a Medical Physics Department, Guy s & St Thomas NHS Foundation Trust and b King s College School of Medicine, St Thomas Campus, London, UK

Transcription

1 Devices and technology 27 Validation of the Omron 705IT (HEM-759-E) oscillometric blood pressure monitoring device according to the British Hypertension Society protocol Andrew Coleman a, Paul Freeman a, Stephen Steel a and Andrew Shennan b Background The Omron 705IT (HEM-759-E, Omron Corporation, Kyoto, Japan) is an automated oscillometric upper arm blood pressure monitor for the professional and home use markets. The aim of this study was to validate the accuracy of this device according to the British Hypertension Society and the Association for the Advancement of Medical Instrumentation SP10 validation criteria. Methods Study participants were recruited until a total of 85 were obtained that filled the blood pressure categories specified by the British Hypertension Society protocol. Recruitment to the study was from the general medical and specialist clinics and from the staff at Guy s & St Thomas Hospital in London, UK. Nine sequential same-arm blood pressure readings were taken from each participant by two trained observers, alternating between mercury reference sphygmomanometers and the Omron 705IT (HEM-759-E). The differences between the reference and test device readings, for both systolic and diastolic pressures, were compared with British Hypertension Society and Association for the Advancement of Medical Instrumentation criteria to determine the outcome of the study. Results The Omron 705IT (HEM-759-E) is graded A for systolic and A for diastolic blood pressures according to the British Hypertension Society criteria. The mean (standard deviation) of the difference between the observer and the device measurements was 0.60 (6.0) mmhg for systolic and 3.15 (6.6) mmhg for diastolic pressures, respectively. The device, therefore, also satisfies the Association for the Advancement of Medical Instrumentation SP10 standard, that requires differences of less than ± 5 (8) mmhg. Conclusions The Omron 705IT (HEM-759-E) achieved an A/A performance classification under the British Hypertension Society criteria and passes the Association for the Advancement of Medical Instrumentation requirements for the study population. It can be recommended for professional and home-use in an adult population. Blood Press Monit 11:27 32 c 2006 Lippincott Williams & Wilkins. Blood Pressure Monitoring 2006, 11:27 32 Keywords: automated devices, blood pressure monitoring, oscillometric measurement, validation studies a Medical Physics Department, Guy s & St Thomas NHS Foundation Trust and b King s College School of Medicine, St Thomas Campus, London, UK Correspondence and requests for reprints to Andrew Coleman, Medical Physics Department, The Rayne Institute, Guy s & St Thomas NHS Foundation Trust, London SE1 7EH, UK Tel: + 44 (0) ; fax: + 44 (0) ; andrew.coleman@gstt.sthames.nhs.uk Sponsorship: The study was funded by Omron Healthcare Europe B.V. Kruisweg 577, 2132 NA Hoofddorp, The Netherlands. Conflicts of interest: None Received 15 April 2005 Revised 17 August 2005 Accepted 17 August 2005 Introduction The traditional clinical use of mercury sphygmomanometers is in decline with increasing numbers of automated blood pressure (BP) monitors available [1]. Although there are now estimated to be over 400 automated devices on the UK market alone, only eighteen automated devices have been validated for accuracy against one or more of the recognized standards published by the British Hypertension Society (BHS) [2], European Society of Hypertension (ESH) [3] and the Association for the Advancement of Medical Instrumentation (AAMI) [4]. Of the eighteen validated devices, only four are recommended for clinical use, on the basis of a critical assessment of the published studies, c 2006 Lippincott Williams & Wilkins as listed by the dables Educational Trust website [5,6]. This study has examined the accuracy of the Omron 705IT following the BHS protocol. The Omron 705IT (HEM-759-E) is a widely used oscillometric BP monitor that is currently marketed for both professional and home use throughout Europe. It has previously been validated using the ESH protocol [7], which involves intercomparison of same-arm BP measurements with a reference mercury sphygmomanometer in a relatively small sample of study participants (n = 33) having systolic blood pressures (SBPs) in a range up to 180 mmhg. The study described here has followed

2 28 Blood Pressure Monitoring 2006, Vol 11 No 1 the BHS protocol to validate the Omron 705IT for clinical use. In common with the AAMI protocol, the BHS specifies a larger number of participants (n = 85) and a wider SBP range, extending above 180 mmhg. These requirements make the BHS and AAMI protocols particularly difficult and time consuming to complete, but also give them greater statistical power, which is appropriate for the evaluation of a clinically used device. Methods The tested device The Omron 705IT (HEM-759-E) is an electronic oscillometric device with a capacitive pressure sensor designed to measure pressure in the range mmhg and pulse rates in the range beats per min. The specified accuracy for the pressure scale is ± 3 mmhg. The device has a memory, which allows recording of 28 measurements, and a data output port, which can download data to a PC and a printer. Systolic and diastolic BPs are displayed, along with pulse rate on a digital readout. Inflation is controlled by an automatic pumping system and deflation is controlled by an automatic pressure release valve. The device is claimed to have the same measurement algorithm as the newer Omron M6 (HEM-7001-E, Omron Corporation, Kyoto, Japan), which has some additional features. The unit is powered by either a mains adaptor or four 1.5 V alkaline batteries (type LR6 AA). It weighs 380 g (without batteries) and measures 115 mm 117 mm 71 mm (length, width and height, respectively) and is supplied with a cuff of dimensions 140 mm 480 mm (width and length, respectively). The standard cuff is applicable to arm circumferences in the range cm. An optional additional large cuff of dimensions 174 mm 617 mm is available for arm circumferences in the range cm. Study design We have adhered to the BHS validation protocol with two modifications: (1) In common with previously published BHS validations [8,9], we have restricted testing in the low SBP range such that instead of eight participants with an SBP of < 90 mmhg and 20 participants with an SBP in the range mmhg, we recruited eight participants with an SBP r 100 mmhg and 20 participants whose SBP fell in the range mmhg. (2) We have used a semi-automated static pressure calibration of the device, as opposed to the manual dynamic pressure calibration approach described in the protocol. This modification is detailed in the following section. Before-use calibration Three identical new Omron 705IT (HEM-759-E) test devices were donated by the manufacturer for the study. Written confirmation that these represented standard production models was obtained. The pressure scales of these three test devices were calibrated on receipt. A semi-automated static pressure calibration was substituted for the manual dynamic pressure described in the protocol. A calibrated Druck 520 (GE Druck, Leicester, UK) pressure actuator, interfaced to a PC, was used to apply and maintain accurately known static pressures to the test device. The pressure scale of this calibrated actuator reads out in steps of 0.01 mmhg and the uncertainty in the absolute applied pressure is ± 0.15 mmhg at the 95% confidence level relative to the UK National Standard. Thirty different static pressures, taken from tabulated random values of pressure specified in the BHS protocol, were automatically applied to the test device in sequence at 30-s intervals. The corresponding pressure readings from the digital scale of the Omron 705IT (HEM-759-E), operated in manual mode, were recorded by an observer who was blinded from the remotely applied pressure values. All 30 control and test measurement pairs were within 3 mmhg for each of the three devices as required. In-use field assessment and after-use calibration The three test devices were then placed in hospital clinics and the static calibration repeated after each had undergone at least 400 inflations in the clinic. Again, all 30 control and test measurements remained within 3 mmhg on each device. One device was arbitrarily selected for inclusion in the subsequent clinical phase of the validation. Static device validation Recruitment of study participants Ethical approval was obtained for the study. Participants were recruited to the trial if they provided signed consent. Recruitment was at random from the population of patients attending routine outpatient clinics in a large teaching hospital and from staff responding to an advertisement posted on the hospital intranet. Recruitment of participants continued until all specified BP categories were filled (n = 76). A further nine participants were recruited to make up the total of 85, without regard to any specified BP category, as allowed under the BHS protocol. Participants excluded were those with atrial fibrillation, frequent extra systoles and sufficiently weak Korotkoff sounds that acceptable auscultation was impossible. In addition, participants were excluded if an incomplete set of BP measurements were obtained. This occurred in participants who could not tolerate the repeated arm compressions or who became anxious during the course of the measurements.

3 Validation of Omron 705IT Coleman et al. 29 Observer training and assessment The study team consisted of a medically qualified expert, who provided clinical oversight, along with two clinical scientists and a technician, who recruited patients and acted as observers, undertaking the clinical intercomparison between the automated and mercury reference devices. All team members were full-time hospital staff and undertook the validation as part of their designated duties. A healthcare assistant was employed to act as an additional observer when required. All those involved as observers were trained using the CD-ROM tutorial detailed on the British Hypertension Society web site [10]. Validation procedure The participants were seated in a warm, quiet room close to the clinic, and measurements started after consent was obtained, allowing at least 5 min rest, during part of which participants were asked their age, height, and weight for the record. The arm circumference was also measured and recorded to allow the appropriate cuff size to be selected. All pressures were measured with the patient seated, with the upper arm at heart level. Sequential same-arm measurements of SBP and DBP were independently recorded by two trained observers simultaneously using two Accoson Dekamet Mk3 mercury sphygmomanometers (manufactured by A. C. Cossor & Son Ltd, London, UK) and the Omron 705IT test device. Care was taken to ensure that observers were blinded to each other s readings. This was achieved by positioning the reference devices so that each observer could view only the scale of their own device and by ensuring they completed separate forms for recording the results. Measurements were carried out and recorded in the protocol-specified sequence, with more than 30 s intervals between measurements to minimize venous congestion, but less than 1 min to minimize variability in BP: BPA mercury (entry value used to categorize participants) BPB device (to check the automated device, not included in the analysis) BP1 mercury (observers 1 and 2) BP2 device (supervisor with test device) BP3 mercury (observers 1 and 2) BP4 device (supervisor with test device) BP5 mercury (observers 1 and 2) BP6 device (supervisor with test device) BP7 mercury (observers 1 and 2) Participants were allocated into BP categories on the basis of measurement BPA, which was subsequently discarded. Measurement BPB was also discarded, and served only to confirm that the automated test device was capable of registering a reading. The measurements BP1 to BP7 were retained for the analysis of the accuracy of the test device. Data analysis The forward and backward differences between the device and the observer measurements are calculated (in mmhg) in each case. The resulting pairs of forward and backward device observer differences are as follows: BP1 BP2, BP2 BP3 BP3 BP4, BP4 BP5 BP5 BP6, BP6 BP7 The three device observer difference values most favourable to the device (i.e. the smallest value of each listed pair, ignoring the sign) are selected for the analysis, giving a total of three device observer differences for each participant, and generating 255 difference values for the complete study of 85 participants. This analysis is done separately for both SBP and DBP. The 255 device observer differences derived from this procedure are then categorized, again ignoring the sign, to determine the percentage with values r 5, r 10 and r 15 mmhg. The protocol specifies the minimum number of device observer comparison values allowed within these classes for the test device to achieve an A, B, C or D rating. The percentage in all three categories should be equal to, or exceed, the specified minimum percentage. An A rating, for example, requires at least 60% of device observer difference values to be in the r 5 mmhg error category, at least 85% in the r 10 mmhg category and at least 95% to be in the r 15 mmhg category. This analysis is done separately for each observer and for both systolic and diastolic BP. The final grade for each systolic and diastolic BP is the better of the grades obtained by the two observers. Results Study participants A total of 143 participants were recruited to fill all the protocol-specified BP categories. Participants excluded consisted of five whose Korotkoff sounds were too weak, four in whom arrhythmias were observed, three who opted out of the study before a complete measurement set was obtained, and two for whom the test device gave repeated error readings. Forty-four participants were excluded because the relevant BP category was already full. Participants were selected consecutively after testing to best fit the required categories. The mean age of the 85 participants included in the study was 47.2 ± 14.9 years (range years), of which 38 were men and 47 women. The mean arm circumference was 30.5 ± 4.0 cm, with a range of cm.

4 30 Blood Pressure Monitoring 2006, Vol 11 No 1 Table 1 Grading criteria, mean and mean differences for test device and a sample analysis for overall pressure levels for both observers Differences between standard and test device (mmhg) Grade r 5 r 10 r 15 Mean ± SD (mmhg) Mean ± SD of differences (mmhg) Observer 1 SBP A ± ± 6.0 DBP A ± ± 6.6 Observer 2 SBP A ± ± 6.3 DBP B ± ± 6.6 Final grade SBP A ± ± 6.0 DBP A ± ± 6.6 Observer comparison SBP DBP Pressure range: SBP, mmhg; DBP, mmhg. On the basis of 255 difference values per observer for SBP, DBP. SBP, systolic blood pressure; DBP, diastolic blood pressure. Observer agreement Observer agreement levels are presented in Table 1. The BHS protocol specifies that at least 80% of measurements by the observers should be within 5 mmhg of each other, and 95% within 10 mmhg. The agreement between observers is well within these limits. Observer device agreement The overall results of the validation are given for both observers in Table 1. The percentage of device observer differences falling in the specified categories is given along with the corresponding grades for SBP and DBP for each observer. The grades are based on the BHS criteria. The mean value and standard deviation of the device and observer values and the mean and standard deviation of the device observer differences are also given. The mean and standard deviation of the device observer differences are 0.6 ± 6.0 mmhg for SBP and 3.15 ± 6.6 mmhg for DBP for the better observer. The data set, therefore, satisfies the AAMI SP10: 2002 standard requirement that SBP and DBP device observer differences be less than 5 ± 8 mmhg. The BHS protocol allows the selection of the results from the best observer, in this case observer 1 for both systolic and diastolic pressures. The Omron 705IT (HEM-759- E), therefore, achieves an overall A rating (the final grade as shown in Table 1) for both systolic and diastolic pressures. Figure 1a and b shows the Bland Altman plots [9] corresponding to the better observer measurements (observer 1) for diastolic and systolic pressures, respectively. Here, the 255 values of the difference between the device and observer (in mmhg), using the better observer values, are plotted against the mean value of the device and observer readings (in mmhg). Reference lines are given at differences of 0, ± 5, ± 10 and ± 15 mmhg. These graphs show a random scatter centred about the mean device observer difference. Fig. 1 (a) Mean diastolic pressure: device and observer 1 (mmhg) Difference: device observer 1 (mmhg) (b) Difference: device observer 1 (mmhg) Mean systolic pressure: device and observer 1 (mmhg) Plots of the pressure difference between the better observer and the test device, and the mean pressure for the test device and that observer, in 85 participants for (a) diastolic pressure and (b) systolic pressure (the number of data points is 255 for each). The two vertical lines divide the pressure scale into low, medium and high regions as defined in Table 2. They also show that there are no trends in the data. These are both required features of the Bland Altman plots for the results to be generalizable. Table 2 gives a breakdown of the results by pressure range. Applying the BHS criteria within these ranges, the Omron 705IT (HEM-759-E) achieves an A/A rating in

5 Validation of Omron 705IT Coleman et al. 31 Table 2 Grading criteria, mean and mean differences for test device at high, medium and low pressure levels using the better observer Differences between standard and test device (mmhg) Grade r 5 r 10 r 15 n Low pressure range ( < 130/80 mmhg) SBP A DBP A Medium pressure range ( / mmhg) SBP A DBP A High pressure range ( > 160/100 mmhg) SBP A DBP C SBP, systolic blood pressure; DBP, diastolic blood pressure, n, number of participants. the low and medium pressure regions, and an A/C rating in the high-pressure region ( > 160/100 mmhg). This analysis provides valuable information suggesting reduced accuracy at high diastolic pressures. These results, however, do not influence the overall A/A rating for the device. Discussion The Omron 705IT (HEM-759-E) achieved an overall A grade for both systolic and diastolic BP according to the BHS protocol. It also fulfilled the AAMI criteria. It should be pointed out that the participant population, designed here to meet BHS protocol, differed slightly from AAMI standard requirements. Specifically, the AAMI standard requires participants representing an adult population to be distributed such that at least 10% have SBP and DBP readings below 100 and 60 mmhg, respectively, and 10% of participants have SBP and DBPs above 160 and 100 mmhg, respectively. Twenty-eight out of 85 (32.9%) participants had SPB and DBPs above the upper thresholds, which is more than the 10% minimum, as required, but only eight participants (9.4%) had SBP and DBPs below the lower thresholds, which represents one participant less than required in this range. The AAMI standard for adult populations also specifies the arm circumference distribution requirements such that at least 10% of participants have an arm circumference above 35 cm, and 10% have an arm circumference below 25 cm. The distribution of participants used in this study satisfies the larger arm circumference requirement, with 11 out of 85 (12.9%) above 35 cm, but does not achieve the smaller arm circumference requirement, having seven participants (8.2%) below 25 cm, two participants less than required. The device can be considered to have passed the AAMI criteria for the study population, which, in practice, deviates only slightly from that specified in the standard to represent an adult population. On the basis of both BHS and AAMI requirements, the Omron 705IT (HEM-759-E) can, therefore, be recommended for professional as well as home-use in an adult population. Only three automated devices currently recommended by the International Advisory Body, Dable Educational Trust, have achieved a BHS A/A grade. This device will usefully increase the currently limited choice of validated automated devices. The device performed well in all ranges except in the diastolic high-pressure range, for which it received a C grade. It is not uncommon for oscillometric monitors to have a tendency to underestimate the systolic pressure. It has been suggested that this could be a function of the vascular compliance, which will be reduced in participants with arterial disease, causing alterations in the waveforms, rendering the algorithms inaccurate [10]. It is also possible that this apparent poorer performance at high pressures is an artefact resulting from the larger fluctuations that occur in BP at higher pressures. This explanation would suggest that larger device observer differences may occur during the > 30-s intervals between sequential mercury and test device readings at higher BPs. One modification to the BHS protocol in this study involved changing the low SBP range from eight participants with SBP < 90 mmhg to eight participants with SBP in the range r 100 mmhg and 20 participants with an SBP in the range mmhg to 20 participants whose SBP fell in the range mmhg. In practice, it is extremely difficult to find participants with a systolic pressure of less than 90 mmhg. This is an accepted modification to the BHS used in other published studies [10,11]. The rationale for this modification is that, given the rarity of such pressures in the general population, it is unlikely that the device will be used clinically over these low ranges. The total number of participants recruited, 143, was larger than that reported in trials completed 2 3 years ago using the same hospital clinics (126 and 95) [10,11]. This resulted from the need to continue unproductive recruitment for a long period in order to find the 20 patients in the DBP category, mmhg, and SBP category, mmhg while other categories were full. The reason for this increased recruitment difficulty in these BP categories is not known, but may reflect the increasingly effective control of BP in the local clinic populations. A further modification to the BHS protocol in this study has been to replace the cumbersome manual dynamic pressure calibration with a simple, fast and accurate semiautomated static pressure calibration. The PCcontrolled electronic calibrated reference pressure actuator has an accuracy exceeding that given by visual readings of a mercury column by at least one order of

6 32 Blood Pressure Monitoring 2006, Vol 11 No 1 magnitude. This approach is particularly appropriate as the digital scales of most modern BP monitors are capable of higher precision than the mercury reference devices [12]. A static calibration, as used here, does not confirm that the dynamic response of the test device mimics that of the falling column of the mercury reference device as intended by the BHS protocol. The dynamic response of the test device, however, is compared with that of the mercury sphygmomanometer as part of the clinical validation, where the systolic and diastolic pressures are determined from a continuously falling pressure. This redundancy of the initial dynamic calibrations in the BHS protocol is reflected in the design of the newer ESH protocol, which no longer specifies the need for calibration of the test device. A static calibration accuracy of better than ± 3 mmhg, however, is a requirement of the European Specification for Sphygmomanometers [13], and the calibration performed in this study usefully serves to confirm that the test device satisfies this standard before embarking on the expensive and timeconsuming clinical phase of the validation. The before and after clinical-use calibrations also serve to test the long-term stability of this calibration, which has important implications for its suitability for clinical use [12]. Acknowledgement Appreciation is expressed to Anne-Marie Reinders for her advice. References 1 O Brien E. Replacing the mercury sphygmomanometer. BMJ 2000; 320: O Brien E, Petrie J, Littler W, de Swiet M, Padfield PL, Altman DG, et al. The British Hypertension Society protocol for the evaluation of blood pressure measuring devices. J Hypertens 1993; 11:S2543 S O Brien E, Pickering T, Asmar R, Myers M, Parati G, Staessen J, et al. Working Group on blood pressure monitoring of the European Society of Hypertension international protocol for validation of blood pressure devices in adults. Blood Press Monit 2002; 7: Association for the Advancement of Medical Instrumentation. American National Standard for manual, electronic, or automated sphygmomanometers. ANSI/AAMI SP10: Arlington, Virginia: AAMI; O Brien E. A web site for blood pressure measuring devices: dableducational.com. Blood Press Monit 2003; 8: Dabl Educational web site. [Accessed February 2005]. 7 El Assaad MA, Topouchian JA, Asmar RA. Evaluation of two devices for selfmeasurement of blood pressure according to the international protocol: the Omron M5-I and the Omron 705IT. Blood Press Monit 2003; 8: British Hypertension Society web site. [Accessed February 2004]. 9 Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986; 1: Jones CR, Taylor K, Chowienczyk P, Poston L, Shennan AH. A validation of the Mobil O Graph (version 12) ambulatory blood pressure monitor. Blood Press Monit 2000; 5: Cuckson AC, Reinders A, Shabeeh H, Shennan A. Validation of the Microlife BP 3BTO-A oscillometric blood pressure monitoring device according to a modified British Hypertension Society protocol. Blood Press Monit 2002; 7: Coleman AJ, Steel SD, Ashworth M, Vowler SL, Shennan A. The accuracy of the pressure scale of sphygmomanometers in clinical use within primary care. Blood Press Monit 2005; 10: BS EN1060-1:1996. Specification for non-invasive sphygmomanometers. Part 1: general requirements. Brussels; 1996.