Reducing Diabetes Disparities Through the Implementation of a Community Health Worker–Led Diabetes Self-Management Education Program

Fam Community Health Vol. 35, No. 2, pp. 161–171 Copyright C 2012 Wolters Kluwer Health | Lippincott Williams & Wilkins Reducing Diabetes Disparities Through the Implementation of a Community Health Worker–Led Diabetes Self-Management Education Program

James W. Walton, DO, MBA; Christine A. Snead, BSN, RN; Ashley W. Collinsworth, MPH; Kathryn L. Schmidt, MPH

Disparities in prevalence of type 2 diabetes and complications in underserved populations have been linked to poor quality of care including lack of access to diabetes management programs. Interventions utilizing community health workers (CHWs) to assist with diabetes management have demonstrated improvements in patient outcomes. Use of CHWs may be an effective model for providing care coordination and reducing disparities, but there is limited knowledge on how to implement this model on a large scale. This article describes how an integrated health care system implemented a CHW-led diabetes self-management education program targeting Hispanic patients and reports lessons learned from the ﬁrst 18 months of operation. Key words: care coordination, community health worker, diabetes management, vulnerable populations

YPE 2 DIABETES is one of the most preva- T lent chronic diseases in the United States, and the burden of this epidemic falls dis- Author Affiliations: Office of Health Equity proportionately on minorities and persons of (Dr Walton), Baylor Community Care (Ms Snead), low socioeconomic status.1 The prevalence and Institute for Health Care Research and Improvement (Mss Collinsworth and Schmidt), Baylor of diabetes in adults in Texas is higher than Health Care System, Dallas, Texas. the national average and is much higher in The authors thank Claudia Chavira, BS, CHW; Hispanics (12.3%) than in white non- Magdalena Lopez, CMA, CHW; Martha Lopez, CCMA, Hispanics (8.5%).2 Despite major advances in CHW; Miriam Lopez, MA, CHW; and Sayra Rojas, MA, CHW, for their contributions to this project and for medical care, technology, and services, His- making the implementation of the Diabetes Equity panics with diabetes continue to experience Project a success. They also thank their community a 50% to 100% higher burden of diabetes- partners Project Access Dallas, Baylor Family Medicine at Worth Street, CitySquare, Healing Hands Ministries, related illness and mortality than non- 3 HOPE Clinic, and Irving Interfaith Clinic. Hispanics. Common comorbidities include This program/initiative was supported by a grant from diabetic retinopathy, lower extremity amputa- The Merck Company Foundation through its Merck tion, and early-stage kidney disease.1,3,4 These Alliance to Reduce Disparities in Diabetes program. complications generally can be avoided with The authors declare no conflicts of interest. proper management of diabetes. However, in- Correspondence: James W. Walton, DO, MBA, Office of dividuals of Hispanic descent are less likely to Health Equity, Baylor Health Care System, Dallas, TX 75246, ([email protected]). receive recommended processes of care for patients with diabetes.4,5 Hispanics are also DOI: 10.1097/FCH.0b013e31824651d3 161

Copyright © 2012 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. 162 FAMILY &COMMUNITY HEALTH/APRIL–JUNE 2012 less likely to accomplish treatment goals such coach patients on effective communication as glycemic control and lowering of choles- with health care providers, explain concepts terol levels and blood pressure.4,6 and answer questions, and provide emotional Observed disparities in diabetes health out- support and a level of care coordination. In ad- comes have been linked to lack of access dition to providing improved care for patients, to preventive care and diabetes management CHWs can reduce the burden on other med- programs as well as poor quality of care.7 ical providers and reduce the cost of health Hispanics report experiencing more barriers care delivery by supporting patient needs that to diabetes care and self-management than do not require the expertise of a clinician17 non-Hispanic white adults and are less likely to and helping patients avoid unnecessary hospi- have access to primary care and medical sup- talizations and other expensive forms of acute plies such as glucometers and testing strips.8 care.18 Although there is increasing evidence Lack of trust in providers and difficulties in regarding the effectiveness of CHW-led inter- communication with clinicians are common ventions on improving diabetes-related health problems cited by this population.8-11 As a outcomes, few studies have described the de- result of these trust and communication bar- velopment and implementation of such pro- riers, patients who do have access to care do grams in detail and their potential contribu- not receive the information they need from tion to the health care delivery system.19-21 visits with health care providers. Hispanic pa- Baylor Health Care System (BHCS) in tients who have a diagnosis of diabetes of- Dallas, Texas, created the Diabetes Equity ten leave physician appointments without an Project (DEP) with funding from a Merck understanding of hemoglobin A1c (HbA1c), Company Foundation grant, with the goal of the risks associated with diabetes, or how to reducing observed disparities in diabetes care manage their disease. Consequently, these pa- and diabetes outcomes in the predominantly tients do not know how to make the lifestyle Hispanic, medically underserved communi- changes needed to achieve control over their ties surrounding BHCS hospitals. The DEP diabetes.8,12 In addition, cultural attitudes represents a unique partnership between a can negatively affect diabetes management. private, nonprofit health care system and com- Hispanic patients are more likely to have a munity clinics. The program features a spe- fatalistic attitude about diabetes and other cially trained, bilingual CHW (diabetes health chronic conditions, believing that there promoter) who delivers a culturally relevant is nothing they can do to control their diabetes education curriculum targeting bar- disease.13,14 riers to diabetes management that Hispan- A systematic review of interventions to ics commonly experience, including lack of improve diabetes care in socially disadvan- knowledge about diabetes, lack of social sup- taged populations concluded that culturally port, poor diets, insufficient physical activity, tailoring interventions and using community and limited access to care. This article de- health workers (CHWs) resulted in better scribes the development and implementation diabetes-related outcomes in Hispanic and of the DEP within a large health care delivery African American populations.15 Community system and reports preliminary findings and health workers can provide time-intensive lessons learned over the first 18 months of and community-based diabetes self-training the program’s operation. and can serve as a bridge between patients and providers.11,16 Patients are more likely to trust a peer from a similar background METHODS who understands the patient’s culture.16 This level of trust allows patients and CHWs to Overview work together to improve diabetes-specific The BHCS Office of Health Equity (OHE) behaviors.4 Community health workers can developed the DEP and facilitated the

Copyright © 2012 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. Implementation of a CHW-Led Diabetes Self-Management Education Program 163 implementation of the program within 5 part- tion). Community health workers assess self- nering community-based charity clinics. Dur- management behaviors and facilitate goal set- ing the 6-month project planning phase, a ting at each visit. The educational intervention program manager was hired to recruit and is grounded in the Adult Learning Theory of manage the CHWs and to implement strate- Knowles24, which states that effective adult gic collaborations with the 5 clinics selected learning should be self-directed, experience- as DEP sites. These clinics were selected on based, goal-oriented, practical and problem- the basis of geographic location, size of pa- centered, relevancy-oriented, and respect- tient panel, and ability to provide support- oriented. ing clinical and care coordination services. All clinics were medical home partners in Project Hiring of community health workers Access Dallas (PAD), a countywide network of We created a CHW position within the providers and care coordination services for BHCS OHE with the official title “diabetes low-income, disadvantaged residents lacking health promoter” to acknowledge the spe- health insurance. PAD’s menu of support ser- cific role this employee would have in dia- vices included transportation, medication as- betes management. Position requirements in- sistance, specialty physician care, health nav- cluded a high school diploma and fluency igation, and access to low-cost medications. in Spanish, with a medical assistant back- The DEP leadership team developed the DEP ground strongly preferred. The position was patient education curriculum on the basis of posted on the BHCS Web site and other Web- the evidence-based Community Diabetes Ed- based recruiting sites and was also shared ucation (CoDETM) program.22 A diabetes reg- with clinic administrators to identify inter- istry, DiaWEB, was created by a local vendor ested parties among their staff or volunteer to allow CHWs to collect data at patient vis- base. Candidates were interviewed by the pro- its and monitor patient progress over time. gram manager and a medical assistant who Institutional review board approval for the had conducted self-management education as conduct of the project was obtained from the part of the CoDETM program. Candidates were BHCS. ranked on the behavioral traits of compassion, Development of program curriculum communication, self-motivation, capacity to learn, integrity, teamwork, and quality. The diabetes education curriculum for the DEP was adapted from CoDETM,apilot program that was implemented in a Dallas Training of community health workers clinic serving a largely uninsured Mexican Once hired, CHWs were required to com- American population.22 Patients who partic- plete the 160-hour Texas Community Health ipated in CoDETM for 12 months experi- Worker Training and Certification Program 22,23 enced a significant reduction in HbA1c. and an additional 50 hours of instruction per- The DEP curriculum consists of (1) 7 one- taining to diabetes management and the DEP hour, one-on-one educational sessions de- protocol. The Texas CHW curriculum focused signed to teach participants about diabetes on skill development in the following ar- and the importance of blood glucose con- eas: communication, interpersonal skills, ser- trol, medication adherence, diet, and exer- vice coordination, capacity building, health cise; (2) routine assessments of HbA1c, blood advocacy, teaching, organizational skills, and pressure, weight, and foot condition (visual chronic diseases. A state-certified CHW in- and monofilament assessment); and (3) rou- structor from the Texas Public Health Train- tine assessments to ensure delivery of pre- ing Center at the University of North Texas ventive services (dilated eye examination, Health Science Center conducted the train- dental examination, influenza and pneumo- ing as a series of twenty 8-hour workshop- nia vaccines, and provider foot examina- style sessions. The 50 hours of additional

Copyright © 2012 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. 164 FAMILY &COMMUNITY HEALTH/APRIL–JUNE 2012 instruction included 12 hours of American DiaWEB containing key outcome and process Diabetes Association (ADA)-recognized group measures for the patient, which was faxed or self-management training at the BHCS Ruth scanned to the patient’s primary care provider Collins Diabetes Center, 8 hours of didac- for inclusion in the patient’s medical record. tic instruction provided by the consulting endocrinologist for the project and certified diabetes educators, and the 6-hour Ameri- Deployment of the Diabetes Equity can Association of Diabetes Educators (AADE) Project intervention Health Care Technician online course. The The DEP intervention was deployed in remaining 24 hours of instruction included 5 collaborating charity clinics after physical BHCS and DEP orientation, protocol discus- space for the DEP program was established, sion, team-building exercises, clinical skills DiaWEB was installed, and the CHWs com- development, and DiaWEB software training. pleted training. Each CHW was matched with The CHWs completed a posttest to assess a community charity clinic on the basis of knowledge of the Texas CHW curriculum geographic location and “fit” with the clinic as well as a posttest developed by a BHCS- culture. Patients were referred to the DEP by certified diabetes educator to assess knowl- providers at the 5 participating clinics and by edge gained pertaining to diabetes manage- clinicians at BHCS facilities following emer- ment and the DEP protocol. gency department visits and hospitalizations related to uncontrolled diabetes. To participate in the DEP, patients had to be 18 years Development of DiaWEB or older with a diagnosis of type 2 diabetes The BHCS OHE worked with a local ven- and be uninsured or underinsured. Although dor, Chiron Data Systems, to create a custom the program targeted Hispanic patients, all pa- HIPAA-compliant Web-based diabetes man- tients who met the inclusion criteria were el- agement system (DiaWEB) to enable CHWs igible to participate regardless of ethnicity or and clinicians to record and monitor patient race. Community health workers explained participation in the program and clinical indi- the program in detail before obtaining con- cators associated with diabetes management. sent from patients who agreed to participate. The CHWs used DiaWEB’s intuitively de- During the first 60- to 90-minute DEP visit, signed Web interface for the input of patient the CHW assessed baseline values of HbA1c, demographic information such as sex, ethnic- blood pressure, BMI, foot condition, and self- ity, primary language, grade level, and literacy management behaviors, and presented the level and clinical indicators such as patient first diabetes education module. The CHW blood pressure, body mass index (BMI), waist also counseled the patient about social bar- circumference, foot examination results, eye riers, medication adherence, and diabetes examination results, and receipt of influenza preventive services at each visit. Depending and pneumonia vaccines. The registry also al- upon urgency, communication with the pa- lowed for the documentation of patient medi- tient’s primary care physician about the visit cations, patient clinical laboratory results, and occurred within 24 hours. A 1-page, user- patient symptoms. In addition, DiaWEB had friendly visit summary was faxed or scanned customizable fields to allow for tracking of to the primary care provider for “routine” vis- various outcome measures such as patient its. If the patient was symptomatic or had scores on diabetes knowledge tests, quality- critical blood glucose or blood pressure mea- of-life assessments, and patient satisfaction surements as defined by program protocol, surveys; patient-reported level of physical ac- the physician was contacted immediately and tivity; and patient-reported emergency depart- patient instructions were relayed. The CHW ment visits and hospital admissions. After each scheduled follow-up visits for the patient, in- patient visit, CHWs generated a report from cluding 2 additional 60-minute visits over an

8-week period to continue the delivery of the CHWs were hired and became certified CHWs DEP curriculum and 4 follow-up visits lasting after completing the 160 hours of Texas CHW 30 to 60 minutes at 3, 6, 9, and 12 months Training and Certification. In addition, all 5 for assessment of HbA1c,bloodpressure,and CHWs completed the 50 hours of additional BMI. The CHW provided the patient with training in diabetes management and DEP im- feedback and coaching at all follow-up visits. plementation and passed a posttest on dia- Data from each DEP patient encounter were betes knowledge with an average score of recorded in the DiaWEB diabetes registry by 96.8%. Infrastructure for the DEP, including the CHW. creation of DEP workspace and planning for Community health workers were managed the integration of the DEP into clinic work- by the DEP registered nurse manager and re- flow, was established. DiaWEB customization ceived clinical support from the clinic primary and testing was completed, and the program care physicians and nurse practitioners. They was installed on DEP computers. also received 35 hours of continuing educa- Enrollment in the DEP began in the 5 clin- tion delivered by certified diabetes educators, ics at the end of September 2009. Within the physicians, a dietician, and a CHW certified first 18 months of rolling enrollment, 806 pa- trainer. Topics included diabetes complica- tients enrolled in the program. The majority tions, relevant evidence-based standards (ADA of patients were female (60%), between 40 and AADE-7), behavior change and motiva- and 59 years of age (64%), and Hispanic (70%) tional interviewing, and setting boundaries. (Table 1). Participants took an average of 1.1 years to complete the program curriculum, Statistical methods used for preliminary and an attrition rate of 15% was calculated for analysis of year 1 results the first 18 months. The mean HbA1c value for We performed a preliminary statistical anal- patients at baseline was 8.7%. A preliminary ysis on a limited set of key outcome variables analysis of the year 1 results revealed a statis- for patients who enrolled in the DEP during tically significant change in mean HbA1c levels the first 18 months of operation to evaluate 1 year post-baseline, where the mean HbA1c the initial impact of the program and to de- value dropped to 7.4%. Results from the anal- termine whether programmatic changes were ysis of covariance are described in Table 2. needed to help patients achieve desired out- Baseline HbA1c levels were significantly comes. All patients who enrolled in the program from September 2009 to March 2011 who had at least 2 visits with the CHW were Table 1. Demographics of Patients Enrolled included in the analysis. The outcome mod- in First 18 Months eled in the analysis was change in mean HbA1c levels from baseline. An analysis of covariance Demographics % model was used to test whether certain factors including program enrollment date, visit Gender number, clinic, age, ethnicity, sex, and base- Male 40 Female 60 line HbA level were significantly associated 1c Age, y with the change from baseline while adjust- Missing 1 ing for random variance in the model. These <30 6 factors were tested at an α level of .0125 to 30-39 15 control for the multiple analyses run over the 40-49 32 course of the program. 50-59 32 60 + 14 RESULTS Ethnicity Hispanic 70 Implementation activities of the DEP were Non-Hispanic 30 completed over the course of 6 months. Five

Table 2. Year 1 Preliminary Results (Analysis line of HbA1c level less than 7%, the point at of Covariance) which risk of diabetes-related comorbidities is believed to be significantly reduced.25 These findings indicate that this CHW-led model of Factors FP care is an effective model for diabetes man- Baseline hemoglobin A1c 334.38 <.0001 agement and can be successfully replicated in Visit 0.74 .5322 clinics serving low-income, Hispanic popula- Ethnicity 1.9 .1704 tions. Age 0.48 .4885 The finding that baseline HbA1c levels were Sex 0.3 .5854 significantly associated with the decrease in Enrollment date 4.42 .0008 HbA from baseline is not unexpected be- Clinic 1.77 .1361 1c cause patients who had better control of their diabetes from the beginning of the study did not need significant reductions in HbA1c to associated with a decrease in HbA1c from meet the ADA guideline of HbA1c level less baseline. In addition, patients who enrolled than 7%. The observation that the change in in the program later in the year experienced HbA1c was greater for patients who enrolled a greater change in HbA1c values than those in the program at a later date suggests that who enrolled earlier in the year. All other fac- the CHWs became better at delivering the tors tested in the model (number of visits with DEP intervention over time. A number of fac- the CHW, age, sex, ethnicity, clinic) did not tors may explain why the number of visits have a significant effect on the mean HbA1c with the CHW was not significantly associ- change from baseline. The analysis of patient ated with HbA1c reduction. Improved access satisfaction surveys revealed that more than to medications and understanding of the im- 98% of DEP participants indicated the high- portance of medication adherence may have est level of satisfaction with the care they re- been the greatest contributor to observed de- ceived (4 on a 4-point Likert scale) in their creases in HbA1c levels from baseline. Patients answers to all 4 questions pertaining to ser- who attended more visits with the CHW may vice delivery posed after each visit with the have done so because they had a more dif- CHW. ficult time managing their diabetes. The fact that age, sex, ethnicity, and clinic were not DISCUSSION significantly associated with changes in mean HbA1c levels from baseline indicate that the Results from the first 18 months of the program was equally effective for all patients, DEP demonstrated the value of integrating a even those who were not Hispanic, and at dif- CHW-led diabetes self-management education ferent clinic sites across Dallas. Patients also program for low-income, historically under- reported a high level of satisfaction with the served, Hispanic patients with diabetes into a care they received, indicating that the DEP health care delivery system. We successfully met their health needs. implemented this model of care coordination, Although many interventions utilizing based on the CoDETM pilot program, in 5 com- CHWs have been linked to improved health munity clinics, providing patients with a struc- outcomes, questions remain regarding how tured, systematic approach to managing dia- to best incorporate the CHW into a health betes. Similar to the findings observed with care delivery team to optimize delivery of CoDETM, DEP patients achieved significant im- care.19,20 Few studies have provided guidance provements in the primary outcome measure, for the implementation of CHW-led interven- glycemic control (HbA1c). The mean HbA1c tions or described the support, training, and for DEP patients decreased from 8.7% at base- resources that should be provided to CHWs to line to 7.4%, approaching the ADA’s guide- maximize the success of the intervention.19

We designed the DEP to be responsive to compensation. We required all CHWs to com- previously identified patient-reported needs plete the Texas State Certification program for CHW interventions (education, commu- as a group. This training provided them with nication and respect, removal of financial a uniform foundation necessary to perform constraints, and access to medication and daily tasks, including skills in communica- transportation)26 by (1) placing an empha- tion, relationship building, service coordina- sis on CHW recruitment and training; (2) tion, advocacy, teaching, and organization, as building on existing community infrastruc- well as a knowledge base in health education. ture through partnerships with local clinics; The interactive workshop format engaged the (3) integrating the CHW role within a health CHWs in problem solving for likely patient care system’s care coordination strategy; and scenarios. The CHWs received an additional (4) developing an electronic diabetes registry 50 hours of training specific to DEP implemen- that tracked patient metrics and facilitated tation and were taught about diabetes-specific communication between CHWs and primary self-management knowledge as if they were care clinicians. patients. Our preliminary results indicate that this education strategy enabled each CHW to Community health worker recruitment effectively transfer diabetes self-management and training knowledge to DEP enrollees while monitor- Because the crux of CHW-led interventions ing blood glucose level, blood pressure, BMI, centers on the interaction between the CHW and self-management behavior adherence. and the patient, the hiring of the right people Continuing education is important in sus- to fill CHW roles is crucial to the success of taining and enhancing CHW knowledge and these programs.16 Patients with chronic dis- performance. Patients who enrolled in the ease have reported that they need to better DEPatalaterdateexperiencedagreaterre- understand their condition in order to appro- ductioninHbA1c on average than patients priately manage it and that CHWs need to who enrolled at earlier dates. We believe that be effective communicators and respectful of this observation is due to the steep learn- the patient.27 The DEP leadership sought to ing curve the CHWs experienced in trans- meet these patient needs through focused re- lating knowledge gained from formal train- cruitment of CHWs within the targeted com- ing to the delivery of patient care. We pro- munities and extensive team-based training. vided CHWs with on-the-job training through We recruited individuals who had excellent continuous monitoring and feedback as well interpersonal skills, the ability to work inde- as structured education. During biweekly pendently, and an aptitude for learning about CHW-management team meetings, CHWs ex- chronic disease management. changed information about patient and pro- A prime function of CHWs is to educate pa- gram challenges, received feedback from DEP tients about diabetes and other health issues. management, and developed solutions. The Thus,itwasimportantthatCHWsbeeducated DEP management team continuously moni- in a manner that would provide them with the tored the CHWs’ performance through pa- appropriate level of knowledge and the ability tient satisfaction surveys providing feedback to confidently and accurately relay health in- to the CHWs’ about good performance or ar- formation to patients. Studies have shown that eas for improvement. Although the state of standardized education for CHWs is important Texas requires CHWs to complete 20 hours to CHW acceptance, use, and professional of continuing education every 2 years to main- development27 and that adequate training en- tain certification, we provided the CHWs with hances CHWs’ competence, confidence, and 35 hours of continuing education in the first overall contribution.28 Texasisoneofthefew year. Community health workers who are states requiring certification (or 1000 hours of hired in the future will be required to shadow previous experience) for CHWs who receive experienced CHWs as part of their initial

Copyright © 2012 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. 168 FAMILY &COMMUNITY HEALTH/APRIL–JUNE 2012 training because we have found that there is administrative and clinical leadership infras- no substitute for real-world experience. tructure of 5 local charity clinics, BHCS also provided centralized management for the DEP Building on existing community with a supervising nurse manager for daily infrastructure oversight. The CHWs were encouraged to take ownership of the project and their pa- By embedding the DEP program within tients; however, the DEP program manage- community charity clinics, the BHCS used ment team and clinic providers provided sup- the existing community-based health care in- port for social and clinical situations that frastructure to remove barriers to care for were beyond the CHWs’ capabilities and underserved patients while enhancing the ef- training. fectiveness of the DEP. Utilizing this com- True integration of CHWs into a health munity “hub” strategy, the DEP was able to care delivery system required changing the build on existing programs that reduced fi- organizational culture to recognize and inte- nancial barriers to care related to co-pays, grate CHWs as members of the primary health transportation, and access to low-cost pre- care team. In an assessment of CHWs work- scriptions. BHCS could also leverage the trust ing to provide chronic disease care manage- that the charity clinics and clinicians had pre- ment, Keller et al19 foundthatCHWswere viously established with the targeted minority often invisible to clinical providers, as most communities. Patient trust in the CHWs was providers did not recognize the CHWs’ ability demonstrated by the willingness of patients to provide diabetes self-management support. to discuss health and other personal concerns In addition, in their study, all communications with CHWs such as self-management adher- from the CHWs had to be filtered through a di- ence, health care costs, disease control, food abetes educator, creating large gaps between security, depression, and anxiety. This open the work of the clinicians and the CHWs. Rec- communication allowed the CHW to work ognizing the importance of effective integra- with the patient and family to identify and tion of the CHWs into the delivery of diabetes overcome barriers to achieving diabetes con- care, we took proactive measures to build trol and enhanced the effectiveness of the DEP relationships between the CHWs and clinic intervention. staff, establishing the CHW as a member of each clinic’s primary care team. To accom- Integrating the CHW’s role within a plish this, we scheduled monthly meetings health care system’s care coordination between the clinic leadership, DEP manager, strategy and CHWs for the first 12 months after pro- Implementation of the DEP involved gram launch to discuss DEP processes and pre- a system-level transformation and demon- liminary outcomes. We also facilitated com- strated how the role of CHWs can be ef- munication between CHWs and clinicians by fectively integrated within the ambulatory providing CHWs with a predefined process care component of a health care system. By and go-to person for clinical questions. In enhancing the role of CHWs to include the addition, the CHWs provided standardized delivery of diabetes education and chronic written, visit summaries to the primary care disease management, in addition to the tra- provider within 24 hours of a patient’s visit. ditional role of CHWs of providing cultur- These reports facilitated coordination of pa- ally sensitive health information and social tient care by providing the primary care physi- support, the DEP empowered CHWs with cian with additional information about the pa- the ability to provide diabetic patients with tient’s condition and alerts about health issues timely, efficient, and coordinated care. Ex- that needed to be addressed during the next panding the role of CHWs required a struc- office visit. tured system of management and support. Providers initially utilized CHWs because While the CHWs were embedded within the there was a lack of available diabetes

Copyright © 2012 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. Implementation of a CHW-Led Diabetes Self-Management Education Program 169 education services for their patients. How- diabetes clinical indicators and outcomes for ever, over time, they grew to embrace the all participants and to identify areas for op- unique value of a CHW. This learning curve portunities for improvement. We believe that likely could have been shortened by engag- the use of DiaWEB fostered clinic and physi- ing the clinicians in the monthly clinic-DEP cian acceptance and utilization of the DEP meetings and using different communication as it generated additional data that was help- strategies to familiarize them with the role ful in treating patients, provided reassurance of CHWs during project implementation. The that the CHWs would seek clinician input for CHWs noted that their perceived value was the management of patients when appropri- enhanced when they had direct interaction ate, and allowed them to observe improve- with providers and could work directly with ments in patient outcomes through summary clinicians to address patient situations that re- reports. quired physician expertise. Communicating guidelines concerning scope of practice, the evidence base for the program, and support CONCLUSION systems for the CHWs created a level of com- fort with clinicians and clinic administrators. The DEP is a low-cost, high-quality model Ultimately, patient feedback and documented for the delivery of patient care to under- improvements in patient outcomes have led served populations that blends care coordi- clinicians and administrators to realize the nation services and disease-specific patient unique value of CHWs. education for patients. The model was con- structed by applying lessons learned from previous studies on the effectiveness of CHW-led Developing an electronic diabetes diabetes education programs for low-income, registry chronically ill patients.22,23 We have refined Supporting technology in the form of the the DEP on an ongoing basis on the basis DiaWEB chronic disease management soft- of feedback from CHWs and program staff ware was key to the successful integration and analyses of participant outcomes. The of the DEP into clinic operations because implementation of the DEP was successful it provided a means to collect and monitor because it featured an evidence-based, team- patient clinical indicators, guide the CHW oriented, patient-centered approach to care. through the patient visit, and evaluate the ef- TheCHWswereintegratedintothehealth fectiveness of the DEP. The software enabled care team and empowered through ongoing the CHWs to view trends in clinical indica- training and support. Systems such as a di- tors such as HbA1c and blood pressure and abetes registry, patient satisfaction surveys, to monitor the patient’s receipt of preven- and dashboards were used for tracking and tive services. DiaWEB was structured to guide reporting of outcomes and to ensure account- the CHW’s visit-specific patient interaction, ability. The program was mission rather than with key decision points clearly identified for task oriented and allowed for the tailoring of when the CHW would seek assistance from treatment protocols to meet individual patient hernursemanagerorthepatient’sprimary needs. care physician. This feature gave the CHWs Outcome data from the first 18 months the confidence to work independently when of the DEP indicate that a CHW-led diabetes appropriate and the necessary decision sup- management program for uninsured Hispanic port to know when to engage other members patients with uncontrolled diabetes can help of the care team. Through DiaWEB-enabled patients achieve improvements in glycemic data collection, the management team had the control (HbA1c) while achieving superior pa- ability to evaluate the overall effectiveness of tient satisfaction. BHCS’s successful adap- the DEP and individual CHWs in improving tation of the CoDETM pilot program has

Copyright © 2012 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. 170 FAMILY &COMMUNITY HEALTH/APRIL–JUNE 2012 provided an affordable strategy for ex- abetes or the disparities in health care access panding chronic disease self-management and outcomes faced by some populations. education and care coordination principles This study does have several limitations. It is to underserved communities and substan- an observational study and there is no control tively increased its capacity to reduce diabetes group. The statistical analysis was conducted health disparities. This research has important using a limited data set of outcome variables policy implications for the design of programs gathered during the ﬁrst 18 months of a 5-year to improve diabetes management in under- project for the purpose of a program evalua- served populations. The DEP’s program per- tion. All patients with 2 or more visits with formance over the ﬁrst 18 months suggests the CHW were included in the analysis even if that its long-term value proposition resides in they had less than a year’s worth of follow-up the fact that CHWs can help patients who data. The true impact of the DEP program on have been unable to manage their diabetes improving patient care and health outcomes with usual care achieve sustainable diabetes will not be evident until we collect data from control. The initial success of the DEP and of additional program participants over time and similar projects that have utilized a CHW to analyze additional process and outcome mea- coordinate care and provide chronic disease sures. However, we believe that the lessons management in underserved populations pro- we have learned to date from the program vides a strong case for the use of this model as implementation are important and should a means to augment the traditional treatment inform the development of new CHW-led in- of chronic disease. The results of this study terventions. BHCS’s successful implementa- were consistent across 5 collaborating char- tion of the DEP in Dallas communities pro- ity clinic sites throughout Dallas with simi- vides a viable roadmap for reducing health lar demographics, indicating that these results disparities and achieving a high-performing may be generalizable to similar populations in health care delivery system. New delivery-of- other cities. The need for new care models has care models such as the DEP that use CHWs been recognized by health care providers and to provide care coordination have the abil- policy makers alike because current health ity to transform health care and improve out- care delivery models cannot combat the ris- comes for patients with chronic conditions ing prevalence of chronic diseases such as di- other than diabetes.

REFERENCES

1. Centers for Disease Control and Prevention. 5. Sequist TD, Fitzmaurice GM, Marshall R, Shaykevich National Diabetes Fact Sheet, 2011. Atlanta, GA: US S, Safran DG, Ayanian JZ. Physician performance Department of Health & Human Services, Centers for and racial disparities in diabetes mellitus care. Arch Disease Control and Prevention; 2011. Intern Med. 2008;168(11):1145-1151. 2. Texas Diabetes Council, Texas Department of 6. Coberley CR, Puckrein GA, Dobbs AC, McGinnis MA, State Health Services. Diabetes: A Comprehensive Coberley SS, Shurney DW. Effectiveness of disease Approach. Austin, TX: Texas Diabetes Council, Texas management programs on improving diabetes care Department of State Health Services; 2010-2011. Pub- for individuals in health-disparate areas. Dis Manag. lication E45-10524. 2007;10(3):147-155. 3. Spencer MS, Kieffer EC, Sinco BR, et al. Diabetes- 7. Davis KL, O’Toole ML, Brownson CA, Llanos P, speciﬁc emotional distress among African Americans Fisher EB. Teaching how, not what: the contribu- and Hispanics with type 2 diabetes. J Health Care tions of community health workers to diabetes self- Poor Underserved. 2006;17(2)(suppl):88-105. management. Diabetes Educ. 2007;33(suppl 6):208S- 4. Heisler M, Smith DM, Hayward RA, Krein SL, 215S. Kerr EA. Racial disparities in diabetes care pro- 8. Heisler M, Spencer M, Forman J, et al. Partic- cesses, outcomes, and treatment intensity. Med Care. ipants’ assessments of the effects of a commu- 2003;41(11):1221-1232. nity health worker intervention on their diabetes

self-management and interactions with healthcare 18. Brownstein JN, Hirsch GR, Rosenthal EL, Rush CH. providers. Am J Prev Med. 2009;37(6)(suppl 1):S270- Community health workers “101” for primary care S279. providers and other stakeholders in health care sys- 9. Collins KS. Diverse Communities, Common tems. J Ambul Care Manage. 2011;34(3):210-220. Concerns: Assessing Health Care Quality for Minor- 19. Keller T, Borges WJ, Hoke MM, Radasa T. Promotores ity Americans New York, NY: Findings from the and the chronic care model: an organizational assess- Commonwealth Fund 2001 Health Care Quality ment. J Community Health Nurs. 2011;28(2):70-80. Survey. the Commonwealth Fund; 2002. 20. Swider SM, Martin M, Lynas C, Rothschild S. Project 10. Cooper-Patrick L, Gallo JJ, Gonzales JJ, et al. Race, MATCH: training for a promotora intervention. Dia- gender, and partnership in the patient-physician re- betes Educ. 2010;36(1):98-108. lationship. JAMA. 1999;282(6):583-589. 21. Rosenthal EL, Wiggins N, Ingram M, Mayfield- 11. Thompson JR, Horton C, Flores C. Advancing di- Johnson S, De Zapien JG. Community health work- abetes self-management in the Mexican American ers then and now: an overview of national studies population: a community health worker model in a aimed at defining the field. J Ambul Care Manage. primary care setting. Diabetes Educ. 2007;33(suppl 2011;34(3):247-259. 6):159S-165S. 22. Culica D, Walton JW, Prezio EA. CoDE: community di- 12. Reinschmidt KM, Hunter JB, Fernandez ML, Lacy- abetes education for uninsured Mexican Americans. Martinez CR, Guernsey de Zapien J, Meister J. Un- Proc (Bayl Univ Med Cent). 2007;20(2):111-117. derstanding the success of promotoras in increasing 23. Culica D, Walton JW, Harker K, Prezio EA. Effective- chronic diseases screening. J Health Care Poor Un- ness of a community health worker as sole diabetes derserved. 2006;17(2):256-264. educator: comparison of CoDE with similar cultur- 13. Gavin JR III, Wright EE Jr. Building cultural compe- ally appropriate interventions. J Health Care Poor tency for improved diabetes care: African Americans Underserved. 2008;19(4):1076-1095. and diabetes. J Fam Pract. 2007;56(9)(suppl):S22- 24. Knowles M. The Adult Learner: A Neglected Species. S28. 3rd ed. Houston, TX: Gulf Publishing; 1984. 14. Brown SA, Blozis SA, Kouzekanani K, Garcia AA, 25. The Diabetes Control and Complications Trial Re- Winchell M, Hanis CL. Health beliefs of Mexican search Group. The effect of intensive treatment of Americans with type 2 diabetes: the Starr County bor- diabetes on the development and progression of der health initiative. Diabetes Educ. 2007;33(2):300- long-term complications in insulin-dependent dia- 308. betes mellitus. NEnglJMed.1993;329(14):977-986. 15. Glazier RH, Bajcar J, Kennie NR, Willson K. A sys- 26. Gimpel N, Marcee A, Kennedy K, Walton J, Lee S, De- tematic review of interventions to improve diabetes Haven MJ. Patient perceptions of a community-based care in socially disadvantaged populations. Diabetes care coordination system. Health Promot Pract. Care. 2006;29(7):1675-1688. 2010;11(2):173-181. 16. McElmurry BJ, McCreary LL, Park CG, et al. 27. Love MB, Legion V, Shim JK, Tsai C, Quijano V, Implementation, outcomes, and lessons learned Davis C. CHWs get credit: a 10-year history of the from a collaborative primary health care program first college-credit certificate for community health to improve diabetes care among urban Latino workers in the United States. Health Promot Pract. populations. Health Promot Pract. 2009;10(2): 2004;5(4):418-428. 293-302. 28. Jackson EJ, Parks CP. Recruitment and training issues 17. Volkmann K, Castanares T. Clinical community from selected lay health advisor programs among health workers: linchpin of the medical home. JAm- African Americans: a 20-year perspective. Health bul Care Manage. 2011;34(3):221-233. Educ Behav. 1997;24(4):418-431.