ASSESSMENT OF THE PERFORMANCE OF 2 OSCILL OMETRIC BLOOD-PRESSURE MONITORS USING A SIMULATOR

Using the PTB simulator, which emits real signals from patients, we ex amined the precision of the oscillometric blood pressure measurement w ith the Dinamap 1846 (Critikon) and the HP M-1008B (Hewlett Packard). For this purpose we simultaneously registered invasive arterial pulsew ave, cuff pressure and cuff pressure oscillations of 20 patients from our intensive care unit and stored them in the database of the simulat or. The invasive reference blood pressure values were determined follo wing the recommendations given by the Association for the Advancement of Medical Instrumentation. The invasive system showed a cut-off frequ ency of 35 Hz; the damping constant was 0.21. With 49 record signals f rom patients we carried out 15 simulated measurements each. From a tot al of 49 bio-signals from patients the Dinamap 1846 was able to proces s 41 signals and the HP M-1008B 47 signals. The mean error of the osci llometric blood pressure measurement of the systolic, diastolic and me an arterial pressure amounted to -2.50 mmHg, 3.35 mmHg (P<0.05) and 1. 51 mmHg with the Dinamap 1846 and to -8.5 mmHg (P<0.001), -5.15 mmHg ( P<0.001) and -5.58 mmHg (P<0.001) for the HP M-1008B. The 95% confiden ce limit for the systolic, diastolic and the mean arterial pressure am ounts to 56 mmHg, 30 mmHg and 35 mmHg for the Dinamap 1846 and 50 mmHg , 38 mmHg and 35 mmHg for the HP M-1008B. The differences between that two instruments could be caused by the different algorithms for the c alculation of blood pressure values and different artefact detection a nd elimination techniques. The results of the performance tests we ach ieved with the PTB simulator correspond to the results of other clinic al examinations. The American Association for the Advanecement of Medi cal Instrumentation recommends a maximum mean error of 5 +/- 8 mmHg. N one of the examined instruments lay within these limits. Due to the sy stematic and stochastic errors, we think that the Dinamap 1846 (Critik on) and the HP M-1008B (Hewlett Packard) do not achieve performance le vels that are adequate for measuring critically ill patients.