How to validate the accuracy of automated blood pressure monitors in children: methodology, protocol, and challenges

Arm movement affects the oscillometric signal leading to inaccurate BP measurement [10]. This is a common issue in validation sessions in younger children, together with talking between and during the measurements, and are common sources of error as children often have difficulty to remain still and silent for 20 min, which is the usual duration of a validation session. Avoiding movement and talking during the validation procedure in children is challenging and requires good communication and instructions, whereas adherence varies with age and behavioral characteristics.

Glenning et al. [7] addressed this issue in two ways. First, participants were allowed to watch videos with noise-canceling headphones aiming at improving compliance, which however is not recommended by the AAMI/ESH/ISO Universal Standard. Second, motion artifacts were identified and categorized as gross motion observed by the investigators, and movement detected by the device’s built-in motion detecting algorithm. The study found frequent movement in children and adolescents, resulting in 47 (15%) measurements discarded by the investigators and 68 (22%) by the device algorithm [7]. The identification and exclusion of these artifacts was important as it improved the device accuracy, with Criterion 1 (mean BP difference ±SD) at −0.1 ± 7.3/−1.6 ± 7.6 mmHg (systolic/diastolic) without considering artifacts, −0.3 ± 6.9/−1.8 ± 7.0 mmHg with visual detection, and −0.1 ± 6.9/−1.5 ± 6.9 mmHg with algorithm-based detection [7].

Auscultatory BP measurement faces unique challenges in children, due to their anatomical and physiological characteristics, including the difficulty in defining K5 for diastolic BP when Korotkoff sounds persist at full cuff deflation or till physiologically impossible low levels [3]. The challenges of auscultatory BP measurements in children are highlighted by the inter-observer disagreement observed in validation studies in children compared to adults. Glenning et al. [7] reported that in 7.1% of the BP readings there was a > 4 mmHg inter-observer difference (threshold for unacceptable agreement by the AAMI/ESH/ISO Universal Standard). In recent pediatric validation studies at the Hypertension Center STRIDE-7, of the Athens University in Greece, the rate of inter-observer disagreement was slightly lower (5.5% of the readings) than in the study by Glenning et al. [7], yet it was higher than in validation studies by the same team in older individuals (3.3%) (Hypertension Center STRIDE-7; unpublished data). It is important to note that 63% of the BP readings with observer disagreement in children were observed in diastolic BP, whereas in older individuals 78% in systolic BP (Hypertension Center STRIDE-7; unpublished data).

留言 (0)

沒有登入
gif