How to ensure personalized accuracy in home blood pressure devices: Should we play it by ear?

Home blood pressure (BP) monitoring has had a dramatic impact on the diagnosis and treatment of individuals with hypertension. It is a technique that is nearly a century old, pioneered by Dr George Brown of the Mayo Clinic in 1930.1 A decade later, Ayan and Goldshine foreshadowed subsequent seminal research studies in the field by concluding that home BP monitoring promoted patient education, encouraged self‐monitoring, enabled assessment of long‐term BP trends, assessed prognosis, and informed response to treatment.2

Accordingly, over the past several decades, a substantial evidence base that supports the use of home BP monitoring (HBPM) over office blood pressure monitoring (OBPM) has accumulated. HBPM can detect white‐coat and masked effect and is superior at predicting cardiovascular events.3, 4 HBPM enhances patient autonomy, enables patients to track their blood pressure, encourages adherence to medication, and provides clinicians with multiple, temporally spaced data points upon which to direct antihypertensive treatment.4 HBPM is also convenient to perform and portable. Finally, when HBPM is enhanced by telemonitoring and case management, blood pressure targets can be achieved more frequently than with usual care.5 These advantages have led multiple national and international guideline committees to strongly endorse the use of HBPM.6, 7, 8, 9

Notwithstanding all of the aforementioned benefits of HBPM, some limitations exist. Twenty‐four‐hour ambulatory monitoring, which is considered the gold‐standard method for hypertension diagnosis, better assesses nocturnal BP, which is most strongly associated with the future risk of cardiovascular events.10 Patients require education on proper technique, as many do not use a standardized measurement approach that includes proper rest, positioning, preparation, cuff size, and measurement technique.11, 12 Correct reporting of all readings by patients to providers is also uncommon; this issue could be rectified by tele‐transmitting measurements to provider electronic records.12

An additional important, and underappreciated, limitation of HBPM is that patient‐specific variability in device accuracy can compromise its usefulness in individual patients. In a study that compared the accuracy of an individual patient's home BP device to blinded, two‐observer, mercury‐based auscultation, nearly 70% of home BP devices were out by ≥5 mm Hg and 30% by ≥10 mm Hg.13 This variability occurs because home BP devices use the oscillometric technique, in which BP is estimated by applying an algorithm to filtered and processed oscillometric waveforms.14 Although oscillometric BP measurement, first described by Etienne Jules Marey in 1860,15 predates the auscultatory technique, it did not gain widespread adoption until automated BP monitors underwent technological advancements in the latter half of the 20th century. Oscillometry has since become the primary method for automated BP measurement, preferred over electronic auscultation because the latter required proper microphone positioning and was prone to sound‐related artifacts.

Oscillometric algorithms are proprietary to each device manufacturer; therefore, the oscillometric technique is less a single, standardized method than a collection of different approaches characterized by a single descriptor. Adding to the confusion is the fact that only a minority of devices (and their algorithms) currently on the market have been clinically validated,16 a process that involves confirming comparability to standardized two‐observer auscultation.17 Oscillometry, in general, performs poorly in about 20% of individuals,18 and devices from different manufacturers yield variable results.19 Therefore, patient‐specific variability in oscillometric‐derived HBPM is an important issue. Some potential explanations the variability demonstrated with oscillometry include differences in arm circumference, arterial stiffness, and pulse pressure.20

To address patient‐specific variability in blood pressure measurements, it has been proposed that patients should bring their home device to their practitioner for “calibration” testing. Since there is no practical way for the patient to have an oscillometric device “recalibrated” to their individual requirement, this is more accurately a form of validation and carries the implicit assumption that poor results would indicate a need for the patient to purchase a different device with a different algorithm from a different manufacturer. Because clinicians do not have the capacity to perform a full clinical validation according to international standards,17 the use of abridged or simplified protocols in the office setting has been proposed.21, 22 These involve performing five same‐arm sequential measurements and following a protocolized analytic plan to calculate mean differences between the device and auscultation. If the mean differences between sequential readings are <5‐10 mm Hg as per the analytic protocol, the monitor is deemed acceptable. The problem with this approach is that clinicians do not often have the time, equipment (including accurate auscultatory sphygmomanometers), or capability to perform even one standardized measurement (let alone five).23 Therefore, the likelihood that even abridged protocols will be followed properly is low.

In this issue of the Journal, Zhang et  al study an alternate proposed solution for assessing oscillometric device accuracy.24 In 99 subjects, they compare the Accutension Stetho, an electronic stethoscope that connects to a smartphone app and records Korotkoff sounds, to each of six oscillometric devices. They consider the Accutension device as a type of auscultatory reference standard because it has been previously validated according to the International Standards Organization protocol.25 The device's digital sound signal coupled with a video and playback feature enables a user to compare auscultation to oscillometrically derived BP. To perform this comparison, the oscillometric device is activated so that it performs an inflation/deflation cycle, allowing the user to perform a same‐arm simultaneous recording of Korotkoff sounds. Captured in the recording is the oscillometric device display window, and this is replayed so that the user can identify the first and last Korotkoff phases by temporally comparing the electronic sounds (or visual sound signal and pattern) with the deflating pressure values on the oscillometric device display. It is proposed that users, either patients or providers, can use the Accutension Stetho to ensure that a patient's home device measures BP appropriately and accurately in that individual. One potential use of the Accutension device is in remote patient management because the Korotkoff sounds can be recorded and sent to a provider.

Zhang and colleagues successfully demonstrate the feasibility of using the Accutension device for this purpose and find that, across the six oscillometric devices, mean auscultatory BP and mean oscillometric BP differ by at least 5 mm Hg in 63% to 79% of cases,24 a result that is consistent with prior findings.13 Does this mean that the Accutension Stetho should be used in clinical practice to validate BP monitors? We suppose it could be used, but, for trained auscultators, we do not generally see any major advantages to the use of this device over simply performing same‐arm simultaneous auscultation or sequential auscultation with a regular stethoscope. After all, the principles are identical, although the Accutension device additionally makes recording of the Korotkoff sounds possible. The major concern with the same‐arm simultaneous method is that the oscillometric device‐controlled deflation rate may be faster than the recommended 2‐3 mm Hg per second,26 and this may underestimate systolic and overestimate diastolic BP27 unless a correction factor is applied.28 In addition, there are the known challenges in performing auscultation properly, particularly if the user is untrained, noise artifact is present, or a wide auscultatory gap exists. For example, even in Figure 1 of the Zhang paper, the stethescope is shown inserted under the cuff because it is too wide for the arm of the individual, violating standardized auscultatory technique, and potentially introducing noise artifact.

Overall, the Accutension device is a reasonable first step toward optimizing HBPM, but does not deliver the type of innovation required to address the challenge of assessing the patient‐specific accuracy of home BP devices. An accurate, automated method that eliminates the need for provider‐performed comparability testing is needed. Accordingly, necessary innovations to accurately determine BP include enabling the device to automatically determine auscultatory BP (without the need to interpret an auditory or visual signal) while adjusting for deflation rate. Ideally, the device could be programmed to calculate the mean of multiple comparisons so that a user needs to only activate sequential measurements and wait for the “final answer.” This type of innovation would represent a substantial advance in ensuring that HBPM is used optimally in individual patients to guide the diagnosis and treatment of hypertension.

CONFLICT OF INTEREST

JR and RP are cofounders of a blood pressure measurement start‐up company, mmHg Inc No products are currently on the market.