The Baltimore Partnership to Educate and Achieve Control of Hypertension (The BPTEACH Trial): A Randomized Trial of the Effect of Education on Improving Blood Pressure Control in a Largely African American Population
Wallace Johnson MD,
From the School of Medicine, University of Maryland, Baltimore, MD;
Hypertension is a major risk factor for developing cardiovascular disease and is more prevalent in African Americans compared with Caucasians. African Americans are often underrepresented in clinical trials. This study was composed of a largely urban African American cohort of hypertensive patients. This was a prospective, 4-arm, randomized controlled trial designed to evaluate the comparative effectiveness of both physician and patient education (PPE), patient education only (PAE), and physician education only (PHE) vs usual care (UC). Hypertension specialists gave a series of didactic lectures to the physicians, while a nurse educator performed the patient education. The mean adjusted difference in systolic blood pressure (SBP) from baseline in the PPE group was an average reduction of 12 mm Hg (95% confidence interval [CI], −4.5 to −19.4) at 6-months, followed by average reductions of 4.6 mm Hg (6.9 to −16.12) in the PAE group, 4.1 mm Hg (3.4 to −11.7) in the PHE group, and 2.6 mm Hg (3 to −8.2) in the UC group. The PPE group achieved a significantly better reduction in SBP compared with the UC group. Additional research should be conducted to evaluate whether the use of certified hypertension educators in collaboration with physicians will result in a similar blood pressure reduction.
Hypertension, one of the most prevalent chronic conditions in the United States, represents a significant public health burden. The age-adjusted prevalence of hypertension is 40.5% among non-Hispanic blacks, 27.4% among non-Hispanic whites, and 25.1% among Mexican Americans.1 It is estimated to account for $76.6 billion in direct health care costs, medications, and missed days of work and remains a major risk factor for stroke, heart, and kidney disease, and an important cause of cardiovascular morbidity.2 Cardiovascular disease (CVD) disparities are particularly evident in Maryland, where 31% of African Americans vs 26% of Caucasians have hypertension,3 and heart disease mortality rates are 305 of 100,000 for African Americans and 245 of 100,000 for Caucasians,4 while stroke mortality rates are 77 of 100,000 in African Americans vs 60 of 100,000 in Caucasians.5 The prevalence of hypertension and other CVD is higher in African Americans compared with Caucasians.
Despite available treatments for hypertension, only one third of all US hypertensive patients achieve blood pressure (BP) control. Many inter-related issues pertaining to the health care system, the patient, and the health care provider impact the control of hypertension. As a result, a variety of strategies have been used to implement procedures for hypertension control, including the education of providers by leaders in the hypertension field. Further, increasing public awareness of hypertension and its link to CVD, coupled with health education in hypertension control, has the potential to impact BP reduction more than either intervention alone.6 Interventions aimed at providers have produced change in provider behavior in terms of adherence to recommended guidelines, but have not necessarily translated into improved BP control in the patients being treated. On the other hand, interventions focusing on patient education suggest potential short- and long-term improvement in BP control.7
The National Institutes of Health (NIH) has recognized that in order for investigators to obtain the research data needed to address the needs of the African American patient population, there should be incentives to increase the diversity of patients and study sites involved in NIH-funded research. The NIH went on to provide grant funding for a community-academic partnership in a community serving mostly African American patients. The purpose of this partnership is to establish community research priorities that would promote physician adherence to guidelines, improve treatment adherence of patients, and assess the impact of these changes on BP control. Targeted health education of patients using community health workers who are drawn from the target population community has been demonstrated to be a successful strategy in several studies.8
Our study uses factorial methodology to implement community based interventions that encourage adherence to the Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (JNC 7) guidelines by physicians and lifestyle modifications by patients. We hypothesized that hypertension-specific continuing education for physicians and office-based education for their patients with documented hypertension would translate into better BP control than usual care or education targeted at physicians alone or patients alone. The strength of this study compared with many prior clinical trials is that it was composed of a largely urban African American population, which is often underrepresented in clinical research. Therefore, very few educational intervention hypertension studies involving this population have been published to date.
Patients and Methods
We conducted this study within the scope of the Baltimore Cardiovascular Partnership, a Community-University collaboration generally aimed at improving communication between research institutions and surrounding communities in Baltimore. The Partnership was funded by a grant from the National Heart Lung and Blood Institute: U01 HL79151, principle investigator (PI): Elijah Saunders, MD; co-PIs: Wallace R. Johnson, Jr, MD, and Fadia T. Shaya, PhD (another parallel study with our community partner, Bon Secours Baltimore Health System [PI: Reed Winston, MD] is reported on and published under a separate cover). We aimed to improve provider and patient approaches to hypertension management. Specifically, we tested the impact of an intensive educational intervention for physicians and their patients, on hypertension control. The study was approved by the institutional review board of the University of Maryland, Baltimore (HP-00040555).
The study was designed as a 4-arm randomized clinical trial, where the educational programs were offered to physicians only (PHE), patients only (PAE), or both physicians and their patients (PPE). In the fourth arm, neither physicians nor their patients (usual care, UC) received any education or intervention. This was a hypothesis-testing, prospective study, with an experimental 2×2 factorial design. Patients were not made aware of whether their physicians were receiving the educational intervention and vice versa (Figure 1).
We enrolled primary care physicians and patients of these physicians. The unit of analysis being the patient, we confined the recruitment of primary care physicians to 2 large practices within the University of Maryland Medical System, the family practice (where physicians received the educational program), and faculty practice clinics (where physicians did not receive the educational program). Five primary care providers were selected from either site and matched by race, sex, and years in practice. We maintained internal validity as factors that influence practice style were consistent across as well as within these two sites. Overall, physicians in the two groups were similar in term of years of practice, and both groups were based in the same academic institution. Physicians in both groups were employees of the University of Maryland, and all physicians were board certified in their respective specialties of family medicine and internal medicine.
The study population consisted of patients with documented hypertension and the ability to sign a written consent form. Within each clinic, patients were randomized to the education or control group, retaining their physician of record. Patients were enrolled on a rolling basis, during a period of 2.5 years, starting on April 1, 2005. Inclusion criteria included uncontrolled hypertension and the absence of medical conditions or treatments that would preclude standard hypertension drug therapies.
BP was obtained during physician visits. The research nurse then measured the participants’ BP 3 times at each data collection visit using a mercury sphygmomanometer and standardized techniques from clinical BP trials.9,10 The second and third values were averaged and used as the clinic BP. A patient was considered “not at goal” if any of the following conditions were met: (1) without diabetes and with systolic BP (SBP) ≥140 mm Hg and/or diastolic BP (DBP) ≥90 mm Hg or (2) with diabetes or renal insufficiency and with SBP ≥130 mm Hg and/or DBP ≥80 mm Hg.
The physician education intervention consisted of an in-depth series of 90-minute interactive lecture sessions presented every 2 months. The physician lectures were given by two American Society of Hypertension–certified hypertension specialists as well as guest lecturers and other members of the research team. The curriculum was implemented over 2.5 years, starting with the first patient enrollment. The topics covered included cardiovascular health disparities, pathophysiology of primary and secondary hypertension, pharmacologic and nonpharmacologic hypertension management, managing comorbidities, and research methods. Case consultations were also made available.
The patient education intervention consisted of up to 30 minutes of personal counseling by the study nurse, at each patient visit, at 6-month intervals. Topics included weight reduction, the Dietary Approaches to Stop Hypertension (DASH) eating plan, sodium and alcohol reduction, physical activity, as well as adherence to visits and medications. The counseling sessions were interactive and customized to the patient and encouraged self-monitoring of BP.
The study nurse was a registered nurse from the community with additional education in hypertension attained by attending the Association of Clinical Research Professionals (ACRP) annual conferences.
The primary outcome was the absolute reduction of SBP within 6 months of follow-up. The secondary outcome was the proportion of patients who achieved BP goal. We also followed the patients’ changes in BP at 6-month intervals.
We built a Microsoft Access relational database for the data repository to provide sufficient functionality and to ensure security, patient confidentiality, and data integrity. Health Insurance Portability and Accountability Act–compliant data, without patient names, were exported to SAS V8 (SAS Institute, Inc, Cary, NC) for analysis.
Sample Size and Power Analysis
The study was powered to detect the proportion of patients who brought their BP to goal. A total of 600 patients balanced over the 4 study arms would have been sufficient to detect a 10%-point change in the proportion of patients at goal attributable to either the patient or physician intervention (with power of 80% and a 1-tailed significance level of 0.05). Our total initial enrollment was 670 patients, and stabilized at 552 after some attrition (Figure 2).
We collected data on patients to include the following: (1) whether they were in the intervention or the control group; (2) demographic data: age, sex, and race/ethnicity; and (3) clinical data: chronic medical conditions, smoking status, BP (mm Hg), presence of diabetes, weight, and diagnosed comorbidities.
A descriptive analysis was performed to compare the characteristics of persons in the study arms. The objective of the analysis was to assess whether participants in each arm were comparable in terms of reaching BP control. Ordinal least square analyses were conducted, with adjustment for potential confounders, to assess the association between the interventions and absolute changes in BP. Logistic regression analysis was used to examine the effect of the interventions on the likelihood of reaching control.
Between April 2005 and July 2007 there were 552 patients randomized into the study (Figure 2). After randomization, drop-out rate was 10% in the PPE group, 8% in the PHE group, 21% in the PAE group, and 14% in the UC group. The main reason for drop-out was patients being lost to follow-up, and 2 patients voluntarily withdrew from the study. There were no significant differences between randomized groups with regard to physician age, sex, and years in practice (Table I). There were 3 African American physicians in the intervention group and no African American physicians in the control group. At baseline, most characteristics of patients between the 4 groups were similar (Table II). Most of patients in the study were African Americans, and the proportion was around 90% across the subgroups. There were more women in the study. Patients in the UC group were older than those in the other subgroups. Approximately 40% of hypertensive patients had concomitant diabetes across the 4 groups. The majority of physicians, namely 4 of 5, attended 90% of hypertension education sessions. For the patient groups, we found a 90% attendance rate for the patient education sessions.
Table I. Baseline Characteristics of Physicians
Abbreviations: JNC 7, Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure; SD, standard deviation.
Age, mean (SD), y
Women, No. (%)
Black, No. (%)
White, No. (%)
Asian, No. (%)
Years in practice, mean (SD)
Specialty certification in family medicine
Specialty certification in internal medicine, No. (%)
Physicians with inpatient duties, No. (%)
Very familiar with JNC 7 guidelines, No. (%)
Any nonpatient administrative responsibilities, No. (%)
Table II. Patient Characteristics (N=552)
Patient and Physician Education
Physician Education Only
Patient Education Only
Abbreviation: SD, standard deviation.
Age, mean (SD), y
SBP decreased in 4 arms of the trial at 6-month follow-up (Figure 3). The mean adjusted difference in SBP from baseline in the PPE group was an average reduction of 12 mm Hg (95% confidence interval [CI], −4.5 to −19.4) at 6 months, followed by average reductions of 4.6 mm Hg (95% CI, 6.9 to −16.12) in the PAE group, 4.1 mm Hg (95% CI, 3.4 to −11.7) in the PHE group, and 2.6 mm Hg (95% CI, 3 to −8.2) in the UC group (Table III). The following baseline characteristics were adjusted: age, sex, arm on which the BP was measured, diabetes, smoking, and drinking alcohol, and the P value was .0007.
Table III. Patient Characteristics at Baseline and 6-Month Follow-Up (N=552)a
Patient and Physician Education
Physician Education Only
Patient Education Only
Abbreviations: BP, blood pressure; SD, standard deviation. aSome information is missing. The information was obtained from the patients’ charts. Thus if missing in the chart, we were not able to record it.
The proportion of patients who achieved their defined goal BP increased in all 4 groups: 38.6% increase, 18.6% increase, 42.1% increase, and 2.7% increase from baseline to 6-month follow-up in the groups of PPE, PHE, PAE, and UC, respectively (P<.0001) (Table III).
The lowest average SBP and DBP at baseline was 143 mm Hg and 84 mm Hg in the UC group. The average cholesterol at baseline was lowest in the UC group compared with other groups, and the average cholesterol level at 6-month follow-up was lowest in the PAE group. Both P values are marginally significant. When patients in the study were divided into two groups, which were patients with and without education, the average SBP and DBP at baseline were 6 mm Hg and 2 mm Hg higher in the group of patients with education compared with those without education, respectively (Table IV). There were no significant differences between the two groups with regard to BP at 6-month follow-up.
Table IV. Patient Characteristics at Baseline and 6-Month Follow-Up (N=552)
While hypertension control has improved in the past decade, with a control rate of 50% noted in the 2007–2008 National Health and Nutrition Examination Survey (NHANES), this is far short of the goals established by Healthy People 2010.6,11 There is a significant need for identifying new models of care, particularly in view of poor access to primary care physicians within minority communities. Labor shortages among primary care physicians means that we must explore methods of improving BP control, which include nonphysicians. Previous research models have used both pharmacists and nurses to improve high BP management with significant success.12 Multicomponent intervention strategies have been found to be more effective than single interventions in most studies.12,13 Although many agree that education for both patients and physicians is an important tool in reducing high BP, the exact best practice model for doing so has not been well established. This study focuses on the patient and the physician since prior investigations have noted hypertension control is often suboptimal despite adequate access to health care.14
This trial actually investigates the comparative effectiveness of various educational interventions in a largely African American cohort. Table II describes the impact of educational intervention on BP control, which is defined as a BP <140/90 mm Hg for nondiabetic uncomplicated patients and BP <130/80 mm Hg in the diabetic or chronic kidney disease subgroups.15 This trial found that the PPE group was twice as likely to reach target BP as the UC subgroup, which received no educational intervention at the physician or patient level. The PAE group was more likely to reach target BP than the PHE group when both subgroups were compared with the UC group. This has particular significance for clinicians and policy makers because these interventions were relatively low in cost yet had considerable potential significance for public health officials. Physician education was easy to incorporate into the practice of all the physicians participating and was very well accepted. Since all physicians have to fulfill continuing medical education (CME) requirements to maintain an active medical license, the administration of CME credits for attending the didactic lecture series provided good incentive for regular physician attendance. The patient education intervention was interactive and individualized with personal counseling as well as group education sessions by the study nurse. The counseling sessions were also designed to be done during the regular office visits at 6-month intervals so that the model could be easily adapted and translated into actual primary care office practice. Overall, the model using both patient and physician education had the largest impact on BP reduction. This model was designed not to be as labor- or resource-intensive as some others that involve the use of large numbers of administrative and health care staff. The investigators think that the efficiency of this model will promote its acceptance by third-party payers and therefore be sustainable.
The educational intervention had a significant impact on SBP. The SBP reduction noted at 6 months in the dual intervention subgroup (PPE) was 12 mm Hg, which exceeds that noted in most team-based care interventions cited in a recent meta-analysis.13 Specifically, when nurses performed the educational study intervention for patients without concomitant physician education, the mean reduction in SBP was only 4.8 mm Hg in the same meta-analysis.13
Study Strengths and Limitations
The study had several strengths including the use of an office-based educational intervention with CME credit for physicians, a largely African American study population of more than 500 patients, the use of standardized clinic BP measurements, and statistical adjustment for potential confounding variables. Our investigators also feel that the relative simplicity of the education model should make it easy to adopt in most primary care office settings. A strong evidence-based curriculum was made possible by the involvement of American Society of Hypertension–certified hypertension specialists in all educational sessions. Overall, the greatest strength of this study is that it is a randomized clinical trial, not a pre-post study.
The limitations of this study include potential sampling bias, a small sample size for the comparator control group, the lack of physician randomization, and “real-world” barriers described by other investigators.14 A significant real-world barrier was the lack of office space for patient screening, which may have had an impact on some of our sampling since the clinics with fewer resources often serve the patients who are “hard to recruit and treat” we were attempting to study. A sampling bias may have resulted from our inability to recruit as many patients when sites had limited space and heavy patient caseloads. Another challenge was the difficulty in both recruitment and retention in the UC group. This led to a significantly smaller sample size in this group. The trial is not “blinded,” but this is unavoidable unless we were to have a separate BP measurement by someone from outside the trial.
This study is significant because of its potential immediate clinical impact on BP management in the office setting. Our study team found that the largest reduction in SBP was 12 mm Hg from baseline to the first follow-up visit in the subgroup where both patients and their physicians received educational intervention (Figure 3). The largest reduction in DBP was 8 mm Hg from baseline to the first follow-up in the same patient/physician intervention subgroup and also in the patient education only subgroup. The reductions in BP are equivalent or better to the expected reductions in SBP and DBP noted with many pharmacologic interventions. Our trial was not specifically designed or powered to address the impact of BP reduction on adverse cardiovascular events, but adequate BP control has been noted to reduce mortality and produce significant cardiovascular benefit. Randomized clinical trials have shown that BP reduction can produce rapid reductions in CVD risk, and a meta-analysis by Lewington and colleagues also provides evidence that even greater difference in risks are likely to be produced by prolonged differences in BP.16 In the above meta-analysis, a 10-mm Hg reduction in SBP or a 5-mm Hg reduction in DBP would, in the long-term, be associated with about a 40% lower risk of stroke death and about a 30% lower risk of death from ischemic heart disease or other vascular causes. In fact, even a 2-mm Hg reduction in SBP would involve approximately 10% lower stroke mortality and a 7% lower mortality from ischemic heart disease or other vascular causes.14 The possibility of a nonpharmacologic intervention resulting in a significant reduction in BP and cardiovascular morbidity and mortality in a largely African American cohort is a significant finding for policy makers and clinicians.
The preliminary assessment by the study physicians is that a low-cost, simple health care education intervention model could be easily adapted to their real-world clinical practice. This model could be reproduced in most primary care offices. A research-intensive medical center is also important in providing a strong evidence-based “train the trainer” model for the physicians. Research has shown that a major contributor to poor BP control in the hypertensive population is poor physician adherence to published guidelines.14 Fortunately, both this trial and a recent interventional study found that academic detailing and giving feedback to participating physicians led to improved BP control rates.14
The previously published NIH-funded Hypertension Improvement Project (HIP) is a community-based, nested, 2×2, randomized, controlled trial of physician intervention vs control and/or patient intervention vs control trial with both significant similarities and differences compared with our trial.17 The important similarity is essentially the same 2×2 trial design with the major difference being a largely African American (91%) cohort vs a 37% African American cohort in the HIP. Another important difference was that the provider training consisted of didactic lectures by a hypertension specialist in the Baltimore Partnership trial while the HIP trial group used online teaching. Both trials emphasized the DASH dietary pattern (JNC 7), weight reduction, alcohol reduction, increased physical activity, medication compliance, and sodium intake reduction.
The trial results were similar in that the group receiving both physician and patient intervention had the largest reduction in SBP, namely a 12-mm Hg reduction in SBP in our study vs an 8.6-mm Hg drop noted in the HIP trial.17 The key finding supported by both trials was that the effect of the patient intervention was significantly enhanced by simultaneous exposure of the study physicians to the education interventions. We acknowledge the previous observation from HIP that the training on lifestyle modification may have increased the physician confidence in their counseling abilities, thus reinforcing the lifestyle advice the patients were receiving in the patient intervention.17 We feel the didactic interactive education method used in the Baltimore Partnership trial was important because it allowed the physicians to engage in individualized case discussion with the hypertension specialist about lifestyle counseling and pharmacologic therapy.
One key aspect to our physician intervention is that it closely resembles that of a real-life model. Our participating physician group stated that in actual practice they attend a didactic CME lecture and then try to incorporate the recommendations of the lecturer into their daily practice. We feel that our simplistic approach to education will allow policy makers to adopt our model on a large scale. Other trials often employed methods that were labor intensive for physicians and their office staff, thus decreasing the probability that health care providers would participate in the identified hypertension intervention.17,18 The concept of sustainability was instrumental in designing this trial because proven interventions must be easy to implement and not labor intensive in order to truly be a long-term solution to uncontrolled hypertension.
Our trial supports many previous investigations that have found that multicomponent interventions can lead to a significant reduction in BP. We therefore make the following recommendations: (1) community physicians should explore opportunities to partner with research-intensive medical centers (and vice versa) so that minority participation in clinical trials can increase, (2) clinicians should employ a multicomponent intervention strategy since it seems to parallel the success of “combination” drug therapy vs monotherapy, (3) all policy makers providing health care delivery systems should address hypertension when targeting African Americans due to the higher prevalence and increased morbidity and mortality associated with this chronic disease, (4) the NIH should consider intensifying the ongoing effort to increase both minority physician and patient participation in clinical trials, (5) certified hypertension educators should be trained and compensated in a manner similar to certified diabetic educators and work with physicians in achieving goal BP, and (6) future research will need to concentrate on determining exactly how educational interventions reduce BP (ie, better physician compliance with guidelines, patient acceptance of lifestyle changes, improved medication adherence, etc). Lastly, more translational research is needed to reduce the “lag time” often noted between clinical guideline publication and guideline implementation into clinical practice.
Acknowledgements and disclosures:
The academic institution and authors involved in this article were supported by grant number (U01 HL79151) from the National Heart Lung and Blood Institute. Its contents are solely the responsibility of the authors and do not necessarily reflect the views and opinions of the National Institutes of Health (NIH). The investigators gratefully acknowledge the valuable contributions of the study participants, research staff members, community health advisors, and physicians and staff at the participating clinics (University of Maryland Medical System, Baltimore, MD). The authors express their special gratitude to the following people for their contributions of time, energy, and support: Dr David Stewart and Dr Louis Domenici from the University of Maryland School of Medicine for allowing the Family Practice and General Internal Medicine staff to participate in this U-01 partnership; Dr Patrice Desvigne-Nickens from the NIH; Mr Clyde Foster from University of Maryland School of Pharmacy as the study nurse patient educator; Dr Hugh Mighty from University of Maryland School of Medicine; Dr Deborah Jones from University of Maryland School of Nursing. The authors of this manuscript report that there are no relevant conflicts of interest to disclose.