UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Improved noninvasive determination of blood pressure by oscillometry in the presence of motion artifacts Bussani, Carlo Robert 1986

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-UBC_1986_A7 B88.pdf [ 6.27MB ]
Metadata
JSON: 831-1.0065057.json
JSON-LD: 831-1.0065057-ld.json
RDF/XML (Pretty): 831-1.0065057-rdf.xml
RDF/JSON: 831-1.0065057-rdf.json
Turtle: 831-1.0065057-turtle.txt
N-Triples: 831-1.0065057-rdf-ntriples.txt
Original Record: 831-1.0065057-source.json
Full Text
831-1.0065057-fulltext.txt
Citation
831-1.0065057.ris

Full Text

IMPROVED NONINVASIVE DETERMINATION OF BLOOD PRESSURE BY OSCILLOMETRY IN THE PRESENCE OF MOTION ARTIFACTS By Carlo Robert Bussani B.A.Sc, The U n i v e r s i t y of B r i t i s h Columbia A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF APPLIED SCIENCE i n THE FACULTY OF GRADUATE STUDIES (Department of E l e c t r i c a l Engineering) We accept t h i s t h e s i s as conforming to the r e q u i r e d standard THE UNIVERSITY OF BRITISH COLUMBIA June 198 6 © Carlo Robert Bussani, 1986 In presenting t h i s thesis i n p a r t i a l f u l f i l m e n t of the requirements for an advanced degree at the University of B r i t i s h Columbia, I agree that the Library s h a l l make i t f r e e l y available for reference and study. I further agree that permission for extensive copying of t h i s thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. I t i s understood that copying or publication of t h i s thesis for f i n a n c i a l gain s h a l l not be allowed without my written permission. Department of /^ecrfV/ogY Ae£,ri<j The University of B r i t i s h Columbia 1956 Main Mall Vancouver, Canada V6T 1Y3 Date j j l i s , m DE-6 r3/8'l^ ABSTRACT Blood pressure measurements provide information regarding the status of the c a r d i o v a s c u l a r system to a i d i n diagnosing p a t h o l o g i c a l c o n d i t i o n s , maintaining p a t i e n t s t a b i l i t y during surgery, and adapting tourniquet pressures. However, the performance of commercially a v a i l a b l e techniques of measuring blood pressure i s often degraded during c l i n i c a l measurements i n the presence of motion a r t i f a c t s . A prototype system based on o s c i l l o m e t r y that uses algorithms which take advantage of a multi-bladder c u f f has the p o t e n t i a l to improve measurements of blood pressure i n noisy environments by d e t e c t i n g noise and r a p i d l y e s t i m a t i n g blood pressure using only data uncorrupted by noise. This system and strategy, implemented i n one v e r s i o n to tr a c k limb o c c l u s i o n pressure (the minimum c u f f pressure r e q u i r e d to occlude u n d e r l y i n g a r t e r i e s ) , was evaluated along with a t y p i c a l commercial, JLV a v a i l a b l e o s c i l l o m e t r i c blood pressure monitor to compare t h e i r a b i l i t y to t r a c k limb o c c l u s i o n pressure i n the presence of v a r y i n g noise c o n d i t i o n s encountered during lab t r i a l s and c l i n i c a l t r i a l s . Results showed that the prototype c o n s i s t e n t l y estimated limb o c c l u s i o n pressure more r a p i d l y , more accurately, and more r e l i a b l y than the o s c i l l o m e t r i c monitor i n noise c o n d i t i o n s t y p i c a l of s u r g i c a l procedures. The r e s u l t s a l s o i n d i c a t e that the prototype i s f e a s i b l e f o r i n c o r p o r a t i o n i n t o an adaptive tourniquet. i i TABLE OF CONTENTS ABSTRACT i i LIST OF TABLES v LIST OF FIGURES v i ACKNOWLEDGEMENT v i i i 1. INTRODUCTION 1 1.1 P h y s i o l o g i c a l S i g n i f i c a n c e of Blood Pressure Measurements 1 1.2 Diagnosing Disease with Blood Pressure Measurements 3 1.3 Blood Pressure Measurements During Surgery 6 1.4 Blood Pressure Measurements Used to Adapt Tourniquets .-. . . 7 1.5 Scope of the Thesis 9 2. REVIEW OF BLOOD PRESSURE MEASUREMENT TECHNIQUES 11 2.1 D i r e c t Blood Pressure Measurements 11 2.2 I n d i r e c t Blood Pressure Measurements 13 2.2.1 Auscultatory Method 14 2.2.2 Ultrasound 15 2.2.3 O s c i l l o m e t r y 17 2.3 I n d i r e c t Blood Pressure Monitors i n Use at Vancouver General H o s p i t a l 20 2.3.1 Dinamap Model 845 A d u l t / P e d i a t r i c V i t a l Signs Monitor 21 2.3.2 Hewlett Packard Model 78354A P a t i e n t Monitor 23 2.3.3 Datascope Accutorr Model 2A Noninvasive Blood Pressure Monitor 24 2.3.4 Physio-Control Model V i t a l Signs Monitor 25 2.3.5 D i s c u s s i o n of Commercial Noninvasive Blood Pressure Monitors 25 3. HARDWARE DEVELOPMENT OF THE PROTOTYPE 28 3.1 General Hardware Design Philosophy 28 3.2 Blood Pressure Cuff 31 3.3 Tubing 35 3.4 Pressure C o n t r o l Means . 36 3.5 Pressure Transducer 39 3.6 A m p l i f i e r s 41 3.7 Microprocessor System 44 3.8 A n a i o g - t o - D i g i t a l and D i g i t a l - t o - A n a l o g C a p a b i l i t y 44 3.9 T o t a l System 44 4. SOFTWARE AND SIGNAL PROCESSING 48 4.1 General Software Design Philosophy 48 4.2 I n i t i a l Data A c q u i s i t i o n 49 4.3 S i g n a l and Noise C h a r a c t e r i z a t i o n 49 4.4 Noise Reduction Techniques 54 4.4.1 Noise Subtraction 54 4.4.2 Adaptive F i l t e r s 56 4.4.3 S i g n a l Averaging 57 4.4.4 S i g n a l Estimation 58 4.5 Noise Detection 59 4.6 S i g n a l Information 60 4.7 B a s e l i n e Wander 61 4.8 Algorithms f o r Estimating Blood Pressure 63 4.8.1 Estimation of S y s t o l i c Pressure .' 65 4.8.2 Estimation of Mean A r t e r i a l Pressure 66 4.8.3 Estimation of D i a s t o l i c Pressure 66 4.9 Implementation of an Algorithm f o r Rapidly Estimating Limb Occlusion Pressure i n Noisy Environments . 67 5. A COMPARATIVE EVALUATION OF DIFFERENT METHODS OF ESTIMATING LIMB OCCLUSION PRESSURE 75 5.1 Overview 75 5.2 The P r o t o c o l f o r Lab T r i a l s 76 5.3 The P r o t o c o l f o r C l i n i c a l T r i a l s 7 8 i i i 5.4 Performance C r i t e r i a 78 5.5 Dis c u s s i o n of Results 81 5.5.1 Results of Lab T r i a l s of the O s c i l l o m e t r i c Monitor 81 5.5.2 Results of Lab T r i a l s of the Prototype 85 5.5.3 Results of C l i n i c a l T r i a l s 85 6. CONCLUSIONS 95 6.1 Summary 95 6.2 Suggestions f o r Further Research 97 6.2.1 A m p l i f i e r Improvements 97 6.2.2 Improvements i n Searching Techniques 98 6.2.3 Improvements i n Safety 100 6.2.4 Improvements i n Noise Detection 101 6.2.5 Improvements i n Threshold Settings 102 6.2.6 Improvements i n Cuff Design 103 6.2.7 Other Flow Detection Methods 10 4 REFERENCES 106 APPENDIX 1 PROGRAM LISTING I l l i v LIST OF TABLES Table 5 . 1 Results of lab t r i a l s used to assess the accuracy of both the o s c i l l o m e t r i c monitor and the prototype. 82 Table 5 . 2 Results of lab t r i a l s used to assess the r e l i a b i l i t y of both the o s c i l l o m e t r i c monitor and the prototype. 82 Table 5 . 3 Results of lab t r i a l s used to assess the s a f e t y of both the o s c i l l o m e t r i c monitor and the prototype. 83 Table 5 . 4 Results of c l i n i c a l t r i a l s used to assess the speed of both the o s c i l l o m e t r i c monitor and the prototype. 83 Table 5 . 5 The d e r i v a t i o n of tourniquet pressure f o r adaptive schemes that use d i f f e r e n t blood pressure estimates. 93 LIST OF FIGURES F i g . 1.1 T y p i c a l ECG and blood pressure waveforms. 2 F i g . 2.1 O s c i l l o m e t r i c p ulses. 18 F i g . 2.2 V a r i a t i o n of o s c i l l o m e t r i c pulses with c u f f pressure. 19 F i g . 3.1 Components of a general o s c i l l o m e t r i c system f o r measuring blood pressure. 30 F i g . 3.2 Noise a r t i f a c t s i n an o s c i l l o m e t r i c system. 32 F i g . 3.3 The multi-bladder c u f f . 32 F i g . 3.4 M u l t i - b l a d d e r c u f f s synthesized from tourniquet c u f f s . 34 F i g . 3.5 Mounting of transducers and valves near the limb. 37 F i g . 3.6 Pressure c o n t r o l u n i t . 38 F i g . 3.7 Cobe pressure transducers. 40 F i g . 3.8 Low-gain D.C. a m p l i f i e r f o r b a s e l i n e c u f f pressure s i g n a l s . 42 F i g . 3.9 Drive c i r c u i t r y f o r r e l a y and v a l v e s . 45 F i g . 3.10 Components of the o v e r a l l prototype system. 46 F i g . 3.11 The prototype system. 47 F i g . 4.1 Histogram of a noise waveform. 51 F i g . 4.2 Histogram of an o s c i l l o m e t r i c s i g n a l . 52 F i g . 4.3 Frequency a n a l y s i s of noise and s i g n a l waveforms. 53 F i g . 4.4 Simultaneous chart recording of waveforms from two bladders. 55 F i g . 4.5 A highpass FIR f i l t e r f o r removal of b a s e l i n e wander r e q u i r i n g 2 m u l t i p l i c a t i o n s and N-l a d d i t i o n s . 62 F i g . 4.6 Further reduction i n computation f o r a FIR f i l t e r used to remove b a s e l i n e wander. 64 F i g . 4.7 T y p i c a l o s c i l l o m e t r i c waveforms observed i n the multi-bladder c u f f . 68 F i g . 4.8 Flowchart f o r the algorithm used to tr a c k limb o c c l u s i o n pressure. 70 F i g . 4.9 D i s p l a y graphics f o r the prototype. 74 F i g . 5.1 The use of the multi-bladder c u f f during c l i n i c a l t r i a l s . 7 9 F i g . 5.2 Results of a c l i n i c a l t r i a l with the o s c i l l o m e t r i c monitor. 87 F i g . 5.3 Results of a c l i n i c a l t r i a l with the prototype. 88 v i F i g . 5.4 V a r i a t i o n of limb o c c l u s i o n pressure with limb p o s i t i o n f o r estimates made by the o s c i l l o m e t r i c monitor during a no i s e - f r e e l a b t r i a l . 89 F i g . 5.5 V a r i a t i o n of limb o c c l u s i o n pressure with limb p o s i t i o n f o r estimates made by the prototype during a n o i s e - f r e e l a b t r i a l . 90 F i g . 6.1 A digitally-programmable high-gain a m p l i f i e r . 99 v i i ACKNOWLEDGEMENT I would l i k e to thank James A. McEwen, Ph.D., P.Eng., C.C.E. f o r h i s h e l p f u l advice throughout the course of t h i s t h e s i s and Brian Day M.D., F.R.C.S. (C) f o r h i s c o l l a b o r a t i o n during c l i n i c a l t r i a l s . I would a l s o l i k e to thank the f o l l o w i n g people and groups f o r t h e i r k i n d a s s i s t a n c e : Mr. Ken Yip, B.A.Sc, P.Eng. f o r h i s e f f o r t s i n proc u r i n g equipment v i t a l f o r the p r o j e c t ; Mr. Rob McNeil, M.Sc., P.Eng. f o r h i s a i d with the hardware and software systems a v a i l a b l e at Vancouver General H o s p i t a l ; Mr. N e i l Cox, M.A.Sc. f o r h i s help with using s i g n a l p r o c e s s i n g programs; the s t a f f of the Biomedical Engineering Department at Vancouver General H o s p i t a l f o r t h e i r p a r t i c i p a t i o n i n lab t r i a l s ; and the engineering group of Andronic Devices Ltd. f o r t h e i r u s e f u l suggestions, support, and p a r t i c i p a t i o n i n p r e l i m i n a r y t r i a l s . I am very g r a t e f u l to the Natural Sciences and Engineering Research C o u n c i l of Canada f o r t h e i r f i n a n c i a l support of my work, and to the Biomedical Engineering Department of Vancouver General H o s p i t a l f o r supplying some of the equipment f o r t h i s p r o j e c t . v i i i -1-CHAPTER 1 INTRODUCTION 1.1 P h y s i o l o g i c a l S i g n i f i c a n c e of Blood Pressure Measurements A r t e r i a l blood pressure i s the pressure exerted by the blood against the walls of the a r t e r i e s as i t i s pumped by the heart. During each heartbeat, the blood pressure waveform traces out a d i s t i n c t i v e pattern (Fig. 1.1). The s y s t o l i c pressure (peak pressure) occurs s h o r t l y a f t e r the heart contracts to pump blood from the heart to the large blood vessels. The d i a s t o l i c pressure (trough pressure) occurs when the heart i s relaxed and being f i l l e d by the blood returning through the veins. Other parameters of i n t e r e s t are the mean a r t e r i a l pressure, which i s the time-weighted average of the pressure waveform, and the pulse pressure, which i s the d i f f e r e n c e between s y s t o l i c and d i a s t o l i c pressures. These four parameters, because of t h e i r d i f f e r i n g p h y s i o l o g i c a l s i g n i f i c a n c e , a l l provide d i f f e r e n t information. By i t s e l f , mean pressure i s not only the si n g l e most important i n d i c a t o r about the blood pressure of a patient, but i s also the best estimator of t i s s u e perfusion since i t reveals the average pressure a v a i l a b l e f o r blood flow [1,2]. S y s t o l i c pressure i s a good i n d i c a t o r of peak a r t e r i a l pressure, oxygen consumption by heart t i s s u e s , and the r i s k of coronary disease [3]. Because a l l major coronary flow occurs during d i a s t o l e , the d i a s t o l i c pressure provides an estimate of the p o t e n t i a l blood flow to the heart t i s s u e s [4]. F i n a l l y , pulse pressure gives information concerning cardiac output, renal blood flow, vasomotor tone during periods of hemorrhagic shock [5], and the condition of the a r t e r i e s [6]. In c l i n i c a l use, the term "blood pressure" r e f e r s only to the s y s t o l i c and d i a s t o l i c pressures [6], measured i n units of millimetres of mercury, or mmHg (where 760 mmHg = 101.3 k i l o P a s c a l s ) . This r e l i a n c e on only s y s t o l i c and d i a s t o l i c measurements has come about because the -2-F i g . 1.1 T y p i c a l ECG a n d b l o o d p r e s s u r e w a v e f o r m s . The u p p e r t r a c e shows a n e l e c t r o c a r d i o g r a m ( E C G ) , a r e c o r d o f e l e c t r i c a l a c t i v i t y i n t h e h e a r t . The l o w e r t r a c e shows t h e a s s o c i a t e d c h a n g e s i n b l o o d p r e s s u r e t h a t m i g h t o c c u r i n a m a j o r a r t e r y . -3-simple and noninvasive manual measurement procedure which was adopted by the American Heart Association i n 1939 could only estimate s y s t o l i c and d i a s t o l i c pressures. Since then, the widespread use and general a p p l i c a b i l i t y of t h i s technique has made i t an accepted standard [7]. Consequently, many diagnostic i n d i c a t o r s of blood pressure are put i n terms of only s y s t o l i c and d i a s t o l i c readings. For example, although hypertension i n defined to be an increased mean blood pressure [3], the World Health Organization categorizes hypertensive conditions by the following conveniently measured s y s t o l i c and d i a s t o l i c values (expressed as s y s t o l i c p r e s s u r e / d i a s t o l i c pressure) [8,9]: 1) normal blood pressure i s 120/80 mmHg; 2) borderline hypertension i s 140/95 mmHg; and 3) dangerous hypertension i s 160/95. As indicated, valuable information regarding the status of the cardiovascular system can be derived from the four blood pressure parameters. Consequently, blood pressure measurements play a valuable r o l e i n diagnosing p a t h o l o g i c a l conditions, maintaining patient s t a b i l i t y during surgery, and adapting tourniquet pressures. 1.2 Diagnosing Disease With Blood Pressure Measurements The importance of diagnosing the onset of cardiovascular problems cannot be understated. Coronary heart disease has reached epidemic proportions i n North America, exceeding a l l other causes of death [10] and d i s a b i l i t y [11] . C r i p p l i n g strokes caused by the rupturing of weakened a r t e r i e s i n the brain bring great personal, s o c i a l , and f i n a n c i a l costs, since stroke patients occupy more h o s p i t a l beds and make more use of s o c i a l welfare services than cancer and accident victims combined [11] . Both of these p a t h o l o g i c a l conditions share the same primary r i s k f a c t o r , namely high blood pressure [12]. The l i s t of ailments that can be a t t r i b u t e d to high blood pressure i s staggering, e s p e c i a l l y when considering that high blood -4-pressure i s the s i n g l e most common chronic disorder i n Western countries, a f f l i c t i n g between 15 - 20 % of the adult population [13] . Sustained hypertension can damage the inner walls of a r t e r i e s to i n i t i a t e a r e p a i r i n g process that tends to narrow the a r t e r i e s [14]; destroy the d e l i c a t e f i l t r a t i o n units of the kidney to produce kidney f a i l u r e [15]; weaken and rupture the unsupported a r t e r i e s i n the b r a i n to cause a brain hemorrhage that can lead to stroke, t i s s u e damage, headaches, b r a i n dysfunction, coma and death [15]; and damage v i s u a l parts of the v i s i o n system to induce a loss of v i s i o n at a rate greater than that of glaucoma [16]. A t h e r o s c l e r o s i s (narrowing of the a r t e r i e s ) can i n turn produce a further e l e v a t i o n i n blood pressure; congestive heart f a i l u r e ; a reduced blood flow to the heart that can starve heart t i s s u e and cause cardiac a r r e s t [15]; and a reduction of the oxygen supply to the b r a i n that can produce memory loss and other phenomena associated with aging [16]. Hypertension during pregnancy has also been associated with a higher rate of p e r i n a t a l m o r t ality, f e t a l growth retardation, impaired childhood development, congenital anomalies, and impaired neurologic conditions [17] . Both the high incidence of hypertension and the consequences of delaying treatment o f f e r incentive to detect high blood pressure r e l i a b l y . However, although casual blood pressure determinations i n the physician's o f f i c e are commonly used to diagnose and monitor hypertension, there i s s u b s t a n t i a l evidence that o f f i c e blood pressure determinations may not always r e l i a b l y r e f l e c t blood pressures at rest, i n other s e t t i n g s , or during periods of p h y s i c a l or emotional stress [18]. The c o r r e l a t i o n between c l i n i c pressures and the average l e v e l over 24 hours i s generally poor (around 0.6) [19]. Reasons f o r t h i s discrepancy are the dependence of blood pressure on d i e t , emotion, posture, d a i l y rhythms, and p h y s i c a l stress [20]; human errors i n measurements [19]; and increased anxiety l e v e l s caused by the mere presence of a doctor [19]. -5-Techniques have emerged f o r recording blood pressure i n ac t i v e , ambulatory patients to provide more accurate information f o r the doctor. The advantages of ambulatory monitoring are: improved c l a s s i f i c a t i o n of high and low r i s k patients whose casual blood pressures appear s i m i l a r ; more r e l i a b l e diagnosis of mild hypertension; better determination of the 24-hour e f f i c a c y of drug therapy; and a higher c o r r e l a t i o n between ambulatory blood pressure and hypertensive organ damage than between casual blood pressure and hypertensive complications [18,19]. Some commercially a v a i l a b l e devices t r y to f u l f i l l the need fo r 24 hour ambulatory monitoring of blood pressure. These portable devices automatically estimate blood pressure throughout the day, r e q u i r i n g no s p e c i a l user t r a i n i n g or i n t e r p r e t a t i o n of r e s u l t s (e.g.[21]). However, due to transducer design, most of these monitors are extremely s e n s i t i v e to external audio noise, transducer placement, and transducer movement caused by limb movement. Thus, during the automated measurements, the device requires that the patient cease a l l p h y s i c a l movements, thereby r e s t r i c t i n g the normal a c t i v i t i e s of the user and l o s i n g information about the rapid instantaneous blood pressure f l u c t u a t i o n s that can occur during p h y s i c a l exertion. C l o s e l y r e l a t e d to the problem of ambulatory blood pressure monitoring i s the measurement of blood pressure during s t r e s s t e s t s . While ambulatory monitoring measures the patient's pressure during a t y p i c a l day, stress t e s t s evaluate the performance of the t o t a l cardiovascular system during maximal loads under the supervision of health p r o f e s s i o n a l s . Over 1.5 m i l l i o n stress t e s t s are performed each year i n the U.S.A. [22], since stress t e s t i n g i s unexcelled as a method fo r diagnosing asymptomatic heart disease, detecting myocardial ischemia (weakening of heart t i s s u e from an inadequate oxygen supply), assessing the s e v e r i t y of impairment, measuring the effectiveness of therapy, and d e f i n i n g the l e v e l at which i t i s safe to exercise f o r heart disease -6-prevention and r e h a b i l i t a t i o n [23]. Throughout the t e s t i t i s imperative that blood pressure measurements be made, preferably as frequently as once every minute [24] . These measurements are needed to ensure that the patient i s e x e r c i s i n g at a safe l e v e l , since extreme elevations i n s y s t o l i c and d i a s t o l i c pressures i n d i c a t e that the t e s t should be terminated. Blood pressure readings can also provide diagnostic information, such as exposing severe 3-vessel coronary disease that produces both a small pulse pressure and a drop i n blood pressure e a r l y i n the exercise stage [25]. Unfortunately, as with ambulatory monitors, most automated blood pressure devices cannot produce r e l i a b l e and accurate estimates when the patient i s moving. The accuracy of the estimates diminishes r a p i d l y as the work load i s increased [24], which i s exactly the time that i s most c r u c i a l i n assessing the patient's condition. The poor performance of commercial monitors i n the presence of movement a r t i f a c t s causes some stress t e s t e r s to resort to manual blood pressure estimates, a procedure that requires another health p r o f e s s i o n a l to perform d i f f i c u l t measurements on a moving subject with loud background noise present. 1.3 Blood Pressure Measurements During Surgery During surgery, a s i g n i f i c a n t goal of the anesthetist i s to maintain patient s t a b i l i t y despite f l u i d l o s s , the administration of anesthetic agents and other drugs, and the trauma of the operation. In order to accomplish t h i s , the anesthetist can monitor the brain a c t i v i t y , r e s p i r a t i o n , and cardiovascular performance of the patient, and act accordingly to counteract any p o t e n t i a l l y damaging changes. Blood pressure measurements are one of the e s s e n t i a l parameters that i s believed necessary f o r safe anesthesia p r a c t i c e , since the anesthetist not only always measures blood pressure before administering the anesthetic, but also c i t e s a change or abnormal l e v e l -7-of blood pressure as the most frequent cause of in t e r v e n t i o n during surgery [26,27]. In fac t i f anesthetists were allowed only a s i n g l e monitored v a r i a b l e , i t has been stated that most would choose blood pressure [28]. However, as previously stated, the performance of automated monitors i s questionable i n the presence of movement a r t i f a c t that could be generated by required operation procedures. While manual measurements are u s u a l l y possible, these can be influenced by human e r r o r [29,30], and at times postponed when the anesthetist i s under pressure to perform several duties at once [31]. 1.4 Blood Pressure Measurements Used to Adapt Tourniquets A f i n a l major use of blood pressure estimates i s i n the s e t t i n g of tourniquet pressure during operations. Pneumatic tourniquets are used on upper and lower extremities to apply s u f f i c i e n t pressure to a limb to occlude the underlying a r t e r i e s . By stopping blood flow into the limb, tourniquets create a bloodless operative f i e l d that makes d i s s e c t i o n e a s i e r , s u r g i c a l techniques l e s s traumatic, and operation time shorter [32] . Pneumatic tourniquets have several other a p p l i c a t i o n s a r i s i n g from t h e i r a b i l i t y to r e s t r i c t c i r c u l a t i o n into and out of the limb. Besides providing a bloodless s u r g i c a l f i e l d , the tourniquet can simultaneously f a c i l i t a t e intravenous regional anesthesia (IVRA), a technique that enables only the limb undergoing surgery to be anesthetized by using a tourniquet which prevents the high concentration of infused l o c a l anesthetic i n the limb from flowing out of the limb u n t i l most of the drug has been absorbed by limb t i s s u e [33]. Tourniquets also enhance the treatment of connective t i s s u e cancer by r e s t r i c t i n g oxygen transport to the cancerous c e l l s or l o c a l i z i n g intravenously i n j e c t e d cytotoxic agents [34]. The b e n e f i t s derived form tourniquet use are recognized to decrease s u r g i c a l and anesthetic r i s k to the patient, as substantiated by -8-the tourniquet's prevalent use i n over one m i l l i o n procedures yearly i n North America [35] . However, associated with the use of pneumatic tourniquets are the reports of tourniquet-related i n j u r i e s to the patient [36]. Factors c o n t r i b u t i n g to these complications are c u f f pressure, cuff c h a r a c t e r i s t i c s and techniques of ap p l i c a t i o n , and excessive periods of i n f l a t i o n [35,36]. The frequency and degree of tourniquet-induced complications can be reduced by using the minimum cuff pressure that s t i l l occludes a r t e r i a l flow into the limb ("limb occlusion pressure") [37] . Such regulation w i l l eliminate not only the underpressurization hazard that can r e s u l t i n shock, blood i n the s u r g i c a l f i e l d , venous congestion, and nerve damage from hemorrhagic i n f i l t r a t i o n , but also the overpressurization hazard that can produce p a r a l y s i s , muscle weakness, and compression i n j u r y to underlying t i s s u e s [35,36]. Furthermore, tourniquet pressure must be maintained above the patient's occlusive pressure during IVRA u n t i l the anesthetic has been absorbed into limb t i s s u e to prevent cardiovascular coll a p s e , convulsions, coma, or death that can r e s u l t from high concentrations of anesthetic entering the c i r c u l a t o r y system [34]. The limb occlusion pressure required to improve tourniquet safety i s a varying parameter that w i l l be d i f f e r e n t f o r each operation, since t h i s optimal pressure s e t t i n g w i l l be af f e c t e d by the r a t i o of the width of the cuff to the circumference of the patient's limb [38], the technique of cu f f a p p l i c a t i o n , limb p o s i t i o n , limb shape, limb physiology, and the patient's s y s t o l i c pressure [32,34]. Because constant-pressure tourniquets must account f o r worst case conditions, they have often been set i n the past to 300 mmHg fo r upper limbs and 500 mmHg f o r lower limbs [36], well above normal s y s t o l i c pressure and with a wide margin to account f o r inherent errors i n mechanical pressure regulators, errors i n cuff s e l e c t i o n and a p p l i c a t i o n , leaks i n the system, and intraoperative changes i n s y s t o l i c pressure [32] . The advent of microprocessor-based tourniquets which accurately maintain a pre-set -9-cuff pressure has g r e a t l y reduced or eliminated hazards occurring from regulator errors and leaks [39] . However, these tourniquets do not compensate f o r changes i n s y s t o l i c pressure. "Adaptive tourniquets" that t r y to maintain the optimal limb occlusion pressure by t r a c k i n g the patient's s y s t o l i c blood pressure have previously been developed [32,40], and the f i r s t commercial version i s now being introduced c l i n i c a l l y . However, adaptive tourniquets have not been widely accepted i n large part because t h e i r a b i l i t y to track pressure accurately, r e l i a b l y , and r a p i d l y i s severely degraded by the presence of limb movement and other noise a r t i f a c t s occurring during surgery. For example, i f the f i r s t commercial implementation, the Richards Pressure. Sentry (Richards Medical Co., Memphis, Tennessee) detects noise during i t s data a c q u i s i t i o n stage ( t y p i c a l l y l a s t i n g more than 30 seconds), the process f o r estimating limb occlusion pressure w i l l e i t h e r be prolonged or aborted. Such time delays and s e n s i t i v i t y to noise severely impairs the e f f i c a c y , safety, and r e l i a b i l i t y of an adaptive algorithm. A l t e r n a t i v e l y , i f noisy data i s not detected, erroneous and hazardous r e s u l t s can be produced. Moreover, devices that use separate occlusive and blood pressure c u f f s require that t h e i r s y s t o l i c estimates be corrected to account f o r d i f f e r e n c e s i n c u f f width, blood pressure, limb physiology, limb shape, and c u f f a p p l i c a t i o n between the two s i t e s [32,34]. 1.5 Scope of the Thesis The problem of measuring blood pressure i n noisy environments i s addressed by t h i s t h e s i s i n the following fashion. a) A review of present methods of measuring blood pressure i s f i r s t presented. b) Following t h i s review, a d e s c r i p t i o n i s given of a prototype system that was used i n the development of a device f o r measuring blood pressure i n moderately noisy environments. c) A d e s c r i p t i o n i s then given of e f f o r t s channelled towards -10-configuring the prototype system to track limb occlusion pressure more rapi d l y , accurately, and r e l i a b l y i n moderately noisy c l i n i c a l environments than presently a v a i l a b l e commercial monitors. d) A set of s p e c i f i c performance c r i t e r i a are defined to evaluate the p o t e n t i a l of using presently a v a i l a b l e blood pressure technology f o r one p a r t i c u l a r l y important a p p l i c a t i o n , adapting tourniquet pressures. e) F i n a l l y , the r e s u l t s of comparative evaluations, using these c r i t e r i a , of the occlusive pressure-tracking performance of both the prototype system and commercial blood pressure monitors i n laboratory and c l i n i c a l environments are presented and discussed. Chapter 2 contains a review of current technology f o r measuring and estimating blood pressure parameters. Besides discussing the theory of operation, advantages and disadvantages of the major methods, the chapter also b r i e f l y describes the automated monitors i n use at a t y p i c a l teaching and research h o s p i t a l , the Vancouver General H o s p i t a l . Chapter 3 describes the development of the prototype system, i n c l u d i n g the general design philosophy and hardware conf i g u r a t i o n . Chapter 4 presents the software decisions made, i n c l u d i n g processing techniques and general algorithms f o r estimating blood pressure. Chapter 5 not only defines the performance c r i t e r i a which w i l l be used to evaluate comparatively the a b i l i t y of automated blood pressure monitors to track limb occlusion pressure, but also presents the r e s u l t s of an assessment of both the prototype system and a t y p i c a l non-invasive blood pressure monitor. Chapter 6 presents the conclusions and some suggestions f o r future research. CHAPTER 2 REVIEW OF BLOOD PRESSURE MEASUREMENT TECHNIQUES 2.1 Di r e c t Blood Pressure Measurements The methods f o r measuring blood pressure can be d i v i d e d into two major categories, namely d i r e c t and i n d i r e c t methods. The d i r e c t method, as described i n d e t a i l elsewhere [41,42], i s an invasive procedure i n which an artery, customarily the r a d i a l or ulnar artery, i s punctured to i n s e r t a small, f l e x i b l e p l a s t i c catheter f i v e to ten centimetres i n s i d e the ves s e l . The catheter can be i n s e r t e d e i t h e r by a "cut-down" procedure, which exposes the arter y to allow i n s e r t i o n and subsequent p o s i t i o n a l adjustment of the catheter, or by means of percutaneous i n s e r t i o n , which involves i n s e r t i n g a hollow needle through the skin i n t o a s u p e r f i c i a l artery to guide the catheter i n t o the ves s e l . D i r e c t measurement of blood pressure with an i n t r a - a r t e r i a l catheter can be done i n two ways. The f i r s t method uses an extracorporeal (outside the body) pressure transducer, r e q u i r i n g the intro d u c t i o n of a s t e r i l e s o l u t i o n i n t o the catheter so that f l u i d pressure i s transmitted from within the arter y to the transducer. The second method measures pressure at the source by p l a c i n g the transducer at the i n t r a - a r t e r i a l t i p of the catheter. For both cases, the pressure transducer measures the pressure-induced d i s t o r t i o n of s t r u c t u r a l components, such as capacitor pla t e s , spring-loaded inductive c o i l s , or r e s i s t i v e s t r a i n gages. There are two major advantages to using t h i s invasive technique. The d i r e c t measurement i s p o t e n t i a l l y the most accurate estimate of blood pressure, since i t measures the pressure phenomenon d i r e c t l y . Furthermore, i t gives a continuous trace of the blood pressure waveform to provide not only instantaneous s y s t o l i c , mean, and d i a s t o l i c values, but also a d d i t i o n a l diagnostic information from the shape of the -12-waveform [43]. However, there are several disadvantages to the d i r e c t measurement of i n t r a - a r t e r i a l blood pressure. Because i t increases the r i s k of i n j u r y to the patient, i t i s too invasive f o r general c l i n i c a l use. Among the p o s s i b l e hazards are i n f e c t i o n , e l e c t r i c a l i n j u r y , vascular i n j u r y , thrombus (clot) formation, cerebral embolization (brain vessel o c c l u s i o n ) , ischemic damage, death of t i s s u e served by the blocked artery, and serious blood los s from disconnected or mechanically fatigued catheter l i n e s [44,45]. It should further be noted that while the d i r e c t method i s p o t e n t i a l l y the most accurate method, c e r t a i n precautions must be taken to avoid m i s i n t e r p r e t i n g the data. Since the pressure pulse varies as i t t r a v e l s along the a r t e r y and i t s propagation i s a f f e c t e d by the presence of any foreign object i n i t s path, the waveform recorded w i l l depend not only on the l o c a t i o n of the t i p of the catheter, but also on the p h y s i c a l c h a r a c t e r i s t i c s of the measurement system [46]. For example, damping of the measuring system can be inadvertently increased to cause a reduction i n the measured peak pressure value by using a smaller bore catheter, increasing the v i s c o s i t y of the s a l i n e s o l u t i o n , increasing the length of the tubing, decreasing the radius of the tubing, or decreasing the s t i f f n e s s of the tubing [47] . Conversely, i f the natural frequency of the recording system l i e s i n the range of s i g n a l frequencies, resonance phenomena can cause erroneously high peak pressures to be recorded [48] . Therefore, any data obtained from the d i r e c t transducer i s incomplete unless i t i s accompaned by a d e t a i l e d d e s c r i p t i o n of the l o c a t i o n of the catheter i n the a r t e r i a l system, the s i z e of the catheter, the type of transducer used, and the frequency response of the e n t i r e transducer-plumbing-recorder system. Any unrecorded changes i n these parameters could lead to waveform d i s t o r t i o n that may be misinterpreted as being c l i n i c a l l y s i g n i f i c a n t . While the use of i n t r a - a r t e r i a l transducers eliminates some of the. problems caused by the external plumbing system required f o r extracorporeal measurements, -13-these c a t h e t e r - t i p devices have tended to be too unstable, c o s t l y , and f r a g i l e to gain widespread acceptance [41]. 2.2 Indirect Blood Pressure Measurements Whereas the d i r e c t method measures the e n t i r e continuous pressure pulse waveform within the artery, the i n d i r e c t method performs non-invasive measurements by r e l a t i n g c e r t a i n blood flow phenomena to d i s c r e t e , c r i t i c a l pressure values within the artery. Because pressure and flow are d i f f e r e n t energy forms, there i s seldom perfect c o r r e l a t i o n between d i r e c t and i n d i r e c t measurements [46] . Furthermore, the fact that d i r e c t measurements are a function of how and where values are taken complicate e f f o r t s to r e l a t e d i r e c t values to i n d i r e c t values. However, i f i n d i r e c t r e s u l t s are consistent and reproducible, blood pressure can be monitored non-invasively to obtain c l i n i c a l l y important information concerning pressure trends, e f f e c t s of drugs or surgery, or values of "normal" blood pressure. Thus, i n d i r e c t methods of blood pressure estimation can be used to diagnose hypertension, evaluate the e f f i c a c y of treatment during ambulatory measurements, in d i c a t e the general condition of the cardiovascular system during stress t e s t s and surgery, and track blood pressure changes during surgery to adapt the pressure of s u r g i c a l tourniquets. Indirect blood pressure measurements are us u a l l y made at the b r a c h i a l artery, between the elbow and shoulder, and require the use of a sphygmomanometer that consists of an occluding c u f f , pressure source, a means to measure cuff pressure, and a transducer to measure a r t e r i a l flow phenomena beneath the cuff [20] . I f proper precautions are taken with the cuff s e l e c t i o n and a p p l i c a t i o n , the measured c u f f pressure w i l l be transmitted to the underlying b r a c h i a l artery [20,38,49]. Depending on both the c u f f pressure and the pressure f l u c t u a t i o n i n the arte r y as the pressure pulse propagates, the arte r y w i l l be i n one of three possible -14-sta t e s : 1) always pinched closed (cuff pressure > s y s t o l i c pressure); 2) sometimes open and sometimes closed ( d i a s t o l i c pressure < cuff pressure < s y s t o l i c pressure); 3) always f u l l y open (cuff pressure < d i a s t o l i c pressure). Determining which of these three states e x i s t , and hence the d i a s t o l i c and s y s t o l i c pressures, can be done by sensing flow phenomena associated with the state of the artery. The p a r t i c u l a r flow c h a r a c t e r i s t i c to be measured w i l l depend on the transducer used. While the transducer output can be in t e r p r e t e d manually, properly designed automated techniques y i e l d greater r e p r o d u c i b i l i t y by removing errors associated with human observers [30] . The major transducers used [17] are l i s t e d i n the following sections. 2.2.1 Auscultatory Method The auscultatory method uses a sound transducer under the cuff to measure Korotkoff sounds that vary with the blood flow i n the artery. With the c u f f pressure above s y s t o l i c pressure, the arter y i s pinched closed, no blood flows, and no sounds are heard. I f the cu f f pressure i s then slowly released from above s y s t o l i c pressure, a v a r i e t y of Korotkoff sounds (K-sounds) w i l l be heard. The f i r s t sound, a thudding noise, occurs during t h i s c uff d e f l a t i o n when the cu f f pressure equals the s y s t o l i c pressure, at which time blood f i r s t begins to s q u i r t through the narrowed ar t e r y . As the cu f f pressure decreases f u r t h e r between s y s t o l i c and d i a s t o l i c pressure, the sounds detected pass through d i s t i n c t stages as the flow passes through d i f f e r e n t regimes of turbulence. When the d e f l a t i n g cuff reaches d i a s t o l i c pressure, the sound becomes r a p i d l y muffled and soon disappears since the cu f f pressure i s too low to deform the a r t e r y and induce turbulent blood flow. In i t s simplest form, a manual blood pressure measurement -15-system consists of a c u f f to apply pressure, a hand pump, a mercury column to measure c u f f pressure, and a stethoscope to l i s t e n f o r these p a r t i c u l a r sounds. Although t h i s system i s extensively used, i t has l i m i t a t i o n s because i t requires a t r a i n e d p r o f e s s i o n a l to properly recognize normal and abnormal K-sounds, apply the c u f f properly, and reduce stethoscope errors [50]. Moreover, t h i s simple manual method i s susceptible to human error, audible noise interference, and patient movement a r t i f a c t s [17,18,19]. The human errors and t r a i n i n g requirements can be eliminated by using a microphone and microprocessor to detect and process the d i s t i n c t sounds as the cuff i s d e f l a t e d from above s y s t o l i c pressure [51] . However, even with automation, the auscultatory method s u f f e r s from the following drawbacks [45,51,52]: 1) s e n s i t i v i t y to audible background noise; 2) requirement that the transducer be placed d i r e c t l y on the skin; 3) dependency on transducer placement and movement, since the microphone's s e n s i t i v i t y to K-sounds w i l l change as i t s distance from the art e r y i s a l t e r e d ; 4) d i f f i c u l t y i n g e t t i n g r e l i a b l e K-sounds from obese patie n t s ; 5) problems with patients who have c a l c i f i e d a r t e r i e s that cannot be narrowed s u f f i c i e n t l y to produce K-sounds; 6) i n a b i l i t y to measure K-sounds f o r patients with abnormally low blood flow, such as infants or those i n shock; 7) absence of a mean a r t e r i a l pressure i n d i c a t o r ; 8) exposure of the transducer to p h y s i c a l abuse; and 9) interference caused by movement a r t i f a c t s . 2.2.2 Ultrasound Instead of measuring the d i f f e r e n t sounds caused by the changing blood flow beneath the occluding c u f f , the flow i t s e l f can be measured. One transducer commonly used to estimate blood flow on the -16-basis of blood v e l o c i t y changes i s the transcutaneous Doppler u l t r a s o n i c flowmeter. I t s operation depends on the Doppler-shift p r i n c i p l e , which re l a t e s the frequency s h i f t of sound back-scattered from moving a r t e r i a l blood p a r t i c l e s to the v e l o c i t y of the blood flow. Processing the Doppler s i g n a l produces an output s i g n a l whose magnitude i s proportional to the v e l o c i t y of the blood c e l l s [53]. I f an occluding c u f f i s used i n conjunction with the Doppler flowmeter, the s y s t o l i c and d i a s t o l i c pressures can be determined [54]. The measureable c u f f pressure i s f i r s t set well above the s y s t o l i c pressure and then decreased slowly. With the cuff pressure above s y s t o l i c pressure, the u l t r a s o n i c transducer detects no flow i n the occluded artery. When flow i s f i r s t detected, the cuff pressure i s equal to the s y s t o l i c pressure. As c u f f pressure i s further reduced, the Doppler s i g n a l w i l l increase i n magnitude as flow increases, but w i l l drop to zero at some time during each pressure pulse cycle, since the s u p r a - d i a s t o l i c c u f f pressure w i l l exceed some portions of the pressure pulse wave to close the arter y and stop flow. When the Doppler s i g n a l f i r s t remains above zero at a l l times to i n d i c a t e continuous flow through an unobstructed artery, the cuff pressure i s at d i a s t o l i c pressure. An a l t e r n a t i v e u l t r a s o n i c technique uses the Doppler s h i f t p r i n c i p l e to detect the movement of the arter y wall under an occluding c u f f [55]. When a decreasing cuff pressure reaches s y s t o l i c pressure, the a r t e r y w i l l begin to snap open momentarily with the propagation of the pulse wave to produce the f i r s t detectable Doppler s i g n a l . This wall movement w i l l continue u n t i l d i a s t o l i c pressure i s reached, at which time flow i s continuous and l i t t l e wall movement i s observed. The u l t r a s o n i c blood pressure monitor s u f f e r s from the following l i m i t a t i o n s [45,56]: ' 1) high s e n s i t i v i t y to transducer and patient movement, since both create a Doppler s i g n a l ; 2 ) requirement that the transducer be placed accurately over the -17-artery; 3) high costs f o r the so p h i s t i c a t e d e l e c t r o n i c s and transducer; 4) exposure of a f r a g i l e , expensive transducer; 5) absence of a mean a r t e r i a l pressure c r i t e r i o n ; and 6) the need f o r d i r e c t skin contact and coupling g e l f o r the transducer. 2.2.3 Oscillometry The f i n a l major i n d i r e c t blood pressure method i s osci l l o m e t r y . S imilar to the auscultatory and u l t r a s o n i c procedures, t h i s method also requires an occlusive cuff around the limb. However, i n osc i l l o m e t r y the important parameter to be measured i s the small-signal pressure v a r i a t i o n within the cu f f that occurs when the pressure pulse induces a displacement of the a r t e r i a l wall [7] (Fig. 2.1). Thus, the o s c i l l o m e t r i c transducer i s a s o l i d - s t a t e pressure transducer that, while connected to the cuff v i a f l e x i b l e tubing, i s conveniently removed from the c u f f . As with the auscultatory and u l t r a s o n i c methods, blood pressure measurements are normally made by i n f l a t i n g the cu f f above the s y s t o l i c pressure and then d e f l a t i n g slowly. T y p i c a l o s c i l l o m e t r i c algorithms then analyze the height of the flow-induced pulses to estimate blood pressure parameters (e.g. [57-60]). With the cu f f pressure above s y s t o l i c pressure, no blood w i l l flow beneath the c u f f . At t h i s time, very small pressure pulses may be recorded within the cuff as a r t e r i a l wall motion i s detected only at the proximal ("upstream") edge of the occluding c u f f . When c u f f pressure i s equal to the s y s t o l i c pressure and blood begins to sq u i r t beneath the e n t i r e cu f f , the observed pulse amplitude i n the cuff w i l l r a p i d l y increase, since the movement of the arter y wall i s now being sensed along the whole length of the cu f f [61-62]. As cu f f pressure i s furthe r reduced, pressure pulsations i n the cuff increase i n magnitude, reaching a maximum s i z e when the cuff pressure i s near the mean a r t e r i a l -18-* 7 2 T 4 5* 5 7 8 9 10 U 12 ?3 14 15 *" TIME (s) F i g . 2.1 O s c i l l o m e t r i c pulses. P u l s a t i l e blood flow i n the a r t e r i e s beneath a p r e s s u r i z e d c u f f induces small pressure p u l s a t i o n s w i t h i n the c u f f . Mean Arteri al Pressure Systol ic Pressure A A A A Diastol ic Pressure L i A L 1 30 120 1 10 100 90 CUFF PRESSURE (mmHg! 80 70 F i g . 2.2 V a r i a t i o n of o s c i l l o m e t r i c pulses with c u f f pressure. A f t e r the p u l s a t i o n s have been separated from the bas e l i n e c u f f pressure s i g n a l / o s c i l l o m e t r i c algorithms t y p i c a l l y analyze the height of these flow-induced pressure pulses i n order to estimate s y s t o l i c pressure, d i a s t o l i c pressure, and mean a r t e r i a l pressure. - 2 0 -pressure [ 7 , 5 8 ] . A further decrease i n c u f f pressure gradually reduces pulse amplitude u n t i l d i a s t o l i c pressure i s reached, at which time the s i g n a l amplitude decreases r a p i d l y . F i g . 2 . 2 shows how pulse height t y p i c a l l y v a ries with cuff pressure. Compared to the other i n d i r e c t methods, o s c i l l o m e t r y o f f e r s the following advantages [ 2 , 4 5 ] : 1) simple system attachment to the patient; 2) no transducer placement problems; 3) a protected, inexpensive transducer; 4) measurements pos s i b l e from obese patients, shock patients, and i n f a n t s ; 5) measurements possible from patients r e q u i r i n g s p e c i a l dressings on t h e i r limbs that w i l l accommodate a c u f f but not a cutaneous transducer; 6) the existence of a mean a r t e r i a l pressure c r i t e r i o n that can be derived e a s i l y and d i r e c t l y from the transducer output; and 7) some noise immunity when measuring the mean a r t e r i a l pressure, since the s i g n a l to be detected i s at i t s maximum (as opposed to the minimum signals that occur at s y s t o l i c and d i a s t o l i c pressures f o r a l l cuff methods). However, as with a l l methods, the o s c i l l o m e t r i c technique s u f f e r s from movement a r t i f a c t s . Movement of the c u f f , connecting hoses, or c u f f - p a t i e n t i n t e r f a c e w i l l cause pressure v a r i a t i o n s i n the system that can mask the small flow-induced pressure p u l s a t i o n s . 2 . 3 Indirect Blood Pressure Monitors i n Use at Vancouver General Hospital Vancouver General Hospital (VGH) i s a large teaching and research h o s p i t a l that has a l l of the special-care areas found i n major r e f e r r a l h o s p i t a l s (including operating rooms, post-anesthesia rooms, int e n s i v e care u n i t s , burn units, cardiology, coronary care, emergency, -21-kidney d i a l y s i s , and diagnostic imaging; but excluding p e d i a t r i c s and o b s t e t r i c s ) . For t h i s reason, a survey of the i n d i r e c t blood pressure monitors i n use at VGH should provide i n s i g h t i n t o the operational requirements that should be met by noninvasive blood pressure monitors f o r c l i n i c a l use i n a wide v a r i e t y of environments. The noninvasive blood pressure monitors at VGH are p r i m a r i l y o s c i llometric-based devices. The h o s p i t a l ' s biomedical engineering department recommended t h i s technology as a standard because these monitors provide a mean a r t e r i a l pressure c r i t e r i o n , eliminate the need for p r e c i s e transducer and cuff placement, avoid the use of exposed and expensive transducers, and provide p r o t e c t i o n against audio and e l e c t r o s u r g i c a l noise t y p i c a l l y found i n c l i n i c a l environments. The choice of p a r t i c u l a r commercially a v a i l a b l e o s c i l l o m e t r i c monitors depends on such factors as cost, ease of s e r v i c i n g , ease of use, accuracy, r e l i a b i l i t y , and the s p e c i f i c features required f o r t h e i r s p e c i f i c uses. There are c u r r e n t l y four d i f f e r e n t types of o s c i l l o m e t r i c noninvasive blood pressure monitors i n use at VGH. These monitors are t y p i c a l of the o s c i l l o m e t r i c monitors which are commercially a v a i l a b l e . The major f u n c t i o n a l c h a r a c t e r i s t i c s of each are summarized below. 2.3.1 Dinamap Model 845 A d u l t / P e d i a t r i c V i t a l Signs Monitor The Dinamap 845 (C r i t i k o n Inc., Tampa FL) , the f i r s t o s c i l l o m e t r i c blood pressure monitor on the market and one of the few to be extensively t e s t e d i n independent studies (e.g. [63-72]), i s designed to monitor s y s t o l i c pressure, mean a r t e r i a l pressure, d i a s t o l i c pressure, and heart rate [57] . Measurements can be i n i t i a t e d e i t h e r manually by depressing a switch, or automatically at chosen i n t e r v a l s between 1 and 160 minutes. When making the f i r s t estimate a f t e r being activated, the monitor w i l l i n f l a t e the cuff to 170 mmHg, generally above the patient's s y s t o l i c pressure. Subsequent measurements w i l l use an -22-i n f l a t i o n l e v e l of 30 mmHg above e i t h e r the previously estimated s y s t o l i c pressure, or the previous i n i t i a l i n f l a t i o n l e v e l i f i t was not s u f f i c i e n t to occlude the artery. The monitor uses incremental d e f l a t i o n , r e l e a s i n g c u f f pressure i n steps of about 8 mmHg to allow i t s noise r e j e c t i o n scheme to be implemented. At each pressure l e v e l , the monitor regulates the cuff pressure before sampling data, which consists of both the cu f f pressure and the o s c i l l o m e t r i c p u l s a t i o n amplitude. Two consecutive pressure pulses of equal amplitude and appropriate time separation (based on past pulses) are required f o r the data to be considered v a l i d . I f matched pulses cannot be found i n two seconds because of noise, the cuff d e f l a t i o n cycle i s halted u n t i l the noise ceases, or u n t i l 60 seconds has passed. At each step, v a l i d pulse height and cu f f pressure information i s stored and analyzed by the microprocessor to determine s y s t o l i c , mean, and d i a s t o l i c pressures based on standard o s c i l l o m e t r i c c r i t e r i a . Stepped d e f l a t i o n continues u n t i l e i t h e r a l l three pressures are determined, or u n t i l the cu f f pressure nears zero. T y p i c a l estimation time i s 40 seconds under noise-free conditions. Among the safety features of the Dinamap 845 are a pressure r e l i e f valve set f o r 275 mmHg and audible and v i s u a l alarms f o r mean pressure l i m i t s , excessive i n f l a t i o n time (due to leaks), excessive time at one pressure l e v e l (due to leaks, hose kinks, and limb motion), and excessive t o t a l determination time (due to leaks, kinks, and motion). According to engineers and technicians i n the VGH biomedical engineering department who must respond to incidents concerning equipment hazards and problems, the Dinamap 845 suf f e r s from the following design l i m i t a t i o n s : a) maximum s y s t o l i c pressure estimation of 210 mmHg; b) i n a b i l i t y to measure blood pressure i n c e r t a i n infants and patients i n shock who have a pulse pressure of le s s than 10 mmHg, -23-as a r e s u l t of the 8 mmHg stepped d e f l a t i o n l e v e l s ; c) inaccurate f i r s t estimate i f the s y s t o l i c pressure i s above 170 mmHg, since the i n i t i a l i n f l a t i o n l e v e l w i l l not occlude the artery; and d) d i f f i c u l t y i n tr a c k i n g r a p i d l y r i s i n g s y s t o l i c pressure because of the time required by the algorithm ( t y p i c a l l y 40 s) and because of the algorithm's f a i l u r e to check f o r occlusion before beginning the estimation c y c l e . 2.3.2 Hewlett Packard Model 78354A Patient Monitor The Hewlett Packard (HP) 78354A Patient Monitor (Hewlett Packard Co., Palo A l t o CA) i s an integrated monitoring system that can measure the ECG waveform, temperature, the d i r e c t pressure waveform, r e s p i r a t i o n , plethysmograph information, and noninvasive blood pressure [58]. It uses the same stepped d e f l a t i o n , noise r e j e c t i o n , and blood pressure estimation algorithm as the Dinamap 845. Consequently, the t y p i c a l estimation time i s s i m i l a r to that of the Dinamap 845 (about 40 s) . However, the HP 78354A d i f f e r s from the Dinamap 845 i n the following ways: a) the pressure r e l i e f valve i s set at 315 mmHg i n the HP 78354A, 40 mmHg above that of the Dinamap 845, allowing f o r higher estimates of s y s t o l i c pressure with the HP 78354A than i s poss i b l e with the Dinamap 845; b) the HP 78354A has manually selectable i n f l a t i o n c y c l e i n t e r v a l s of 2, 5, 10, 15, 30, and 60 minutes; and c) the algorithm of the HP 78354A checks that the c u f f pressure i s s u f f i c i e n t to occlude the underlying a r t e r i e s before beginning the c u f f d e f l a t i o n and data a c q u i s i t i o n c y c l e . -24-2.3.3 Datascope Accutorr Model 2A Noninvasive Blood Pressure Monitor The Datascope Accutorr (Datascope Corporation, Paramus NJ) measures s y s t o l i c pressure, d i a s t o l i c pressure, mean pressure, and heart rate noninvasively using an o s c i l l o m e t r i c algorithm. I t has the following features [59]: a) continuous d e f l a t i o n at about 4 mmHg/second that allows the monitor to be used as an e l e c t r o n i c sphygmomanometer f o r simultaneous manual auscultatory determinations, provides better r e s o l u t i o n of estimates [73], and measures pressures even when the pulse pressure i s les s than 10 mmHg; b) i n i t i a l i n f l a t i o n pressure of 200 mmHg; c) 40 mmHg i n f l a t i o n increment above the l a s t s y s t o l i c estimate or the l a s t i n i t i a l i n f l a t i o n l e v e l i f i t was not enough to occlude the artery; d) automatic r e p e t i t i o n of the measurement c y c l e i f the previous attempt was not successful i n estimating a l l parameters, up to a maximum of four r e p e t i t i o n s ; e) 320 mmHg pressure r e l i e f valve; f) a weak pul s a t i o n alarm that sounds when, f o r e i t h e r p h y s i o l o g i c a l reasons or problems with the monitor, the amplitudes of the detected o s c i l l o m e t r i c pulses are not s u f f i c i e n t to ensure an accurate analysis by the monitor; g) manually adjustable alarms f o r s y s t o l i c values; h) manually adjustable alarms f o r excessive estimation times; i) matched pulse c r i t e r i o n f o r noise r e j e c t i o n ; j) 15 second time-out l i m i t to f i n d 2 matched pulses before aborting the measurement attempt; k) an algorithm that suspends d e f l a t i o n when noise i s detected; 1) manually adjustable i n f l a t i o n cycle i n t e r v a l s of 1, 2, 3, 5, 15, 30, and 60 minutes; -25-m) t y p i c a l estimation time of 40 seconds i n noise-free conditions; n) an optional b u i l t - i n p r i n t e r to record trends and produce a copy of the pul s a t i o n amplitude vs. cu f f pressure information; and o) a continuously-repeating estimation mode l a s t i n g 5 minutes. 2.3.4 Physio-Control Model V i t a l Signs Monitor The Physio-Control V i t a l Signs Monitor (Physio-Control Corporation, Redmond WA) i s a monitor that w i l l record the ECG waveform and estimate s y s t o l i c , d i a s t o l i c , and mean pressures noninvasively using an o s c i l l o m e t r i c algorithm. It has the following c h a r a c t e r i s t i c s [60]: a) i n i t i a l i n f l a t i o n l e v e l of 180 mmHg; b) i n f l a t i o n increment of 45 mmHg; c) continuous, versus stepwise, d e f l a t i o n ; d) manually adjustable alarms f o r s y s t o l i c , d i a s t o l i c and mean pressure values; e) manually adjustable i n f l a t i o n cycle of 1, 2, 3, 5, 10, 15, and 30 minutes; f) allowance f o r manual s e t t i n g of the i n i t i a l i n f l a t i o n l e v e l f o r better t r a c k i n g of r a p i d l y r i s i n g s y s t o l i c pressures; g) automatic r e p e t i t i o n of the estimation cycle i f the f i r s t attempt f a i l s to measure a l l three parameters. 2.3.5 Discussion of Commercial Noninvasive Blood Pressure Monitors Although the noninvasive blood pressure monitors which are commercially a v a i l a b l e have a v a r i e t y of d i f f e r e n t c h a r a c t e r i s t i c s intended to meet s p e c i f i c needs, a l l share some common l i m i t a t i o n s . For example, because most manufacturers do not f u l l y d i s c l o s e t h e i r o s c i l l o m e t r i c and noise r e j e c t i o n algorithms f o r competitive reasons, i t i s d i f f i c u l t to compare adequately the performance of d i f f e r e n t monitors - 2 6 -without a thorough c l i n i c a l evaluation. Another l i m i t a t i o n with a l l major i n d i r e c t methods of estimating blood pressure i s that the estimates are based on data c o l l e c t i o n over many heartbeats. Since large beat-to-beat v a r i a t i o n s occur due to arrhythmias and 15-20 mmHg peak-to-peak v a r i a t i o n s occur due to r e s p i r a t i o n [73], the blood pressure i s not stable over the measurement period. Thus, t h e i r accuracy i s l i m i t e d by the blood pressure f l u c t u a t i o n s that can occur during the data a c q u i s i t i o n c y c l e . The c l i n i c i a n does not generally know whether the i n d i r e c t estimate considers the highest, lowest, or some average of the varying blood pressure values that have occurred during the measurement period. Algorithms which have long pressure-estimation cycles are also poorly s u i t e d f o r tra c k i n g r a p i d l y changing pressures which occur i n a v a r i e t y of c l i n i c a l s i t u a t i o n s . A f a c t o r that can a f f e c t the r e l i a b i l i t y of these noninvasive blood pressure monitors i s the effectiveness of t h e i r noise r e j e c t i o n schemes. Most o s c i l l o m e t r i c devices implement a simple pattern recognition scheme using a template based on received pulses. Consequently, a f t e r a time of p e r i o d i c noise, the devices may form t h e i r templates from the i n t e r f e r i n g s i g n a l and i n t e r p r e t these noise pulses as being v a l i d s i g n a l s . The f i n a l l i m i t a t i o n associated with commercially a v a i l a b l e monitors r e l a t e s to recommended c a l i b r a t i o n schemes. Most devices have both a s e l f - d i a g n o s t i c routine i n c l u d i n g automatic c a l i b r a t i o n - c h e c k i n g which i s i n i t i a t e d at start-up, and a procedure o u t l i n e d i n the t e c h n i c a l manuals f o r manually checking c a l i b r a t i o n and r e - c a l i b r a t i o n as required. However, current methods of checking c a l i b r a t i o n and r e - c a l i b r a t i o n are uns a t i s f a c t o r y and may be hazardous, since they only check microprocessor and transducer accuracy, but do not assess the dynamic functions of the complete system. One independent c l i n i c a l t r i a l of the Dinamap 845 [70] showed how inaccurate readings were obtained from a machine with a -27-defective O-ring washer at the cuff-hose connector. However, repeated c a l i b r a t i o n checks, performed as s p e c i f i e d i n the manual f o r the device, d i d not i n d i c a t e any malfunctions. Only by comparing the estimates of the device with estimates obtained from an alternate technique was the device found to be f a u l t y . -28-CHAPTER 3 HARDWARE DEVELOPMENT OF THE PROTOTYPE 3.1 General Hardware Design Philosophy The previous review of blood pressure measurement techniques can be summarized as follows: 1) d i r e c t blood pressure measurements are too invasive f o r general use; 2) i n d i r e c t noninvasive blood pressure determinations that are consistent and reproducible are s u i t a b l e f o r general use to derive information obtained from values of s y s t o l i c pressure, d i a s t o l i c pressure, mean pressure, and pulse pressure; 3) manual i n d i r e c t blood pressure estimation by auscultation i s extensively used and i s considered to be a standard against which other jnethods can be compared, but i t i s most s u i t a b l e only when a t r a i n e d health p r o f e s s i o n a l i s a v a i l a b l e to make measurements i n a noise-free environment; 4) i n d i r e c t methods can be automated to eliminate the need f o r a health p r o f e s s i o n a l and to improve the p o t e n t i a l f o r better r e p r o d u c i b i l i t y through the design of decision-making algorithms that w i l l not make the errors associated with human observers; 5) automated i n d i r e c t methods that use u l t r a s o n i c or auscultatory transducers s u f f e r from problems associated with transducer placement, motion a r t i f a c t , exposed transducers, and the lack of a c r i t e r i o n f o r mean pressure based d i r e c t l y on received flow-induced s i g n a l s ; 6) automated i n d i r e c t methods based on os c i l l o m e t r y have only one major f a u l t , namely a v u l n e r a b i l i t y to the e f f e c t s of motion a r t i f a c t s , which has made oscillometry the recommended technology at VGH, a large teaching and research h o s p i t a l that -29-has most major s p e c i a l care u n i t s ; and 7) commercially a v a i l a b l e o s c i l l o m e t r i c monitors i n use at VGH could benefit from improvements i n t h e i r algorithms to y i e l d a quicker estimation process and a better noise detection scheme. The primary objective i n assembling a prototype system was to provide the means to improve the measurement of blood pressure i n the presence of motion a r t i f a c t s . The best way of achieving t h i s was determined to be to s t a r t with a preferred technology, oscillometry, and to choose system components that would allow the protype to perform be t t e r i n noisy c l i n i c a l environments than commercially a v a i l a b l e devices. With t h i s objective i n mind, a general o s c i l l o m e t r i c blood pressure measurement system was assembled using the following components (Fig. 3.1): a) an i n f l a t a b l e blood pressure cuff to occlude the arter y and sense o s c i l l o m e t r i c pressure pulsations; b) tubing to connect the c u f f to pressure transducers and pressure c o n t r o l means; c) pressure c o n t r o l means to i n f l a t e and d e f l a t e the cu f f ; d) pressure transducers to convert both the baseline c u f f pressure and the superimposed o s c i l l o m e t r i c pressure pulsations to e l e c t i c a l s i g n a l s ; e) a m p l i f i e r s to separate and amplify the baseline and o s c i l l o m e t r i c e l e c t r i c a l s i g n a l s ; f) a microprocessor system to i n t e r p r e t the acquired data, provide a user i n t e r f a c e , and c o n t r o l other system components; and g) A/D and D/A systems to allow the microprocessor system to communicate with other components. What remains to be solved i s oscillometry's major problem, -30-Mi c r o p r o c e s s o r S y s t e m A/D D/A P r e s s u r e R e g u l a t o r T u b i n g C u f f A m p l i f i e r s P r e s s u r e T r a n s d u c e r s F i g . 3.1 Components of a general o s c i l l o m e t r i c system f o r measuring blood pressure. -31-namely the movement a r t i f a c t s that are produced when the c u f f , hoses, or limb are moved or bumped. Whatever i t s o r i g i n , t h i s interference can cause large pressure f l u c t u a t i o n s i n the cuff pressure to mask the small o s c i l l o m e t r i c s i g n a l (Fig. 3.2). Trying to eliminate the c u f f and limb noise at the source would involve developing and evaluating a completely new blood pressure cuff , which i s beyond the scope of t h i s t h e s i s . Instead, developing a system that could reduce hose noise at the source and deal with the observed e f f e c t s of noise from the c u f f , limb and hoses was defined to be a goal of the t h e s i s . Primary concerns when choosing components f o r the system were the performance of each component i n meeting t h i s goal, the a v a i l a b i l i t y of components, and f l e x i b i l i t y of the o v e r a l l system to allow adjustments of p a r t i c u l a r design parameters. Size and c o s t - e f f e c t i v e n e s s of the system were considered to be aspects to be addressed only a f t e r the "optimal" design parameters could be more p r e c i s e l y s p e c i f i e d . A d e t a i l e d d e s c r i p t i o n of the components of the prototype system i s contained i n the following sections. 3.2 Blood Pressure Cuff One major unique feature of the prototype system i s a multi-bladder blood pressure cuff , which was synthesized from commonly a v a i l a b l e c u f f s . Commercial blood pressure monitors use a standard single-bladder occlusive c u f f , which i s i n e f f e c t a lone pneumatic transmission channel that contains both the o s c i l l o m e t r i c s i g n a l and i n t e r f e r i n g noise. However, a multi-bladder cuff provides i n d i v i d u a l bladders that can s p e c i f i c a l l y function to occlude the artery, detect a r t e r i a l blood flow, or act as separate noise and signal-plus-noise transmission channels. One configuration i s i l l u s t r a t e d i n F i g . 3.3. When the proximal bladder i s v a r i a b l y pressurized, i t w i l l contain both noise and the o s c i l l o m e t r i c s i g n a l . It can also act as a "flow valve", F i g . 3 . 3 The multi-bladder c u f f . -33-syst e m a t i c a l l y blocking and permitting blood flow to the other two bladders through an adjustment of i t s p r e s s u r i z a t i o n to supra-occlusive and sub-occlusive values, r e s p e c t i v e l y . The middle bladder can not only serve as a tourniquet c u f f when maintained at occlusive pressures, but can also i n d i c a t e whether the proximal bladder i s occluding the artery, since the middle bladder can detect the flow past the proximal bladder by the presence of flow-induced o s c i l l o m e t r i c pulses. The f i n a l bladder i s a two-piece component that detects motion a r t i f a c t s , with one section located at the d i s t a l end of the cuff to sense the r e l a t i v e movement between the cuff and limb, and the other positioned over the proximal and middle bladders to detect external impacts against the c u f f . Because of i t s p o s i t i o n over and d i s t a l to the occluding bladder, no flow-induced pulses w i l l be present i n t h i s bladder. Thus, any pulses seen here can be a t t r i b u t e d to noise a r t i f a c t s that also d i s t u r b the signals i n the other bladders. The problems associated with constructing a multi-bladder blood pressure c u f f were avoided by modifying commercially a v a i l a b l e tourniquet c u f f s (manufactured by Aspen Laboratories, Inc., Englewood CO). The two types of c u f f s used were a si n g l e tourniquet cuff which i s 34 inches long with a 3 3/4-inch wide backing that supports a 3 1/4-inch wide sheath holding a 3-inch bladder; and a dual tourniquet c u f f (normally used f o r procedures i n v o l v i n g intravenous regional anesthesia) that i s 24 inches long with a 4-inch wide backing that supports two 2-inch wide sheaths, each containing a 1 3/4-inch wide bladder. The sheaths were structured to permit the i n s e r t i o n of an a d d i t i o n a l bladder, made from 0.5-inch "Penrose" late x tubing which was sealed with epoxy at one end and attached to a standard cuff f i t t i n g at the other end. Two 34-inch cuffs were attached with Velcro fasteners along one-third of t h e i r length to complete f a b r i c a t i o n of the prototype 34-inch multi-bladder c u f f (Fig. 3.4). The use of these narrow tourniquet c u f f s i s required i n the f a b r i c a t i o n of the multi-bladder c u f f s to l i m i t the o v e r a l l width of the -34-Latex Tubing (a) 9 §>§> Tubing Connectors - - Q Q Dista l (Noise J Detection) ] Middle (Occlusion and Flow Detection) ~[ Proximal (Flow Control) 0 Tourniquet Cuffs ~j External (Noise Detection) Fasteners (b) 9 6>6> (d) F i g . 3.4 Multi-bladder c u f f s synthesized from tourniquet c u f f s . -35-c u f f . A narrow c u f f i s desirable f o r two reasons: 1) when used i n tourniquet a p p l i c a t i o n s , a c u f f needs to be narrow to avoid entering the s u r g i c a l f i e l d ; and 2) while conventional c u f f s are shaped to f i t a c y l i n d r i c a l form as opposed to a tapered limb, a narrow cuff may better f i t those limbs that are nearly c y l i n d r i c a l over a l i m i t e d region. Although convenient f o r purposes of the work described i n t h i s t h e s i s , these multi-bladder c u f f s do have some l i m i t a t i o n s . Since a four-inch overlap i s recommended when applying such a c u f f to a limb [74], the 24-inch and 34-inch c u f f s should only be used f o r limbs with a circumference of le s s than 20 inches and 30 inches, r e s p e c t i v e l y . Furthermore, because the multi-bladder c u f f s are r e i n f o r c e d to prevent ballooning of the i n s e r t e d bladders, the c u f f s tend to be b u l k i e r than normal blood pressure c u f f s , and thus more d i f f i c u l t to apply snugly to tapered limbs. A f i n a l e f f e c t of using narrow bladders i s the reduction i n the o s c i l l o m e t r i c s i g n a l amplitude. With a smaller section of the a r t e r y being e n c i r c l e d by a bladder, a smaller pressure p u l s a t i o n i s transmitted to the cuff , and a diminished pressure v a r i a t i o n i s measured i n the bladder [61,62]. Greater a m p l i f i c a t i o n of the s i g n a l i s required to compensate f o r t h i s e f f e c t • 3.3 Tubing P o l y v i n y l tubing with 1/4-inch outside diameter (Clippard Instrument Laboratory, Inc., C i n c i n n a t i OH) was chosen to connect the various pneumatic components. It i s inexpensive, r e a d i l y a v a i l a b l e , and e a s i l y heated to form a i r - t i g h t connections with a v a r i e t y of pneumatic f i t t i n g s . Moreover, i t i s both f l e x i b l e enough to allow convenient manipulation of interconnected components and r i g i d enough to r e s i s t kinking. A fur t h e r advantage of the r i g i d tubing i s i t s good transmission of pneumatic pressure pulses. For example, common s u r g i c a l -36-tubing tends to attenuate p u l s a t i l e pressure signals due to i t s e l a s t i c walls. F l e x i b l e walls increase the capacitance of the pneumatic transmission l i n e and hence absorb high frequency acoustic energy [47]. Other advantages are r e a l i z e d by using as l i t t l e tubing as po s s i b l e . Decreasing the length of tubing between the c u f f , transducers, and pressure c o n t r o l l e r w i l l minimize the r e s i s t i v e attenuation of the s i g n a l along the l i n e , decrease the damping of the system [47], decrease the t o t a l volume of the system to reduce the i n f l a t i o n time, and reduce the incidence of impacts against the hose that can produce noise a r t i f a c t s . While most commercial blood pressure units supply over 10 feet of tubing to connect the cuff to t h e i r transducers, the prototype uses about 3 feet of tubing between the cuff and the shielded transducers that can be conveniently mounted near the limb (Fig. 3.5). 3.4 Pressure Control Means The pressure i n the pneumatic system i s generated and v a r i e d by the pressure c o n t r o l unit (Fig. 3.6). The source of the pressure i s an Aspen Model ATS 1000 Automatic Tourniquet System (Aspen Laboratories, Englewood CO), containing a line-powered pump that i s capable of supplying pressures of up to 450 mmHg with a t y p i c a l unloaded flow rate of 2.35 1pm. Miniature valves (Clippard EVO-3-12, Cl i p p a r d Instrument Laboratory Inc., C i n c i n n a t i OH), configured as single-pole, single-throw pneumatic switches, receive 5 V co n t r o l signals to open the normally-closed pneumatic l i n e e i t h e r to the pressure source to increase pressure, or to the atmosphere to decrease pressure. The valves are normally-open 3-way poppet valves having a poppet t r a v e l of 0.015 inches, pressure range of 0 - 105 p s i , a i r f l o w of 0.5 cfm at 100 p s i , and a power consumption of 0.65 W. For coarse pressure regulation (+ 5 mmHg), a s i n g l e pressure and sensing l i n e i s s u f f i c i e n t , with the transducer sensing pressure near the valves to approximate the actual pressure i n the c u f f . However, -37-F i g . 3.5 Mounting of transducers and valves near the limb. -38-E n c l o s e d i n Box and Mounted C l o s e to Limb ATS 1000 I n c r e a s e P r e s s u r e T r a n s d u c e r s C o n t r o l S i g n a l s D e c r e a s e P r e s s u r e TX TX TX I M D C u f f F i g . 3.6 Pressure c o n t r o l u n i t . -39-because of the i n e r t i a and pneumatic capacitance of the a i r i n the system, separate sensing and pressure l i n e s are needed to provide f i n e pressure regulation (+_ 2 mmHg). This configuration allows the transducer to sense pressure i n the l i n e f a r removed from the pressure source or sink to get a better approximation of the actual c u f f pressure. The i d e a l case of having sensors at the cuff i s not f e a s i b l e with the present choice of components, since there i s no convenient way of mounting, connecting and p r o t e c t i n g the transducers, or properly applying an excessively bulky cuff to a limb. 3.5 Pressure Transducer Disposable pressure transducers (Cobe Laboratories Inc., Lakewood CO) are used i n the prototype to sense the pressure within the pneumatic system. These inexpensive and r e a d i l y a v a i l a b l e transducers are p i e z o r e s i s t i v e measuring devices that use a balanced Wheatstone-bridge sensing element (Fig. 3.7(a)). With one side of the element open to the atmosphere and the other side exposed to the pressure chamber, an applied pressure w i l l stress the element to unbalance the r e s i s t o r bridge and create an output voltage that i s p r o p o r t i o n a l to the applied pressure. The Cobe pressure transducer i s widely used f o r d i r e c t , i n t r a - a r t e r i a l measurement of blood pressure i n h o s p i t a l s , and i s designed to function with a s e n s i t i v i t y of 5 m i c r o v o l t s / e x c i t a t i o n voltage/mmHg, a resonant frequency exceeding 200 kHz i n a i r , and e x c i t a t i o n voltage between 2.5 - 10 V, an operating l i f e greater than 300 hours, and an operation range of -50 to 300 mmHg. A modification was attempted to eliminate the problems associated with a capacitively-coupled a m p l i f i e r with the very low frequency (0.1 Hz) high-pass cutoff required f o r separating the o s c i l l o m e t r i c s i g n a l from the baseline c u f f s i g n a l . O r d i n a r i l y , the pressure transducer compares the ambient atmospheric pressure through i t s -40-Cuf f P r e s s u r e C u f f P r e s s u r e Grn RT = t e m p e r a t u r e - c o m p e n s a t i n g r e s i s t o r RN = l a s e r - t r i m m e d r e s i s t o r s f o r o f f s e t a d j u s t m e n t RS = l a s e r - t r i m m e d r e s i s t o r f o r s e n s i t i v i t y a d j u s t m e n t R1-R4 = p i e z o r e s i s t o r s f o r W h e a t s t o n e b r i d g e Red = e x c i t a t i o n v o l t a g e B i k = g r o u n d wht = p o s i t i v e s i g n a l o u t Grn = n e g a t i v e s i g n a l o u t F i g . 3.7 Cobe pressure transducers. A Cobe pressure transducer i s shown: (a) unmodified; and (b) modified with a feedback pressure loop to cancel the b a s e l i n e c u f f pressure. -41-plug with the pressure i n i t s chamber, sensing pressure values by the d e f l e c t i o n and concurrent resistance change of a r e s i s t i v e bridge element. However, the transducer can be modified so that the pressurized chamber i s connected to the ambient pressure l i n e (Fig. 3.7(b)). This d i f f e r e n t i a l configuration cancels the baseline cuff pressure s i g n a l , but allows the small o s c i l l o m e t r i c s i g n a l to be detected since the small pressure pulsations are attenuated s u f f i c i e n t l y along the feedback path. With t h i s d i f f e r e n t i a l transducer, a direct-coupled a m p l i f i e r could be used f o r the o s c i l l o m e t r i c s i g n a l s , f o r there i s no need to remove a low-frequency baseline pressure s i g n a l . Unfortunately, t h i s modification could not be repeated uniformly to ensure that two d i f f e r e n t modified transducers would have the same response, since the response of the d i f f e r e n t i a l transducers depends on the feedback loop length, feedback loop diameter, and general construction techniques. This standardization problem l e d to the use of unmodified transducers, one f o r each bladder, with 5 V e x c i t a t i o n voltages and t h e i r outputs connected to the separate amplifying c i r c u i t s required f o r the baseline and o s c i l l o m e t r i c s i g n a l s . 3.6 Amplifiers Magnification of the baseline cuff pressure s i g n a l i s achieved through a low-gain d.c. a m p l i f i e r (Fig. 3.8). It uses a standard 3-chip instrumentation a m p l i f i e r as the f i r s t stage, allows f o r o f f s e t and gain adjustment, removes the o s c i l l o m e t r i c s i g n a l with a 0.5 Hz RC low-pass f i l t e r (LPF) at the output, and uses op-amps that are compatible with CMOS I.C.'s that w i l l be necessary f o r a n t i c i p a t e d battery-powered use i n the future. It i s presently adjusted so that the output voltage i s Vout = Pressure x (1.0 V/50 mmHg) - 4.0 V (3.1) corresponding to a gain of about 800. A stable, high-gain a.c. a m p l i f i e r i s required to i s o l a t e and -42-"sv ( N a t i o n a l Offset S e m i c o n d u c t o r Adjust C o r p . , S a n t a C l a r a CA) F i g . 3.8 Low-gain D.C. a m p l i f i e r f o r b a s e l i n e c u f f pressure s i g n a l s . -43-magnify the o s c i l l o m e t r i c s i g n a l that rides on the baseline low frequency pressure s i g n a l . S p e c i f i c a l l y , the a m p l i f i e r should provide a gain of about 100,000, although the exact value w i l l depend on transducer performance, pneumatic attenuation, and factors that a f f e c t the s i z e of the pressure pulse received by the c u f f such as bladder s i z e , c u f f a p p l i c a t i o n , c u f f s i t e , cardiovascular condition, and limb physiology beneath the c u f f . Because the o s c i l l o m e t r i c waveform resembles the d i r e c t pressure waveform [75,76], the 0.1 - 20 Hz bandwidth needed f o r f a i t h f u l reproduction of the d i r e c t waveform [46] was thought to be s u f f i c i e n t f o r the o s c i l l o m e t r i c waveform, with exact bandwidth l i m i t s to be determined by analyzing the frequency content of the s i g n a l . An a d d i t i o n a l desired feature i s a reset c a p a b i l i t y to quicken the recovery time so that the amplifer, having a large RC time constant associated with the high-pass f i l t e r (HPF), does not take a long time to recover from saturation when subjected to large swings i n c u f f pressure s i g n a l s . Fortunately, an a m p l i f i e r with adjustable gains, v a r i a b l e bandwidth, and reset c a p a b i l i t y i s a v a i l a b l e i n the form of a commercial device, a Beckman Accutrace 200 EEG a m p l i f i e r (Beckman Instruments, Inc. F u l l e r t o n CA) with the following features: a) 18 channels; b) chart recorder f o r hard copies of waveforms; c) single-pole HPF settings with 3 dB frequencies of 0.16, 0.5, 1, and 5 Hz; d) single-pole LPF settings with 3 dB frequencies of 15, 35, 50, and 70 Hz; e) 60 Hz notch f i l t e r ; f) a u x i l i a r y output of the amplified s i g n a l s ; g) 10-position master s e n s i t i v i t y c o n t r o l of 2,3,5,7,10,15,20, 30,50, and 75 microvolt/mm chart d e f l e c t i o n (or microvolt i n / 0.5 V out); h) i n d i v i d u a l s e n s i t i v i t y m u l t i p l i e r pushbuttons of 20,4,1,0.5, -44-and 0.25; and i) reset input r e q u i r i n g a normally-open r e l a y connection. An a d d i t i o n a l amplifying stage, c o n s i s t i n g of simple non-inverting op-amp c i r c u i t s with a gain of about 3.5, was added at the output of the EEG am p l i f i e r to increase i t s range from 2.8 V peak-to-peak: to around 10 V peak-to-peak. 3.7 Microprocessor System A standard IBM PC was used to provide a programmable microprocessor system with user i n t e r f a c e f o r the prototype. 3.8 Analog-to-Digital and Digital-to-Analog C a p a b i l i t y A Tecmar Labtender Board (Tecmar Inc., Cleveland OH) provides the microprocessor system with the c a p a b i l i t y of not only c o n t r o l l i n g the reset r e l a y and pressure valves, but also acquiring c u f f pressure and o s c i l l o m e t r i c data. It i s a multifunction board with a n a l o g - t o - d i g i t a l inputs, d i g i t a l - t o - a n a l o g outputs, d i g i t a l I/O, timers, and counters. Thirty-two single-ended or sixteen d i f f e r e n t i a l A/D channels are av a i l a b l e , each with 8 b i t r e s o l u t i o n , 18 microsecond conversion time, allowed input range of + 5 V, and 1 Megohm input impedance. Sixteen D/A channels are present, each with 8 b i t r e s o l u t i o n , 3 microsecond conversion time, -5 V to +4.96 V output range, 100 ohm output impedance, and -1/2 LSB l a g i n 400 milliseconds. Because these D/A channels can only supply a l i m i t e d current, a d d i t i o n a l relay and valve drive c i r c u i t r y i s needed (Fig. 3.9). 3.9 T o t a l System The frequency c h a r a c t e r i s t i c s of the pneumatic system were determined by introducing a step pressure input and observing the step response as follows. The 24 inch cuff was wrapped around a 3 inch diameter c y l i n d e r and pressurized to 150 mmHg, the EEG a m p l i f i e r b a n d w i d t h was m a x i m i z e d ( L P F s e t t o 0 . 1 6 H z , HPF s e t t o 70 H z , a n d 60 Hz n o t c h f i l t e r d i s a b l e d ) , a n d a s y r i n g e was a t t a c h e d t o t h e b l a d d e r s y s t e m b e i n g t e s t e d t o i n j e c t a v o l u m e o f 2 c c . A n a n a l y s i s o f t h e s t e p r e s p o n s e d a t a i n d i c a t e d t h a t t h e p n e u m a t i c s y s t e m b e h a v e d as a s e c o n d - o r d e r u n d e r d a m p e d s y s t e m , w i t h t h e d a m p i n g f a c t o r n e a r 0 . 2 a n d a n a t u r a l f r e q u e n c y n e a r 50 H z . T h e r e f o r e , no u n w a n t e d r e s o n a n t c o n d i t i o n s o r d a m p i n g e f f e c t s a r e c r e a t e d b y t h e p n e u m a t i c s y s t e m i n t h e f r e q u e n c y r a n g e o f i n t e r e s t (0 - 20 H z ) . The c o m p o n e n t s o f t h e o v e r a l l p r o t o t y p e s y s t e m ( F i g . 3 . 1 0 ) were m o u n t e d o n t o t h e m o b i l e EEG m a c h i n e f o r t r a n s p o r t i n g t o c l i n i c a l a r e a s ( F i g . 3 . 1 1 ) . F i g . 3 . 9 D r i v e c i r c u i t r y f o r r e l a y a n d v a l v e s . -46-Prcssure Source Valve C o n t r o l S i g n a l s Valves I I Pressure Transducer Valves Pressure Transducer Valves Pressure Transducer Orlve C i r c u i t r y High Gain A m p l i f i e r D/A A m p l i f i e r Reset S i g n a l Cuff Low Gain A m p l i f i e r O s c i l l o m e t r i c S i g n a l s A/D B a s e l i n e S i g n a l s IBM PC F i g . 3.10 Components of the prototype system. -47-F i g . 3.11 The prototype system. CHAPTER 4 SOFTWARE AND SIGNAL PROCESSING 4.1 General Software Design Philosophy The previous chapter described the components of a prototype system capable of making estimates of blood pressure i f programmed with appropriate o s c i l l o m e t r i c algorithms. The system's major hardware advantages over commercial devices are: decreased pneumatic volume r e s u l t i n g i n decreased i n f l a t i o n time, decreased s i g n a l attenuation, and reduced damping; a multi-bladder c u f f that provides a d d i t i o n a l information and s p e c i f i c i t y of bladder functions; an adjustable o s c i l l o m e t r i c a m p l i f i e r ; and an inherently f l e x i b l e microprocessor system. The programmability of the microprocessor system allows f o r experimentation with d i f f e r e n t s i g n a l processing and blood pressure estimation algorithms. The development of algorithms generally involved i n i t i a l data a c q u i s i t i o n , c h a r a c t e r i z a t i o n of the s i g n a l , noise, or other parameters of i n t e r e s t , formulation of hypotheses, examination of the f e a s i b i l i t y based on a d d i t i o n a l a n a l y s i s , and implementation as a C-language program (C86 C Compiler, Computer Innovations Inc., Red Bank NJ). This programming language was chosen f o r i t s e f f i c i e n t code and speed of execution that made real-time operation poss i b l e , the existence of a l i b r a r y of useful routines and documentation, and the a v a i l a b i l i t y of a cross-assembler to convert the C-language program in t o I n t e l 8085 executable machine code f o r future development of a dedicated 8085-based microprocessor system. This chapter d e t a i l s the r e s u l t s of the waveform c h a r a c t e r i z a t i o n study, i d e n t i f i e s some general s i g n a l processing techniques that could be applied to the p a r t i c u l a r waveforms produced by the prototype system and indicates the projected degrees of success of each, o u t l i n e s some algorithms f o r estimating blood pressure based on -49-o s c i l l o m e t r i c waveforms, and describes the algorithm f o r estimating limb occlusion pressure which was implemented with the prototype system. 4.2 I n i t i a l Data A c q u i s i t i o n I n i t i a l noisy and noise-free o s c i l l o m e t r i c data was acquired from 10 subjects (3 female and 7 male aged between 19 and 36 with no known h i s t o r y of c i r c u l a t o r y problems), chosen from the s t a f f of the Vancouver General Hospital's Biomedical Engineering Department. An unmodified 24" dual tourniquet c u f f (Aspen Laboratories, Inc., Englewood CO) was connected to the prototype system and wrapped around the upper arm of a supine subject, between the elbow and shoulder. The proximal c u f f was f i r s t used to obtain reference estimates of limb occlusion pressure with a Dinamap 845. Then with the d i s t a l c u ff manually i n f l a t e d and held at supra-occlusive pressures to provide a "noise-only" channel, the proximal cuff was manually i n f l a t e d to a supra-occlusive pressure and allowed to bleed slowly, at a rate of about 5 mmHg/s, to provide a " s i g n a l + noise" channel. The EEG machine, set to operate with a 0.16 - 15 Hz bandwidth, provided a chart recording of the o s c i l l o m e t r i c signals from both c u f f s . The analog o s c i l l o m e t r i c and baseline s i g n a l s produced by the am p l i f i e r s from both c u f f s were also recorded on a Hewlett Packard 4-channel FM tape recorder f o r l a t e r a n a l y s i s . Runs were repeated under noise-free conditions, with passive limb movement, during external banging on the c u f f , and while the subject a c t i v e l y moved about. 4.3 Signal and Noise Characterization The analog data stored on the Hewlett-Packard 4-channel tape recorder was sampled at 50 Hz and stored on disk by the VGH biomedical engineering department's RT-11 operating system. ILS s i g n a l processing software (available from Signal Technology Inc., Santa Barbara CA) was used to analyze the data i n order to d i s p l a y the frequency spectrum and -50-histograms of both the o s c i l l o m e t r i c s i g n a l and noise. The r e s u l t s are summarized as follows: 1) the noise t y p i c a l l y followed a Gaussian d i s t r i b u t i o n (Fig. 4.1), with zero mean and a variance that depended on the type of motion a r t i f a c t being generated; 2) the o s c i l l o m e t r i c s i g n a l was not Gaussian (Fig. 4.2), c o n s i s t i n g of a se r i e s of unipolar pulses whose d i s t r i b u t i o n could be af f e c t e d by o s c i l l o m e t r i c amplitude f a c t o r s (e.g. c u f f pressure, c u f f a p p l i c a t i o n , and limb physiology) and the pulse duty cycle ( i . e . the heart r a t e ) ; 3) both the s i g n a l and noise had powers concentrated i n the 0 - 5 Hz region (Fig. 4.3); 4) the frequency spectrum of the s i g n a l showed two dominant components, one corresponding to the heart rate and the other being i t s second harmonic; 5) the noise was not white, implying that some c o r r e l a t i o n e x i s t e d between successive noise samples; and 6) the s i g n a l - t o - n o i s e - r a t i o was t y p i c a l l y l e s s than -15 dB according to the chart recordings, although t h i s value would vary according to the type of noise present and o s c i l l o m e t r i c amplitude f a c t o r s . These r e s u l t s l e d to the following preliminary conclusions: 1) noise c h a r a c t e r i s t i c s can change very q u i c k l y depending on the type of movement a r t i f a c t s being generated; 2) s i g n a l c h a r a c t e r i s t i c s can experience both i n t e r - s u b j e c t and int r a - s u b j e c t v a r i a t i o n ; 3) a simple bandpass f i l t e r w i l l not s i g n i f i c a n t l y help to improve the SNR, since both s i g n a l and noise occupy s i m i l a r frequency bandwidths; 4) the 18.2 Hz i n t e r n a l clock of the IBM PC w i l l o f f e r a convenient timing s i g n a l f o r sampling the waveforms, since the sign a l s can -51-Number of Points = 1500 >-200 100 •264 •ILS-generated Gaussian F i t to Data .is 306 BrNS F i g . 4.1 Histogram of a noise waveform. -52-Number of Points = 1500 200' cr * 100t -100 ILS-generated Gaussian F i t to Data BINS 211 F i g . 4.2 Histogram of an o s c i l l o m e t r i c s i g n a l . -53-100 Signal Noise CO -o ~ 50 o a . \ , u K * *•* %\ W U 1 A A V V y f 't i • fKCOUCNCV <HZ> 100, co -o - 5^ 0 cn UJ 3 o a . \ ft w Hi J - j jyj j y™ i n * ' in i! II IM t i L * I H SHI 1 »Jf J1 •i ! * • i | i rREQUCNCY <H2> F i g . 4.3 Frequency a n a l y s i s of noise and s i g n a l waveforms. -54-be passed through an 8 Hz low-pass a n t i - a l i a s i n g f i l t e r without l o s i n g s i g n i f i c a n t pulse information; and 5) the low SNR, l i m i t e d r e s o l u t i o n of the 8-bit A/D converter, and importance of pulse height information f o r implementing any o s c i l l o m e t r i c algorithm create a condition whereby the high a m p l i f i e r gains required to obtain the necessary pulse height information w i l l also y i e l d a high incidence of noise-induced amplifer saturation that " c l i p s " the s i g n a l and destroys a l l pulse information. 4.4 Noise Reduction Techniques While the use of short hoses helps to reduce the v u l n e r a b i l i t y of the prototype system to incidents of hose kinking that may occur with longer lengths of exposed tubing, the processing of the a d d i t i o n a l si g n a l s from the multi-bladder c u f f was i n v e s t i g a t e d to see i f the prototype system could y i e l d a method f o r r e a l i z i n g f u r t h e r improvements i n SNR. 4.4.1 Noise Subtraction The use of a multi-bladder cuff creates separate noise and signal-plus-noise channels. If the noise i n the two channels i s well co r r e l a t e d , a simple subtraction of the two channels should y i e l d a new signal-plus-noise waveform with an improved SNR. This noise subtraction technique has been used with a Korotkoff-sound transducer to produce a modest SNR improvement of 30 % [77] . However, preliminary chart recordings of the proximal (signal-plus-noise) and supra-occlusive d i s t a l (noise) s i g n a l s (e.g. F i g . 4.4) i n d i c a t e d the following problems with t h i s approach: 1) periods of noise-induced a m p l i f i e r saturation y i e l d e d perfect subtraction but t o t a l l o s s of s i g n a l information; 2) d i f f e r e n c e s i n bladder pressure created varying amplitudes of -55-Di s t a l C u f f ( n o i s e + smal l s i g n a l component) : j F i g . 4.4 Simultaneous chart recording of waveforms from two bladders. - 5 6 -noise interference; and 3) d i f f e r e n c e s i n the l o c a t i o n of the bladders on the limb produced v a r i a t i o n s i n phase and amplitude between the two noise waveforms fo r the observed limb movements and external c u f f impacts. These l i m i t a t i o n s preclude the p o s s i b i l i t y of implementing a simple noise subtraction pre-processing stage i n hardware or software to improve SNR. 4.4.2 Adaptive F i l t e r s Adaptive noise c a n c e l l a t i o n , explained i n d e t a i l elsewhere [78], i s used to cancel stationary or slowly varying a d d i t i v e i n t e r f e r e n c e . Without a p r i o r i knowledge of s i g n a l and noise, an adaptive f i l t e r can learn the s t a t i s t i c s i n i t i a l l y and then track them. A t y p i c a l adaptive noise c a n c e l l a t i o n system requires a primary input containing the s i g n a l plus uncorrelated a d d i t i v e noise, and a reference input containing noise which i s also uncorrelated to the s i g n a l but c o r r e l a t e d i n some way to the noise i n the primary input. The multi-bladder cuff described e a r l i e r could provide the required primary and reference inputs to the adaptive noise c a n c e l l a t i o n system, which would then adaptively f i l t e r the reference input to create a waveform that i s a best least-squares estimate of the noise i n the primary input. Subtraction of the f i l t e r e d reference noise from the primary input then produces a best least-squares estimate of the s i g n a l . While adaptive noise c a n c e l l a t i o n may appear to o f f e r a way to compensate f o r the unknown phase and amplitude d i f f e r e n c e s between the noise i n two separate bladders that made simple subtraction i n f e a s i b l e , there are l i m i t a t i o n s i n the adaptive scheme that prevent i t from being u s e f u l i n the a p p l i c a t i o n considered here. In p a r t i c u l a r , convergence rates may not be rapid enough, since adaptive f i l t e r s that have been implemented elsewhere [78] t y p i c a l l y require over 200 i t e r a t i o n s to converge. With the waveforms being sampled at 18.2 Hz, an adaptive -57-f i l t e r w i l l l i k e l y require a learning phase exceeding 10 seconds whenever presented with new noise conditions that produce waveforms with new c h a r a c t e r i s t i c s . Consequently, while an adaptive f i l t e r may be appropriate f o r stationary or slowly varying noise (e.g. 60 Hz i n t e r f e r e n c e ) , i t w i l l not be able to track c h a r a c t e r i s t i c s that can vary s i g n i f i c a n t l y within seconds as d i f f e r e n t limb movements occur or as changes i n c u f f pressure take place. E f f o r t s to increase the convergence rates can lead to other problems, since increasing the convergence rate parameter can lead to i n s t a b i l i t y and poor t r a c k i n g as the f i l t e r c o n t i n u a l l y overshoots the target set of c o e f f i c i e n t s . Oversampling the waveform i n an e f f o r t to reduce the learning time w i l l increase the number of f i l t e r c o e f f i c i e n t s required to achieve the same f i l t e r r e s o l u t i o n , which i n turn w i l l n u l l i f y the benefit by i n c r e a s i n g the time constant of the adaptive process. F i n a l l y , periods of a m p l i f i e r saturation w i l l create unrecoverable loss of information and also a f f e c t the performance of an adaptive noise c a n c e l l a t i o n process. 4.4.3 Signal Averaging Since the o s c i l l o m e t r i c pulses repeat themselves at every heart beat, the noisy waveform can be d i v i d e d i n t o a s e r i e s of data blocks, each containing a si n g l e pulse aligned i n the same l o c a t i o n within the block. Then N number of blocks can then be added together to increase the amplitude of the s i g n a l by a f a c t o r of N and the standard deviation of the random uncorrelated noise by a f a c t o r of /N , leading to an increase i n SNR by VW [80] . However, s i g n a l averaging has the f o l l o w i n g shortcomings: 1) there i s a need to a l i g n and s t r e t c h the o s c i l l o m e t r i c pulses that are neither exactly p e r i o d i c nor constant; 2) t h i s scheme w i l l be slow, since a large number of data blocks may be needed to achieve a u s e f u l increase i n SNR; -58-3) since the noise contains s i g n i f i c a n t energy at low frequencies, i t w i l l tend to be c o r r e l a t e d between successive data blocks to y i e l d an improvement i n SNR of le s s than f$f ; and 4) data blocks containing periods of noise-induced a m p l i f i e r saturation could not be used f o r the averaging process. 4.4.4 Signal Estimation Signals corrupted by a d d i t i v e noise can be recovered using time-domain s i g n a l estimation techniques that require the knowledge of s i g n a l and noise c h a r a c t e r i s t i c s . Common estimation schemes make use of the following: information about the p r o b a b i l i t y density functions to a s s i s t i n choosing the most probable value of the s i g n a l (maximum a p r i o r i estimate) [80]; knowledge about the a u t o c o r r e l a t i o n of the samples and the c r o s s c o r r e l a t i o n between the samples and the s i g n a l (Wiener estimate) [79]; or a model of the dynamics of the process generating the s i g n a l (Kalman f i l t e r ) [79] . However, t h e i r dependence on s i g n a l and noise c h a r a c t e r i s t i c s makes these methods i n f e a s i b l e i n t h i s a p p l i c a t i o n which has nonstationary signals and periods of a m p l i f i e r s a turation. 4.5 Noise Detection Although there are problems with the implementation of noise reduction techniques due to a m p l i f i e r saturation and nonstationary s i g n a l s , i t i s possible to improve the performance of the prototype system i n the presence of motion a r t i f a c t s . Rather than processing the s i g n a l i n an attempt to improve the SNR, the prototype can detect the presence of noise and suspend processing to avoid using corrupted data. Such noise r e j e c t i o n delays blood pressure estimation processes, but allows f o r consistent performance i n varying noise conditions, provided that noise can be r e l i a b l y detected. Commercially a v a i l a b l e o s c i l l o m e t r i c monitors use simple -59-pattern recognition schemes to r e j e c t noisy data (e.g. [57-60]). However, these noise r e j e c t i o n algorithms can mistakenly accept c y c l i c noise as being v a l i d o s c i l l o m e t r i c signals [73]. Because the prototype system uses a multi-bladder cuff , a "noise-only" channel can be conveniently created to y i e l d another way of checking f o r the presence of noise. In p a r t i c u l a r , "noise-only" channels can be formed i n a v a r i e t y of ways, in c l u d i n g : an external bladder located over the multi-bladder c u f f and separated from the limb where the flow-induced pulses o r i g i n a t e ; a bladder that receives no a r t e r i a l flow by being placed d i s t a l to an occluding bladder; or a supra-occlusive bladder placed proximal to the occlusive c u f f , although the l a t t e r could s t i l l produce small flow-induced o s c i l l o m e t r i c pulses that could be mistaken f o r noise i f some a d d i t i o n a l pattern recognition scheme i s not used. The s i g n a l from any one of these channels could then be compared to a threshold l e v e l , set to be s u f f i c i e n t l y d i f f e r e n t than the background e l e c t r i c a l noise present i n the system even when the limb i s motionless (for example +_ 3 standard deviations of background noise samples). When the noise s i g n a l i s within allowable l i m i t s , no noise would be assumed to be present and processing would be permitted to continue; but when the noise s i g n a l exceeds the threshold, a noise condition would be considered to e x i s t and processing would stop. By sampling signals only when noise i s not detected, the prototype can lessen the e f f e c t that i n t e r f e r i n g noise may have on the accuracy and r e l i a b i l i t y of blood pressure estimates. Because noise channels can be analyzed separately to determine whether noise i s present, t h i s noise r e j e c t i o n technique has the p o t e n t i a l to peform better than both the pattern recognition schemes that o s c i l l o m e t r i c monitors use to discriminate between s i g n a l and noise, and f i l t e r i n g techniques that auscultatory monitors use to reduce i n t e r f e r i n g audio noise. -60-4.6 Signal Information Two types of s i g n a l information must generally be obtained from sampled o s c i l l o m e t r i c waveforms i n order to estimate blood pressure. One i s pulse height, which can be derived by l o c a t i n g the peaks of v a l i d s i g n a l s , accomplished e i t h e r by examining slope changes of the s i g n a l or by measuring l o c a l maximal data points that e x i s t between zero crossings of the waveform. The second type of s i g n a l information involves s i g n a l detection, required i n order to e s t a b l i s h whether or not a r t e r i a l blood flow i s occurring by determining whether or not an o s c i l l o m e t r i c pulse i s present. Pulse detection with a matched f i l t e r can be used to improve the SNR; however, t h i s would e n t a i l the added complexities associated with forming and updating the template of the expected pulse, which could vary. Furthermore, the maximum improvement i n SNR of v/Energy i n the Template (4 .1 ) peak s i g n a l value w i l l never be attained, since the noise i s not white [79]. Calculations made using the r e s u l t s from the i n i t i a l data a c q u i s i t i o n suggest that while a SNR improvement by a fac t o r of 2 may be r e a l i z e d i f the noise were white, the actual improvement may be less than a f a c t o r of 1.2 for the c o r r e l a t e d noise that has been recorded by t h i s system. Such small gains are not worth the cost of t r y i n g to implement a matched f i l t e r . Instead, a pulse detection scheme s i m i l a r to the noise detection scheme described e a r l i e r can be used. With t h i s method, a d e c i s i o n concerning whether or not flow i s present i s based on whether or not the o s c i l l o m e t r i c s i g n a l from the flow-detecting channel i s above or below a flow detection threshold (set, f o r example, at 3 standard deviations of background noise above b a s e l i n e ) . -61-4.7 Baseline Wander With any of the pulse height or s i g n a l detection schemes o u t l i n e d e a r l i e r , i t i s important to remove the baseline wandering of the high-gain EEG a m p l i f i e r that may a f f e c t the s i g n a l analysis without t r i g g e r i n g the noise detector. This can be achieved by pre-processing the sampled data before analyzing i t . The pre-processor should have the following c h a r a c t e r i s t i c s : a highpass f i l t e r with a sharp c u t o f f below 1 Hz to remove the low frequency wandering without s i g n i f i c a n t l y reducing the waveform amplitude; l i n e a r phase c h a r a c t e r i s t i c s to avoid introducing d i s t o r t i o n and phase s h i f t s that may a f f e c t processing schemes that r e l y on synchronizing s i g n a l s ; and an a b i l i t y to operate i n real-time. An e f f i c i e n t f i n i t e impulse response (FIR) f i l t e r that meets these s p e c i f i c a t i o n s can be implemented by taking advantage of symmetry pr o p e r t i e s . For example, an FIR f i l t e r designed by the frequency sampling method [79] from N points (N odd) with 0, K = 0 |H(K)| = (4.2) 1, K = 1, 2, 3, . . ., N-l w i l l have an impulse reponse of 0, n<0, n>N-l h(n) = (N-l)/N, n = (N-l)/2 (4.3) -1/N, elsewhere as shown i n F i g . 4.5. Consequently, t h i s N-l order f i l t e r only needs 2 m u l t i p l i c a t i o n s and N-l additions. The number of additions can i n turn be reduced by using the p e r i o d i c i t y of the sampled frequency response. Increasing the time i n t e r v a l between impulse response c o e f f i c i e n t s by a f a c t o r of 2 reduces the spacing between frequency response samples by a f a c t o r of 2. This produces f i n e r frequency r e s o l u t i o n (and hence sharper cutoff) but also -62-h(r>r 0.88 •0.11 ~ < — i — i — i — i — i — I — i — x — x x x x x x x H(K)|-F i g . 4.5 A highpass FIR f i l t e r f o r removal of ba s e l i n e wander r e q u i r i n g 2 m u l t i p l i c a t i o n s and N-l ad d i t i o n s . -63-reduces the sampling frequency (and the frequency range of i n t e r e s t ) . Besides having f i n e r r e s o l u t i o n , the p e r i o d i c frequency response of the expanded impulse response w i l l have a s i m i l a r bandstop notch at the same frequencies as the o r i g i n a l highpass f i l t e r . Thus, padding the o r i g i n a l impulse response with zeros between samples y i e l d s the b e n e f i t s of f i n e r frequency r e s o l u t i o n provided by the expanded impulse response, but maintains the desired sampling frequency ( f s ) , as shown i n F i g . 4.6. In other words, an impulse response with every second sample being zero w i l l produce a N-l order bandpass f i l t e r with a notch at f = 0 and f = fs/2 that has a s i m i l a r low frequency c u t o f f as the o r i g i n a l N-l order highpass f i l t e r , but w i l l require only (N-l)12 additions and 2 m u l t i p l i c a t i o n s . The f i l t e r i n g of s i g n a l components near fs/2 w i l l not be s i g n i f i c a n t , since the analog waveform would be passed through an a n t i - a l i a s i n g lowpass f i l t e r before sampling. With t h i s f i l t e r , i ncreasing the order tends to decrease the r i p p l e of the frequency response and sharpen the c u t o f f , since the frequency samples would be c l o s e r together to define more p r e c i s e l y the desired frequency response. 4.8 Algorithms f o r Estimating Blood Pressure By suspending processing when detecting noise, the prototype can r e j e c t noise and analyze o s c i l l o m e t r i c waveforms to provide improved estimates of blood pressure i n the presence of motion a r t i f a c t s . Values that can be measured include s y s t o l i c pressure, limb occlusion pressure, mean a r t e r i a l pressure, and d i a s t o l i c pressure. Since the prototype uses bladders that are l i k e l y to be narrower than the standard blood pressure c u f f recommended f o r most limbs, the s y s t o l i c , d i a s t o l i c , and mean pressure measurements made by the prototype w i l l have to be corrected to account f o r the "narrow c u f f e f f e c t " that increases a l l blood pressure estimates [38] . The limb occlusion pressure, which i s the c u f f pressure needed to occlude the underlying art e r y regardless of c u f f s i z e , i s -64-Hz) F i g . 4.6 Further reduction i n computation f o r a FIR f i l t e r - u s e d to remove b a s e l i n e wander. This f i g u r e shows: (a) a FIR lowpass f i l t e r r e q u i r i n g 2 m u l t i p l i c a t i o n s and N-l a d d i t i o n s ; (b) the FIR lowpass f i l t e r with i t s sampling frequency halved; and (c) the bandpass f i l t e r having a s i m i l a r low-frequency c u t o f f as the highpass f i l t e r i n (d) but r e q u i r i n g few a d d i t i o n s ((N-D/2) . -65-i d e n t i c a l to the uncorrected " s y s t o l i c " pressure estimate made with the multi-bladder c u f f , since both ind i c a t e the cuff pressure at which blood w i l l f i r s t flow past a d e f l a t i n g occluding c u f f [32]. 4.8.1 Estimation of S y s t o l i c Pressure S y s t o l i c pressure can be estimated by two methods. One method can use a v a r i a b l y pressurized proximal bladder (using the other bladders f o r noise detection) to obtain a table of the o s c i l l o m e t r i c pulse heights measured at d i f f e r e n t proximal pressures s e t t i n g s . The proximal pressure at which the maximum increase i n pulse height i s obtained i s considered to be the (uncorrected) s y s t o l i c pressure. While t h i s scheme i s used by many commercial o s c i l l o m e t r i c monitors (e.g. [57-60]), i t i s slow because i t must search through a large range of pressures. Since noise r e j e c t i o n techniques suspend processing and lengthen the time needed to make a measurement, i t i s desirable to f i n d a f a s t e r estimation process that requires l e s s data and therefore le s s time to make an estimate. Quick estimates are us e f u l f o r t r a c k i n g s y s t o l i c blood pressure with continuous estimates, reducing the time that a limb i s occluded during a si n g l e estimate, and eli m i n a t i n g the unknown e f f e c t s that blood pressure changes may have on an estimate made with a lengthy, prolonged data a c q u i s i t i o n stage. A p o t e n t i a l l y f a s t e r estimation process f o r continuous measurements involves detecting flow past the proximal bladder. When the proximal bladder pressure i s occluding underlying a r t e r i e s , no flow w i l l be detected at the more d i s t a l bladders. Conversely, flow w i l l be detected i n these d i s t a l bladders when the proximal bladder no longer occludes the a r t e r i e s . Thus, the t r a n s i t i o n of flow conditions can in d i c a t e when the proximal pressure has passed through the limb occlusion pressure (or uncorrected s y s t o l i c pressure). Rapid continuous estimates can then be made by d i t h e r i n g the proximal pressure above and below estimated limb occlusion pressure values. - 6 6 -4.8.2 Estimation of Mean A r t e r i a l Pressure Mean a r t e r i a l pressure i s considered to be the minimum (corrected) c u f f pressure at which o s c i l l o m e t r i c pulses reach a maximum amplitude. Commercial o s c i l l o m e t r i c monitors use a slow search through a large pressure range accomplished with a steady cuff i n f l a t i o n . The prototype can make the search with the proximal bladder through the same range more e f f i c i e n t by using optimization techniques. For example, while steady stepped d e f l a t i o n over a 105 mmHg range requires 15 steps to obtain a re s o l u t i o n of 7 mmHg i n the estimate of mean pressure, a golden section search [81] can f i n d the cuff pressure at which a maximum pulse i s achieved i n the same range and with the same r e s o l u t i o n using only 7 steps. 4.8.3 Estimation of D i a s t o l i c Pressure D i a s t o l i c pressure i s considered by other monitors to be estimated by the (corrected) cuff pressure at which a maximum decrease i n o s c i l l o m e t r i c pulses i s observed during a steady d e f l a t i o n of an occluding c u f f . A f a s t e r algorithm can be a t t r i b u t e d to De Dobbeleer [82], who suggested that d i a s t o l i c pressure can be determined with a mul t i - c u f f system by n o t i c i n g the phase s h i f t s between the o s c i l l o m e t r i c s i g n a l s i n the d i f f e r e n t c u f f s . He reasoned that while a proximal cuff receives an o s c i l l o m e t r i c s i g n a l when an a r t e r i a l pressure pulse h i t s the proximal edge of the bladder, the d i s t a l c u f f w i l l record the s i g n a l only a f t e r the p u l s a t i l e flow has forced i t s way through the ar t e r y being compressed by the proximal c u f f . When the proximal c u f f i s below s y s t o l i c pressure but above d i a s t o l i c pressure, the delayed flow reaches the d i s t a l c u f f about 30 milliseconds a f t e r a pulse i s detected at the proximal c u f f . When the proximal cuff i s below d i a s t o l i c pressure, the arte r y remains f u l l y open and the unobstructed flow produces aligned pulses i n both c u f f s . Therefore, d i a s t o l i c measurements can be done with -67-the prototype by measuring phase s h i f t s between the signals received from two bladders and determining at what proximal bladder pressure the phase s h i f t i s zero or nonzero. While De Dobbeleer used a steady proximal cuff d e f l a t i o n , the prototype can d i t h e r the proximal pressure about estimated d i a s t o l i c values to y i e l d a rapid measurement algorithm f o r t r a c k i n g d i a s t o l i c pressure with continuous estimates. 4.9 Implementation of an Algorithm f o r Rapidly Estimating Limb Occlusion Pressure i n Noisy Environments While t h i s t h e s i s addresses improving the measurement of a l l blood pressure values i n the presence of noise, the major e f f o r t of t h i s work has been to improve the c l i n i c a l performance of adaptive tourniquets by developing a safer, f a s t e r , more accurate, and more r e l i a b l e method f o r t r a c k i n g limb occlusion pressure i n moderately noisy s u r g i c a l environments. The approach taken involves analyzing only data obtained from the b r i e f noise-free i n t e r v a l s that often separate r e c u r r i n g noisy periods i n r e a l i s t i c c l i n i c a l s i t u a t i o n s , and then using t h i s data to examine flow conditions past a v a r i a b l y pressurized c u f f kept near the limb occlusion pressure. This approach i s made pos s i b l e with the m u l t i -bladder c u f f that provides the means to detect noise, estimate limb occlusion pressure r a p i d l y , act as an occluding c u f f , and reduce the errors associated with the use of separate occluding and pressure-estimation c u f f s on d i f f e r e n t limbs. While the proximal bladder of the cuf f i s used as a "flow valve" to block and permit blood flow to the other two bladders through an adjustment of i t s p r e s s u r i z a t i o n , the middle bladder serves as both the occluding device and a flow detector, and the f i n a l d i s t a l and surrounding bladder i s used to detect noise. F i g . 4.7 shows the r e l a t i o n s h i p between t y p i c a l waveforms seen i n the various bladders. When the middle bladder i s properly occluding underlying a r t e r i e s , pulses i n the noise bladder can only be caused by noise a r t i f a c t s that also d i s t u r b the signals i n the other bladders. The F i g . 4.7 T y p i c a l waveforms observed in the m u l t i - b l a d d e r c u f f : (a) w i th the prox ima l b l adder p re s su re above the l imb o c c l u s i o n p r e s s u r e ; and (b) with the proximal b ladder pres sure below the l imb o c c l u s i o n p r e s s u r e . In the a l g o r i t h m , when pu l ses are de tec ted from the no i se b l a d d e r , p r o c e s s i n g i s suspended to prevent no i se a r t i f a c t s from be ing mis taken f o r f l o w - i n d u c e d pu l ses at the middle b l a d d e r . When no s i g n a l i s d e t e c t e d from the no i se b l a d d e r , the absence or presence of f l o w - i n d u c e d pu l se s from the m idd le b l adder r e v e a l whether the proximal p r e s s u r e i s above or below the r e q u i r e d l imb o c c l u s i o n p r e s s u r e . -69-m i s i n t e r p r e t a t i o n of noise a r t i f a c t s f o r flow-induced s i g n a l s i n the middle bladder i s avoided by suspending processing when any pulses are detected i n the noise bladder. The absence of noise-induced pulses i n the noise bladder indicates that the s i g n a l i n the middle bladder i s not corrupted by noise and accurately r e f l e c t s the actual state of a r t e r i a l flow past the proximal bladder, the c l i n i c a l condition that i s to be detected. Under noise-free conditions, the absence or presence of flow-induced pulses i n the middle bladder determines i f the proximal bladder pressure i s above or below the required limb occlusion pressure, r e s p e c t i v e l y . It was a n t i c i p a t e d that an algorithm incorporating t h i s integrated c u f f f o r estimating limb occlusion pressure would have the p o t e n t i a l to be more accurate than current methods because i t would be based d i r e c t l y on flow past a c u f f , more r e l i a b l e since data a c q u i s i t i o n i s suspended when noise i s detected on a separate noise channel, and more rap i d because the pressure i n the proximal bladder i s maintained near the limb occlusion pressure to narrow the range of pressures being searched. The flowchart shown i n F i g . 4.8 outlines the algorithm that was developed and explains how, once the proximal c u f f pressure approaches the limb occlusion pressure, i t i s adjusted to d i t h e r across t h i s target pressure to track changes i n limb occlusion pressure and provide during noise-free periods r a p i d estimates of the required tourniquet pressure. The use of i d e n t i c a l bladders on the same c u f f ensures that the limb occlusion pressure estimated at the proximal bladder i s s i m i l a r to the minimum tourniquet pressure that could be used i n the middle occluding bladder, since f a c t o r s a f f e c t i n g limb occlusion pressure such as bladder s i z e , c u f f a p p l i c a t i o n , and limb physiology are s i m i l a r f o r the two bladders. This technique of tr a c k i n g limb occlusion pressure i s s i m i l a r i n some respects to a method developed by Krueger [83], who tracked s y s t o l i c (or d i a s t o l i c ) pressure by sensing the presence and amplitude of Korotkoff sounds to decide whether a s i n g l e occluding c u f f was above or below - 7 0 -F i g . 4.8 Flowchart f o r the algorithm used to t r a c k limb o c c l u s i o n pressure. -71-s y s t o l i c (or d i a s t o l i c ) pressure, and to make appropriate c u f f pressure changes based on t h i s d e c i s i o n . However, Krueger's method i s not s u i t a b l e f o r use i n an adaptive tourniquet because i t requires blood flow past a c u f f to produce Korotkoff sounds f o r a n a l y s i s , because i t would be s e n s i t i v e to noise and transducer placement, and because i t o f f e r s no r e l i a b l e way to ensure that the data being processed i s not corrupted by noise. Appendix 1 contains a l i s t i n g of the C-language program used to implement the algorithm f o r t r a c k i n g limb occlusion pressure. The following provides an abbreviated d e s c r i p t i o n of the program: 1) i n i t i a l i z e v a r i a b l e s ; 2) accept from the user the tourniquet pressure to be used i n the middle c u f f and an estimate of the limb occlusion pressure; 3) c a l i b r a t e the s t a t i c pressure a m p l i f i e r s using zero pressure and the known source pressure; 4) enter a routine which allows the user to set the EEG a m p l i f i e r gain that w i l l depend on limb physiology, c u f f a p p l i c a t i o n , limb s i z e and other factors that a f f e c t the magnitude of o s c i l l o m e t r i c waveforms; 5) accept inputs from the user to adjust the p o s i t i v e and negative noise thresholds that depend on the a m p l i f i e r gain; 6) accept inputs from the user to adjust the flow detector threshold, which w i l l depend on the a m p l i f i e r gain; 7) set-up the IBM PC screen to d i s p l a y waveforms, thresholds, bladder pressures, and trends of limb occlusion pressure; 8) enter the main c o n t r o l loop to a) refresh the screen as required; b) read the keyboard to check f o r user i n t e r v e n t i o n ; c) sample the o s c i l l o m e t r i c waveform ( f i l t e r e d with an a n t i - a l i a s i n g 8 Hz f i l t e r ) from the d i f f e r e n t bladders (with sampling synchronized to the IBM 18.2 Hz i n t e r n a l -72-c l o c k ) ; d) f i l t e r the sampled waveforms with an eighth-order d i g i t a l FIR bandpass f i l t e r , having 3 dB frequencies of 1.4 Hz and 7.6 Hz, and requ i r i n g 4 additions and 2 m u l t i p l i c a t i o n s ; e) p l o t the f i l t e r e d waveforms onto the screen; f) check f o r the presence of noise, i n d i c a t e d by a f i l t e r e d noise s i g n a l that exceeds e i t h e r noise threshold l e v e l , a f i l t e r e d flow detection s i g n a l that i s below baseline by an amount equal to the flow-detection threshold (the s i g n a l should be unipolar), or an u n f i l t e r e d flow detection s i g n a l that indicates that the EEG am p l i f i e r i s i n a state of noise-induced saturation; g) i f noisy, reset the EEG am p l i f i e r , regulate bladder pressures, check that the maximum allowable time between estimates has not been exceeded because of the delaying e f f e c t of i n t e r f e r i n g noise, and get a new block of data; h) i f quiet and no f i l t e r e d flow s i g n a l has exceeded the flow detection threshold f o r 2 seconds, assume that the proximal bladder i s above the limb occlusion pressure, decrease the proximal bladder pressure by 6 mmHg, and regulate other bladder pressures; i ) i f quiet before and a f t e r the f i l t e r e d flow s i g n a l exceeded the flow detection threshold, assume that the flow s i g n a l i s v a l i d , determine that the proximal bladder i s below the limb occlusion pressure, increase the proximal bladder pressure by 6 mmHg, and regulate other bladder pressures; j) i f a t r a n s i t i o n i n flow conditions has been observed, estimate the current limb occlusion pressure to be equal to the average of the two successive proximal pressures between which the t r a n s i t i o n occurred; and k) store the new estimates of limb occlusion along with the -73-time they occurred f o r updating the p l o t of pressure trends; 9) store the trend r e s u l t s on a disk; and 10) shut-down the system. While estimating limb occlusion pressure, the program provides the user with the following features: an alarm that sounds when the time between estimates exceeds a time period s p e c i f i e d by the user, a l e r t i n g the user that the observed delay i n updating limb occlusion pressure may a f f e c t the pressure-tracking process; a leak and kink detector that n o t i f i e s the user when a bladder i s taking over 15 seconds to reach i t s desi r e d pressure; a p l o t of the pressure trend; an optional 3-point moving average f i l t e r applied to estimates to help reduce the short-term v a r i a b i l i t y that e x i s t s i n successive determinations; and an a b i l i t y while processing waveforms to accept commands from the user to change the tourniquet pressure, adjust thresholds, a l t e r the maximum allowable time between estimates, enable or disable the 3-point moving average f i l t e r , and end the t r i a l . F i g . 4.9 shows how o s c i l l o m e t r i c waveforms, pressure trends, thresholds, alarm s e t t i n g s , and baseline pressures are displayed on the IBM PC screen. - 7 4 -F i g . 4.9 D i s p l a y g r a p h i c s f o r t h e p r o t o t y p e . -75-CHAPTER 5 A COMPARATIVE EVALUATION OF DIFFERENT METHODS OF ESTIMATING LIMB OCCLUSION PRESSURE 5.1 Overview The a b i l i t y of the prototype system, described i n chapter 3 and programmed with the algorithm o u t l i n e d i n chapter 4, to track limb occlusion pressure r a p i d l y i n d i f f e r e n t noise conditions was evaluated by comparing the performance of the prototype system with that of a t y p i c a l commercially a v a i l a b l e o s c i l l o m e t r i c monitor under s i m i l a r conditions. The evaluation consisted of two phases: lab t r i a l s were performed to evaluate accuracy and r e l i a b i l i t y under c o n t r o l l e d conditions; and c l i n i c a l t r i a l s were then performed to evaluate the speed, t r a c k i n g a b i l i t y , and safety of a v a i l a b l e methods f o r updating tourniquet pressures during s u r g i c a l procedures. The monitor chosen f o r comparison was the Hewlett Packard 78354A Patient Monitor, which estimates blood pressure using an o s c i l l o m e t r i c algorithm that i s representative of o s c i l l o m e t r i c algorithms widely used i n commercially a v a i l a b l e devices [58] . The bandwidth of the prototype system f o r these t e s t s was determined at the high frequency end by the a n t i - a l i a s i n g f i l t e r used to f i l t e r the o s c i l l o m e t r i c waveform before sampling (8 Hz LPF), and at the low frequency end by the HPF s e t t i n g used on the EEG a m p l i f i e r . It was discovered e m p i r i c a l l y during t e s t i n g that the recovery time of the high-gain a m p l i f i e r employed could be reduced during periods of high-amplitude noise a r t i f a c t s that saturate the a m p l i f i e r by s e t t i n g i t s HPF to 5 Hz. Although t h i s s e t t i n g d i s t o r t e d the waveform, i t improved the performance of the prototype i n noisy environments and s t i l l permitted s u f f i c i e n t higher frequency components of the s i g n a l to be passed f o r the simple analysis of flow detection. This d i s t o r t i o n of the -76-o s c i l l o m e t r i c should be avoided i f using algorithms that depend on pulse amplitude or phase e i t h e r by improving the design of the high-gain a m p l i f i e r , or by using an A/D converter with higher r e s o l u t i o n that would allow f o r comparable pulse amplitude d i s c r i m i n a t i o n with a lower gain s e t t i n g on the a m p l i f i e r . 5.2 The Protocol f o r Lab T r i a l s Lab t r i a l s were performed at the Vancouver General Hospital to compare, under s i m i l a r and reproducible conditions, the accuracy and r e l i a b i l i t y of estimates of limb occlusion pressure made i n a noisy environment by both the prototype system and a Hewlett Packard 78354A Patient Monitor. Records were made only of the o s c i l l o m e t r i c monitor's estimates of s y s t o l i c pressure, which correspond to limb occlusion pressure when the estimates are made using a tourniquet c u f f instead of a standard blood pressure c u f f . The o s c i l l o m e t r i c monitor was tested on 15 i n d i v i d u a l s who were positioned comfortably i n a supine p o s i t i o n . The multi-bladder tourniquet cuff shown i n F i g . 3.3 was wrapped on the subject's r i g h t arm between the elbow and shoulder. The proximal bladder was then used to obtain limb occlusion pressure estimates with the o s c i l l o m e t r i c monitor while an u l t r a s o n i c a r t e r i a l flow detector (Arteriosonde 1010, Roche Medical E l e c t r o n i c s Inc., Cranbury NJ), with i t s transducer placed d i s t a l to the cu f f over the b r a c h i a l artery, simultaneously i n d i c a t e d the onset of flow. Both the cu f f pressure, as in d i c a t e d by the o s c i l l o m e t r i c monitor's disp l a y , at which flow-induced sounds were f i r s t heard and the limb occlusion pressure estimate provided by the o s c i l l o m e t r i c monitor were recorded. This scheme of simultaneous measurements on the same limb eliminates the temporal and s p a t i a l pressure v a r i a t i o n s that make i t d i f f i c u l t to compare blood pressure done at d i f f e r e n t times on d i f f e r e n t limbs [84]. Three measurements, each occurring during d i f f e r e n t noise conditions (noise-free, low-noise, and moderate-noise conditions), were - 7 7 -taken on each subject. Noisy runs consisted of a l t e r n a t i n g 5 second periods of noise and noise-free i n t e r v a l s to f a c i l i t a t e the simultaneous use of the u l t r a s o n i c device during the noise-free sections, since limb movement generally created a Doppler s i g n a l that would obscure the flow-induced s i g n a l . Movement a r t i f a c t s were generated by rocking the arm of a passive subject at a frequency of about 1 Hz, with 0.5 inch movement at the elbow cre a t i n g low-noise conditions and 2 inch movement at the elbow producing moderate-noise conditions. While a rocking motion was permitted, the o r i e n t a t i o n of the limb remained constant f o r the duration of the lab t r i a l run. This simulation of c l i n i c a l a r t i f a c t conditions was defined a f t e r evaluating the nature and range of limb movements t y p i c a l of common s u r g i c a l procedures i n which tourniquets are employed, such as diagnostic and therapeutic arthroscopies, j o i n t replacements, and trauma surgery. The a b i l i t y of the prototype to detect the onset of flow past an occluding c u f f and estimate limb occlusion pressure under varying noise conditions was tested using the same pr o t o c o l . During lab t r i a l s , the algorithm of the prototype was modified to employ a stepped d e f l a t i o n scheme s i m i l a r to that used by the o s c i l l o m e t r i c monitor, instead of the d i t h e r i n g pressure regulation employed i n c l i n i c a l s i t u a t i o n s , because such a d e f l a t i o n not only made flow detection with the u l t r a s o n i c device eas i e r , but also standardized the conditions under which both the o s c i l l o m e t r i c monitor and the prototype estimated limb occlusion pressure. Furthermore, the middle bladder was set to sub-occlusive pressures i n the lab t r i a l s to permit the u l t r a s o n i c device to detect a r t e r i a l flow at a s i t e d i s t a l to the multi-bladder c u f f . The d i f f e r e n c e s between paired simultaneous measurements ("paired estimates") were then analyzed to examine the accuracy and r e l i a b i l i t y of the estimates of limb occlusion pressure made by the o s c i l l o m e t r i c monitor and the prototype under varying noise conditions i n a manner s i m i l a r to that recommended by the Assoc i a t i o n f o r the -78-Advancement of Medical Instrumentation (AAMI) f o r assessing the accuracy and r e l i a b i l i t y of.commercially a v a i l a b l e automated sphygmomanometers [84]. Even though the number of subjects used i n t h i s i n i t i a l study f a l l s short of the number recommended by the proposed AAMI standard f o r evaluating the performance of commercially a v a i l a b l e sphygmomanometers, i t i s s u f f i c i e n t f o r providing an i n i t i a l i n d i c a t i o n of the prototype's f e a s i b i l i t y f o r the proposed a p p l i c a t i o n . 5.3 The Protocol f o r C l i n i c a l T r i a l s C l i n i c a l t r i a l s were performed i n the Health Sciences Centre Hospital at the U n i v e r s i t y of B r i t i s h Columbia to compare the speed and safety of the prototype with the o s c i l l o m e t r i c monitor used f o r reference purposes during s i m i l a r s u r g i c a l procedures performed by the same surgeon. Data was c o l l e c t e d during s i x s u r g i c a l arthroscopies of the knee. With the multi-bladder c u f f wrapped around the limb undergoing surgery (Fig. 5.1), the middle bladder was i n i t i a l l y maintained at the constant tourniquet pressure normally selected f o r the procedure. The proximal bladder was then used by e i t h e r the prototype or the o s c i l l o m e t r i c monitor to make estimates of limb occlusion pressure. These estimates were recorded, together with simultaneous estimates of s y s t o l i c pressure made by the anesthetist on a limb not a f f e c t e d by the s u r g i c a l procedure, when a v a i l a b l e . The r e s u l t i n g data was c o l l e c t e d to permit a comparative evaluation of the safety and speed of adaptive tourniquet schemes that might use estimates of s y s t o l i c pressure from another limb, or estimates of limb occlusion pressure derived from the same limb by means of an o s c i l l o m e t r i c monitor, with estimates of limb occlusion pressure derived from the same limb by the prototype. 5.4 Performance C r i t e r i a A comparative evaluation of the prototype and the o s c i l l o m e t r i c monitor was based on the following c r i t e r i a : -79-I F i g . 5.1 The use of the multi-bladder c u f f during c l i n i c a l t r i a l s . -80-1) accuracy, i . e . the average d i f f e r e n c e of pa i r e d estimates c o l l e c t e d during lab t r i a l s ; 2) r e l i a b i l i t y , i . e . the standard deviation of dif f e r e n c e s of pair e d estimates c o l l e c t e d during lab t r i a l s ; 3) speed, i . e . the average estimation time and maximum time between estimates during c l i n i c a l t r i a l s ; and 4) safety, i . e . the l i k e l i h o o d that an occlusive c u f f w i l l be maintained above the limb occlusion pressure, without excessive pressure a p p l i c a t i o n . Accuracy and r e l i a b i l i t y are two important parameters, since they i n d i c a t e to the user how confident he can be that estimated values t r u l y r e f l e c t the limb occlusion pressure being measured and used f o r adapting tourniquet pressures. Limits f o r accuracy and r e l i a b i l i t y l e v e l s were obtained from the proposed AAMI standard [84], which recommends that the mean dif f e r e n c e between estimates made with a non-invasive blood pressure monitor and simultaneous manual measurements has a magnitude les s than 5 mmHg, and that the standard d e v i a t i o n of diff e r e n c e s of paired estimates i s less than 8 mmHg. An adaptive scheme f o r tourniquets must be ra p i d enough to track the s i g n i f i c a n t changes i n limb occlusion pressure that may occur during surgery so that appropriate changes can be made i n the tourniquet pressure. While estimates i d e a l l y should be updated "as often as possibl e " , a r e a l i s t i c minimum speed value was derived by observing the t y p i c a l changes i n limb occlusion pressure that occurred during surgery. Because safety considers the l i k e l i h o o d that an occlusive cuff w i l l be maintained above the limb occlusion pressure, i t encompasses many d i f f e r e n t parameters such as accuracy, r e l i a b i l i t y , speed, and range of estimates, a l l of which may a f f e c t the a b i l i t y of an adaptive scheme to maintain occlusive pressures. The safety of d i f f e r e n t adaptive schemes was evaluated by considering these f a c t o r s . - 8 1 -5 . 5 Discussion of Results Tables 5 . 1 - 5 . 3 summarize the lab t r i a l r e s u l t s , c o n s i s t i n g of an analysis of paired estimates measured from 15 subjects to compare the average d i f f e r e n c e , standard deviation of the d i f f e r e n c e s , and range of d i f f e r e n c e s obtained with both the o s c i l l o m e t r i c monitor and the prototype during v a r i e d noise conditions. Table 5 . 4 shows the r e s u l t s of c l i n i c a l t r i a l s concerning estimation speed i n a noisy environment that were obtained during 6 knee arthroscopies, 3 each f o r both the o s c i l l o m e t r i c monitor and the prototype. In general, the r e s u l t s summarized i n Tables 5 . 1 - 5 . 4 showed that the presence of noise a f f e c t e d the performance of both the o s c i l l o m e t r i c monitor and the prototype. However, the degree to which noise a f f e c t e d the a b i l i t y of each monitor to track limb occlusion pressure was noticeably d i f f e r e n t . 5 . 5 . 1 Results of Lab T r i a l s of the O s c i l l o m e t r i c Monitor The o s c i l l o m e t r i c monitor i s based on a widely employed technique f o r estimating mean a r t e r i a l pressure, s y s t o l i c pressure, and d i a s t o l i c pressure (e.g. [ 5 7 - 6 0 ]). It r e j e c t s noise by using a "matched pulse" scheme f o r accepting data, processing only p a i r s of pulses that are s i m i l a r i n amplitude and time separation. Thus, the monitor requires a ser i e s of noise-free periods, t y p i c a l l y 2 seconds each and t o t a l l i n g 40 seconds, i n order to c o l l e c t enough data to construct an adequate table of pulse heights and cuff pressures f o r estimating these three parameters of pressure. Because the monitor waits f o r a matched set of pulses before continuing with i t s d e f l a t i o n process, the presence of detected noise f o r prolonged periods can t r i g g e r alarms (after 60 seconds at a given pressure l e v e l or 2 minutes of t o t a l elapsed time) that w i l l abort the estimation c y c l e . As i n d i c a t e d i n Tables 5 . 1 - 5 . 3 , the r e s u l t s suggest that both -82-Table 5.1 Results of l a b t r i a l s used to assess the accuracy of both the o s c i l l o m e t r i c monitor and the prototype. Condition Average D i f f e r e n c e Between Pa i r e d Estimates (mmHg) O s c i l l o m e t r i c Prototype Monitor Noise-free 4.6 ' -0.1 Low noise 9.7 -0.3 Moderate noise 12.1 0.8 Table 5.2 Results of lab t r i a l s used to assess the r e l i a b i l i t y of of the o s c i l l o m e t r i c monitor and of the prototype. Condition Standard D e v i a t i o n of D i f f e r e n c e s Between P a i r e d Estimates (mmHg) O s c i l l o m e t r i c Prototype Monitor Noise-free Low noise Moderate noise 4.7 8.3 9.8 3.9 3.5 5.4 -83-Table 5.3 Results of lab t r i a l s used to assess the s a f e t y of both the o s c i l l o m e t r i c monitor and the prototype. Condition Range of D i f f e r e n c e s Between P a i r e d Estimates (mmHg) O s c i l l o m e t r i c Prototype Monitor Noise-free 16 16 Low noise 31 12 Moderate noise 39 19 Table 5 . 4 Results of c l i n i c a l t r i a l s used to assess the speed of both the o s c i l l o m e t r i c monitor and the prototype algorithm. Average Estimation Time Maximum Time Between (s/estimate) Estimates (s) O s c i l l o m e t r i c Prototype O s c i l l o m e t r i c Prototype Monitor Monitor 5310/50 = 106 3198/256 = 12.5 228 95 -84-low-amplitude and moderate-amplitude noise a f f e c t e d the accuracy, r e l i a b i l t y , and safety of the o s c i l l o m e t r i c monitor. While noise-free measurements s a t i s f i e d the proposed AAMI standards by d i f f e r i n g from the simultaneous manual measurements by less than 5 mmHg with a standard de v i a t i o n of less than 8 mmHg, the noisy estimates f a i l e d to meet these standards. Not only were measurements made i n the presence of motion a r t i f a c t s l e s s accurate and r e l i a b l e , but these estimates also spanned a lar g e r range of estimates to y i e l d a p o t e n t i a l l y hazardous condition i f these estimates were being used to adapt a tourniquet because blood might flow under the occlusive c u f f i n t o the s u r g i c a l s i t e . The data which was c o l l e c t e d also suggested that the o s c i l l o m e t r i c monitor at times accepted c y c l i c noise as being v a l i d o s c i l l o m e t r i c matched pulses, r e s u l t i n g i n erro r s i n estimated limb occlusion pressure i n which the p o l a r i t y and magnitude of the er r o r depended on when i n the data a c q u i s i t i o n cycle the er r o r occurred. For example, i f the erro r occurred e a r l y i n the cycle of the monitor, with the cuff pressure greater than the limb occlusion presssure, the monitor would record a large increase of pulse s i z e ( i . e . , i t s a r t e r i a l occlusion c r i t e r i o n would be met) well above the actu a l limb occlusion pressure, and consequently the monitor would overestimate the value of limb occlusion pressure. Conversely, i f a s i m i l a r e r r o r occurred l a t e i n the cycle at a low cuff pressure, the limb occlusion pressure would be underestimated. It was also noted that the o s c i l l o m e t r i c monitor f a i l e d to occlude the arter y of two subjects before beginning i t s data a q u i s i t i o n cycle, as flow was detected by the u l t r a s o n i c flow detector even before the o s c i l l o m e t r i c monitor began to de f l a t e the c u f f . These runs, excluded from the study, underestimated the limb occlusion pressure by more than 30 mmHg. This e r r o r probably r e s u l t e d from the use of a standard tourniquet cuff instead of a standard blood pressure c u f f , since the padded, narrower tourniquet c u f f generates smaller flow-induced o s c i l l o m e t r i c pulsations than the blood pressure c u f f . It was noted that these errors occurred despite the fac t that the -85-algorithm of the o s c i l l o m e t r i c monitor i s designed to check that the cuff i n f l a t i o n l e v e l i s s u f f i c i e n t f o r occluding blood flow i n underlying a r t e r i e s before beginning i t s data a c q u i s i t i o n c y c l e . 5.5.2 Results of Lab T r i a l s of the Prototype The prototype system uses a multi-bladder c u f f to sense the occurrence of noise-free periods when i t can detect flow past a v a r i a b l y pressurized proximal bladder to make estimates of limb occlusion pressure. Thus t h i s system requires a se r i e s of b r i e f noise-free periods, of duration le s s than 2 seconds, as i t tracks the limb occlusion pressure by making appropriate adjustments of the proximal pressure based on flow conditions past the proximal bladder. The number of noise-free data periods needed f o r an estimate depends on the v a r i a b i l i t y of the limb occlusion pressure and on the dif f e r e n c e between the limb occlusion pressure and the pressure i n the proximal bladder at measurement times. As these v a r i a b l e s increase, the time spent searching f o r the limb occlusion pressure increases. As summarized i n Tables 5.1-5.3, the r e s u l t s i n d i c a t e that the prototype performed c o n s i s t e n t l y w ell f o r d i f f e r e n t noise conditions, with the average d i f f e r e n c e between simultaneous estimates and the standard deviation of these differences meeting the proposed AAMI guidelines f o r a l l cases. Moreover, the range of noisy estimates made by the prototype i s comparable to the range of noise-free estimates made by the o s c i l l o m e t r i c monitor. La s t l y , the prototype was able to provide limb occlusion pressure estimates f o r the two subjects described above fo r whom the o s c i l l o m e t r i c monitor f a i l e d to occlude blood flow. 5.5.3 Results of C l i n i c a l T r i a l s As i n d i c a t e d i n Table 5.4, the r e s u l t s show that the prototype was able to track limb occlusion pressure s i g n i f i c a n t l y b e t t e r than the -86-o s c i l l o m e t r i c monitor; t h i s can be seen by comparing both the average estimation time and the maximum time between estimates. The importance of r a p i d pressure t r a c k i n g i s g r a p h i c a l l y displayed i n F i g . 5 . 2 and F i g . 5 . 3 , which show the intraoperative limb occlusion pressure estimates made at the tourniquet cuff by both the prototype and the o s c i l l o m e t r i c monitor, and compare these to the intr a o p e r a t i v e s y s t o l i c pressure estimates derived from another limb. Not only were large limb occlusion pressure changes of 40 mmHg/minute observed, but these changes were not associated with a s i m i l a r change i n s y s t o l i c estimates measured on another limb. Such rapid changes appear to represent v a l i d v a r i a t i o n s i n limb occlusion pressure, and do not appear to be err o r s caused by motion a r t i f a c t s , i n view of the following: these wide v a r i a t i o n s i n limb occlusion pressure could be reproduced i n a lab environment by taking noise-free measurements on a l e g f i x e d i n d i f f e r e n t o r i e n t a t i o n s , as shown i n F i g . 5 . 4 and F i g . 5 . 5 ; and e a r l i e r lab t r i a l s with the prototype e s t a b l i s h e d the accuracy and r e l i a b i l i t y of the prototype i n noisy environments. Consequently, while s y s t o l i c pressure i s one of the factors that a f f e c t s limb occlusion pressure, changes i n limb p o s i t i o n can be a more s i g n i f i c a n t f a c t o r during periods of stable s y s t o l i c pressure l e v e l s . A change i n limb o r i e n t a t i o n could have caused these v a r i a t i o n s by changing the elev a t i o n of the cuff with respect to the heart, by s h i f t i n g the cu f f p o s i t i o n along the limb to a f f e c t the cuff-limb i n t e r f a c e , or a l t e r i n g the circumference of the limb as limb t i s s u e was stretched or relaxed. Besides d i s p l a y i n g r e s u l t s of the c l i n i c a l t r i a l s , F i g . 5 . 2 and F i g . 5 . 3 also i l l u s t r a t e the pressures that might have been used i n the c u f f s of adaptive tourniquets based on adaptive schemes employing (a) estimates of s y s t o l i c pressure derived from another limb, (b) estimates of limb occlusion pressure made at the s i t e of the tourniquet c u f f by the o s c i l l o m e t r i c monitor, and (c) estimates of limb occlusion pressure made at the s i t e of the tourniquet c u f f by the prototype. An adaptive scheme -87-300 Constant Tourniquet Pressure Employed r -1 T I i L I ^Adaptive Tourniquet (Different Limb; k f *"* Oscillometric Monitor) r j : "1 . X e E g 200 to E ^-Adaptive Tourniquet (Same Limb; ^ Oscillometric Monitor) Limb Occlusion Press Systolic Pressure , "AN 100 k 8 12 16 20 2k 28 32 36 kk kB TIME (min.) F i g . 5.2 Results of a c l i n i c a l t r i a l with the o s c i l l o m e t r i c monitor. This graph shows the r e s u l t s of a c l i n i c a l t r i a l , i n c l u d i n g (a) s y s t o l i c pressure estimates from the p a t i e n t ' s arm made by an o s c i l l o m e t r i c monitor and (b) limb o c c l u s i o n pressure estimates made by an o s c i l l o m e t r i c monitor at the s i t e of the tourniquet c u f f . Also shown are: the constant tourniquet pressure a c t u a l l y used; the p r o j e c t e d adaptive tourniquet pressure d e r i v e d from a d i f f e r e n t limb; and the p r o j e c t e d adaptive tourniquet pressure d e r i v e d from the same limb. - 8 8 -300 /• Constant Tourniquet Pressure Employed Adaptive Tourniquet (Different Llmbj Oscillometric Monitor) ( Adaptive Tourniquet (Same Llmbj Prototype) to w 200 K tn A - - A - , A Systolic Pressure 100 • • • • — • • r i • • - — 100 200 300 kOQ 500 600 700 800 900 1000 1100 1200 TIME (s) F i g . 5.3 Results of a c l i n i c a l t r i a l with the prototype. This graph shows the r e s u l t s of a c l i n i c a l t r i a l , i n c l u d i n g (a) s y s t o l i c pressure estimates from the pa t i e n t ' s arm made by an o s c i l l o m e t r i c monitor and (b) limb o c c l u s i o n pressure estimates made by the prototype at the s i t e of the tourniquet c u f f . Also shown are: the constant tourniquet pressure a c t u a l l y used; the p r o j e c t e d adaptive tourniquet pressure d e r i v e d from a d i f f e r e n t limb; and the p r o j e c t e d adaptive tourniquet pressure d e r i v e d from the same limb. -89-270 Level Elevated Level Lowered Level Limb 1 Limb 1 Limb 1 Limb 1 Limb 220 -170 t 1 I 1 ; 100 200 300 400 500 600 700 800 900 1000 1 >00 1200 T I M E ( s ) F i g . 5.4 V a r i a t i o n of limb o c c l u s i o n pressure with limb p o s i t i o n f o r estimates made by the o s c i l l o m e t r i c monitor during a no i s e - f r e e lab t r i a l . -90-Level Elevated Level Lonered Level T I M E ( s ) F i g . 5.5 V a r i a t i o n of limb o c c l u s i o n pressure with limb p o s i t i o n f o r estimates made by the prototype during a n o i s e - f r e e lab t r i a l . -91-that uses s y s t o l i c pressures estimated from another limb requires that the estimates be corrected to account f o r d i f f e r e n t c u f f s i z e s and types, limb circumferences, and techniques of cu f f a p p l i c a t i o n . Such a scheme must also include an additi v e constant to account f o r err o r s i n pressure regulation, s y s t o l i c pressure estimations and c u f f a p p l i c a t i o n s , and to account f o r the maximum increase i n the patient's s y s t o l i c pressure that i s l i k e l y to occur between estimates [32]. For F i g . 5.2 and F i g . 5.3, t h i s c o r r e c t i v e f a c t o r was estimated from measured limb circumference information and the previously derived r a t i o between c u f f width and limb circumference [38]; and f o r the same data the ad d i t i v e constant was estimated to be about 50 mmHg [32] . In contrast to the adaptive scheme based on s y s t o l i c estimates from another limb, both adaptive schemes shown i n F i g . 5.2 and F i g . 5.3 that are based on measuring limb occlusion pressure at the s i t e of the tourniquet cuff require only the addi t i o n of a constant f a c t o r to account f o r errors i n estimation, errors i n pressure regulation, and maximum increases i n s y s t o l i c pressure between estimates, t o t a l l i n g about 40 mmHg. Less modification i s required f o r limb occlusion estimates made at the tourniquet s i t e because i d e n t i c a l bladders from the same cuff are used to measure pressure and occlude underlying a r t e r i e s on the same limb. A l l adaptive schemes would r e s u l t i n the use of a lower tourniquet pressure than the constant pressure a c t u a l l y employed, i n d i c a t i n g that any adaptive algorithm can help to reduce the average tourniquet pressure used during surgery. However, as shown i n F i g . 5.2 and F i g . 5.3, the use of s y s t o l i c pressure estimates from another limb would have r e s u l t e d i n higher mean tourniquet pressures than those projected from the use of limb occlusion pressure estimates at the actual s i t e of the tourniquet c u f f . Although blood pressure estimates from another limb were advantageously unaffected by motion a r t i f a c t s generated at the tourniquet by the s u r g i c a l procedure, these estimates required large c o r r e c t i v e factors to account f o r diff e r e n c e s i n c u f f s , limbs and -92-c u f f a p p l i c a t i o n . Furthermore, as shown i n F i g . 5.3, measurements on another limb d i d not track the wide v a r i a t i o n s i n limb occlusion pressure caused by changes i n limb p o s i t i o n , a shortcoming that could produce p o t e n t i a l l y hazardous underpressurization errors i f only s y s t o l i c pressure estimates were used. Therefore, the ad d i t i v e constant should be increased to account f o r changes i n limb occlusion pressure r e s u l t i n g from v a r i a t i o n s i n limb p o s i t i o n (Table 5.5). The use of the o s c i l l o m e t r i c monitor to estimate limb occlusion pressure at the s i t e of the tourniquet c u f f o f f e r e d some improvements by taking i n t o account the e f f e c t s of limb o r i e n t a t i o n . However, the monitor was unable to estimate the limb occlusion pressure of some subjects during lab t r i a l s when a narrow tourniquet c u f f was employed. Also, the observed e f f e c t s of noise i n delaying the estimation process, increasing the range of estimates, and reducing the l e v e l of accuracy and r e l i a b i l i t y r e s u l t e d i n measurements by the o s c i l l o m e t r i c monitor that were slow and hazardous. I f these e f f e c t s are ignored, then underpressurization errors can occur. However, i f an ad d i t i v e constant larger than 40 mmHg i s used to account f o r both the observed degradation i n performance and the maximum increase i n limb occlusion pressure that can occur during the time between estimates (about 60 mmHg), the p o s s i b i l i t y of an underpressurization e r r o r may be reduced but there may be no s i g n i f i c a n t reduction i n the mean tourniquet pressure compared to the constant pressure normally used (Table 5.5). These problems were not observed when using the prototype, which was able to estimate limb occlusion pressure at the s i t e of the tourniquet c u f f r a p i d l y and c o n s i s t e n t l y i n varying noise conditions. Consequently, the addit i v e constant required f o r the prototype i s smaller than the constant required f o r the slower o s c i l l o m e t r i c monitor (e.g. about 50 mmHg f o r the prototype versus about 150 mmHg f o r the o s c i l l o m e t r i c monitor), r e s u l t i n g i n a lower mean tourniquet pressure without s a c r i f i c i n g safety. In other words, even when both the -93-Table 5.5 The derivation of tourniquet pressure for adaptive schemes that use different blood pressure estimates. Method of Deriving Tourniquet Pressure Errors Systolic Estimates on Another Limb Limb Occlusion Pressure Estimates by the Oscillometric Monitor at the Tourniquet Cuff Site Limb Occlusion Pressure Estimates by the Prototype at the Tourniquet Cuff Site ADDITIVE CONSTANTS (mmHg) Errors in Pressure Regulation 5 Errors in Pressure Estimation 16 (2 standard deviations) Errors in Cuff Application 10 Maximum Increase in Limb Occlusion Pressure Due to 60 Changes in Limb Position (Observed During C l i n i c a l Trials) Maximum Increase in Systolic Pressure (Observed During 15 C l i n i c a l Trials) TOTAL ADDITIVE CONSTANT 106 5 5 20 10 0 0 60 (228 s) 20 (30 s) (60 s) 60 (228 s) 10 (30 s) 145 45 MULTIPLICATIVE CONSTANT (for limb circumference:cuff width 1.4 1 1 ratio - 22/3 at the tourniquet cuff) -94-o s c i l l o m e t r i c monitor and the prototype produce i d e n t i c a l estimates of limb occlusion pressure, the tourniquet s e t t i n g derived from the o s c i l l o m e t r i c monitor w i l l be higher than that derived from the prototype because the o s c i l l o m e t r i c monitor requires a l a r g e r a d d i t i v e constant to account f o r i t s l a r g e r t r a c k i n g errors r e s u l t i n g from i t s slow, degraded performance i n noisy environments (Table 5.5). However, i t should be noted that the c a l c u l a t i o n s shown i n Table 5.5 assume that the maximum time between estimates by the prototype w i l l be about 30 s. With an average estimation time of 12.5 s, the prototype c e r t a i n l y appears capable of meeting t h i s d e s i r e d l e v e l of performance. But of the 256 estimates made by the prototype during c l i n i c a l t r i a l s , 13 estimates took over 30 s. The average d i f f e r e n c e i n pressure values between successive estimates separated by more than 30 s was about 28 mmHg, i n d i c a t i n g that such delays could be a t t r i b u t e d to both the occurrence of noise which suspends processing, and a presently i n e f f i c i e n t routine f o r searching f o r flow t r a n s i t i o n s , which requires an excessive number of pressure steps. It i s a n t i c i p a t e d that a better search routine can help to eliminate the infrequent occurrence of large delays between estimates, thereby improving the t r a c k i n g a b i l i t y of the prototype. Furthermore, the use of a smaller allowable time between estimates f o r the programmable audio alarm (about 20 s) may help to reduce the delays r e s u l t i n g from long periods of continuous noise by a l e r t i n g the surgeon that he should suspend the s u r g i c a l procedure b r i e f l y ( t y p i c a l l y f o r about 12 s) at the next convenient moment to allow the tourniquet pressure to be updated. However, even without these improvements, the prototype has the p o t e n t i a l to track limb occlusion pressure more accurately, more r e l i a b l y , and more r a p i d l y than commercially a v a i l a b l e o s c i l l o m e t r i c monitors. Thus, the r e s u l t s i n d i c a t e that a system f o r measuring limb occlusion pressure that i s based on the prototype system i s f e a s i b l e f o r incorporation into an adaptive tourniquet to y i e l d a s i g n i f i c a n t reduction i n mean tourniquet pressure without increasing the r i s k of underpressurization. -95-CHAPTER 6 CONCLUSIONS 6.1 Summary This t h e s i s has addressed the problem of measuring blood pressure i n the presence of motion a r t i f a c t s . A review of common techniques of measuring blood pressure i n d i c a t e d that o s c i l l o m e t r y o f f e r s a p r e f e r r e d method of automating blood pressure measurements because i t i s non-invasive, does not s u f f e r from transducer placement or transducer cost problems, does not require d i r e c t skin contact, o f f e r s a mean pressure c r i t e r i o n , i s immune to audio noise, and i s able to determine the blood pressure of patients who have low blood flow. A prototype system based on oscillometry has been assembled to deal with noise i n two ways: the use of a multi-bladder cuff y i e l d s a s p e c i a l i z e d noise channel that r e l i a b l y detects noise to prevent the processing of corrupted data; and the implementation of r a p i d estimation algorithms with the multi-bladder c u f f produces new estimates using only a small number of b r i e f noise-free periods that separate r e c u r r i n g noisy i n t e r v a l s i n r e a l i s t i c c l i n i c a l environments. This combination of r e l i a b l e noise detection and r a p i d estimation gives the prototype the p o t e n t i a l to perform better than commercially a v a i l a b l e blood pressure monitors by t r a c k i n g blood pressure changes more accurately, increasing r e l i a b i l i t y i n d i f f e r e n t noise conditions, decreasing the patient discomfort associated with prolonged blood pressure estimates, and reducing the unknown errors i n blood pressure estimates caused by processing data over many heart beats when the blood pressure i s f l u c t u a t i n g . This strategy was r e a l i z e d i n one form f o r evaluation, namely a blood pressure monitor f o r t r a c k i n g limb occlusion pressure. While noise detection was accomplished by checking f o r the absence or presence -96-of a s i g n a l from a noise bladder, rapid estimation of limb occlusion pressure was achieved by checking f o r the absence or presence of a flow-induced s i g n a l on a middle bladder, a condition that would depend on the pressure s e t t i n g of a proximal bladder which was maintained near the limb occlusion pressure. The prototype was comparatively evaluated against a t y p i c a l commercially a v a i l a b l e o s c i l l o m e t r i c blood pressure monitor during lab t r i a l s and c l i n i c a l t r i a l s . The r e s u l t s of estimates made during the c o n t r o l l e d noise conditions of the lab t r i a l s with 15 subjects i n d i c a t e that the prototype performed c o n s i s t e n t l y during d i f f e r e n t noise conditions without s u f f e r i n g the degraded performance i n accuracy, r e l i a b i l i t y and range of estimates observed i n the o s c i l l o m e t r i c monitor during s i m i l a r conditions. For a l l noise conditions the prototype met the proposed AAMI standards f o r automated sphygmomanometers by producing estimates that d i f f e r e d from simultaneous measurements by less than 5 mmHg on average and y i e l d i n g standard deviations of these di f f e r e n c e s that were les s than 8 mmHg. The re s u l t s of estimates made during the c l i n i c a l t r i a l s i n v o l v i n g 6 knee arthroscopies show that the prototype was able to estimate limb occlusion pressure more r a p i d l y than the o s c i l l o m e t r i c monitor (on average about 7 times f a s t e r ) . The r e s u l t s of c l i n i c a l t r i a l s also show that changes i n limb p o s i t i o n can produce changes i n limb occlusion pressure (up to 60 mmHg) that are observed i n estimates of limb occlusion pressure made at the s i t e of the tourniquet cuff but are not r e f l e c t e d by corresponding changes i n estimates of s y s t o l i c pressure made on another limb. Consequently, i n i t i a l r e s u l t s i n d i c a t e that the algorithm and strategy employed i n the prototype o f f e r a better way to adapt tourniquets than the use of e i t h e r estimates of s y s t o l i c pressure made on another limb or estimates of limb occlusion pressure made by an o s c i l l o m e t r i c monitor at the s i t e of the tourniquet c u f f . The prototype functions more r e l i a b l y , more accurately, more r a p i d l y , and more s a f e l y than a representative o s c i l l o m e t r i c monitor f o r t h i s purpose. Because -97-the prototype o f f e r s a better way to track limb occlusion pressure, when used to set tourniquet pressures i t s estimates require a smaller additive constant to account f o r t r a c k i n g errors (about 50 mmHg) than do adaptive schemes based on e i t h e r estimates of limb occlusion pressure at the tourniquet c u f f made by an o s c i l l o m e t r i c monitor (requiring an additive constant over 100 mmHg) or estimates of s y s t o l i c pressure made on another limb (requiring a m u l t i p l i c a t i v e constant based on cu f f and limb s i z e s , plus an add i t i v e constant over 100 mmHg). Thus, with smaller t r a c k i n g e r r o r s and lower allowable pressures, the prototype can y i e l d a reduction i n the mean tourniquet pressure during s u r g i c a l procedures to help reduce overpressurization hazards without introducing underpressurization e r r o r s . 6.2 Suggestions f o r Further Research Although the prototype was able to track limb occlusion pressure b e t t e r than the o s c i l l o m e t r i c monitor, the prototype requires some improvements to permit i t s routine use i n c l i n i c a l environments. 6.2.1 Amplifier Improvements The high-gain EEG am p l i f i e r used i n the prototype system should be replaced by an am p l i f i e r with s i m i l a r gain and bandwidth s p e c i f i c a t i o n s and reset c a p a b i l i t i e s . However, the new am p l i f i e r should be smaller to improve p o r t a b i l i t y and decrease the space i t occupies i n the operating room, and should have a greater peak-to-peak output to decrease the incidence of noise-induced saturation that plagues the EEG am p l i f i e r with i t s l i m i t e d 2.8 V peak-to-peak output. One poss i b l e replacement i s the Hewlett Packard Model 8811A B i o p o t e n t i a l Amplifier (Hewlett Packard Co., Palo A l t o CA), which has the following features: 1) maximum gain = 100,000; 2) remote reset; -98-3) LPF settings of 1, 10, 30, 100, 300, 1000, 3000, and 10000 Hz; 4) HPF settings of 0.05, 0.1, 0.15, 1, 1.5, 5, 15, 50, and 150 Hz; 5) dimensions of 2 inches x 7 inches x 16 inches; 6) operation with a Hewlett Packard multichannel recording system power supply or with a Hewlett Packard 8811A s i n g l e channel power supply; and 7) 10 V peak-to-peak output. Three of these b i o p o t e n t i a l a m p l i f i e r s would be required f o r use with the three bladders of the multi-bladder c u f f . Another option would be to design and assemble a high-gain a m p l i f i e r using d i s c r e t e op-amps (e.g. F i g . 6.1). The advantages of such an a m p l i f i e r are as follows: 1) low power consumption i f CMOS op-amps are used; 2) a c a p a b i l i t y f o r changing gains and f i l t e r s under computer c o n t r o l by using d i g i t a l l y programmable potentiometers (Xicor X9MME, Xicor Inc., M i l p i t a s CA), thereby o f f e r i n g the user both the opportunity to experiment with d i f f e r e n t bandwidths and the a b i l i t y to use the computer to implement an automatic gain c o n t r o l scheme that may help to reduce the incidence of noise-induced saturation; and 3) improved resistance to noise-induced saturation with the c i r c u i t shown i n F i g . 6.1, since the d i f f e r e n t i a l inputs to the f i n a l high-gain stage consist of the o s c i l l o m e t r i c s i g n a l r i d i n g on the baseline cuff pressure and the f i l t e r e d baseline c u f f pressure s i g n a l , meaning that large noise-induced pressure disturbances w i l l appear to some extent i n both inputs and thus p a r t i a l l y cancel before being amplified. 6.2.2 Improvements i n Searching Techniques It was observed during c l i n i c a l t r i a l s with the prototype that -99-o . l l B a s e l i n e C u f f P r e s s u r e D i g i t a l l y P r o g r a m m a b l e AmpT1f i e r Osc i 1 1 o m e t r i c S i g n a l s Up/Oown C o n t r o l W i p e r Change C o n t r o l X 1 c o r D i g i t a l l y C o n t r o l l e d P o t e n t i o m e t e r 7 I8 C h i p S e l e c t F i g . 6.1 A digitally-programmable high-gain a m p l i f i e r . -100-some delays i n updating estimates could have been compounded by an i n e f f i c i e n t use of constant pressure changes i n the proximal bladder when searching f o r a limb occlusion pressure that was f a r from the proximal pressure. A table of past pulse heights sensed i n the proximal bladder could be used to adjust the step s i z e dynamically and optimize the search. For example, when the proximal pressure i s above the (uncorrected) mean a r t e r i a l pressure but below the limb occlusion pressure, the proximal pulse height w i l l decrease as the proximal pressure approaches the limb occlusion pressure. Thus, proximal pulse height provides a d d i t i o n a l information concerning the d i f f e r e n c e between proximal pressure and limb occlusion pressure that can be used to adjust the step s i z e of pressure changes. A l t e r n a t e l y , the amplitude of the middle pulse height can be used to provide information concerning how f a r the proximal pressure i s from the limb occlusion pressure. The onset of blood flow past the proximal bladder occurs when the proximal pressure f a l l s just below the limb occlusion pressure, at which time the observed middle pulse should just exceed the flow detection threshold. However, when the proximal pressure i s fa r t h e r below the limb occlusion pressure, the proximal bladder o f f e r s l e s s resistance to blood flow to the middle bladder, r e s u l t i n g i n a greater flow to the middle bladder that produces larger o s c i l l o m e t r i c pulses. Thus, the di f f e r e n c e between the observed pulse height i n the middle bladder and the flow detection threshold can be used to adjust the step s i z e , with l a r g e r pressure changes being used when la r g e r d i f f e r e n c e s are measured. 6.2.3 Improvements i n Safety The safety of the prototype when i t i s used as a tourniquet can be improved by adding flow-detecting c a p a b i l i t i e s at the d i s t a l bladder. E i t h e r the addition of an independent flow detector or the development of an algorithm to d i f f e r e n t i a t e between noise pulses and -101-flow-induced o s c i l l o m e t r i c pulses at the d i s t a l bladder could increase safety by providing a redundant check of the performance of the occluding middle bladder and i n i t i a t i n g appropriate system responses when flow i s detected. In t h i s way, the prototype could o f f e r a t e s t of i t s estimation process and dynamic operation, which i s not c u r r e n t l y a v a i l a b l e with other commercial blood pressure monitors or tourniquets. Possible suggestions f o r better noise detection are given i n the next sec t i o n . 6.2.4 Improvements i n Noise Detection The performance of a flow detection algorithm (either at the middle bladder or at the d i s t a l bladder) depends on the a b i l i t y to detect noise. When the middle bladder i s properly occluding the underlying a r t e r i e s and preventing blood from flowing to the d i s t a l bladder, the d i s t a l s i g n a l can r e l i a b l y detect noise, as i n d i c a t e d by the consistent performance of the prototype i n d i f f e r e n t noise conditions during lab t r i a l s . However, i f i t i s desired to d i f f e r e n t i a t e f u r t h e r between noise and flow-induced signals i n order to r e j e c t noise (as with a d i s t a l flow detector), then the following schemes could be implemented: 1) the simple recognition of matched-pulses, employed by other o s c i l l o m e t r i c monitors, could be used to r e j e c t noise that i s not c y c l i c ; 2) a narrow search window can be created to exclude noise by accepting pulses as being v a l i d flow-induced pulses only i f they occur within an allowable time a f t e r a measured QRS complex from from an ECG waveform; and 3) a synchronization of pulses can be u t i l i z e d to r e j e c t noise by accepting pulses as being v a l i d only i f a s i m i l a r pulse i s recorded i n other (more proximal) bladders at a s i m i l a r time. -102-6.2.5 Improvements i n Threshold Settings Because of differences i n cuff a p p l i c a t i o n techniques, the use of a d d i t i o n a l p r o t e c t i v e f a b r i c beneath the tourniquet c u f f , and va r i a t i o n s i n limb physiology, the reception of o s c i l l o m e t r i c pulses i n the c u f f r e s u l t i n g from flow-induced a r t e r i a l wall movement v a r i e d from subject to subject during t r i a l s . For t h i s reason, the gain of the EEG am p l i f i e r had to be adjusted from t r i a l to t r i a l i n order to prevent the o s c i l l o m e t r i c signals recorded under d i f f e r e n t conditions from saturating the a m p l i f i e r during times when the proximal bladder pressure might f a l l s i g n i f i c a n t l y below the limb occlusion pressure while an estimate was being made. During t r i a l s , a l l t h i s was done manually. The gain was adjusted so that the f i r s t flow-induced pulse seen by the user while decreasing the supra-occlusive proximal pressure was about 1 V i n magnitude, a value emperically found to permit s a t i s f a c t o r y s i g n a l r e s o l u t i o n and a m p l i f i e r operation. The thresholds were then manually adjusted so that background noise would not t r i g g e r events by observing the displayed waveforms and d i s c r i m i n a t i n g between the background noise and what appeared to be motion a r t i f a c t s i n order to set the thresholds s u f f i c i e n t l y above the background noise. (Although the thresholds could have been based on the energy of the background noise, such a s e t t i n g would have required sampling waveforms that only had background noise present, meaning that the subject would have had to remain absolutely motionless during t h i s i n i t i a l i z a t i o n period. Generally, t h i s requirement was not met during lab t r i a l s with unanesthetized subjects or during c l i n i c a l t r i a l s when the patient was being prepared f o r surgery p r i o r to the operation.) These inconvenient, i r r e p r o d u c i b l e , and time-consuming manual adjustments of gains and thresholds could be eliminated by employing a digitally-programmable high-gain a m p l i f i e r as previously o u t l i n e d . A scheme to adjust gains automatically could then be implemented so that the signals are always maintained at a convenient -103-magnitude to produce signals with s u f f i c i e n t r e s o l u t i o n that would not saturate the a m p l i f i e r . By keeping account of the gains, the computer could then compensate f o r d i f f e r e n t a m p l i f i c a t i o n s , thereby recording the true r e l a t i v e s i z e s of d i f f e r e n t pulses. Furthermore, a routine to measure noise at one s p e c i f i e d gain could then be run before a measurement t r i a l to generate the thresholds required f o r t h i s one gain s e t t i n g . This threshold information could then be adjusted automatically to compensate f o r any changes i n gains that might be i n i t i a t e d by the computer. 6.2.6 Improvements i n Cuff Design A p r a c t i c a l problem concerning the multi-bladder c u f f was the d i f f i c u l t y i n properly applying the c u f f , which was wider than conventional tourniquet c u f f s , and which f i t poorly onto tapered limbs and could encroach into the operative f i e l d f o r some s u r g i c a l procedures. Experimentation with c u f f f a b r i c a t i o n techniques could y i e l d a s p e c i a l i z e d cuff that i s comparable to conventional c u f f s i n width, and that has an improved a b i l i t y to conform to tapered limbs without i n t e r f e r i n g with s u r g i c a l procedures. The incorporation of sensors and pressure c o n t r o l valves into the c u f f could also improve the system's performance. In p a r t i c u l a r , with a smaller pneumatic volume and with pressure sensing at the cuff i t s e l f , improvements i n pressure regulation, speed of pressure changes, p o r t a b i l i t y , noise immunity, and signal-to-noise r a t i o could be r e a l i z e d . The use of a u x i l i a r y sensors to detect motion at the cuff could also replace the noise bladders to reduce the c u f f width and bulkiness. A l t e r n a t i v e l y , a narrower c u f f could be used i f a d i f f e r e n t scheme f o r adapting tourniquets i s used. For example, a dual bladder c u f f having a standard proximal width, a very narrow d i s t a l width, and a d d i t i o n a l motion-detecting transducers could be incorporated with an adaptive scheme that would keep the proximal pressure at the minimum -104-pressure required to prevent flow-induced o s c i l l o m e t r i c pulses from being detected at the d i s t a l bladder when noise i s not detected by the a u x i l i a r y sensors. While flow i s not detected, the proximal pressure could be slowly decreased, or maintained at a pressure below the constant tourniquet pressure normally used and above a pressure derived from an e a r l i e r estimate of limb occlusion pressure made by the system. But at the onset of flow detected by the d i s t a l bladder, the proximal pressure could be incremented to a supra-occlusive pressure. The advantages of t h i s system are a standard cuff s i z e , fewer number of channels and components, and an o s c i l l o m e t r i c flow detector that could provide information about the occlusion performance of the tourniquet that i s not given by commercially a v a i l a b l e devices. The major disadvantage of such a scheme i s that some blood i s allowed to flow past the occluding bladder before a change i n pressure i s i n i t i a t e d . Thus, r a p i d detection of the onset of flow past the proximal cuff i s c r u c i a l f o r t h i s a p p l i c a t i o n , whereas f o r the t r i - b l a d d e r cuff configuration a middle occluding cuff could be set to s u f f i c i e n t l y high pressures to account f o r t r a c k i n g errors made while the proximal bladder i s searching f o r the limb occlusion pressure. 6.2.7 Other Flow Detection Methods The present configuration of the prototype detects during noise-free periods the onset of flow past a proximal bladder to estimate limb occlusion pressure r a p i d l y . The prototype determines whether or not flow i s present by deciding whether or not a flow-induced s i g n a l i s present i n the d i s t a l bladder. However, the same strategy of noise detection and flow detection f o r r a p i d l y estimating limb occlusion pressure can be implemented with d i f f e r e n t transducers that may be more s e n s i t i v e to flow t r a n s i t i o n s past the proximal bladder than the o s c i l l o m e t r i c system. Other flow detection methods include auscultation, ultrasound, l a s e r Doppler, and e l e c t r i c a l impedance plethysmography. As -105-described e a r l i e r , auscultatory and u l t r a s o n i c transducers s u f f e r from problems with transducer movement, transducer placement, and movement a r t i f a c t . For example, with these two techniques, the common s h i f t i n g of the transducer from i t s optimal p o s i t i o n over a major a r t e r y could reduce t h e i r s e n s i t i v i t y to flow that would be wrongly i n t e r p r e t e d as being caused by an absence of blood flow, which could lead to hazardous underestimations of limb occlusion pressure. Laser Doppler methods only measure c a p i l l a r y flow at the skin surface, which may not provide accurate information concerning flow i n major a r t e r i e s . However, e l e c t r i c a l impedance plethysmography may provide a. f e a s i b l e a l t e r n a t i v e to o s c i l l o m e t r y that i s more s e n s i t i v e to blood flow. Although e l e c t r i c a l impedance plethysmography, l i k e oscillometry, i s a f f e c t e d by motion a r t i f a c t s , i t need not be a f f e c t e d by p o s i t i o n i n g d i f f i c u l t i e s . Electrodes have been applied to the skinward side of a standard blood pressure c u f f to f a c i l i t a t e the determination of blood pressure by measurements of e l e c t r i c a l impedance [85]. Thus, i n t h i s implementation, as with oscillometry, only a c u f f needs to be applied to a limb; but the c u f f may be e a s i e r to apply uniformly than the bulky multi-bladder cuff proposed f o r use i n the prototype o s c i l l o m e t r i c system. Furthermore, whereas the o s c i l l o m e t r i c system cannot sample waveforms during the large pressure disturbances caused by the regulation of c u f f pressure, an impedance system could detect flow while the c u f f pressure was being regulated. Such simultaneous sampling and pressure regulation could lead to estimation schemes that are quicker than the algorithm used by the o s c i l l o m e t r i c prototype. With these modifications i n c u f f design, hardware components, and software programming, the prototype may be f e a s i b l e f or routine c l i n i c a l use to adapt tourniquet pressures and reduce the incidence of tourniquet-related i n j u r i e s . -106-REFERENCES [ 1] A.K. Ream, " S y s t o l i c , d i a s t o l i c , mean or pu l s e , " E s s e n t i a l  Noninvasive Monitoring i n Anesthesia, J.S. Gravenstein, R.S. Newbower, A.K. Ream, and N.T. Smith ( e d i t o r s ) . New York: Grune & S t r a t t o n , 1980, p. 63. [ 2 ] J . Looney J r . , "Blood pressure by o s c i l l o m e t r y , " Medical E l e c t r o n i c s , A p r i l , pp. 57-63, 1978. [ 3] A.K. Ream, p. 64. [ 4] A.K. Ream, p. 66. [ 5] A.K. Ream, pp. 66-68. [ 6] G.J. Tortora and N.P. Anagnostakos, P r i n c i p l e s of Anatomy and  Physiology, 4th ed. New York: Harper & Row, 1984, p. 492. [ 7] A.K. Ream, p. 56. [ 8 ] L. Galton, The S i l e n t Disease: Hypertension. New York: Crown P u b l i s h e r s , 1973, p. 56. [ 9] W.C. McCall and V.R. McCall, "Evaluation of hypertension by means of f u l l y automated blood-pressure monitoring," Health Care  Instrumentation, pp. 22-26, 1985. [10] M.H. E l l e s t a d , Stress T e s t i n g : P r i n c i p l e s and P r a c t i c e . P h i l a d e l p h i a : F.A. Davis, 1975, p. 97. [11] L. Galton, p. 4. [12] L. Galton, p. 5. [13] E. O'Brien and K. O'Malley, High Blood Pressure. New York: Arco P u b l i s h i n g , 1982, pp. 9-11. [14] G.J. Tortora and N.P. Anagnostakos, pp. 513-515. [15] L. Galton, pp. 35-36. [16] L. Galton, p. 6. [17] P.W.J, van Dongen, "Methodology of blood pressure measurements," C l i n i c a l and Experimental Hypertension - Hypertension i n Pregnancy, v o l . B2, No. 3, pp.371-403, 1983. [18] M.J. Horan, "Role of ambulatory blood pressure r e c o r d i n g i n the diag n o s i s , prognosis, and management of hypertension," C l i n i c a l  and Experimental Hypertension, v o l . A7, No.2&3, pp. 205-216, 1985. [19] T.G. P i c k e r i n g , G.A. H a r s h f i e l d , and J.H. Laragh, "Ambulatory versus c a s u a l blood pressure i n the diagnosis of hypertensive p a t i e n t s , " C l i n i c a l and Experimental Hypertension, v o l . A7, No.2&3, pp. 257-266, 1985. [20] W.M. K i r k e n d a l l , M. F e i n l e i b , E.D. F r i e s , and A.L. Mark, "American Heart A s s o c i a t i o n Committee Report: Recommendations f o r human blood pressure determination by sphygmomanometers," C i r c u l a t i o n , v o l . . 62, No. 5, pp. 1146A-1155A, 1980. -107-[21] B.A. Gould et a l . , "Evaluation of the Remler M2000 Blood Pressure Recorder," Hypertension, v o l . 6, No. 2, pp. 209-215, 1984. [22] M.H. E l l e s t a d , p. VII. [23] A.A. Kathus, Foreward to Stress T e s t i n g : P r i n c i p l e s and P r a c t i c e , author M.H. E l l e s t a d . P h i l a d e l p h i a : F.A. Davis, 1980, p. V. [24] M.H. E l l e s t a d , p. 124. [25] M.H. E l l e s t a d , pp. 125-127. [26] J.S. Gravenstein, " I n t r o d u c t i o n , " E s s e n t i a l Noninvasive Monitoring  i n Anesthesia, J.S. Gravenstein, R.S. Newbower, A.K. Ream, and N.T. Smith ( e d i t o r s ) . New York: Grune & S t r a t t o n , 1980, p. 2. [27] M. Ramsey, "Noninvasive blood pressure monitoring methods and v a l i d a t i o n , " E s s e n t i a l Noninvasive Monitoring i n Anesthesia, J.S. Gravenstein, R.S. Newbower, A.K. Ream, and N.T. Smith ( e d i t o r s ) . New York: Grune & S t r a t t o n , 1980, p. 37. [28] A.K. Ream, p. 53. [29] N.T. Smith and J.E.W. Beneken, "An overview of a r t e r i a l pressure monitoring," E s s e n t i a l Noninvasive Monitoring i n Anesthesia, J.S. Gravenstein, R.S. Newbower, A.K. Ream, and N.T. Smith ( e d i t o r s ) . New York: Grune & S t r a t t o n , 1980, n. 79. [30] A.K. Ream, "Automated i n d i r e c t blood pressure measurement - a response," Medical Instrumentation, v o l . 18, No. 5, pp. 286-287. 1984. : [31] H.P. Apple, "Automatic noninvasive blood pressure monitors," E s s e n t i a l Noninvasive Monitoring i n Anesthesia, J.S. Gravenstein, R.S. Newbower, A.K. Ream, and N.T. Smith ( e d i t o r s ) . New York: Grune & S t r a t t o n , 1980, p. 16. [32] J.A. McEwen and R.W. McGraw, "An adaptive tourniquet f o r improved s a f e t y i n surgery," IEEE Trans. Biomed. Eng., v o l . BME-29, pp. 122-128, 1982. [33] ECRI, "Pneumatic tourniquets used f o r r e g i o n a l anesthesia," Health  Devices, v o l . 12, pp. 48-49, Dec. 1982. [34] ECRI, "Pneumatic to u r n i q u e t s , " Health Devices, v o l . 13, pp. 299-316, Oct. 1984. [35] J.A. McEwen and G.F. Auchinleck, "Advances i n s u r g i c a l tourniquets," AQRN Journal, v o l . 36, pp. 889-896, 1982. [36] J.R F l e t c h e r and T.E.J. Healy, "The a r t e r i a l t o u rniquet," Annals of the Royal College of Surgeons of England, v o l . 65, pp. 409-417, 1983. [37] R. Sanders, "The tourniquet: Instrument or weapon?" Hand, v o l . 5, pp. 119-123, 1973. [38] L.A. Geddes and S.J. Whistler, "'The e r r o r i n i n d i r e c t blood pressure measurement with the i n c o r r e c t s i z e of c u f f , " American Heart Journal, v o l . 96, pp. 4-8, 1978. [39] J.A. McEwen, "Pneumatic tourniquet," United States Patent #4,469,099, Sept. 4, 1984. - 1 0 8 -[40] J . A . M c E w e n , " A d a p t i v e p n e u m a t i c t o u r n i q u e t , " U n i t e d S t a t e s P a t e n t # 4 , 4 7 9 , 4 9 4 , O c t . 3 0 , 1 9 8 4 . [41] L . C r o m w e l l , F . J . W e i b e l l , a n d E . A . P f e i f f e f , B i o m e d i c a l I n s t r u m e n t a t i o n a n d M e a s u r e m e n t s , 2 n d e d . E n g l e w o o d , N . J . : P r e n t i c e H a l l , 1 9 8 0 , p p . 1 3 5 - 1 3 8 . [42] R . A . P l e u r a , " B l o o d p r e s s u r e a n d s o u n d , " M e d i c a l I n s t r u m e n t a t i o n - A p p l i c a t i o n a n d D e s i g n , J . G . W e b s t e r ( e d i t o r ) . B o s t o n : H o u g h t o n M i f f l i n , 1 9 7 8 , p p . 3 3 6 - 3 3 8 . [43] J . M . R . B r u n e r , Handbook o f B l o o d P r e s s u r e M o n i t o r i n g . L i t t l e t o n , M a s s . : PSG P u b l i s h i n g , 1 9 7 8 , p p . 6 9 - 8 0 . [44] J . M . R . B r u n e r , p p . 1 1 9 - 1 2 1 . [45] D . A . P a u l u s , " N o n i n v a s i v e b l o o d p r e s s u r e m e a s u r e m e n t , " M e d i c a l  I n s t r u m e n t a t i o n , v o l . 1 5 , N o . 2 , p p . 9 1 - 9 4 , 1 9 8 1 . [46] J . M . R . B r u n e r , L . J . K r e n i s , J . M . Kunsmen , a n d A . P . S h e r m a n , " C o m p a r i s o n o f d i r e c t a n d i n d i r e c t m e t h o d s o f m e a s u r i n g a r t e r i a l b l o o d p r e s s u r e , p a r t I I I , " M e d i c a l I n s t r u m e n t a t i o n , v o l . 1 4 , N o . 3 , p p . 1 8 2 - 1 8 8 , 1 9 8 1 . [47] J . M . R . B r u n e r , p . 6 5 . r48] J . M . R . B r u n e r e t a l . , " C o m p a r i s o n o f d i r e c t a n d i n d i r e c t m e t h o d s o f m e a s u r i n g a r t e r i a l b l o o d p r e s s u r e , p a r t I I , " M e d i c a l I n s t r u m e n t a t i o n , v o l . 1 5 , N o . 2 , p p . 9 7 - 1 0 1 , 1 9 8 1 . [49] H . A l e x a n d e r , M . L . C o h e n , a n d L . S t e i n f e l d , " C r i t e r i a i n t h e c h o i c e o f a n o c c l u d i n g c u f f f o r t h e i n d i r e c t m e a s u r e m e n t o f b l o o d p r e s s u r e , " . M e d i c a l a n d B i o l o g i c a l E n g i n e e r i n g a n d C o m p u t i n g , v o l . 1 5 , p p . 2-10, J a n . 1977. [50] S . L o n d e a n d T . K l i t z n e r , " A u s c u l t a t o r y b l o o d p r e s s u r e measuremen t -e f f e c t o f p r e s s u r e o n t h e h e a d o f t h e s t e t h o s c o p e , " W e s t e r n J o u r n a l  o f M e d i c i n e , v o l . 1 4 1 , N o . 2 , p p . 1 9 3 - 1 9 5 , A u g . 1 9 8 4 . [51] D . P . G o l d e n e t a l . , " D e v e l o p m e n t o f a K o r o t k o v s o u n d p r o c e s s o r f o r a u t o m a t i c i d e n t i f i c a t i o n o f a u s c u l t a t o r y e v e n t s , " I E E E T r a n 3 . B i o m e d . E n g . , v o l . B M E - 2 1 , N o . 2 , p p . 1 1 4 - 1 2 4 , M a r c h 1 9 7 4 . [52] J . E . Wes t e t a l . , " F o i l e l e c t r e t t r a n s d u c e r f o r b l o o d p r e s s u r e m o n i t o r i n g , " J o u r n a l o f t h e A c o u s t i c a l S o c i e t y o f A m e r i c a , v o l . 7 4 , N o . 3 , p p . 6 8 0 - 6 8 6 , S e p t . 1 9 8 3 . [53] R . A . P l e u r a , p p . 4 0 2 - 4 0 9 . [54] T . M . K a z a m i a s e t a l . , " B l o o d p r e s s u r e m e a s u r e m e n t w i t h D o p p l e r u l t r a s o n i c f l o w m e t e r , " J o u r n a l o f A p p l i e d  P h y s i o l o g y , v o l . 3 0 , N o . 4 , p p . 5 8 5 - 5 8 8 , A p r i l 1 9 7 1 . [55] H . L . M a s s i e e t a l . , " U l t r a s o n i c measu remen t o f i n f a n t b l o o d p r e s s u r e , " M e d i c a l I n s t r u m e n t a t i o n , v o l . 7 , N o . 4 , p p . 2 4 0 -2 4 4 , 1 9 7 3 . [56] [57] H . P . A p p l e , p p . 1 7 - 1 8 . D i n a m a p 845 S e r v i c e M a n u a l , C r i t i k o n I n c . , Tampa F L . - 1 0 9 -[58] H e w l e t t P a c k a r d 78354A P a t i e n t M o n i t o r S e r v i c e M a n u a l , H e w l e t t P a c k a r d Company, P a l o A l t o C A . [59] D a t a s c o p e A c c u t o r r S e v i c e M a n u a l , D a t a s c o p e C o r p o r a t i o n , Pa ramus NJ. [60] P h y s i o - C o n t r o l V i t a l S i g n s M o n i t o r S e r v i c e M a n u a l , P h y s i o - C o n t r o l C o r p o r a t i o n , Redmond WA. [61] G . W . Mauck e t a l . , "The m e a n i n g o f t h e p o i n t o f maximum o s c i l l a t i o n s i n c u f f p r e s s u r e i n t h e i n d i r e c t m e a s u r e m e n t o f b l o o d p r e s s u r e - P a r t I I , " J o u r n a l o f B i o m e c h a n i c a l  E n g i n e e r i n g , v o l . 1 0 2 , p p . 2 8 - 3 3 , F e b r u a r y 1 9 8 0 . [62] W . T . L i n k e t a l . , " A p p a r a t u s a n d p r o c e s s f o r d e t e r m i n i n g s y s t o l i c p r e s s u r e , " C a n a d i a n P a t e n t # 1 , 0 6 0 , 2 2 6 , O c t . 1 4 , 1 9 7 9 . [63] P . H u t t o n a n d C . P r y s - R o b e r t s , "The o s c i l l o t o n o m e t e r i n t h e o r y a n d p r a c t i c e , " B r i t i s h J o u r n a l o f A n a e s t h e s i a , v o l . 5 4 , p p . 5 8 1 - 5 9 1 , 1 9 8 0 . [64] R . H . F r i e s e n a n d J . L . L i c h t o r , " I n d i r e c t m e a s u r e m e n t o f b l o o d p r e s s u r e i n n e o n a t e s a n d i n f a n t s u t i l i s i n g a n a u t o m a t i c n o n i n v a s i v e o s c i l l o m e t r i c m o n i t o r , " A n e s t h e s i a a n d A n a l g e s i a , v o l . 6 0 , p p . 7 4 2 - 7 4 5 , 1 9 8 1 . [65] K . M . B o r o w a n d J . W . N e w b u r g e r , " N o n i n v a s i v e e s t i m a t i o n o f c e n t r a l a o r t i c p r e s s u r e u s i n g t h e o s c i l l o m e t r i c m e t h o d f o r a n a l y z i n g s y s t e m i c a r t e r y p u l s a t i l e b l o o d f l o w . C o m p a r a t i v e s t u d y o f i n d i r e c t s y s t o l i c , d i a s t o l i c , a n d mean b r a c h i a l a r t e r y p r e s s u r e w i t h s i m u l t a n e o u s d i r e c t a s c e n d i n g a o r t i c p r e s s u r e m e a s u r e m e n t s , " A m e r i c a n H e a r t J o u r n a l , v o l . 1 0 3 , p p . 8 7 9 - 8 8 6 , 1 9 8 2 . [66] J . H . S i l a s , A . T . B a r k e r , a n d L . E . Ramsay , " C l i n i c a l e v a l u a t i o n o f D i n a m a p 845 a u t o m a t e d b l o o d p r e s s u r e r e c o r d e r , " B r i t i s h H e a r t  J o u r n a l , v o l . 4 3 , p p . 2 0 2 - 2 0 5 , 1 9 8 0 . [67] M . Y e l d e r m a n a n d A . K . Ream, " I n d i r e c t m e a s u r e m e n t o f mean b l o o d p r e s s u r e i n t h e a n e s t h e t i z e d p a t i e n t , " A n e s t h e s i o l o g y , v o l . 5 0 , p p . 2 5 3 - 2 5 6 , 1 9 7 9 . [68] K . J . K i m b l e , R . A . J . D a r n a l l , M . Y e l d e r m a n , R . L . A r i a g n o , a n d A . K . Ream, " A n a u t o m a t i c o s c i l l o t o n o m e t r i c t e c h n i q u e f o r e s t i m a t i n g mean a r t e r i a l p r e s s u r e i n c r i t i c a l l y i l l n e w b o r n s , " A n e s t h e s i o l o g y , v o l . 5 4 , p p . 4 2 3 - 4 2 5 , 1 9 8 1 . [69] A . W a l d , " C o m p a r i s o n o f b l o o d p r e s s u r e w i t h two n o n - i n v a s i v e m o n i t o r s , " I E E E F r o n t i e r s o f E n g i n e e r i n g a n d C o m p u t i n g i n H e a l t h  C a r e . New Y o r k : I E E E , 1 9 8 3 , ppT 2 5 6 - 2 6 0 . [70] C . J . H . J o h n s o n a n d J . H . K e r r , " A u t o m a t i c b l o o d p r e s s u r e m o n i t o r s -A c l i n i c a l e v a l u a t i o n o f f i v e m o d e l s i n a d u l t s , " A n e s t h e s i a , v o l . 4 0 , p p . 4 7 1 - 4 7 8 , 1 9 8 5 . [71] D . F . G l o y n a , P . H u b e r , P . A b s t o n , a n d J . F . A r e n s , " A c o m p a r i s o n o f b l o o d p r e s s u r e measu remen t t e c h n i q u e s i n t h e h y p o t e n s i v e p a t i e n t , " A n e s t h e s i a a n d A n a l g e s i a , v o l . 6 3 , p . 2 2 2 , 1 9 8 4 . -110-[72] E. Nystrom, K.H. Reid, R. Bennett, L. Couture, and H.L. Edmonds, "A comparison of two automated i n d i r e c t a r t e r i a l blood pressure meters: with recordings from a r a d i a l a r t e r i a l c a t h e t e r i n ane s t h e t i z e d s u r g i c a l p a t i e n t s , " Anesthesiology, v o l . 62, pp. 526-530, 1985. [73] H.P. Apple, pp. 14-15. [74] Aspen ATS 1000 Tourniquet System Service Manual, Aspen Labs Inc., Englewood CO. [75] K.R. B h a t t a r a i , P.I. Chen,and P.J. Reynolds, "Pressure measurement using a h y d r a u l i c o c c l u s i v e c u f f , " Advances i n  Bioengineering. New York: ASME, 1981, pp. 55-58. [76] F.R. Borkat et a l . , "An approach to the continuous non-invasive measurement of blood pressure," Proceedings of the  San Diego Biomedical Symposium, v o l . 15, 1976, pp. 9-12. [77] P. Kantrowitz, "A review of non-invasive blood pressure measurement using a c u f f , with p a r t i c u l a r respect to motion a r t i f a c t , " Biomedical Engineering, v o l . 8, pp. 480-481, 1973. [78] B. Widrow et a l . , "Adaptive noise c a n c e l l i n g : P r i n c i p l e s and a p p l i c a t i o n s , " Proceedings of the IEEE, v o l . 63, No. 12, pp. 1692-1716, Dec. [79] P.A. Lynn, An Introduction to the A n a l y s i s and Processing of  S i g n a l s , 2nd ed. Hong Kong: MacMillan Press Ltd., 1982. [80] M. Schwartz and M. Shaw, S i g n a l Processing: D i s c r e t e S p e c t r a l  A n a l y s i s , Detection, and Estimation. New York: McGraw H i l l , 1975, pp. 286-298. [81] D.J. Wilde, Optimum Seeking Methods. New Jersey: P r e n t i c e H a l l , 1964, pp. 11-35. : [82] G.D.P. De Dobbeleer, "Measurement of s y s t o l i c and d i a s t o l i c pressure by means of phase 3 h i f t , " World  Medical E l e c t r o n i c s and Instrumentation, v o l . 3, pp. 122-126, 1965. [83] D. Krueger, "Method and system f o r estimation of a r t e r i a l pressure," United States Patent #3,552,383, Jan.5, 1971. [84] (Proposed) Standard f o r E l e c t r o n i c or Automated Sphygmomanometers, A s s o c i a t i o n f o r the Advancement of Medical Instrumentation, A r l i n g t o n VA, Aug. 1980. [85] H. H e r s c o v i c i and D.H. R o l l e r , "Noninvasive determination of c e n t r a l blood pressure by impedance plethysmography," IEEE  Transactions on Biomedical Engineering, v o l . BME-33, No. 6, June 1986. - 111 -APPENDIX 1 PROGRAM LISTING •include "o_f3.c" /........................................,........,...,»........,....,..,,.... /» o_p.c • / * This f i le contains the main program to estimate occlusive pressure. * / * When quiet ( i .e . distal signal between positive and negative noise * / * thresholds), the middle signal is checked to see i f blood is flowing * / * past the proximal bladder and hitting the middle bladder, as indicated • / • when the middle signal exceeds a flow-signal threshold. * I* A l l signals are fi ltered with a digital FIR HPF prior to analysis, and • / * the raw middle signal is also examined for amplifier saturation * / * conditions (i .e. middle signal exists beyond allowable boundaries). • / * The setting of the proximal bladder pressure depends on whether flow • / * has been detected or not (flow -> Increase proximal pressure; * / ' n o flow -> decrease proximal pressure). A changing flow condition * / * A changing flow condition between successive proximal pressure levels * / * indicates that the proximal pressure has passed through the occlusive * / * pressure (estimated by the average of the two successive pressure values * / * for which a change was observed. *. FILE «fp_data, "fopenO; mainO ( int proximal occl<mlddle_occl,pp0,pp200,pd0,pd200,psrc; int pm0,pm200"; int auto cal,nflag,ok,signal; int a,c,3,t_p,beep; int o_p, prox_prss,mid_prss,dist_prss,past op,past_prox; int i , j ,k ,adata; int num_est; int pos,neg,screen_cnt,quiet_cnt,noise_cnt; int threshold,detect,lag ncheck; ~ int dst_dyn(100],mld_dynTl00),prx dyn(100]; int result(1000),recount,avg_ceauTt,var_result; lht dst_pt,mid_pt,prx D t ; ~ — int trendycnt,alarm_tlme,a^flag,r—index,p_t,transition; int o^p_raw,o_p_r jLW~ml,o_p~raw_ra27f l i t er ; float h0,h2; unsigned long total_tlme, sample_t line, current_time, est time, start_time; unsigned long r_time(1000), long_var, past_est_time; ~" unsigned short key_prssed, duration; ~* char file_name(10); / • DECLARATIONS * / /.............................,.,.,,,.........,............................... / * proximal_occl - i n i t i a l guess of occlusive pressure (mmHg) / * middle_occl • tourniquet pressure for middle occluding bladder (mmHg) / * ppO - zero pressure calibration constant for baseline proximal pressure /« pmO - »• middle / * pdO - • distal /* pp200 - source pressure calibration constant for baseline prox. pressure /• pm200 - • middle /» pd200 - " distal / * psrc - source pressure (mmHg) / * auto_cal - flag set to 1 i f baseline pressure has been automatically / * calibrated before estimating pressure (else - 0) /* ' nflag - noise flag set to 1 i f noisy (else - 0) / * ok - user response flag set to 1 i f O.K. (else 0) / * signal - signal flag set to 1 i f flow signal is present (else 0) / * a , c , d , i , j , k - miscellaneous temporary variables / * t_p - target baseline bladder pressure / * beep » audio flow indicator flag, set to 1 to enable (else - 0) /» o_p • limb occlusion pressure estimate / * prox_prss - proximal baseline cuff pressure / * mid_prss - middle " /*- dist_prss - distal " / * past_op - past estimate of limb occlusion pressure / * past_prox - past proximal baseline cuff pressure / • adata - sampled A/D data value / * num_est " number of limb occlusion pressure estimates made / * pos - positive noise threshold for distal signal / * neg - negative noise threshold for distal signal / * acreen_cnt " counter for screen plotting purposes / * quiet_cnt - counter of quiet data samples / * noise'cnt * counter of noisy data samples / * tnresKold - threshold for flow signal from middle bladder /* detect - flow detection flag, set to 1 i f flow is detected (else - 0) / • lag_ncheck - counter of data samples taken after flow is detected to / • check for lagging noise / • dst dynU - array to hold sampled distal oscillometric waveform /»• mid~dyn(] • * middle / * prx~dyn(J - " proximal -112-/* r e s u l t ( ] « array to hold estimates and event markers */ /* r_count » counter for r e s u l t s array */ /* avg_result *» average of limb occlusion estimates */ /* var_result » variance of limb occlusion estimates */ /* dst_pt = f i l t e r e d d i s t a l o s c i l l o m e t r i c sampled point */ /* mid_pt - middle " */ /* prx_pt «" proximal " */ /* trend_cnt = counter for trend recorder of estimates */ /* alarm_time » maximum allowable time between estimates */ /* a _ f l a g » alarm f l a g that enables alarm i f = 1 (disables i f = 0) */ /* r_index » index for r e s u l t s array */ /* t r a n s i t i o n flow t r a n s i t i o n f l a g (=1 for t r a n s i t i o n , else » 0) */ /* p_t « proximal cuff target baseline pressure */ /* o_p —raw = raw (unfiltered) occlusion pressure value */ /* o_p2raw_ml "* past raw occlusion pressure value */ /* o p~raw~m2 - past past raw occlusion pressure value */ /* f i l t e r - f i l t e r f l a g - 1 i f 3-pt moving average i s being done (else = 0) •/ /* hO - f i l t e r c o e f f i c i e n t */ /* h2 - f i l t e r c o e f f i c i e n t */ /* total_time « t o t a l time elapsed since the s t a r t of the program */ /* sample_time *• time at which a data sample was l a s t taken */ /* current_time •» .current time */ /* est_time « time at which an estimate was l a s t made */ /* past_est_time - past estimation time */ /* start_time - s t a r t i n g time of the program */ /* r_time[] - array to hold the time of estimates and event markers */ /* long_var = variance of estimates (large magnitude needed for intermediate */ /* c a l c u l a t i o n s */ /* key_prsaed * f l a g that • 1 i f a key has been pressed (else *> 0) */ /* duration « duration of sound tones */ /**************************•***************************************************/ /* INITIALIZATION */ nflag - 0; screen_cnt psrc » 0; ppO -pp200 pmO -pm200 pdO -pd200 0; - 0 0; - 0 0; - 0 num_est " 0 ; close_valve(dec_prox); c l o s e ^ a l v e j i n c ^ p r o x ) ; close_valve(dec_mid); close_valve(inc~mid); close_valve(dec^dist); c l o s e _ v a l v e ( i n c ~ d i s t ) ; beep - 1; f i l t e r - 1; r count - 0; alarm time - 90; a f l a g - 1; h0" - -0.2; h2 - 0.8; outportb(BASE, st_prox); while (inportb(BASE) < 128) adata - inportblBASE + 1); c r t mode(6); /* RECEIVE PARAMETERS PROM USER */ i n t l z n ( t p r o x i m a l _ o c c l , smiddle_occl) i c r t mode(6); /* AUTO-CALIBRATION OF STATIC PRESSURE */ /*********••************•**********************************»*******************/ p r i n t f C A u t o-calibration of s t a t i c pressure desired ? (1 - yes) " ) ; scanf("%d", fiauto c a l ) ; i f <auto_cal -- IT ( p r i n t f C " Set pressure source for calibration.")•; p r i n t f ( " Enter any number to continue."); scanf("%d", tok); auto_calibrate(lnc_prox,dec_prox,sppO,Ipp200, st_prox, Cpsrc); auto c a l i b r a t e ( i n c mid,dec mid,&pm0,fipm200, st mid,cpsrc); -113-auto c a l i b r a t e ( i n c dist,dec dist,SpdO,Spd200, at di s t , S p a r c ) ; ) /******»»«*************************«*************************************«***»*/ /* INCREASE PRESSURE SOURCE FOR ESTIMATIONS */ p r i n t f ( n Increase pressure source to 400 mmHg."); pause(20); crt_mode(6); /* SET UP GAIN AND THRESHOLD CHECK «/ /* ./ cal_check: print£( " Quiet please: threshold c a l i b r a t i o n . " ) ; sound(50,18); timer(LIMIT); gain_thresholds(Spos,tneg,&o_p,(threshold,psrc,middle_occl,proximal_occl, Pp0,pp200,pm0,pm200,pd0,pd200,auto_cal); /***************«*******************************************************•******/ /* STARTING POINT FOR REPEATED RUNS (WITH SAME GAINS AND THRESHOLD SETTINGS) */ t o t a l _ s t a r t : /» SET UP SCREEN */ screen_set(pos,neg,threshold,«screen_cnt); crt_srcp(1,1,0); p r i n t £ ("alarm time - %3d s e c " , alarm_time); crt_srep(0,5,0); p r i n t f ( " 3 - p t f i l t e r " ) ; i f ( f U ' o r - 1) printf("ON"); else p r i n t f ( " O F F " ) ; crt_srcp(0,0,0); /* CLEAR FLAGS AND COUNTERS V /. ./ quiet_cnt - 0; noise cnt » 0; i - 0; signal " 0; t r a n s i t i o n - 0; lag_ncheck - 0; detect - 0; duration » 0; trend_cnt - 0; nflag - 0; duration - beep*4; /« RESTORE CUFF PRESSURES TO OCCLUSIVE PRESSURES */ /**************»******•****«***************«**•*******«**********************«*/ t_p «« middle_occl; q v a l v e — c o n t r o l (t_p,auto_cal, inc_mid,dec_mid,pm0,pm200, st_mid,psrc); t_p - o p ; qvalve control(t_p,auto_cal,inc —dist,dec_dlst,pd0,pd200, s t ^ d i s t , p s r c ) ; p_t - o_p + 20; - ~ - _ qvalve_control(p_t, auto_cal,inc_prox,dec_prox,pp0,pp200,st_prox,pare); past prox - p t ; ~ / * * * T * * * « * * * * T # * # * * * * * * * * * * * * * * » * * * * * * • * » * » * * * # » * * * * * * * * * * * * * * * « * * * * * * * * • * » * * * / /* AUDIO INDICATION OF START-UP, WITH STARTING TIME SAVED »/ / . / sound(50,2); sound(100,2) sound(150,2) sound(200,2) sound(250,2) st a r t time - get_the_time_of_day 0 j est time - s t a r t time; -114-/ f t * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * / /* NOW ENTER MAIN CONTROL LOOP •/ s t a r t — l o o p : aample_tima ** get_the_time_of_day () ; while (screen cnt < 620) /* CONTINUE WHILE THE SCREEN IS NOT FULL */ ( i f ( i s _ a )cey_ready to be_read() — 1) /* CHECK FOR USER */ /* KEY PRESSES THAT »/ /• INITIATE SYSTEM */ /* CHANGES OR MARK V /* EVENTS */ /****************************************«**«»********«*****/ { switch (get a key with echoO) { case 'X': /* EXIT PRESSURE ESTIMATING ROUTINE */ curfent_time = get_the_time_of_day(); goto menu; ~* case 'a': /* ENABLE/DISABLE AUDIO FLOW INDICATOR */ beep - (beep+1)%2; goto sample; case 'n': /* DECREASE NEGATIVE NOISE THRESHOLD */ —-neg; goto sample; case 'N': /* INCREASE NEGATIVE NOISE THRESHOLD */ ++neg; goto sample; case 'p': /* DECREASE POSITIVE NOISE THRESHOLD */ — p o s ; goto sample; case 'P': /* INCREASE POSITIVE NOISE THRESHOLD */ ++pos; goto sample; case 'c': /* EVENT MARKER */ res u l t Ir_count] - 0; r_time(r count] - {get_the time of day() -start_tTme)*T0/T82; ++r_count; goto sample; case 'V : /* CHANGE 3-PT FILTER STATUS */ f i l t e r - ( f i l t e r + l ) % 2 ; c r t srcp(0,5,0); i f T f i l t e r — 1) pr i n t f ( " 3 - p t f i l t e r ON • ) ; else p r i n t f ( N 3 - p t f i l t e r OFF ">; crt_srcp<0,0,0); goto next; case ' t ' : /* DECREASE FLOW SIGNAL THRESHOLD */ — t h r e s h o l d ; goto sample; case 'T': /* INCREASE FLOW SIGNAL THRESHOLD */ ++threshold; goto sample; case 'm': /* DECREASE TOURNIQUET PRESSURE */ middle^occl - middle_occl - 10; goto next; ~ -115-case 'M' : /* INCREASE TOURNIQUET PRESSURE */ mlddle_occl » middle_occl + 10; goto next; /* CHANGE ALARM TIME */ alarm time « alarm_tlme + 30; i f (alarm_time > 300) alarm_time = 30; c r t _ s r c p ( l , 1 , 0 ) ; p r i n t f ( " a l a r m time » %3d s e c " , alarm_time); crt_srcp(0,0,0); goto next; case 'a': /* DISABLE ALARM */ a_f l a g - 0; goto sample; case 'v': /» DECREASE PROXIMAL CUFF PRESSURE */ p_t - p_t - 10; goto next; case 'V : /* INCREASE PROXIMAL CUFF PRESSURE «/ p_t - p_t + 10; goto next; d e f a u l t : /*******************************************»*********«***** sample; /» USE INTERNAL IBM CLOCK FOR TIMING SAMPLING (18.2 HZ). */ /* SAMPLE AT "START" OF EVERY CLOCK TICK BY WAITING FOR THE NEXT CLOCK */ /* TICK BEFORE SAMPLING. •/ /. ,/ current_time • gct_the_time of_day(); while (current_time « sampTe_time) current_time - get_the_time_of_day(); sample time • current time; ~ /. ~ ~ ./ /* SAMPLE OSCILLOMETRIC WAVEFORMS FROM THE THREE BLADDERS */ mld_dyn(i] - msamplc;) - 128; dst_dyn[i] - dsampleO; prx_dyn(i] - psample(); /***«********************«*****•#«* ************************* *************** i i f (1 < 8) /* NOT ENOUGH DATA FOR FILTER. */ /* GET MORE DATA. */ ( ++i; goto end loop; ) /* ELSE FILTER DATA */ /. ./ dst_pt - h2*dst_dyn[i-4]+h0»(dst_dyn[i]+dst_dyn(i-21+dst_dyn[i-6]+dst_dyn(i-8)) prx_pt - h2*prx_dyn(i-4 j +h0»(prx_dyn(ij +prx_dyn(i-2)+prx_dyn[1-6]+prx_dyn 11-81) mi d p t - h2*mld dyn(1-4)+h0*(mid dyn[i]+mid dyn(i-2)+mid dyn[i-6]+mid dyn(i-8]) /. , Z. ~ . /* NOW CHECK FOR NOISE ON THE DISTAL BLADDER; LIMIT PROXIMAL AND DISTAL */ /* SIGNALS PRIOR TO GRAPHING TO AVOID PLOTTING OFF THE SCREEN; AND PLOT »/ /* THE FILTERED DISTAL SAMPLE, FILTERED MIDDLE SAMPLE, FILTERED PROXIMAL */ /» SAMPLE, AND RAW MIDDLE SAMPLE */ /***«********************•******•**********••************«****************/ i f (dst_pt > pos I I dst_pt < neg) nflag - 1; i f (dst_pt < -30) dst_pt - -30; /» 1YNDIS MOU ON dO 3H0S 38 OI VIVO HOuON3 ION »/ (SC > quo -qafnb) j f /.#»#«######ft###MUHI##t«»»f»l#fflHfH#»fl##»#*####l############»»/ /» IVNOIS Hou ON JO aans aa oi <IV3SIHV3H i < ) v i v a HOIIONS s x v i ./ /» i a i aaioaisa MOI3 ON >/ ) ( / » # l « # # M M i # I M M M M M i » l # » # M # M » M M # l l t M M I I M M M # * f M M i M M I . / .'9 + 5 d _ 3 d / . ( a a i o a i a a SVH HOU) ONinas a nn s s aua i v w i x o a a asvanoNi «/ ?I - {puS^s f(uo*rq?3np'oc)punos /» D V U TVNSIS a i v a a n »/ /. ooi awn i s n MOU avH »/ -'o - uo-rqT«u«3 /« NOIIISNVHI UCTli 3AVH »/ 'T - uof^ T B " " ! / » i O O I 3WII ISV/I MOU ./ t t - i T»u6T«) JT /» ' (MOU J I > 'UOli ON .31 0-) 9V/I3 TVNSIS 1SY3 3HI » / /» DNiNiwvxa x a oananDDO SVH NOIIISNVHI MOIJ V J I aas ox xoaHD »/ /» N33S 3S I0N ON H U H 03133130 M0T.3 3AVH 3ST3 •/ ?dooi~pua oqo6 / » 3S I0N SNIOSVT ON SI 3M3HI 3HnS 33IW 01 VIVO SHOW 139 , £ (6 > X 0 8 U O U bv\) J f fxoeqou bv\++ /» SVT.3 N O I I 3 3 I 3 0 I 3 S » / - Jaeqep / * f * f l l t t « t » l l « t l l t t i l l l t t l « t t « t t « t l « t l l t i t < l t t l ( l l ( * < l « « t t » « t t t t t f l t > / /. -xavssaoaN ai a s i o N SNISSVT HOJ XOSHO o i v i v a SHOW SXYI ./ / • * I 3 A N33S i N a s a u a asioN ON ONV (MOIJ 3AVH) a a a a a o x a a i o H s s H H i >/ _ ) (I — qosqop I I pioqesjuq < i d p^m) J J /» M01J HOd X03H0 NOII33I30 T3A3T WHOJHSd . / .'quo~qe7nb++ :qBfnb /. H013 HOJ X03H0 : SN0III0N00 13100 3S13 . / ( ?qxau oqo6 •' (V/IIWn)du"e~3Bi!83 :ac-rou /» v i v a H3N axvx ONV a a u n a w v lasau :XSION »/ ( (0 -I &*TJU) || ( 5-i.pxoqssiqq- > ttTpfa) | | (06 < (I-T)uAp _pTin)) jf /. v i v a ONV 'JHV I3s3a <- aaioaiaa Naae SVH asioN N3HI »/ /» TVNSIS TVISIO 3HI XB I3S SI 0V13 3SI0N SHI .31 ./ /» HO »/ /» NOIIYHDIVS 3AIIVS3N HO 3SI0N 01 ana) »/ /. aAiivoaN ATINVOIJINSIS s i i v N s i s araaiw aauaxu^ ani J I »/ /» HO */ /» 'Nonvanivs S A i n s o a saivoiaNi TVNSIS aiaaiw HVH 3HI SI :SMO3HO SSION »/ _ ./ •'T + + , / UaiNOOO N33H0S ONV H3IN0.OO VIVO 3IV0dn »/ / * * * * * * * * * * * * * * * * * * J I X * « * * * « « * .'(I'quo U88J3G 'zl [T ] "Ap PTi»-09) 3°P"_5^' ; (-[ 'quo~u88Joe 'j/qd-x-id-oz) qop«_qjo .' (I 'quo_U88Jos ,Z/qd_PT"'-00't) qop«_qjo i (I 'quo U88JDs'j/qd qcp-QH) TOp« qio .'OE ~ qd~xjd (OE < qtTxid) j-f - 9 L I --117-goto end_loop; /» SURE OF NO FLOW SIGNAL BEING PRESENT */ /» CHECK TO SEE IF A FLOW TRANSITION HAS OCCURRED BY EXAMINING */ /« THE PAST SIGNAL FLAG (-0 IF NO FLOW, -1 IF FLOW). »/ i f (signal !=0> /* NO FLOW LAST TIME TOO? */ t r a n s i t i o n - 1; /* HAVE FLOW TRANSITION */ else t r a n s i t i o n - 0; /* HAD NO FLOW LAST TIME TOO * sound(1000,duration); si g n a l - 0; /• DECREASE PROXIMAL PRESSURE SETTING (NO FLOW) •/ p_t - p_t - 6; / *< t t t« t t«< i l t t*««» l t«< t l* l t«« t l< t l t t l t t t i « l t« t t« t«< t*«<t<« i* l t» t t<«** / ) /* GET CURRENT PROXIMAL CUFF PRESSURE */ ouff_pressure(tprox_prss,st_prox,pp0,pp200,auto_cal,psrc); /* IF A FLOW TRANSITION OCCURRED, UPDATE AND RECORD ESTIMATES */ i f ( t r a n s i t i o n — 1) t ++num est; o_p_raw_m2 - o_p_raw_mi; /* SAVE PAST 0_P VALUES »/ o_p~raw_ml - o_p_raw; o p raw - (prox_prss + past_prox)/2; i i Tnum_est < 3 I I f i l t e r — 0) o_p - o_p_raw; /* NO FILTER HERE */ else ~ /* 3-PT AVG. FILTER */ o_p - (o_p_raw+o_p_raw_ml+o_p_raw_m2)/3; r_time[r_count] - (get_the_time_of_day() - start_time)*10/182; result[recount] » o_p; a _ f l a g - 1; /* ENABLE ALARM FLAG AFTER EST. */ past_est_tlme • est_time; est_time - get_the_time_of_day(); /* GET TIME OF NEW EST. */ trend_cnt •» trend_cnt + — (est_tlme-past_est_time) *10/182; /* UPDATE TREND PLOT IF THERE IS ROOM ON THE GRAPH •/ i f (trend_cnt < 300) cr t _ l i n e ( t r e n d _ c n t , 198, trend_cnt, 198-o_p/iu, 1)"; ++r count; crt-srcp(22,50,0); p r i n t f ( " % 3 d " , o p ) ; c r t srcp(24,50,0); p r i n t f ( " ( % d ) " , nuraest); crt_srcp(0,0,0); ~ /. ./ ) /*• SAVE PROXIMAL PRESSURE (BEFORE PRESSURE CHANGES) */ past_prox » prox_prss; /* FALL THROUGH TO SECTION WHERE PRESSURE IS REGULATED BEFORE OBTAINING »/ /« NEW DATA. */ next: /* RESET FLAGS */ i- 0 ; quiet_cnt - 0; noise_cnt » 0; lag_ncheck - 0; detect - 0; nflag - 0; duration - beep*4; /* CHECK ALARM */ -118-i f ((get the time of day() - eat time) •10/182 > alarm time t t a f l a g — 1 ) ( ~ ~ crt_srop(2,2,0) ; p r i n t f < " * * * ALARM * * * " ) ; sound(2000,18); c r t srcp(2,2,0); p r i n t f ! " »); crt_srcp(0,0,0); ) /* REGULATE PRESSURES */ cuff_pressure(tdist_prss,3t_dist,pd0,pd200,auto_cal,psrc); t o - o_p; i f (dist_prss - t o > 6 I I t_p - dis t _ p r s s > 6) qvalve_controlTt_p, auto_cal, inc_diat,dec_dist,pd0,pd200, at_diat,psrc) ; cuf f_prea3ure (tmid_prss, st_mid,pm0,pm200, auto_cal,parc); t p » middle_occl; IF (mid praa - t_p > 6 I I t_p - mid__praa > 6) qvaXve_control(t_p,auto_cal,inc_mid,dec_mid,pm0,pm200, st —mid, pare); cuff_pressure(Cprox_prss,at_prox,pp0,pp200,auto_cal, pare); i f (prox_prss - p t > 2 II p_t - prox_pras > 2) qvalve_controlTp_t,auto_cal,inc^prox,dec_prox,pp0,pp200,at_prox,parc); cuff_pressure(<prox_prss,3t_prox,ppO,pp200,auto_cal,parc); cuff_pressure(tdist_prss,at_diat,pd0,pd200,auto_cal,pare); cuf f_presaure (&mid_prsa, at_mid,pm0,pm200,auto_cal,psrc) ; /* UPDATE PRESSURE VALUES ON SCREEN */ crt_arcp(22,70,0); p r i n t f ( " % 3 d " , prox_prss); crt_srcp<22,65,0); p r i n t f ( " % 3 d " , mid_pras); crt_arcp(22,60, 0); p r i n t f ( " % 3 d " , d i s t _ p r a s ) ; crt_arcp(0,0,0); reaet_amp(LIMIT/5); end_loop: /* GO BACK TO START OF SCREEN LOOP */ /•%%%%%%tt%«4%%%t%%%%%%%tt%%%%%t%%t%%%%%«%%t%t%t«t%%%%««%*%«%%*%%%%%%%%%%« ) /* SCREEN FULL */ /* RESET FLAGS •/ detect =0; lag ncheck •» 0; 1-0; quiet_cnt - 0; noise_cnt - 0; /* REFRESH SCREEN »/ screen_set(pos,neg,threahold,&screen_cnt); crt_srcp(22,50,0); p r i n t f ( " % d " , o p ) ; crt_srcp(1,1,0J: p r i n t f ("alarm time - %3d s e c " , alarm_time) ; crt_srcp(0,5,0); p r i n t f ( " 3 - p t f i l t e r " ) ; i f ( f i l t e r — 1) printf("ON"); else p r i n t f ( " O F F " ) ; crt_srcp(24, SO, 0); p r i n t f ( " ( % d ) " , num_est); crt_srcp(0,0,0); ~" /» UPDATE TREND PLOT, INCLUDING PAST */ /* EST.'S OVER THE LAST 2 MIN. */ r index -'r_count - 2; wHile (r time(r_count-l] - r_time[r_index] < 120) — r_Tndex; ~ ~ ++r_index; for (i-r_index; i<r_count; ++i) trend_cnt - r time(i] - r_time[r_index]; c r t _ l i n e ( t r e n 3 _ c n t , 198, trend_cnt, 198-result[i]/10, 1); -119-i /* GO TO START OF MAIN LOOP */ goto start_loop; /*eeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeegeeeeeeeeeeeeeeeeeeeeeeeeeeeeeee*/ /*****************************************************************************/ /* THIS SECTION CONTAINS THE PROGRAM TERMINATION SECTION */ / • A * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * / menu: current_time = get_the__time_of_day (); crt_mode(6); ~ p r i n t f ( H Select the desired function"); p r i n t f C 0) Shut down (deflate and data store) •*); p r i n t f ( " 1) End t r i a l (deflate and data dump) " ) ; ' p r i n t f C 2) Continue new t r i a l " ) ; p r i n t f C " ) ; key_loop: i f (is_a_key_ready_to_be_read () =- 0) goto key~loop; e l s e ( switch (get_a key with echoO) { case '0': crt_mode(6); sound(50,18); p r i n t f C * system shut down? (1-yes) " ) ; scanf("%d", *ok) ; i f (ok ! - l ) goto menu; /* DEFLATE CUFFS */ qvalve_control(5.auto_cal* inc_prox,dec_prox,ppO,pp200, st_prox*psrc); qvalve_control(5,auto_cal,inc_dist,dec_dist,pdO,pd200, s t _ d i s t , p s r c ) ; qvalve_control(5,auto_cal,inc__mid,dec_mid,pmO,pm20Q, st_mid, pare); crt_mode(6); p r i n t f (" enter f i l e name <« 7 characters " ) ; scanf ("%s", file_name); /* OPEN FILE */ fp_data - fopen(file_name,"w"); /* WRITE TO FILE */ for ( i - 0 ; i<r_count; ++i) f p r i n t f ( f p _ d a t a , • %d % l u " , r e s u l t [ i ] , r _ t i m e [ i ] ) ; /*CLOSE FILE */ fclos e ( f p _ d a t a ) ; goto e n d _ t r i a l ; case '1': qvalve_control(5,auto_cal,inc_mid,dec_mid,pm0,pm200, st_mid,psrc); qvalve_control(5 f auto_cal,inc_prox,dec_prox,ppO,pp200, "~ st_prox,psrc) ; qvalve_control(5,auto_cal,inc_dist,dec_dist,pdO,pd200, — s t _ d i s t , p s r c ) ; crt_mode(6); p r i n t f C number of estimates - %d", num_est); p r i n t f C t o t a l time - % l u s e c " , (current time-start_time) •10/187); p r i n t f ( " data dump to follow"); -120-pause(23); crt_mode(6); avg_result - 0; for (i»0; i<r_count; ++i) ( i f ( r e s u l t [ i ] !=0) avg_result •* avg_result + result [ i j ; p r i n t f ! " %d % l u " , r e s u l t ! ! ] , r t i m e d ] ) ; i f ( i !=0 it 1%20 ~ 0) ( pause(23); crt_mode(6); ) ) avg_result " avg_result/num_est; long_var - 0 ; for (1=0; K r count; ++i) i f ( r e s u l t ( i ] !- 0) long —var = long_var + (unsigned long) ( ( r e s u l t [ i ] - a v g _ r e s u l t ) * ( r e s u l t l i ] - a v g _ r e s u l t v a r _ r e s u l t = (int) (long_var/num e s t ) ; p r i n t f ( " mean - %d variance •» %3", avg_result, v a r _ r e s u l t ) ; pause(23); goto e n d _ t r i a l ; case '2': crt_mode(6); p r i n t f ( " enter a f i l e name <- 7 characters " ) ; scanf("%a", f i l e name) ; fp_data - fopen(?ile name,"w"l; for ( i - 0 ; i<r_count;~++i) f p r i n t f ( f p data, " *d % l u " , r e s u l t [ 1 ] , r_time[i)> fclose (fp_dataT*; ~* nflag - 0; num_est - 0; est_time - 0; r_count - 0; goto t o t a l _ s t a r t ; d e f a u l t : goto menu; -121-(include "stdio.h" (define LIMIT 15000 tdefine BASE 0x330 (define st_prox 1 (define st_mid 5 (define s t _ d i s t 9 (define dyn_prox 13 (define dyn_mid 17 (define dyn_dist 21 (define inc_prox 0 (define dec_prox 2 (define inc_mid 4 (define dec_mid 6 (define i n c _ d l s t 8 (define dec_dist 10 (define auto__reset 4 (define timing_check 6 / f t * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * /« O F3.C /* Th~is f i l e contains functions for the occlusive pressure program /**«**»********«**********************«•«******************************* /* A/D channel 0 */ /• A/D channel 2 */ /* A/D channel 4 •/ /* A/D channel e */ /* A/D channel 8 «/ /* A/D channel 10 */ /* D/A channel 0 */ /» D/A channel 2 */ /* D/A channel 4 */ /* D/A channe1 6 */ /* D/A channel 8 •/ /* D/A channel 10 */ /* D/A channel 12 */ /* D/A channel 14 */ **** I «/ */ unsigned long get_the_time_of_day() /* THIS FUNCTION RETURNS THE NUMBER OF CLOCK TICKS SINCE MIDNIGHT */ /* (written by R. MacNeil) */ unsigned long unsigned short clock t i c k s ; peek(7; ) i /* HIGH WORD */ disable i n t e r r u p t s ( ) ; clock t i c k s - peek( 0x006e, 0x0040 c l o c k ~ t i c k s - clock ticks*0xl0000L + peek"< 0x006c, 0x0040 ); /* LOW WORD */ enable_interrupts(); r e t u r n ( c l o c k _ t i c k s ) ; sound (frequency,duration) unsigned short frequency, duration; /* THIS FUNCTION TURNS ON THE IBM PC SOUND GENERATOR, WITH */ /* FREQUENCY IN HZ, DURATION IN CLOCK TICKS (18.2 TICKS PER SECOND) »/ /* (written by R. MacNeil) */ ( (define TIME_OF_DAY 0x001a struct regvals (unsigned short ax,bx,cx,dx,si,di,ds,es;) r e g i s t e r s ; outportb(0x0043,0xb6); /• LOAD 8253 REGISTER */ frequency - 1190000L/(unsigned long) frequency; outportb(0x0042,frequency £ OxOOff); •utportb(0x0042,frequency » 8); registers.ax - 0; /* READ THE TIME •/ sysint(TIME_OF_DAY, ( r e g i s t e r s , ( r e g i s t e r s ) ; duration +>~registers.dx; frequency - inportb(0x0061); /* SAVE SPEAKER CONTROL VALUE •/ outportb(0x0061,frequency I 3); /* TURN ON SPEAKER «/ while (registers.dx < duration) /* WAIT OUT DURATION '/ { ) registers.ax « 0; sysint(TIME_OF_DAY, ( r e g i s t e r s , ( r e g i s t e r s ) ; outportb(0x0061,frequency); /« RESTORE SPEAKER CONTROL */ wait(duration) unsigned short duration; /* WAIT OUT DURATION CLOCK TICKS */ /* (written by R. MacNeil) */ (define TIME_OF_DAY 0x001a stru c t regvals (unsigned short ax,bx,cx,dx,si,di,ds,es;} r e g i s t e r s ; registers.ax - 0; /* READ THE TIME */ sysint(TIME_OF_DAY, ( r e g i s t e r s , ( r e g i s t e r s ) ; duration +*"registers.dx; -122-while (registers.dx < duration) ( /* WAIT OUT DURATION */ registers.ax - 0; sysint(TIME_OP_DAY, ( r e g i s t e r s , ( r e g i s t e r s ) ; unsigned short is —a_key_ready_to_be_read () /* CHECKS TO SEE IF A KEY HAS BEEN PRESSED */ /* (written by R. MacNeil) «/ i f ( (bdos( 0x0b ) ( OxOOff ) -= OxOOff ) return (1); else return (0 ) ; /» KEY PRESSED */ /* KEY NOT PRESSED •/ unsigned short get_a_key_with_echo() /* RETURNS KEY PRESSED WITH ECHO TO SCREEN »/ /« (written by R. MacNeil) «/ unsigned short key_pressed; key_pressed » bdos(l) ( OxOOff; /* GET KEY WITH ECHO */ i f (key_pressed — 0) /* CHECK FOR EXTENDED KEYS •/ key_pressed - (bdos(l) ( OxOOff) « 8; return(key_pressed); /* /* /* /* /* /* /« /* /* /* /* THE ROUTINE: intlzn(p_o, m_o) This routine i s used to i n i t i a l i z e the i n i t i a l proximal pressure and the middle tourniquet pressure. A l l values are derived from the user. THIS ROUTINE CALLS: nothing GLOBALS: none STATIC LOCALS: none AUTOMATIC LOCALS: INT ok ARGUMENTS: INT *p o, *m o CONSTANTS: none intlzn(p_o,m_o) i n t *p o, *m o; ( in t ok; s t a r t : crt_mode(6); printf("INITIALIZATION OF PARAMETERS " ) ; p r i n t f ( " Input guess of occlusive pressure (about 200 mmHg): " ) ; scanf("%d", p_o); p r i n t f ( * Input middle tourniquet pressure (mmHg): " ) ; scanf("Id", mo); p r i n t f ! " " ) ; printf("ECHO INPUT"); p r i n t f ( " Guess of occlusive pressure - Id", *p_o); p r i n t f ( " Tourniquet pressure - Id", *m_o); p r i n t f ( " O.K. ? (yes - 1) " ) ; scanf("Id", (ok); i f (ok .!- 1) goto s t a r t ; p r i n t f ( " INITIALIZATION FINISHED"); timer(LIMIT); /. . /* THE ROUTINE: screen set(p,n,threshold,screen set) * /* /* This routine refreshes the screen display (axes, t i t l e s , and thresholds), * /* and also resets the screen counter. * /« /* THIS ROUTINE CALLS: functions available In the C86 l i b r a r y to plot l i n e s * /* (crt l i n e ) , set the cursor p o s i t i o n !crt_srcp), and * /* set Ene graphics mode (crt_mode) * /* GLOBALS: none * -123-/* STATIC LOCALS: none */ /* AUTOMATIC LOCALS: INT i »/ /* ARGUMENTS: INT p,n,threshold,*screen_cnt «/ /* CONSTANTS: none «/ screen_set(p,n,threshold,screen_cnt) int p,n,threshold,'screen cnt; < i n t i ; c r t mode(6); /* PLOT AXES «/ c r t line(0,20,620,20,1); crt~line(0,60,620,60,1); crt_line(0,100,620,100,1); crt_line(0,140,620,140,1); crt_line(0,159,620,159,1); c r t line(0,160,620,160,1); crt~line(300,160,300,199,1); crt~line(0,168,305,168,1); crt_line(0,178,305,178,1); crt_line(0,188,305,188,1); c r t line(0,198,639,198,1); crt~line(0,199,0,197,1); c r t line(59,199,59,197,1) ; crt~llne(119,199,119,197,1); c r t llne(179,199,179,197,l); crt~line(239,199,239, 197,1) ; /* PLOT THRESHOLDS FOR */ /* POSITIVE NOISE DETECTION, */ /* NEGATIVE NOISE DETECTION, */ /* AND FLOW DETECTION. «/ for ( i - 0 ; K620; 1 - i+10) < crt_line(i,140-p/2,1+5,140-p/2,1); c r t line(i,140-n/2,1+5,140-n/2,1); crt~line(i,100-threshold/2,1+5,100-threshold/2,1); ) /* PRINT LABELS ONTO GRAPHS */ crt_srcp(0,40,0); printf("Dyn. Pressure vs. Time"); crt_srcp(19,12,0); p r i n t f ( " P r e s s u r e Trend"); crt_srcp(19.SO,0); p r i n t f ( " B a s e l i n e Pressures"); crt_srcp(4, 75, 0); p r i n t f ("P"); crt_srcp(8, 75, 0) ; printf("RM"); crt_srcp(12,75,0); p r i n t f ( " M " ) ; crt_srcp(16,75,0) ; p r i n t f ( " D " ) ; c r t srcp(23,47,0); printf("OCCL PRSS"); crt_srcp(23, 6"1, 0) ; p r i n t f ( " D " ) ; crt_srcp(23, 66, 0); p r i n t f ( " M " ) ; c r t srcp(23,71,0); p r i n t f ( " P " ) ; c r t srcp(23,39,0) ; pr i n t f ( " 1 0 0 " ) ; c r t srcp(22,39,0) ; pr i n t f ( " 2 0 0 " ) ; crt_srcp(21,39,0) ; pr i n t f ( " 3 0 0 " ) ; crt_srcp(24,0,0) ; p r i n t f ( " 0 " ) ; crt_srcp(24,7,0); p r i n t f ( " 1 " ) ; crt_srcp(24,14,0); p r i n t f ( " 2 " ) ; c r t srcp(24,22,0); p r i n t f ( " 3 " ) ; c r t srcp(24,29,0); p r i n t f ( " 4 " ) ; c r t srcp(24,36,0); p r i n t f ( " 5 " ) ; c r t srcp(0,0,0); /* RESET SCREEN COUNTER */ 'screen cnt » 0; -124-/* */ /* THE ROUTINE: c a l screen(screen cnt) */ /. ./ /* This routine refreshes the screen d i s p l a y (axes and t i t l e s ) and also */ /* resets the screen counter for the gain-threshold c a l i b r a t i o n routine. */ /* */ /* THIS ROUTINE CALLS: functions a v a i l a b l e in the C86 l i b r a r y to plot l i n e s «/ /* ( c r t _ l i n e ) , set the cursor p o s i t i o n (crt_srcp), and */ /* set the graphics mode (crt_mode) */ /* GLOBALS: none «/ /* STATIC LOCALS: none */ /* AUTOMATIC LOCALS: INT i «/ /* ARGUMENTS: INT *screen_cnt */ /* CONSTANTS: none »/ ca 1 —screen (screen_cnt) " i n t *screen cnt; c r t mode(6); c r t line(299,20,600,20,1); crt~line(299,60,600,60,1); c r t line(299,100,600,100,1); crt~line(299,140,600,140,1); crt"line(299.0.299,199,1); crt_srcp(0,50,0); printf("Dyn. Pressure vs. Time"); crt_srcp(4,75,0); p r i n t f ( " P " ) ; crt_srcp(8,75,0); printf("RM"); c r t srcp(12,75,0); p r i n t f ( " M " ) ; crt_srcp(16,75,0); p r i n t f ("D"); crt_srcp(0,0,0); ++screen cnt; /» PLOT AXES */ /* PRINT LABELS */ /* RESET SCREEN COUNTER */ /* /» THE ROUTINE: squareroot(val) 7* /• This routine returns the integer square root of the integer passed to i t /* (rounded up). /* /* THIS ROUTINE CALLS: nothing /* GLOBALS: nothing /* STATIC LOCALS: none /« AUTOMATIC LOCALS: INT root /* ARGUMENTS: INT v a l /* CONSTANTS: none int squareroot(val) i n t v a l ; ( int root; root - 0; while (root*root < val) ++root; return(root); ) /. * /* THE ROUTINE: dsampleO * /* * /* This routine samples the o s c i l l o m e t r i c waveform from the d i s t a l bladder * /* and returns t h i s sampled value. * /* * /* THIS ROUTINE CALLS: functions i n the C86 l i b r a r y for writing to a port * /• (outport) and reading from a port (inport) * /* GLOBALS: none * /• STATIC LOCALS: none * /* AUTOMATIC LOCALS: INT adata * /* ARGUMENTS: none * /« CONSTANTS: BASE - base loc a t i o n of A/D and D/A channels and con t r o l * /* r e g i s t e r s i n memory * /* dyn_dist - the A/D channel for the d i s t a l o s c i l l o m e t r i c s i g n a l » int dsampleO < in t adata; outportb(BASE,dyn d i s t ) ; /* SET UP A/D CHANNEL. */ while (inportb(BASE) < 128) /* WAIT FOR CONVERSION AND */ ; /* SAMPLING BY A/D BOARD. */ -125-adata •> inportb(BASE+1); /* GET SAMPLE V return(adata); /* RETURN SAMPLE */ /. ./ /* THE ROUTINE: psampleO »/ /* •/ /* This routine samples the o s c i l l o m e t r i c waveform from the proximal bladder */ /* and returns t h i s sampled value. "/ /* */ /* THIS ROUTINE CALLS: functions i n the C86 l i b r a r y for writing to a port */ /* (outport) and reading from a port (inport) to */ /* f a c i l i t a t e use of the A/0 board */ /* GLOBALS: none */ /» STATIC LOCALS: none «/ /» AUTOMATIC LOCALS: INT adata */ /* ARGUMENTS: none */ /* CONSTANTS: BASE - base loc a t i o n of A/D and D/A channels and control •/ /* r e g i s t e r s i n memory */ /* dyn_prox » A/D channel for the proximal o s c i l l o m e t r i c signal */ in t psampleO ( in t adata; outportb(BASE,dyn_prox); /* SET UP A/D CHANNEL. */ while (inportb(BASE) < 128) /• WAIT FOR CONVERSION AND •/ /* SAMPLING BY A/D BOARD. */ adata - inportb(BASE+1); /• GET SAMPLE «/ return(adata); /* RETURN SAMPLE »/ /. ./ /* THE ROUTINE: msampleO */ /* »/ /* This routine samples the middle c u f f waveform and returns the sampled */ /* value. */ /* */ /* THIS ROUTINE CALLS: functions i n the C86 l i b r a r y for writing to a port «/ /* (outportb) and reading to a port (inportb) to */ /* f a c i l i t a t e use of the A/D. board •/ /* GLOBALS: none */ /* STATIC LOCALS: none */ /* AUTOMATIC LOCALS: INT adata */ /* ARGUMENTS: none •/ /* CONSTANTS: BASE - base l o c a t i o n of A/D and D/A channel and control */ /* r e g i s t e r s l n memory */ /* dyn^mid - A/D channel for the middle o s c i l l o m e t r i c s i g n a l */ int msampleO ( in t adata; outportb(BASE,dyn_mid); while (inportb(BASE) < 128) adata - inportb(BASE+1); return(adata); /. /* THE ROUTINE: timer(d) /* /* This routine i s used to in s e r t a delay into the program (d-15000 => 1 s) /* /* THIS ROUTINE CALLS: nothing /• GLOBALS: none /* STATIC LOCALS: none /* AUTOMATIC LOCALS: INT n,a /* ARGUMENTS: INT d /* CONSTANTS: none timer(d) i n t d; ( int n,a; for (n-0; n<d; ++n) a-0; f o r (n-0; n<d; ++n) a-0; / / . */ /* THE ROUTINE: pause(line_cnt) */ /. */ /* This routine suspends the program u n t i l the user presses a key. A prompt */ -126-/* i s issued to the user on the screen, being printed on l i n e l i n e cnt. */ /* _ */ /* THIS ROUTINE CALLS: cr t _ s r c p , a l i b r a r y function to set the cursor posn. */ /* is_a_key_ready_to_be_read, a function that returns 1 */ /* i f a key was pressed (returns 0 otherwise. */ /* get_a_key_with_echo(), a function that returns the */ /* value of the key pressed. */ /. ./ /* GLOBALS: none */ /* STATIC LOCALS: none */ /* AUTOMATIC LOCALS: none •/ /* ARGUMENTS: INT li n e _ c n t */ /* CONSTANTS: none */ pause(line_cnt) int l i n e cnt; ( c r t s rcp(line_cnt,0,0); p r i n t f C " ) ; crt_ s r c p ( l i n e _ c n t , O , 0 ) ; p r i n t f C H i t any key to continue n ) ; pause_loop: i f (is_a_key_ready_to_be_read() -- 0) goto pause_loop; — switch (get a key with echoO) { - - - -d e f a u l t : 1 /*-/* . / /* THE ROUTINE: qvalve control(p new,a c,inc prss.dec prss,p0,p200,st s i g , */ /* ~ psrc) ~ */ / * • . * / /* This routine sets the bladder s p e c i f i e d by st s i g to a new pressure, */ /• p_new. It regulates proximal pressure more f i n e l y than middle and d i s t a l */ /* pressures. It changes pressures by sending pulsed signals to the pressure*/ /* control valves (to open tbr ^ r rrmatic system to the source or s i n k ) . The */ /* length of the pulses i s - p r o p o r t i o n a l to the difference between the target */ /* pressure and the current pressure. A leak/kink detector signals when an */ /* attempt to change pressure takes more than IS s. */ /* */ /* THIS ROUTINE CALLS: c r t srcp (sets cursor position) «/ /* cuff_pressure (reads bladder pressure) */ /* r e s e t _ s t a r t (disables EEG amplifier) */ /* close_valve (closes valve) */ /* open_valve (opens valve) */ /* timer (causes a delay) */ /* reset stop (enables EEG amplifier) */ /* get_tHe_time_of day (gets reference time) */ /* GLOBALS: none ~ */ /* STATIC LOCALS: none */ /• AUTw^TIC LOCALS: INT adata,count,diff,old_pold,p old,error,d c,window •/ /* UL start_time _ »/ /* ARGUMENTS: INT p new,a c,inc pras,dec_prss,pO,p200,st sig,psrc */ /* CONSTANTS: TIMER -(gives 1 s delay) ~ •/ qvalve_control(p_new,a_c,inc_prss,dec prss,p0,p200,st_sig,psrc) i n t p_new, a_c, inc_prss,dec_prss,p0",p200, s t _ s i g , p s r c ; i n t adata,count, d i f f , old_pold,p_old,error,d_c, window; unsigned long s t a r t time; — d i f f - 0; count - 0; i f (st s i g — s t p r o x ) window - 2; /* SET WINDOW SIZE FOR PRESSURE REG. */ else window - 5; /* GET CURRENT PRESSURE OF BLADDER •/ /* BEING CHANGED */ cuff_pressure(Sp_old,st_sig,pO,p200, a_c,psrc); start_time - get3the_tiie_of_day () ; chkprss: while (p new-p o l d > window I I p_old-p_new > window) ( /* WHILE PRESSURE IS NOT WITHIN WINDOW,*/ /* USE VALVES TO PRODUCE A CHANGE */ reset s t a r t 0; /* DISABLE EEG AMP. */ e r r o r - - (p_new - p_old); /* GET PRESSURE DIFF. */ i f (error < 0) ( /* DECREASE PRESSURE */ --127-close_valve(lnc_prss); open_valve(dec_prss); /* IF ERROR IS "LARGE", KEEP VALVE OPEN TILL */ /« NEXT CHECK. */ /* IF ERROR IS "SMALL", USE PULSED, FINER */ /* TUNING, WITH A DUTY CYCLE PROPORTIONAL TO «/ /* TO THE MAGNITUDE OF THE ERROR. */ i f (error > -12) /* SMALL MAGNITUDE */ ( /* DUTY CYCLE */ /* PROPORTIONAL */ d c - -5»error; /* TO ERROR */ timer! (LIMIT/1000)*d_c); /« DELAY */ close_valve(dec_prss); /'END PULSE*/ ) e l s e ( /• INCREASE PRESSURE •/ close_valve(dec_prss); open —valve(inc_prss); /* PULSE CONTROL FOR SMALL MAGNITUDES */ i f (error < 12) /* SMALL MAGNITUDE */ ( /* DUTY CYCLE »/ /• PROPORTIONAL */ /* TO ERROR */ d c - 5*error; timer!(LIMIT/1000)*d_c); /*DELAY*/ close_valve (inc_prss); ) /* UPDATE PRESSURES AND DIFFERENCE */ old pold - p_old; cuff_pressure(tp old, st sig,p0,p200,a c p s r c ) ; d i f f - p_old - oTd_pold; /* UPDATE DISPLAY OF BASELINE PRESSURES */ i f ( s t _ s i g — st dist) c r t srcp(22760,0); i f ( s t _ s l g — st mid) c r t srcp(22765,0); I f ( s t _ s l g st_prox) c r t srcp<22,70,0); printf(«*3d",p_old); /* LEAK/KINK CHECK »/ i f ((get_the_time of_day()-start time)"10/182 > 15 it p new > 15) ( /* LEAK/KINK */ crt_srcp(4,1,0); printf("EXCESSIVE TIME"); c r t srcp(5,l,0); i f Tst_sig — st _ d i s t ) printf("DISTAL BLADDER"); i f ( s t _ s l g st_mid) printf("MIDDLE BLADDER"); i f (st s i g ~ st_prox) printf("PROXIMAL BLADDER"); ) c r t srcp(0,0,0); cufT_pressure(tp_old,st_sig,pO,p200,a_c,psrc); /• PRESSURE CHANGE FINISHED */ close v a l v e ( i n c _ p r s s ) ; close valve(dec_prss); /« CLOSE VALVES •/ timerTLIMIT/5); reset_stop(); /* ENABLE EEG AMPLIFIER »/ cuff^pressure(Cp_old,st_sig,p0,p200,a_c,psrc) ; i f (st s i g -= s t _ d i s t ) c r t srcp(22,60,0); /* UPDATE DISPLAY */ i f ( s t _ s l g — st_mid) c r t srcp(22,65,0); i f ( s t _ s l g — st_prox) c r t srcp(22,70,0); printf("¥3d", p_old); -128-i f (<get_the_time_of_day()-start_time)*10/182 > 15) { ) crt_srcp(4,1,0); p r i n t f ( " crt_srcp(5,1,0); p r i n t f ( " ort_srcp(0,0,0); THE ROUTINE: cloae_valve(channel) This routine closes the s p e c i f i e d pressure control valve. THIS ROUTINE CALLS: outportb, a l i b r a r y function to write to a port (D/A board) GLOBALS: none STATIC LOCALS: none AUTOMATIC LOCALS: none ARGUMENTS: INT channel CONSTANTS: none close_valve(channel) int channel; i f (channel < 8) < outportb(BASE outportb(BASE outportb(BASE outportb(BASE + 4, channel); /* SET UP D/A MUX. */ + 5, 128); /* SET UP VALUE */ + 4, channel + 8); /• ENABLE CHANNEL */ /* DISABLE CHANNEL */ + 4, channel); else channel a channel - 8; outportb(BASE + 4, channel); outportb(BASE + 5, 128); outportb(BASE + 4, channel + outportb(BASE + 4, channel); /* SET UP D/A MUX. */ /* SET UP VALUE */ 16); /* ENABLE CHANNEL */ /« DISABLE CHANNEL •/ -*/ -*/ */ */ */ */ */ */ */ */ */ */ */ /«-I* /* /* /* /* /* /* /* /* /« /* THE ROUTINE: open_valve(channel) This routine opens the s p e c i f i e d pressure control valve THIS ROUTINE CALLS: outportb, a l i b r a r y function to write to a port (A/D board) GLOBALS: none STATIC LOCALS: none AUTOMATIC LOCALS: none ARGUMENTS: INT channel CONSTANTS: none open valve(channel) Xnt channel; i f (channel < 8) ( outportb(BASE outportb(BASE outportb(BASE outportb(BASE 4, channel); 5, 255); 4, channel + 4, channel); else channel ™ channel - 8; outportb(BASE + 4, channel); outportb(BASE + 5, 255); outportb (BASE + 4, channel + outportb(BASE + 4, channel); /» SET UP D/A MUX. •/ /* SET UP VALUE */ 8); /• ENABLE CHANNEL •/ /* DISABLE CHANNEL */ /* SET UP D/A MUX. »/ /» SET UP VALUE »/ 16); /* ENABLE CHANNEL •/ /* DISABLE CHANNEL »/ /• THE ROUTINE: reset amplr time) /* /* This routine resets (and disables) the EEG amplifier for a time period /* r time (r time - 15000 -> 1 s ) . /* ~ /• THIS ROUTINE CALLS: outportb, a l i b r a r y function to write to a port /• (D/A board) /• GLOBALS: none -129-/* STATIC LOCALS: none */ /* AUTOMATIC LOCALS: INT i */ /» CONSTANTS: none */ reset_amp(r time) int r tTme; f int i ; outportb(BASE + 4,auto_reset); /» SET UP MUX. */ outportb(BASE + 5,255); /* SET UP VALUE <5V) */ for ( i - 0 ; K r time; ++i) /« ENABLE CHANNEL TO */ outportbTBASE + 4,auto_reset + 16); /* ENERGIZE RESET */ /» RELAY. */ outportb(BASE + 4,auto re s e t ) ; /* SET UP MUX. */ outportb(BASE + 5,128); I* SET UP VALUE (OV) . »/ outportblBASE + 4,auto_reset + 16); /» ENABLE CHANNEL TO */ /• DE-ENERGIZE RESET */ /* RELAY. »/ outportb(BASE + 4,auto re s e t ) ; /• DISABLE CHANNEL V ) /. «/ /. »/ /* THE ROUTINE: reset start!) */ /* " */ /• This routine disables the EEG am p l i f i e r by a sending a 5V signal to the "/ /* relay c i r c u i t r y that resets the a m p l i f i e r . */ /* */ /* THIS ROUTINES CALLS: outportb, a l i b r a r y function to write to a port »/ /» (D/A board) */ /» GLOBALS: none »/ /* STATIC LOCALS: none */ /* AUTOMATIC LOCALS: none */ /* ARGUMENTS: none */ /* CONSTANTS: none */ reset sta.L () r outportblBASE + 4,auto r e s e t ) ; /* SET UP MULTIPLEXER */ outportblBASE + 5,255); /* SET UP 5V SIGNAL TO RELAY. •/ outportblBASE + 4,auto_reset + 16); /• ENABLE CHANNEL (ENERGIZE */ /• RESET RELAY) •/ outportblBASE + 4,auto r e s e t ) ; /• DISABLE CHANNEL */ ) /. ./ / . */ /* THE ROUTINE: reset stop!) «/ /* _ */ /* This routine enables the EEG amplifier by sending a OV signal to the */ /* relay c i r c u i t r y that resets the a m p l i f i e r . */ /* */ /* THIS ROUTINE CALLS: outportb, a l i b r a r y function to write to a port */ /* (D/A board) */ /* GLOBALS: none */ /* STATIC LOCALS: none */ /* AUTOMATOC LOCALS: none */ /* ARGUMENTS: none */ /* CONSTANTS: none */ reset_stop() ( outportblBASE + 4,auto r e s e t ) ; /* SET UP MULTIPLEXER »/ outportblBASE + 5,128); /* SET UP OV SIGNAL TO RELAY. »/ outportblBASE + 4,auto reset + 16); /• ENABLE CHANNEL (DE-ENERGIZE*/ /• RESET RELAY) */ outportblBASE + 4,auto_reset); /* DISABLE CHANNEL */ ) /. ./ /. ./ /* THE ROUTINE: auto c a l i b r a t e ( i n c prss,decprss,pO,p200, st sig.psrc) */ /. - -r -r - v /" This routine sets the c a l i b r a t i o n constants (pO,p200,psrc) for the */ /* baseline pressure of the bladder s p e c i f i e d by st s i g . */ /• */ /* THIS ROUTINE CALLS: outportb ( l i b r a r y function for writing to a port), V /* inportb ( l i b r a r y function for reading a port), */ /* close_valve (closes s p e c i f i e d valve), */ /* open_valve (opens s p e c i f i e d valve), and */ /* timer (i n s e r t s a delay) */ /* GLOBALS: none */ /* STATIC LOCALS: none */ -130-/» AUTOMATIC LOCALS: INT i , j «/ /* ARGUMENTS: INT inc_prss,dec prss,*p0,*p200,st sig,*psrc »/ /* CONSTANTS: LIMIT (timing constant) */ auto calibrate(inc_prss,dec_prss,p0,p200,st sig,psrc) Tnt inc_prss,dec_prss,*p0,*p200,st_sig, Ypsrc; int i , j ; /* GET ZERO PRESSURE IN THE CUFF */ /* BY OPENING SYSTEM TO ATMOSPHERE »/ close v a l v e ( i n c p r s s ) ; for (T=0; i<4; ++i) { for (j-0; j<LIMIT; ++j) open_valve(dec_prss); p r i n t f C Zero pressure s e t " ) ; outportb(BASE,st_sig); while (inportb(BASE) < 128) *p0 -'inportb(BASE+1); p r i n t f ! " %d",*p0); /» GET SOURCE PRESSURE FROM USER */ i f C p s r c — 0) < p r i n t f C Enter source pressure"); scanf("%d",parc); ) close valve(dec_prss); for (1-0; i<4; ++i) ( for (j-0; j<4; ++j) open_valve(inc_prss); p r i n t f C Source pressure s e t " ) ; /• GET SOURCE PRESSURE CONSTANT BY */ /* OPENING SYSTEM TO PRESSURE SOURCE */ outportb(BASE,st_sig); while (inportb(BASE) < 128) *p200 - inportb(BASE + 1); p r i n t f ( " %d",»p200); close v a l v e ( i n c _ p r s s ) ; /* END WITH BOTH VALVES CLOSED */ timerlLIMIT) ; p r i n t f ! " " ) ; /. /* THE ROUTINE: cuff_pressure(cff_prss,st_sig,pO,p200,a_c,psrc) /* This routine updates the desired cuff pressure, as determined by values /* passed to the function. /* /* THIS ROUTINE CALLS: outportb and inportb, l i b r a r y functions for writing /* * to and reading from a port (A/D board) /* GLOBALS: none /* STATIC LOCALS: none /* AUTOMATIC LOCALS: INT adata /* DL lng_pold,lng pO,lng_p200,lng_psrc,lng_adata /» ARGUMENTS: INT *cf f_prss, st_sig,p0~,p200,a_c,psrc /* CONSTANTS: none int c uff pressure(cff_prss, st s i g , pO, p200, a_c, psrc) int T c f f_prss, st_sig,p0,p20~0, a_c,psrc; i n t adata; double lng_pold,lng_p0,lng_p200,lng_adata,lng_psrc; outportblBASE,st s i g ) ; while (inportb(BASE) < 128) /* GET A/D VALUE OF CUFF PRESSURE */ adata - inportb(BASE+1); i f (a c -- 0) /* IF NO AUTO CALIBRATION, */ T c f f _ p r s s - (adata - 26)*2; /* ASSUME THIS RELATIONSHIP */ /* BETWEEN A/D VALUE AND */ /* ACTUAL CUFF PRESSURE. */ else ( l n g p s r c - (double) psrc; /* OTHERWISE, USE AUTO-CAL */ IngpO - (double) pO; /* CONSTANTS TO CONVERT A/D*/ lng_p200 - (double) p200; /« VALUES TO CUFF PRESSURE */ lng_adata - (double) adata; -131-lng_pold - lng_psrc*(lng_adata-lng_pO)/ (lng_p200-lng pO); * c f f _ p r s s = (int) lng_poTd; /. ./ /* THE ROUTINE: gain thresholds(pos,neg,o_p,threshold,pare,middle_occl, */ /« proximal Occl,pp0,pp200,pm0,pm200,pd0,pd200,auto cal) */ /* ~ */ /* This routine i n i t i a l l y i n f l a t e a the middle bladder to the tourniquet */ /* pressure and the other bladder3 to a gueaa of the limb occlusion pressure,*/ /* and then allows the user the a b i l i t y to adjust pressures and set */ /* thresholds for noise and flow detection. */ /* */ /* THIS ROUTINE CALLS: l i b r a r y functions c r t _ l i n e , c r t _ s r c p , a n d crt_wdot; and */ /* qvalve_control,timer,cal screen,cuff_pressure, */ /* i s a Icey_ready_to be_rea3, and get_a_key with echo. */ /* GLOBALS: none " */ /* STATIC LOCALS: none */ /* AUTOMATIC LOCALS: INT ok,i,u,k,1,mx,mn,prox prss,dst_dyn[310],prx_dyn(310]*/ /* mid_dyn(310],dst_pt,mld pt,prx ot,dl,d2,map, */ mid_dyn[310],dst_pt,mid pt,prx Dt,dl,d2,map, "/ screen_cnt,sys_peak,peakl,peak?,vl,v2,t_p,pmx,mmx, */ /* dist_prss,mid_prss */ /* UL current_time,sample_time */ /* FLOAT h0,h8 */ /* ARGUMENTS: INT *pos,*neg,*op,'threshold,psrc,middle occl,proximal_occl, */ /* ppO,pp200,pmO,pm200,pdO,pd200,auto_caT */ /* CONSTANTS: LIMIT (timing constant) */ gain_thresholda(pos,neg,o_p,threshold,psrc,middle occl,proximal_occl,ppO, pp200,pmO,pm200,pdO,pd200, auto_caI) i n t *pos,*neg,*o_p,*threshold,psrc,middle_occl,proximal_occl,ppO, pp2 0 0,pmO, pm2 0 0, pdO, pd2 0 0, auto_ca1; int ok,i,k,l,mx,mn,prox prss; int dst_dyn[310],prx dyn[310],mid_dyn[310],dst_pt,prx_pt,mid_pt; int screen_cnt,t_p,dlst_prss, mid_jprss; unsigned long current_time, sample_time; f l o a t h0,h8; s t a r t _ g t : /* INFLATE TO INITIAL PRESSURES t_p - proximal o c c l ; qvalve_controlTt_p,auto_cal,inc_prox,dec_prox,pp0, pp200,at~_prox,pare); qvalve_control(t p,auto cal,inc_dist,dec_dist,pdO, pd200, s t ~ d i s t , p s r c ) ; t_p " middle_occl; qvalve_control(tjp,auto_cal,inc_mid,dec_mid,pm0, " pm200,st_mid,psrc); timer(LIMIT); •threshold - 10; /* INITIAL THRESHOLDS SETTING */ •pos - 6; •neg • -4; hO - -0.2; /* FILTER COEFFICIENTS •/ h8 - 0.8; repeat: /• SET UP SCREEN •/ c a l screen(cacreen c n t ) ; crt~line(320,100-*threahold/2,600,100-*threahold/2,1); c r t line(320,140-«pos/2,600,140-«pos/2,1); crt_line(320,140-«neg/2,600,140-*neg/2,1); crt_srcp(5,0,0); p r i n t f ( " Press ' i ' to i n f l a t e " ) ; p r i n t f ( " Press 'd' to de f l a t e " ) ; p r i n t f (" Press 'T to increase threshold ") ; p r i n t f ( " Press ' t ' to decrease threshold " ) ; p r i n t f ( " Press 'P' or 'p' for +ve noise " ) ; p r i n t f ( " Press 'N' or 'n' for -ve noise " ) ; i - 0; screen_cnt » 0; sample_time - get_the time_of_day0; while (screen_cnt~< 28"0) /* FILL SCREEN WITH SAMPLED WAVEFORM •/ { -132-current_time = get_the_time of_day(); while (current time =- sampTe time) current_tTme = get_the_tIme_of_day(); 3ample_time - current time; dst_dyn[i] » dsample(7; prx_dyn(i ] - psample(); mid_dyn[i] • msamplel) - 128; If (i>8) < ) ++i; /* FILTER »/ dst pt - h8*dst_dyn[i-4]+h0*(dst_dyn(i]+ dst_dyn[i-2]+dst dyn(i-6]+dst_dyn(i-8]) prx_pt = h8*prx_dynTi-4]+hO*<prx_dyn[i)+ prx dyn(i-2]+prx dyn[i-6]+prx_dyn[i-8]) mid_pt~- h8*mid_dynTi-4] +hO* (mid_dyn(i] + mid_dyn(i-2]+mid_dyn[i-6]+mid_dyn[i-8]) i f (dst_pt < -30) dst_pt - -SO.-i f (prxjpt > 30) prx_pt - 30; crt_wdot(140-dst_pt/2,screen_cnt+300,1); crt_wdot(100-mid pt/2,screen_cnt+300,1); crt~wdot(60-mid_3yn[i]/2,screen_cnt+300, 1); crt_wdot(20-prx_pt/2,screen_cnt+300,1); ++screen cnt; /* ACCEPT USER INPUT TO CHANGE PRESSURE AND "/ /* THRESHOLDS */ i f ( is a key ready to be read() 1) <~ ~ ~ - - -switch(get a key with echoO) ( - - -case 'd': /* DEFLATE */ cuff_pressure(fiproxprss,st_prox,pp0,pp200,auto_cal,psrc); t_p • prox_prss - 10; ~ qvalve_control(t_p,auto_cal,inc_prox,dec_prox,ppO,pp200,st_prox,psrc); qvalve_control(t p,auto~cal,inc_dist,dec_dist,pdO,pd200, s t _ d i s t , p s r c ) ; t_p - middle_occl; qvalve c o n t r o l ( t p , a u t o c a l , i n c mid,dec mid,pm0, pm200, st mid,psrc); i-0; - _ goto end_lp; /* INFLATE •/ cuff_pressure((prox prss,st_prox,pp0,pp200,auto_cal,psrc); t_p « prox_prss + 10; qvalve_control(t_p,auto_cal,inc_prox,dec_prox,ppO,pp200,st_prox,psrc); qvalve_control(t p,auto~cal,inc_dist,dec_dist,pd0,pd200,st_dist,psrc); t_p - middle_occX; ~ qvalve_control (t_p, auto_cal, inc_mid,dec_mid, pm0,pm200, st_mid,psrc) ; goto end —lp; — — — crt_line(600,100-»threshold/2,610,100-*threshold/2,0); — ' t h r e s h o l d ; crt_line(600,100-»threshold/2,610,100-«threshold/2,l); goto end_lp; crt_line(600,100-*threshold/2,610,100-*threshold/2,0); ++*threshold; crt_line(600,100-«threshold/2,610,100-*threshold/2,1)i goto end_lp; /* DECREASE FLOW */ /* THRESHOLD */ /* INCREASE FLOW •/ /* THRESHOLD */ crt_line(600,140-*pos/2,610,140-*pos/2,0); — * p o s ; crt_line(600, 140-*pos/2,610,140-*pos/2,1); goto end_lp; crt_line(600,140-*pos/2,610,140-»pos/2,0)j ++*pos; crt_llne(600,140-»pos/2,610,140-*pos/2,1)j goto end_lp; case *p': /» DECREASE +VE */ /* NOISE THRESHOLD */ /* INCREASE +VE «/ /* NOISE THRESHOLD */ -133-case n' crt_line(600,140-*neg/2,610,140-*neg/2,0); —*neg; crt_line(600,140-»neg/2,610,140-*neg/2,1); goto end —lp; crt_line(600,140-«neg/2,610,140-»neg/2,0); ++*neg; crt_line(600,140-*neg/2,610,140-«neg/2,1); goto end l p ; cuff_pressure(o_p,st_prox,pp0,pp200,auto_cal,psrc); goto end__rtn; d e f a u l t : goto end l p ; /* DECREASE -VE */ /* NOISE THRESHOLD »/ /* INCREASE -VE */ /* NOISE THRESHOLD */ /« SET PRESSURES */ /• AND THRESHOLDS */ /* TO CURRENT VALUES */ ) end_lp: goto repeat; /» FINISHED GAIN-THRESHOLD ROUTINE */ end r t n : ) 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0065057/manifest

Comment

Related Items