|Home | About | Journals | Submit | Contact Us | Français|
Recent studies suggest that involving pharmacists is an effective strategy for improving patient adherence and blood pressure (BP) control. To date, few controlled studies have tested the cost-effectiveness of specific models for improving patient adherence and BP control in community pharmacies, where most Americans obtain prescriptions. We hypothesized that a team model of adherence monitoring and intervention in corporately owned community pharmacies can improve patient adherence, prescribing, and BP control among hypertensive African American patients.
The Team Education and Adherence Monitoring (TEAM) Trial is a randomized controlled trial testing a multi-step intervention for improving adherence monitoring and intervention in 28 corporately owned community pharmacies. Patients in the 14 control pharmacies received “usual care” and patients in the 14 intervention pharmacies received TEAM Care by trained pharmacists and pharmacy technicians working with patients and physicians. Data collectors screened 1,250 patients and enrolled 576 hypertensive African American patients. The primary endpoints were the proportion of patients achieving BP control and reductions in systolic and diastolic BP measured after 6 and 12 months. Secondary endpoints were changes in adherence monitoring and intervention, patient adherence and barriers to adherence, prescribing, and cost-effectiveness. Researchers also will examine potential covariates and barriers to change.
Involving pharmacists is a potentially powerful means of improving BP control in African Americans. Pharmacists are in an excellent position to monitor patients between clinic visits and to provide useful information to patients and physicians.
The global prevalence of hypertension (HTN), and the associated morbidity and mortality rates, is among the highest for African Americans in the U.S. (1). Barriers to HTN control in African Americans have been identified at multiple levels, including: patient nonadherence to regimens (2), suboptimal prescribing (3–4), and ineffective health systems (5). A recent review of quality improvement (QI) strategies (5) suggests that a team change involving pharmacists may be one of the most effective QI strategies for improving BP control. However, it is not known whether pharmacy interventions can be implemented successfully in corporately owned pharmacies, where most Americans now obtain their prescriptions.
While nonadherence with antihypertensive regimens is found in all groups, African Americans often have higher nonadherence according to objective measures (6). Studies show that minority populations experience more barriers to adherence, even after controlling for socioeconomic status. For example, African Americans are more likely than Caucasians to experience the “core” barriers to adherence, including: problematic side effects, difficulty remembering, doubts or concerns about drug efficacy, and difficulty paying for drugs (2). Higher priority must be placed on better tools or models of care for detecting and reducing these modifiable barriers to adherence in African Americans.
Reviews of adherence-enhancing interventions (7) suggest that many educational and behavioral interventions are efficacious; however, practitioners generally lack formal training in these techniques and lack user-friendly tools for assessing and changing patient behavior in busy practices. As a result, they often lack confidence in changing patient behavior and view evidence-based interventions as too complicated, labor-intensive, or expensive to be implemented widely (8–9). Higher priority must be placed on the design of user-friendly tools and strategies that require minimal training and start-up costs and are easy to implement in busy settings. We designed our interventions with these issues in mind.
Over the past 10–15 years, pharmacy education has emphasized pharmacists’ role in patient care and disease management, including hypertension (10). Pilot studies generally show that pharmacists who provide patient education and monitoring of BP can have significant effects on patient adherence (range 28% to 95%) and BP control (range 38% to 70%) (11–17). Studies also show that physicians frequently change therapy based on pharmacist recommendations (18–21). Zillich and others (21) conducted a trial that randomized 12 Iowa pharmacies to high-intensity versus low-intensity pharmacist intervention in predominantly white patients. Their results showed that 75% of the pharmacist’s recommendations were accepted by physicians.
While these trends are promising, past studies examining the effects of pharmacist intervention have limitations. First, many studies have been conducted by pharmacy faculty or residents in university-affiliated sites, raising questions about generalizability . Especially needed are studies in corporately owned pharmacies, which now account for over 50% of the prescriptions dispensed in the U.S. (22). Second, most studies have involved a small number of patients referred by physicians, resulting in biased samples. Third, studies suggest that pharmacist interventions can be cost-effective in hypertension clinics (23) and Canadian pharmacies (24). It is unknown whether similar results can be obtained in the community, where a larger number of people could be reached. Our study therefore evaluates a team approach to hypertension management for African American patients in corporately-owned pharmacies.
The Team Education and Adherence Monitoring (TEAM) Trial evaluates the outcomes of a team-oriented intervention that is evidence-based, patient-centered, and designed to be relatively brief and easy to implement in a community pharmacy. The interventions are focused on pharmacists and pharmacy technicians working with patients and physicians to improve BP control, the primary outcome. Secondary outcomes are changes in pharmacy-based adherence monitoring and intervention, patient adherence and barriers to adherence, prescribing, and cost-effectiveness. The cost-effectiveness results can be compared to findings from other settings to identify the potential role of community pharmacy-based interventions in optimizing BP control. It is possible that a community pharmacy-based program may be less expensive than previously evaluated settings, offering a wider scale of intervention for fewer resources.
The proposed intervention builds on two models which help us predict and address barriers to patient adherence and changes in pharmacy practice. First, the intervention builds on Svarstad and Bultman’s Health Collaboration Model (HCM) (25–27). The HCM identifies five “core” adherence barriers amenable to change by pharmacists. These “core barriers” include: regimen knowledge barriers (poor understanding of the drug regimen and its elements such as dosage schedule, treatment duration, purpose); recall barriers (difficulty remembering multiple drugs and doses); motivational barriers (doubts regarding drug efficacy, benefits, or need for therapy); side effect barriers (bothersome side effects or concerns about long-term effects); and access barriers (difficulty paying or obtaining refills). The HCM posits that these “core barriers” can be detected and reduced by using certain tools for: a) monitoring adherence and barriers to adherence; b) eliciting patient involvement; c) “tailoring” interventions to address patient-specific concerns and barriers; and d) providing regular follow-up and support to patients and physicians. We used these concepts when designing our system interventions, TEAM tools, and measures of “intervention fidelity “ (staff adherence to the TEAM protocol). It is hypothesized that our intervention will lead to predicted changes in adherence monitoring and intervention by pharmacists which, in turn, lead to predicted reductions in the “core” barriers to adherence, prescribing changes, and, ultimately, improved patient adherence and BP control (Figure 1).
Second, the intervention build on organizational strategies recommended in Rogers’ Diffusion of Innovation Model (28). His work suggests several strategies for promoting the adoption of new tools and practices. First, the change agent (researcher) must seek input from users (pharmacists) to make sure that the proposed innovations (BP program and tools) are “compatible” with existing routines and work flow in the user’s organization (pharmacy). The change agent also must recognize the influence of other stakeholders (pharmacy district managers or supervisors) who can facilitate change and provide the support needed to achieve changes in practice. The change agent also must make sure that the new tools and practices are relatively simple, easy to use, and viewed by stakeholders as having relative advantages or benefits for them, their clients, and the larger organization (corporation). It is for these reasons that we developed a pharmacy-based BP program that is relatively short (6 months), involves pharmacists and pharmacy technicians, requires minimal training (8 hours), and includes time-saving tools for more rapid adherence monitoring and intervention. In addition, we established a management advisory group to facilitate collaboration between researchers and pharmacy district managers and supervisors to make sure our TEAM program and tools were compatible with pharmacy routines, resources, corporate missions, and plans. Rogers’ work also helped identify potential barriers to change such as workload, turnover, staff beliefs, and perceived barriers to change. These variables have been added to the study model to guide our analysis of intervention fidelity and outcomes (Figure 1).
This study evaluates the outcomes of TEAM Care (TC) versus Usual Care (UC) in a randomized controlled trial approved by each investigator’s institutional review board (IRB), with the University of Wisconsin IRB assuming primary responsibility. The study involves 28 community pharmacies owned by Walgreens and Aurora Pharmacy, Inc. Walgreens is the nation’s largest drugstore chain with over 5,000 pharmacies nationwide and Aurora Pharmacy, Inc. is a regional chain with over 100 pharmacies in Wisconsin. The study statistician used computer software to randomize the 28 sites into an intervention group (TEAM Care or TC) and a control group (Usual Care or UC). See Figure 2. Data collectors screened patients, obtained consents, and collected baseline patient and staff surveys. Researchers also reviewed pharmacy records at baseline. The goal was to enroll 25 hypertensive African American patients, one pharmacist, and 1–2 pharmacy technicians per site. Primary and secondary outcomes were assessed 6 and 12 months after enrollment using blinded data collectors, prescription profiles, and program records.
The study pharmacies had to be owned by Walgreens or Aurora Pharmacy, Inc., located in a zip code area with a significant African American population, and have one licensed pharmacist and one pharmacy technician willing to implement a BP clinic and TEAM Care program for 25 hypertensive patients over a 6 month period. The pharmacies were identified by their pharmacy supervisor or district manager in collaboration with the principal investigator (BLS) and located in five Wisconsin cities: Milwaukee, Racine, Kenosha, Beloit, and Madison. The majority of pharmacies were located in zip code areas with ≥30% African Americans.
To be eligible for the TEAM trial, patients had to: be African American and 18 years or older, have an active prescription for antihypertensive medication, get their antihypertensive medication filled at the study pharmacy, have a mean screening BP ≥ 140/90 mm Hg and ≤ 210/115 mm Hg (using the second and third BP measurements), and be able to read and fill out a self-administered questionnaire and come to the study pharmacy for 6 monthly visits if needed. Patients were excluded if they reported kidney dialysis, liver disease, organ transplant, memory loss that interferes with daily activities, terminal illness, pregnancy, alcohol or substance problem, or symptoms of heart failure (short of breath or chest pain after walking one-half block, feet and legs so swollen that you can’t put on your shoes, or difficulty sleeping due to shortness of breath). Patients also were excluded if their arm circumference was larger than recommended for the available cuffs (> 16.5 inches) or if their physician recommended exclusion for medical reasons. Patients were distributed across all physician practice settings.
Patients were recruited by trained data collectors using study posters, flyers, monetary incentives, and help from pharmacy staff, pharmacy students, and African American health professionals who sponsored a small number of church-based screenings in one of the cities. The flyer included a brief study description, a toll-free number, and a tear-off section for patients to indicate interest and contact information. The data collectors then contacted interested patients about a convenient time for screening at their pharmacy. Patients were offered a $5 gift card for completing the BP screening, a $20 card for completing enrollment questionnaires, a $25 card for the 6-month follow-up questionnaire and BP check, a $25 card for the 12-month follow-up questionnaire and BP check, and up to $10 in gift cards to offset certain transportation costs. Researchers also distributed a staff information sheet and consent form to participating pharmacists and pharmacy technicians before completing a baseline and 6-month follow-up survey. Pharmacy staff received no monetary incentives or compensation for participation. Intervention staff received 7.0 hours of continuing education credit at no cost. Employers received compensation for staff retrieval of patient records and other study-related activities.
All pharmacists and pharmacy technicians received an oral and written description of the study, one copy of a reference card with JNC 7 guidelines (29), and copies of two HTN brochures that were mailed to their patients. After the pharmacy met enrollment goals, the principal investigator (PI) mailed a packet to all patients. This packet explained the patient’s assigned group and encouraged the patient to read the enclosed brochures titled “High Blood Pressure in African Americans” (American Heart Association, 2001) and “Your Guide for Lowering Blood Pressure” (National Heart, Lung, and Blood Institute, 2003). The brochures explained the risks of high BP, how to interpret BP numbers, and ways to lower BP (tips for managing drugs, being active, cutting salt, following DASH plan, losing weight). The PI also mailed a packet to the patient’s primary physician. The physician packet included a description of the study and eligibility criteria, a form for recommending patient exclusion (if needed), and a reference card with JNC 7 guidelines for treating hypertension. No additional interventions were delivered to the usual care pharmacy staff, patients, or their physicians.
Table 1 lists the multi-step intervention targeting pharmacy staff and patients at the TC sites. The first step involved training pharmacists and pharmacy technicians. Pharmacy staff who completed this training received 7 credit hours of continuing education for self-study (1 hour) and attending a 7-hour interactive workshop conducted by pharmacy and medical educators using diverse methods (lecture, slides, handouts, demonstration and practice, role play, case studies, break out discussion). Three modules addressed: the rationale and guidelines for measuring and managing hypertension in African Americans; the rationale and guidelines for improving patient adherence and involvement in care; and guidelines for improving patient follow-up, retention, and physician collaboration. [While our trial targeted African Americans, the TEAM materials and tools were designed so they can be applied in any population or pharmacy.]
The second intervention step involved the distribution of clinical tools to intervention staff at TC sites (Table 1) these user-friendly tool kits included: automatic BP monitoring equipment, portable furniture and a privacy screen for setting up a BP clinic in the pharmacy waiting area, and tools for more rapid adherence monitoring and intervention with patients and physicians. These tools included: a 2-page Brief Medication Questionnaire (BMQ) for screening patient adherence and barriers to adherence (30); a 4-page Health and Lifestyle Questionnaire (HLQ) for assessing patient risk factors, BP awareness, lifestyle, and patient goals; adherence counseling tips with an algorithm for “tailored intervention” using the BMQ and HLQ; a 1-page Team Action Plan (TAP) for documenting the patient’s BP, adherence and lifestyle issues, interventions, and plans; a 1-page BP Goal Check (BGC) for brief monitoring at follow-up visits; a 1-page TEAM Log for brief charting of BP and progress notes; a tool and guidelines for informing and referring patients with severely elevated BP; and appointment tools (appointment cards, reminders, scheduling book). The clinical tools also included an innovative check list called the Request for Medication Review (RMR). The RMR enabled pharmacists to provide systematic and rapid feedback to physicians and included space for BP readings, adherence and lifestyle issues, and recommendations for change in the drug regimen, if appropriate. The RMR also included space and instructions for the physician to fax a response to the pharmacy if appropriate. Pharmacists were encouraged to send specific recommendations to physicians based on JNC-7 guidelines if appropriate.
The third intervention step involved the distribution of patient tool kits to intervention staff. These user-friendly tools included adherence aides and colorful leaflets that we designed specifically for our pharmacy-based BP clinics. The content was based on JNC-7 recommendations, established behavioral interventions (7), and the Health Collaboration Model (25–27). They included: a pill organizer, BP tracker, pedometer, and easy-to-read leaflets or tips sheets for increasing patient awareness of BP goals, patient use of pharmacists and other strategies for managing medications, and tips for making lifestyle changes (Table 1).
The fourth intervention step involved the establishment of BP clinic hours and implementation of a 6-month TEAM Care Program at each TC site. Intervention staff selected BP clinic hours based on available staffing; however, all TC sites aimed for 1 monthly visit with each enrolled patient over a 6-month period. Table 2 lists the components of the TEAM Care program and roles performed by intervention staff and patients. In preparation for all visits, the pharmacy technician made and confirmed appointments, printed patient profiles, and set up the BP equipment and furniture. During Visit 1, the pharmacy technician welcomed the patient, took BP readings, and introduced the BMQ and HLQ forms to be completed by patients. The pharmacist met with the patient to assess BP, adherence and lifestyle issues, and patient goals for the next 6 months. The pharmacist also offered BP counseling, tailored intervention, and the patient tool kit. During Visits 2–6, the pharmacy technician welcomed the patient, took BP readings, and showed the patient how to complete parts of the BGC form for assessing patient progress and concerns at follow-up visits. The pharmacist met with the patient to monitor progress, reinforce adherence, and support change efforts. Pharmacists also faxed written feedback with recommendations for change to the physician as needed. These follow-up visits were designed to be brief and coincide with the patient’s regular visits to the pharmacy for picking up refills, if possible.
Based on past studies with a similar protocol and > 30% African Americans, we estimated an intervention effect size ranging from .40 to .55. Due to the hierarchical nature of the data (patients within pharmacies), we projected varying pharmacy dependencies in patient responding (ρ= .02 to .08). We used the OPDES (Optimal Design Version 0.19) computer program (31) to estimate the optimal sample size to maintain statistical power at .80, with α = .05 (two-tailed). A study of power based on various size configurations of pharmacies, patients, and pharmacy dependencies indicated that a configuration of 28 pharmacies would require approximately 18 patients per site or 504 patients overall to maintain statistical power with an effect size of .40 and a pharmacy dependency of ρ=.08. We oversampled patients (25 patients/site) to allow for attrition. The final sample included 576 patients. The follow-up evaluations were performed by the University of Wisconsin Survey Center, a nationally recognized center that employs experienced interviewers and experts in locating and tracking subjects in longitudinal studies. As a result, we retained 546 patients (approximately 20 patients/site), yielding a preliminary response rate of 94.8%. This means we should have sufficient power.
Table 3 lists the measures and data collected. The primary outcomes (BP control and reductions in BP) were measured 6 and 12 months after enrollment by data collectors blinded to the patient’s assigned group and trained to obtain accurate BP readings using appropriately sized cuffs, an adjustable table, and a monitor validated by the British Hypertension Society (Microlife Model # 3AA1-2). After measuring the patient’s BP, data collectors asked patients to complete two questionnaires: a Brief Medication Questionnaire or BMQ (30) for assessing adherence and barriers to adherence and a questionnaire for measuring patient background characteristics and other outcome measures.
In addition to self-reported patient adherence, we are measuring changes in refill adherence using pharmacy prescription profiles. Changes in adherence monitoring and intervention will be examined using program records, staff surveys, and patient survey items. Changes in prescribing will be measured using prescription records to determine whether the prescriber intensified drug therapy, whether the regimen was consistent with JNC 7 prescribing guidelines, and how prescribers responded to any pharmacist recommendations. Cost-effectiveness will be analyzed for TC versus UC patients using program records, prescription profiles, and patient-reported use of health services. Costs to deliver the intervention include primarily pharmacist and pharmacy technician time. We also anticipate changes in antihypertensive medication use and changes in the use of other health care services in the short run. We will apply the societal and provider perspectives although we will only include short-run elements.
We plan to construct 2-level hierarchical regression models in our analysis of intervention effects. Our plan includes identifying and controlling any pre-existing patient differences that can impact patient involvement and response (e.g. patient age, gender, education, income, co-morbidities, payment source, drug regimen). We also will analyze variations in “intervention fidelity” or the extent to which intervention staff actually use the TEAM tools and adhere to TEAM guidelines for adherence monitoring and intervention with patients and physicians. It is hypothesized that greater intervention fidelity will be associated with greater reductions in the “core” barriers to adherence, prescribing changes, patient adherence, and BP control (Figure 1). Finally, we will examine whether any baseline pharmacy and staff characteristics are linked to changes in adherence monitoring, intervention, and intervention fidelity. This analysis will help identify barriers to change in pharmacy practices. For example, higher workloads may reduce the extent to which intervention staff is willing or able to fully implement the TEAM protocol and tools for adherence monitoring and intervention.
We have experienced several challenges. The first challenge was to identify pharmacy corporations able to participate in a large controlled trial. We first identified a corporation that was willing to participate and gave access to pharmacies in two states. Unfortunately, the corporation experienced delays when installing a new information system and later notified us that they planned to sell their pharmacies before the program could be implemented. We then approached Walgreens and Aurora Pharmacy, Inc. Fortunately, they were interested and provided exceptional support during the study. A key lesson in our partnership with Walgreens and Aurora Pharmacy, Inc. was the importance of regular communication with district supervisors who believed strongly in the need for improved pharmacy care in this area and provided valuable help in recruiting pharmacies, encouraging staff to participate, and addressing barriers to implementation.
Second, we anticipated the lack of private space in a busy pharmacy. Despite these challenges, data collectors and pharmacy staff were able to create semi-private spaces using portable tables, chairs, and privacy screens in pharmacy waiting areas. Another set of challenges involved the review of protocols by multiple IRBs. The IRBs had limited experience with pharmacy trials and often had different concerns or points of view about consent procedures, whether and how to inform physicians, and other issues.
Another set of challenges related to the staff time, resources, and coordination required to recruit and orient pharmacy staff and patients at 28 pharmacies in five cities. Fortunately, many pharmacists and pharmacy technicians were eager to participate and provided valuable help in enrolling nearly 600 patients over an 8 month period. As in many trials, intervention staff later experienced difficulties re-locating and contacting patients due to disconnected telephones and changes in home address and insurance. As a result, a number of patients never attended any pharmacy visits or had a low rate of attendance despite heroic outreach efforts by pharmacy staff. Whether patient attendance at a pharmacy-based clinic is associated with certain patient-related or pharmacy-related barriers to change will be examined along with sources of variation in staff adherence to the TEAM guidelines (“intervention fidelity”). These findings will help researchers identify the predictors of program success or failure across different patient populations, providers, and sites.
We gratefully acknowledge the contributions of our colleagues, including: our project staff and consultants; the University of Wisconsin Survey Center project director, interviewers, and staff; the Walgreens and Aurora corporations, and all participating pharmacy supervisors and district managers, pharmacists, pharmacy technicians, and patients. The project would not have been possible without them.
Funding Sources: This study was funded by National Heart, Lung, and Blood Institute (NHLBI) Grant # R01 HL78580 (P.I.: Bonnie L. Svarstad, Ph.D.).
Journal subject codes: hypertension, African American, compliance/adherence, pharmacy
Clinical trial registration information: NCT00205153
Disclosures: There is no actual or potential conflict of interest.