Professors: The World Is Not Flat


Steven A. Yarrows, MD, Chelsea International Medicine, 128 Van Buren Street, Chelsea, MI 48118

The cornerstone of hypertension treatment is lowering the blood pressure (BP), which clearly has resulted in improvement in cardiovascular outcomes. The decades of research have validated this principle, yet the “bench to practice” translation of this research is hampered by different techniques of BP measurement. The most abundant hypertensive research has been in a clinical research setting in which the patient has been seated for a minimum of 5 minutes and triplicate readings are measured and averaged. This average BP is the basis of clinical decisions about treatment and clinical cardiovascular outcomes studies.

In contrast, the vast majority of BP measurements occur outside of research settings in an office by a single measured BP. In this setting in the United States, the BP is measured by a medical assistant (MA) shortly after the patient is escorted to an examination room. We need to address the “elephant in the room,” that BP measurement within nonresearch settings is not the same as within academic institutions that utilize research protocols. Perhaps this is often not understood by some academic physicians whose office setting may be quite different. A recent journal submission regarding BP measurement had the following criticism by a reviewer: “What is more, in these types of settings, blood pressure measurements are generally made by nurses and medical support staff trained in this field, not by medical assistants.” Certified MA are initially trained in proper BP measurement techniques, however this may not translate into later actual practice. Additionally, many practices do not employ personnel that have this specific BP measurement training. There is scant discussion by hypertension specialists about this difference in techniques and the clinical importance of this.

Some observations have indicated higher measurements from “usual” methods by 6.2 to 15.5/4.7 to 11.6 mm Hg compared to “research nurses,”1–3 which is a potentially clinically significant variance. The common doctrine of physician-measured BP is higher by 6 to 10 mm Hg for systolic BP and 8 mm Hg for diastolic BP than “nurse” measured readings.4,5 Therefore, it is believed that physician-measured BP will be higher than MA-measured BP in nonacademic offices. Is this correct?

My 28 years of clinical experience indicate otherwise. I performed a retrospective, electronic chart review of 888 patients from my practice over an approximate 10 month period to investigate this observation. I found that the MA average BP was 9.4/6.8 mm Hg higher than the physician readings (P<.05). The MA-measured, systolic BP readings were ≥140 mm Hg whereas the physician-measured readings were under this threshold 27.0% of the time. The MA-measured diastolic was ≥90 mm Hg, whereas the physician-measured readings were also under this criteria in 16.0% of patients. There was a ≥15 mm Hg difference between the MA and physician readings in 38.5% of systolic and 34.8% of diastolic measurements. If the treatment decisions were based on the MA readings, a significant amount of patients would have incurred an inappropriate medication change.

This was not a randomized, controlled trial and there were many potential factors that could have resulted in the differences. The most likely factors were that the physician readings were performed after a longer period of rest in the examination room, and use of a proper-sized, large cuff. I believe that this is not unique to my practice and may in fact occur in most nonresearch practices.

Additionally, we should acknowledge that primary care physician practices are designed for high volume, expense-restricted visits due to significant overhead and current low reimbursement rates which may be compounded by the anticipated shortage of providers. A proper BP measurement should take approximately 1 minute (assuming a range of 130 mm Hg at 2 mm Hg/second deflation), which would be 7% to 10% of the allotted office visit, not accounting for the suggested minimum of 5 minutes of rest and triplicate readings (53%–80% of the visit). This is not feasible due to limited examination room space. Using calibrated, automatic devices that take repetitive readings has been demonstrated to be accurate compared to awake ambulatory BP monitoring.6–8 Unfortunately, these devices are currently expensive and would increase the patient time within the examination room.

Could future BP measurement devices be used prior to the visit in the lobby during the patient wait? I asked 235 patients to measure BP before and after a physician visit in the waiting room of my office using an automatic BP monitor and found no significant difference in the average readings (0/0 mm Hg, P>.05) before and after the visit, indicating that the readings were reproducible. These waiting room devices would need to have a minimum 5 minute relaxation period prior to the first reading and automatically average the triplicate readings. Proper cuff size would have to be assured. The readings could be transmitted to an electronic medical record or recorded for paper charts. Conceptually this is feasible, as most patients wait prior to being taken to the examination room. The question that remains: is this accurate?

When Columbus tried to tell other learned people that the “World is not flat,” this was met with skepticism. We need to drop any skepticism that BP is measured in primary care offices the same as in research settings and determine how to address this problem. So how do we validate the best method to accurately measure BP within a clinical primary care office? Now is an excellent time to look at workflow within offices, as progressively more primary care offices are converting to electronic medical records, greatly necessitating workflow changes. I would like to organize interested people to design and implement appropriate research to address this issue. If you have interest, please contact me at: Dr Thomas Pickering stated in 1994 that “the measurement of blood pressure is much too serious to be left to physicians.”9 Perhaps in a nonresearch clinical practice this should be amended to “the measurement of blood pressure is much too serious to be left to medical assistants” that usually do not have time to properly measure the BP. A 30% overdiagnosis of hypertension for systolic BP is unacceptable.