37. The EuroQol Group. EuroQol--a new facility for the measurement of health-related quality of life.Health Policy. 1990; 16:199–208. [PubMed: 10109801]38. He Y. Missing data analysis using multiple imputation: getting to the heart of the matter. CircCardiovasc Qual Outcomes. 2010; 3:98–105. [PubMed: 20123676]39. Shojania KG, Ranji SR, McDonald KM, Grimshaw JM, Sundaram V, Rushakoff RJ, Owens DK.Effects of quality improvement strategies for type 2 diabetes on glycemic control: a meta-regression analysis. JAMA. 2006; 296:427–440. [PubMed: 16868301]40. Clark CE, Smith LF, Taylor RS, Campbell JL. Nurse-led interventions used to improve control ofhigh blood pressure in people with diabetes: a systematic review and meta-analysis. Diabet Med.2011; 28:250–261. [PubMed: 21309833]41. Walsh JM, McDonald KM, Shojania KG, Sundaram V, Nayak S, Lewis R, Owens DK, GoldsteinMK. Quality improvement strategies for hypertension management: a systematic review. MedCare. 2006; 44:646–657. [PubMed: 16799359]42. Skyle JS, Bergenstal R, Bonow RO, Buse J, Deedwania P, Gale EA, Howard BV, Kirkman MS,Kosiborod M, Reaven P, Sherwin RS. Intensive glycemic control and the prevention ofcardiovascular events: implications of the ACCORD, ADVANCE, and VA Diabetes Trials: aposition statement of the American Diabetes Association and a Scientific Statement of theAmerican College of Cardiology Foundation and the American Heart Association. J Am CollCardiol. 2009; 53:298–304. [PubMed: 19147051]43. Adler AI, Stratton IM, Neil HA, Yudkin JS, Matthews DR, Cull CA, Wright AD, Turner RC,Holman RR. Association of systolic blood pressure with macrovascular and microvascularcomplications of type 2 diabetes (UKPDS 36): prospective observational study. BMJ. 2000;321:412–419. [PubMed: 10938049]44. Sherbourne CD, Hays RD, Ordway L, DiMatteo MR, Kravitz RL. Antecedents of adherence tomedical recommendations: results from the Medical Outcomes Study. J Behav Med. 1992;15:447–468. [PubMed: 1447757]45. Wagner EH, Austin BT, Von Korff M. Organizing care for patients with chronic illness. MilbankQ. 1996; 74:511–544. [PubMed: 8941260]Allen et al. Page 11Circ Cardiovasc Qual Outcomes. Author manuscript; available in PMC 2012 November 1.NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

46. Cooley WC. Providing a primary care medical home for children and youth with cerebral palsy.Pediatrics. 2004; 114:1106–1113. [PubMed: 15466117]47. American Academy of Family Physicians, American Academy of Pediatrics, American College ofPhysicians, American Osteopathic Association. [Accessed January 28, 2011.] Guidelines forPatient-Centered Medical Home (PCMH) Recognition and Accreditation Programs. Guidelines forPatient-Centered Medical Home (PCMH) Recognition and Accreditation Programs. 2011.http://www.acponline.org/running_practice/pcmh/understanding/guidelines_pcmh.pdf48. American College of Physicians. Nurse Practitioners in Primary Care. Philadelphia, PA: AmericanCollege of Physicians; 2009. Policy Monograph. (Available from American College of Physicans,190 N. Independence Mall West, Philadelphia, PA 19106.)49. Parsons Schram A. Medical Home and the Nurse Practitioner: A Policy Analysis. The Journal forNurse Practitioners. 2010; 6:132–139.50. Dower, C.; Know, M.; Lindler, V.; O’Neil, E. Advancing Community Health Worker Practice andUtilization: The Focus on Financing. San Francisco, CA: National Fund for Medical Education;2006.Allen et al. Page 12Circ Cardiovasc Qual Outcomes. Author manuscript; available in PMC 2012 November 1.NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

Figure 1.CONSORT diagram of COACH TrialAllen et al. Page 13Circ Cardiovasc Qual Outcomes. Author manuscript; available in PMC 2012 November 1.NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author ManuscriptAllen et al. Page 14Table 1Baseline sample characteristicsCharacteristic Intervention (n = 261) Usual Care (n = 264) PAge, y, mean (SD)* 54.3 (12.0) 54.7 (11.5) 0.692Female, n (%) 187 (71.7) 187 (70.8) 0.837Marital status, n (%) 0.591  Single 61 (23.4) 69 (26.1)  Married 86 (33.0) 80 (30.3)  Separated 19 (7.3) 28 (10.6)  Widowed 39 (14.9) 37 (14.0)  Divorced 56 (21.5) 50 (18.9)Race, n (%) 0.946  Non-black 54 (20.7) 54 (20.5)  Black 207 (79.3) 210 (79.6)Education, n (%) 0.051  <High School 76 (29.1) 94 (35.6)  High school/GED 118 (45.2) 92 (34.9)  Some college 67 (25.7) 78 (29.6)Employment status, n (%) 0.318  Employed 110 (42.2) 100 (37.9)  Not employed 151 (57.9) 164 (62.1)Type of insurance, n (%) 0.403  Private 112 (42.9) 105 (39.8)  Medicare and/or Medicaid 106 (40.6) 101 (38.3)  Uninsured 43 (16.5) 55 (20.8)  Unknown 0 (0) 3 (1.1)Annual income, n (%) 0.223  <$20,000 137 (52.5) 149 (56.4)  ≥$20,000 120 (46.0) 105 (39.8)  Unknown 4 (1.5) 10 (3.8)Comorbidity score, mean (SD) 1.6 (1.3) 1.8)1.4) 0.193Diastolic Blood Pressure, mean (SD) 83.1 (12.6) 82.3 (13.0) 0.442Systolic Blood Pressure, mean (SD) 139.7 (23.8) 138.7 (19.9) 0.587Total Cholesterol, mean (SD) 199.7 (46.0) 191.3 (45.0) 0.036LDL-C†, mean (SD) 121.6 (40.0) 116.3 (40.5) 0.132HDL-C‡, mean (SD) 50.8 (14.7) 50.9 (13.6) 0.92Triglycerides, median (IQR)§ 113 (85) 105 (76) 0.220Hemoglobin A1c, mean (SD) 8.9 (2.2) 8.3 (1.9) 0.006PACIC||, mean (SD) 1.5 (0.9) 1.6 (0.9) 0.883*SD = standard deviation;†LDL-C = low density lipoprotein cholesterol;Circ Cardiovasc Qual Outcomes. Author manuscript; available in PMC 2012 November 1.

NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author ManuscriptAllen et al. Page 15‡HDL-C = high density lipoprotein cholesterol;§IRQ = interquartile range;||PACIC = Patient Assessment of Chronic Illness CareCirc Cardiovasc Qual Outcomes. Author manuscript; available in PMC 2012 November 1.

NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author ManuscriptAllen et al. Page 16Table 2Changes in Primary Outcomes by GroupOutcome Intervention Group (n=261) Usual Care Group (n=264) P Value* Estimated Between Group Difference (95% CI)Change ChangeSystolic BP, mmHg  Baseline 139.7±23.8 8.9±25.1 138.7±19.9 2.7±22.0 0.003 −6.2 (−10.2, −2.1)  One year 130.8±20.7 135.9±20.5Diastolic BP, mmHg  Baseline 83.0±12.7 5.6±13.6 82.3±13.0 2.6±12.1 0.013 −3.1 (−5.3, −0.9)  One year 77.4±12.5 79.7±12.6Total Cholesterol†, mg/dL  Baseline 199.7±46.0 27.0±50.7 191.3±45.0 7.3±44.5 <0.001 −19.7 (−27.9, −11.5)  One year 172.7±44.5 184.1±41.9LDL†, mg/dL  Baseline 121.6±40.0 21.6±44.0 116.3±40.5 5.7±38.9 <0.001 −15.9 (−23.0, −8.8)  One year 100.1±39.2 110.6±36.8Triglycerides †‡, mg/dL  Baseline 138.1±93.4 20.1±88.3 126.8±71.5 3.8±65.6 0.013 −16.3 (−29.6, −3.0)  One year 121.3±81.6 123.1±72.2HDL†, mg/dL  Baseline 50.8±14.7 1.4±10.5 50.9±13.6 1.0±10.9 0.497 −0.4 (−2.2, 1.4)  One year 49.4±13.5 49.9±12.9Hemoglobin A1c, %  Baseline 8.9±2.2 0.6±2.3 8.3±1.9 0.1±1.8 0.034 −0.5 (−0.9, −0.2)  One year 8.3±2.2 8.2±2.1PACIC  Baseline 1.6±0.9 −1.4±1.0 1.6±1.0 −0.2±0.9 <0.001 1.2 (1.0, 1.3)  One year 2.9±0.9 1.8±1.0*Intention to treat analysis using general linear mixed model with group, time, groupxtime effects and covariates age, sex, race, education, body mass index, insurance and an indicator of in-control forclinical outcome at baseline.Circ Cardiovasc Qual Outcomes. Author manuscript; available in PMC 2012 November 1.

NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author ManuscriptAllen et al. Page 17†To convert total cholesterol, LDL cholesterol, and HDL cholesterol to mmol/L, multiply by 0.0259; to convert triglycerides to mmol/L, multiply by 0.0113.‡Computed using a trimmed mean of triglyceride values less than 1000. P-value based on log-triglycerides due to non-normality.CI indicates Confidence Interval; BP Blood Pressure; LDL Low-density lipoprotein; HDL High-density lipoprotein; PACIC Patient Assessment of Chronic Illness Care.Circ Cardiovasc Qual Outcomes. Author manuscript; available in PMC 2012 November 1.

NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author ManuscriptAllen et al. Page 18Table 3Changes in Patient Assessment of Chronic Illness Care (PACIC) Scores by GroupOutcome Intervention Group (n=261) Usual Care Group (n=264) P Value*Change ChangeTotal PACIC  Baseline 1.6±0.9 −1.4±1.0 1.6±1.0 −0.2±0.9 <0.001  One year 2.9±0.9 1.8±1.0Patient Activation Subscale  Baseline 1.5±1.2 −1.3±1.5 1.4±1.2 −0.2±1.3 <0.001  One year 2.7±1.2 1.6±1.2Delivery System Subscale  Baseline 2.2±1.1 −1.3±1.3 2.2±1.1 −0.1±1.2 <0.001  One year 3.5±0.8 2.4±1.1Goal Setting Subscale  Baseline 1.4±1.1 −1.5±1.3 1.5±1.1 −0.2±1.1 <0.001  One year 3.0±0.9 1.7±1.2Problem Solving Subscale  Baseline 1.9±1.2 −1.3±1.3 1.8±1.3 −0.2±1.3 <0.001  One year 3.2±1.0 2.0±1.3Follow-Up Subscale  Baseline 1.1±1.0 −1.2±1.3 1.1±1.1 −0.2±1.0 <0.001  One year 2.4±1.0 1.4±1.1Intention to treat analysis using general linear mixed model with group, time, group × time effectsCirc Cardiovasc Qual Outcomes. Author manuscript; available in PMC 2012 November 1.