Background: Automated measurement of blood pressure (BP) in designated BP kiosks have in recent years been introduced in primary care. If kiosk blood pressure (BP) monitoring provides results equivalent to in-office BP or daytime ambulatory BP monitoring (ABPM), follow-up of adult patients could be managed primarily by self-checks. We therefore designed a comparative trial and evaluated the diagnostic performance of kiosk- and office-based BP (nurse- versus physician-measured) compared with daytime ABPM.
Methods: A trial of automated BP monitoring in three settings: a designated BP kiosk, by nurses and physicians in clinic, and by ABPM. The primary outcome was systolic and diastolic BP, with respective diagnostic thresholds of ≥135 mmHg and/or ≥ 85 mmHg for daytime ABPM and kiosk BP and ≥ 140 mmHg and/or ≥ 90 mmHg for office BP (nurse- and physician-measured).
Results: Compared with daytime ABPM, mean systolic kiosk BP was higher by 6.2 mmHg (95% confidence interval [CI] 3.8–8.6) and diastolic by 7.9 mmHg (95% CI 6.2–9.6; p < 0.001). Mean systolic BP taken by nurses was similar to daytime ABPM values (+ 2.0 mmHg; 95% CI − 0.2–4.2; p = 0.071), but nurse-measured diastolic values were higher, by 7.2 mmHg (95% CI 5.9–9.6; p < 0.001). Mean systolic and diastolic physician-measured BPs were higher compared with daytime ABPM (systolic, by 7.6 mmHg [95% CI 4.5–10.2] and diastolic by 5.8 mmHg [95% CI 4.1–7.6]; p < 0.001). Receiver operating characteristic curves of BP monitoring across pairs of systolic/diastolic cut-off levels among the three settings, with daytime ABPM as reference, demonstrated overall similar diagnostic performance between kiosk and nurse-measured values and over the curve performance for physician-measured BP. Accuracy with nurse-measured BP was 69.2% (95% CI 60.0–77.4%), compared with 65.8% (95% CI 56.5–74.3%) for kiosk BP.
Conclusions: In this study kiosk BP monitoring was not comparable to daytime ABPM but could be an alternative to in-office BP monitoring by trained nurses. The diagnostic performance of kiosk and nurse-measured BP monitoring was similar and better than that of physician-measured BP.