Doctors Rely on Clinic BP, Automated Tools at Patient’s Peril

Two studies presented at Hypertension 2020 highlight the importance of out-of-office monitoring for diagnosing high BP.

Doctors Rely on Clinic BP, Automated Tools at Patient’s Peril

Two new studies presented this week at the virtual Hypertension 2020 Scientific Sessions emphasize the importance of ambulatory blood pressure monitoring (ABPM) for diagnosing hypertension—one showed that health professionals rely too much on highly variable office measurements and another cast doubt on the reliability of automated office blood pressure (AOBP) readings.

In the first study, led by Beverley Green, MD (Kaiser Permanente Washington Health Research Institute, Seattle), researchers found that among the health professionals they surveyed, most rely on clinic measurements of blood pressure to make a diagnosis of hypertension despite guideline recommendations to use either 24-hour ABPM or home blood pressure monitoring.

Across 10 primary care medical centers part of Kaiser Permanente Washington health system, 96% of respondents said they “always or almost always” rely on clinic measurements for diagnosing hypertension, although the majority of physicians and midlevel providers said they would prefer using ABPM if it was readily available.

The 2017 American College of Cardiology/American Heart Association (ACC/AHA) hypertension guidelines recommend that healthcare providers use out-of-office blood pressure measurements to confirm the diagnosis of hypertension (class 1A recommendation). The US Preventive Services Task Force (USPSTF) also recommends out-of-office blood pressure measurements before making a new diagnosis of hypertension, either with 24-hour ABPM or home monitoring (grade A).

To TCTMD, Green said the survey was designed to evaluate knowledge about the clinical guidelines and to understand current practice patterns about BP diagnostic tests. The survey included 119 physicians, physician assistants, and advanced practice registered nurses, all of whom could not only diagnose hypertension but also prescribe medication, as well as 168 medical assistants, licensed practical nurses, and registered nurses who do not prescribe antihypertensive therapy.

“We found that among both groups, most of the time they made a hypertension diagnosis based on office measurements,” said Green. “They also noted that they trusted office measurements taken with a stethoscope the most.” Of the physicians and midlevel providers, 61% did say that if 24-hour ABPM was available they would prefer this for making a diagnosis of hypertension.

When it came to diagnosing hypertension using 24-hour ABPM, 49.6% of physicians and midlevel providers said the threshold for diagnosis was 140/90 mm Hg as opposed to the cut-off of 130/80 mm Hg set out in the most recent ACC/AHA clinical guidelines. Just 6.7% of physicians and midlevel providers surveyed correctly identified the 24-hour ABPM threshold for hypertension diagnosis.     

In their review of the evidence, the USPSTF stated that 24-hour ABPM is the “gold-standard” noninvasive blood pressure measurement because it is the best predictor of future cardiovascular events.

“The reason for that is because blood pressure is really variable,” said Green. “Systolic blood pressure can vary, on average, by 30 to 35 mm Hg throughout the day for an average person. If your blood pressure is higher, it can be more than that.” Clinic measurements capture only a slice in time and the concern with relying only on clinic measurements is misdiagnosing individuals with white-coat hypertension. “It goes both directions, though,” she said. “It could lead to under- and overdiagnosis.”

Leaving the Room in SPRINT

The second study, also led by Green, examined the impact of having a health professional in the room when taking measuring blood pressure using an automated office blood pressure (AOBP) machine. Although the AOBP, which takes three to five measurements in a row that are then averaged, is intended to reduce the risk of white-coat hypertension, Green said there is controversy about whether the presence of a physician, nurse, or other healthcare provider still impacts blood pressure recordings, a question that arose during the landmark SPRINT trial.

In SPRINT, the protocol required health professionals to leave the room during a 5-minute rest period before the automated machine started up to record blood pressure, but it later emerged that attendants didn’t always leave the room. This left some confusion about the impact of attendance on the measurements, said Green.

To look at this question, the researchers enrolled 133 participants aged 18 to 84 years without a confirmed hypertension diagnosis but who had elevated blood pressure identified using electronic health records from Kaiser Permanente. All patients had received 24-hour ABPM (mean daytime ABPM 138.2 mm Hg) and were randomized to one of two groups when they returned the device to clinic: 1) randomized first to “attended” automated blood pressure measurements where the health professional stayed in the room and then to “unattended” measurements; or 2) randomized first to unattended automated blood pressure measurements and then to attended readings. The researchers also tested the effects of giving patients 5 or 15 minutes of rest before starting the automated machine.    

Overall, AOBP was 3.7 and 3.9 mm Hg lower with attended and unattended measurements compared with daytime ABPM (P < 0.001 for both). However, there was no significant within-person difference in automated systolic or diastolic blood pressure when measured with providers in or out of the room. Additionally, there was no significant difference in AOBP when patients were given 5 or 15 minutes of rest before initiating the measurement, although both measurements were significantly lower than daytime ABPM.

“We found that attendance didn’t matter at all,” said Green. Additionally, “the 15 minutes of rest made almost no difference.” Given that the automated readings—either with or without a provider in the room or with 5- or 15-minutes of rest prior to starting the machine—were lower than the measurements obtained with ABPM, “that meant that a lot of cases of hypertension were missed,” said Green.

Overall, the sensitivity and specificity for hypertension diagnosis based on a daytime mean ABPM systolic ≥ 135 mm Hg or diastolic ≥ 85 mm Hg was 71% and 57% with unattended AOBP, respectively. For the attended AOBP measurements, the sensitivity and specificity was 68% and 51%, respectively. Automated readings missed 21% of patients with hypertension and overdiagnosed 12% of patients enrolled in the study.

To TCTMD, Green said their study shows that automated blood pressure measurements shouldn’t be used to replace ABPM or home monitoring, given its relatively poor performance and lack of correlation with long-term outcomes.

Daichi Shimbo, MD (Columbia University Irving Medical Center, New York, NY), who moderated the poster session, highlighted the negative predictive values of 12% and 14% with unattended and attended AOBP sessions, saying “it tells me we’re going to be missing a lot of people with masked hypertension and that is a concern with AOBP, especially if the rest periods are extended to 5 minutes or more.” This highlights “just how important ABPM is” for diagnosing hypertension, he said.

Given that, Shimbo asked just where automated office blood pressure measurements should fit in clinical practice. For her part, Green said she’d discourage its use, agreeing with Shimbo that AOBP will miss not only people with masked hypertension but also those with high blood pressure in general.   

Michael O’Riordan is the Managing Editor for TCTMD. He completed his undergraduate degrees at Queen’s University in Kingston, ON, and…

Read Full Bio
  • Green B, Anderson ML, Cook AJ, et al. Blood pressure checks and diagnosing hypertension: provider knowledge, beliefs, and practices. Presented at: Hypertension 2020. September 10, 2020.

  • Green B, Anderson ML, Cook AJ, et al. Automated office blood pressure: impact of attendance and time on blood pressure and hypertension diagnosis. Presented at: Hypertension 2020. September 10, 2020.

  • Green reports no relevant conflicts of interest.