Barriers to blood pressure control initiatives: Regional diversity, inadequate measurement techniques, guideline inconsistencies, and health disparities

Community and regional health improvement projects have had widely varying degrees of success. Impactful voluntary efforts are not common, making the 5-year results of the Rochester High Blood Pressure Collaborative (Fortuna, et al in this issue of the Journal) noteworthy. The primary goals of the Rochester program were to (1) improve hypertension control rates and (2) reduce health disparities. This collaborative effort involved academic physicians, community practices, hospitals, and insurers and was funded through a special hospital charge. The authors used the words “lessons learned” in their title and admonished those planning future health initiatives to (1) anticipate a response plateau, (2) identify the specific needs of important subpopulations and develop specific strategies to address and reduce health disparities, (3) recognize variation among practices treating patients with low socioeconomic status (SES), (4) anticipate new outcome measures, and (5) continually seek best practices and barriers to success. The initiative apparently succeeded: In 226 843 patients from 193 practices followed from 2011 to 2016, overall hypertension control rates improved from 61.9% to 69.5% but more so in whites (73.7%-81.5%) than blacks (58.8%-64.7%) and the effort deserves a measure of respect. However, many issues cloud the present effort and it is not entirely clear that the experience can be easily reproduced elsewhere. My concern can be arbitrarily divided into four critical areas: First, whether Rochester is different from other communities, and then whether the three most universal and substantial barriers to better BP control were successfully overcome: (1) inadequate blood pressure measurement techniques, (2) the proliferation and vagaries of ever-changing treatment guidelines, and (3) the actual impact of health disparities. Any one of these concerns presents an enormous challenge, and none were squarely addressed by Fortuna and colleagues. This not to say that other locales should not try to organize health care demonstration projects or that elements of the Rochester program cannot be emulated. What is most important is to thoroughly evaluate their methods and results in the context of their region and the known barriers to better blood pressure control. The population of Western New York is not different from the general U. S. population so it is conceivable that the experience in Rochester can be directly applied elsewhere. But the health care climate in Rochester is different from many other areas and even though it includes many different groups and institutions, I believe it behaves more like a closed than an open system. In closed systems, blood pressure (BP) control initiatives have borne fruit, early on with the Veterans Administration and more recently with Kaiser Permanente.1, 2 In the Rochester region, there are two major hospital systems but overall health care practice and policy are dominated by the influence of the University of Rochester, also the region's largest employer, and the overwhelmingly predominant regional health insurer, Excellus. The region demonstrates high standards of care and an unusual level of cooperation that most communities may not be able to match easily, in part because a very large number of local practitioners are Rochester alumni and faculty. There are at least three major areas in which “dirty secrets” of real-world hypertension practice have created formidable barriers to better blood pressure control; these issues were not adequately developed in the present work. Perhaps the quintessential barrier to controlling hypertension is competent detection of the condition. More than a century after the establishment of sphygmomanometry, rational measurement techniques are not being applied in the United States. The most fundamental issue is the intrinsical variability of blood pressure, which changes to meet varying physiological and psychological demands and therefore cannot really be treated as a constant value. The use of traditional office BP measurements to approximate real-world blood pressures is also clearly inadequate.3, 4 We have known for decades that hypertension-attributable end-organ damage is more strongly associated with nonoffice BP measurements (either 24-hour monitoring or self-measured home blood pressures) than office BP measurements.4-6 On top of this is the failure of most health care professionals to adhere to the rules of blood pressure measurement.7 A major part of the Rochester initiative was to improve substandard office BP measurement techniques. An obvious benefit of better measurement is the minimization of the “white coat effect,” which not only confounds the diagnosis of true hypertension but is not closely associated with end-organ damage.7 Standard office technique involves using proper cuff size, leaving patients undisturbed in quiet rooms for 5 minutes or more before measurement, having patients sit in a chair with back supported and feet on floor, avoiding talking, and recording multiple measurements and their mean at each visit.7 These elements are rarely met in busy offices. But there is another hidden carryover of our traditional sloppiness in BP measurement: Simply improving blood pressure measurement techniques actually leads to lower office BP values, which also increases apparent blood pressure control rates. In the Canadian experience, the difference between proper and improper measurement techniques is about 10/7 mm Hg,8 and in the Spanish experience, there is a difference of 17/8 mm Hg between daytime ambulatory BP values and office measurements.9 Because of its superior reproducibility and predictive power, automated office BP (AOBP, in which self-actuated measurements are performed in the absence of a practitioner) is now the BP measurement standard in Canada.10 This technique yields BP values very similar to ambulatory daytime values, which are about 5-10 mm Hg less than usual office BP values.8, 11 With respect to how such difference in measured BP will affect hypertension control rates, the 2017 American College of Cardiology/American Heart Association guideline recommended a 10 mm Hg decrease in the hypertension treatment threshold (from 140/90-130/80 mm Hg); accordingly, the nominal US BP control rates decreased from almost 70% to about 47%.7, 12 It should thus be readily apparent that the apparent improvement in BP control in Rochester can be attributed, at least in part, to improved blood pressure measurement. Another core issue in hypertension management is the acceptance of a universally accepted blood pressure target from which a diagnosis can be made and a control rate established (despite the obvious uncertainty caused by different patterns of BP variation in individual patients). Somewhat surprisingly, practicing physicians often decide on their own BP standards rather than adopt guideline values. Even those who follow guidelines are confused by the variation in published blood pressure thresholds over the last 5 years, for reasons more political than evidentiary. From 1977 until 2013, hypertension guideline development was the province of National High Blood Pressure Educational Program and its effector arm, the Joint National Committee on the Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (the JNC), a collaborative effort of the National Heart, Lung, and Blood Institute and eventually over 30 professional organizations.13 Keeping diastolic BP <90 mm Hg was the major focus of their first three reports but from JNC IV (1988) to JNC7 (2003), a combined BP threshold of 140/90 mm Hg was recommended.14 Only since 2000 was it firmly established that systolic hypertension was a greater public health concern in adults than diastolic hypertension.15 The NIH abandoned guideline development in 2013 and since then, there has been wide gyration in recommended blood pressure targets. The Rochester program adopted the recommendations of the problematic and unsponsored “JNC8” initiative,16 whose errors were compounded by their extension to the 2015 HEDIS standards. The “JNC8 panel report” was anything but an “evidence-based guideline” as was advertised; their recommendations were predominantly based on opinion, which was split. A slight majority of the panel decided that benefit of therapy had not been specifically proven for individuals over age 60 if systolic BP was <150 mm Hg. Parenthetically, this relaxation of the systolic threshold increased apparent BP control rates substantially since systolic blood pressure increases so dramatically with age. The new threshold was so controversial, however, that a minority report was issued urging practitioners to use the traditional blood pressure threshold of 140/90 mm Hg at any age.17 In most of the US hypertension guidelines, the apparent philosophy has followed the latest large NIH-sponsored trial; for JNC8, the ACCORD study had suggested that BP values below 120 mm Hg (in diabetics) did not improve outcomes.18 Contrast that document to the 2017 ACC/AHA guidance and the results of the SPRINT study, which drew the opposite conclusion of ACCORD: low BP is beneficial (at least in nondiabetics) over age 60.19 As a result, BP targets in older adults in the last 2 US guidelines have swung from 150/90 to 130/80 mm Hg. Not only do such gyrations engender a lack of confidence, but also they cause headaches for patients and researchers alike. Hypertension control rates in the Rochester program are now hypothetically obsolete, although the relative changes still have interpretative value. Health disparities are essentially universal, whether considered from a local, regional, or worldwide perspective.20 The Rochester group reported that there were substantial disparities in practice among the providers who cared for patients with health disparities but did not elaborate much further. Clearly, patients with low socioeconomic status (SES) are at increased risk of excess cardiovascular disease1-11 yet low SES also interacts with many other psychosocial and sociocultural factors to create this high-risk phenotype. The negative impact of low SES in childhood may even extend to future generations, further suggesting that a variety of nongenetic familial and cultural influences should be considered.21 Actual income may not be the most critical factor because individuals who subjectively report low SES have greater cardiovascular risk even after adjustment for actual income.22 Similarly, the strong negative correlation between income and blood pressure found in the United States is not found in Canada.23 The resources for low SES individuals to successfully cope with their situations are obviously diminished; low SES is associated with greater workplace “strain” and higher ambulatory blood pressure,24 and low educational achievement may exacerbate the problem in developed and underdeveloped countries.25, 26 The interaction of SES and race is complex.21, 27-30 With better SES assessment tools, low SES itself often supersedes the impact of race or ethnicity, a concept we included in guidance to New York State Medicaid providers about treating hypertension in minority populations.20 Still, young African Americans who believe they are treated unfairly have greater “trait anger” and higher blood pressures31 and those who believe themselves to be victims of discrimination have higher nocturnal blood pressures.32 The Rochester experience may contain clues about health disparities, such as the lesser impact of the program on blacks compared to whites, but the causes and solutions are not clear. Their major thrust was probably provider education, which they apparently achieved in varying degrees across diverse practices. Were the apparent racial differences in blood pressure control related to both providers and patients, or perhaps to their interactions? It has been my impression for many years that a massive “disconnect” exists between many providers and their patients with respect what patients actually take away from health encounters. If a provider loudly berates a disadvantaged patient for not following lifestyle modification instructions, is the message “I really care about you” or “You are an inferior person”? We need much more insight regarding what is sent and what is received. For my money, the core commodity is trust. In hypertension care in minority populations, enhancing trust can improve blood pressure control.33, 34 It is also my experience that trust is infrequently considered by providers and the importance of trust in patient-provider relationship is concept is extremely under-represented in the medical literature and in provider training. Trust does not automatically result from learning stylistic communication tricks alone because generic training in communication skills does not improve health behaviors, especially in minority populations.34 What is also needed is greater understanding of patient beliefs and attitudes, even if they are difficult to understand or incorrect.35 Personal engagement is easily discerned by patients, and this is probably the first step toward trust because the underlying message is “I really do care.” Given the formidable barriers to success, can any local or regional initiative be truly useful? Perhaps surprisingly, my answer is “definitely yes, we need more!” By virtue of the “Hawthorne effect” alone, most local health initiatives will achieve some measure of improvement. But can they provide more? These observational experiments, flawed as they may be, create a real-world learning laboratory and by virtue of their intrinsic differences, are reasonably likely to yield preliminary insights into causality and best practices. By examining the variation in their methods and the corresponding differences in their outcomes, trends can be discerned and, if necessary, tested experimentally later. Vital data will be needed about the local population and health care system and reasonable measurement techniques and goals are essential for any health indicator. Divergences in methods and results are thus not intrinsically negative. Organizers must also continually ask the overarching question “Can our experience be exported and applied beyond our community?” If the ultimate goal is to establish best practices, how else can we obtain enough real-world information to move forward? No conflict of interest.