 |
Description  |
|
|
BACKGROUND OF THE INVENTION
This invention relates to a vital signs monitor which combines the signals
from multiple sensors to generate more reliable measurements of vital
signs such as heart rate, respiration rate, and vasomotor activity.
Conventional vital signs monitors use single sensors to measure individual
vital signs. For example, an ECG sensor can be used to measure heart rate,
or a chest strap connected to a strain gauge can be used to measure
respiration rate. Such single sensor techniques are prone to measurement
errors when the sensor generates artifacts as a result of body motion,
poor sensor placement, and the like. At least in part for this reason,
vital signal monitors are often restricted to use by highly trained
personnel, as for example in an intensive care unit.
There is a need for a vital signs monitor which is less dependent on the
error-free operation of individual sensors. Such a monitor could be used
effectively by less highly trained personnel, and would be better suited
for use by a subject at home, without the assistance of a health care
professional.
SUMMARY OF THE INVENTION
The vital signs monitor of this invention combines information from
multiple sensors to improve the reliability and the accuracy with which
vital signs are measured. It has been found that the signals generated by
sensors commonly used in vital sign monitors can be cross-referenced to
reduce measuring errors and increase accuracy, without significantly
increasing the cost or the electronic complexity of the monitor.
For example, an ECG sensor can be used to generate a first series of pulses
indicative of heart rate, but chest motion of the subject may interfere
with the ECG sensor. Similarly, an oximetric finger probe can be used to
generate a second series of pulses indicative of heart rate, but hand
motion of the subject may interfere with the finger probe signal. The
monitor described below responds to both the first and second pulses by
identifying those pulses with more regularly repeating pulse intervals and
by using the identified pulses to determine heart rate of the subject. In
this way, proper operation continues, even when the ECG signal or the
finger probe signal is interrupted.
As another example, a conventional electrical impedance sensor can be
attached to the chest of a subject to generate a series of impedance
pulses correlated with respiration rate. However, chest motion such as
that associated with a cough will often create additional impedance pulses
which can result in an over estimate of respiration rate. The monitor
described below suppresses these additional impedance pulses by using a
chest wall motion sensor (such as a chest strap coupled to a mechanical
strain gauge) to detect chest motion artifacts. Impedance pulses
associated with such chest motion artifacts are then suppressed.
As a third example, vasomotor activity can be monitored by automatically
comparing core blood pressure with the waveform generated by an oximetric
finger probe. The ratio of pulse amplitude as measured with an occluding
cuff over a major artery to pulse amplitude of the finger probe waveform
is an excellent measure of the extent to which the vascular bed at the
finger is dilated or constricted. As described below, this ratio can be
automatically measured and checked for trends that may give advance
warnings of impending changes in core blood pressure.
In each of these three examples, two separate sensor signals are correlated
to generate an accurate measure of the desired parameter. The heart rate
measurement approach described above is relatively undemanding with
respect to sensor placement, since either of the two sensors can supply
information to the monitor if the other fails to provide a reliable
signal. The respiration rate measuring approach is relatively immune to
chest motion artifacts since the chest wall motion sensor provides an
excellent indication of such artifacts. The vasomotor activity monitoring
approach provides an accurate measure of dilation and constriction of a
peripheral vascular bed.
The invention itself, together with further objects and attendant
advantages, will best be understood by reference to the following detailed
description, taken in conjunction with the accompanying drawings.
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1 is a block diagram of a vital signs monitor which incorporates the
presently preferred embodiments of this invention.
FIGS. 2a, 2b; 3a, 3b; 4a, 4b; 5a, 5b; 6a, 6b; and 7a, 7b, are signal
waveforms related to the heart rate monitoring system of the monitor of
FIG. 1.
FIGS. 8a-8c are flow charts of a first version of the heart rate monitoring
system of the monitor of FIG. 1.
FIG. 9 is a flow chart of a second version of a heart rate monitoring
system suitable for use in the monitor of FIG. 1.
FIGS. 10a and 10b are signal waveforms related to the respiration rate
monitoring system of the monitor of FIG. 1.
FIG. 11a is a flow chart of the respiration rate monitoring system of the
monitor of FIG. 1.
FIG. 11b is a flow chart of an alternate respiration rate monitoring
system.
FIGS. 12a and 12b are signal waveforms related to the blood pressure
monitoring system of the monitor of FIG. 1.
FIG. 13 is a graph of the time delay between the blood pressure cuff pulse
and the finger probe pulse as a function of the pressure of the blood
pressure cuff.
FIG. 14a and 14b are signal waveforms related to the blood pressure
monitoring system of the monitor of FIG. 1.
FIG. 15 is a flow chart of the blood pressure monitoring system of the
monitor of FIG. 1.
FIGS. 16a-16e are flow charts of the vasomotor activity detection and blood
pressure waveform display system of the monitor of FIG. 1.
FIGS. 17-20 are detailed schematic diagrams of electronic circuits suitable
for use in the monitor of FIG. 1.
DETAILED DESCRIPTION OF THE PRESENTLY PREFERRED EMBODIMENTS
Turning now to the drawings, FIG. 1 shows a block diagram of a vital signs
monitor 10 that incorporates presently preferred embodiments of this
invention. This monitor 10 includes a microcomputer 12 which controls a
display 14. A number of sensors, including an ECG sensor 16, an oximetric
finger probe 18, an impedance sensor 20, a strain gauge 22, and a blood
pressure cuff pressure sensor 24, generate waveforms. These waveforms,
after suitable signal processing, are interrelated as described below by
the microcomputer 12 in order to generate reliable measures of heart rate,
respiration rate, and blood pressure.
The ECG sensor 16 is a conventional system used to monitor electrical
voltages associated with cardiac activity. The waveforms generated by the
ECG sensor 16 are processed in a pulse detector 26 to generate a train of
ECG pulses which are applied as signal inputs to the microcomputer 12.
The oximetric finger probe 18 is a conventional sensor which optically
monitors the transmission or reflection of light through the finger of the
subject to generate a periodic signal. The AC component of this periodic
signal can be used as a measure of pulse rate of the subject. The periodic
signal generated by the oximetric finger probe 18 is applied to a pulse
detector 28, which generates a series of finger probe pulses applied as
inputs to the microcomputer 12.
Sensors such as the electrical impedance sensor 20 of FIG. 1 are
conventionally attached to the chest of a subject to provide a measure of
respiration. Rhythmic chest motion associated with breathing results in an
AC component of the signal generated by the sensor 20. This signal is
digitized in an A/D converter 30, and the digitized signal (including the
periodic component) is applied as an input to the microcomputer 12.
Similarly, chest straps are conventionally used to monitor respiration, and
such chest straps typically contain strain gauges such as the strain
gauges 22 of FIG. 1. The strain gauges 22 generate a waveform having a low
amplitude periodic component correlated with the periodic chest motion of
breathing and high amplitude peaks associated with artifacts such as
coughing. This signal waveform is applied to a pulse detector 32 which is
selectively responsive only to large peaks to generate pulses which are
applied as an input to the microcomputer 12. Pulse detector 32 responds
only to large pulses, and therefore the output of the pulse detector 32 is
indicative of artifacts.
As shown in FIG. 1 the monitor 10 also includes a blood pressure cuff 34.
This cuff 34 is of the conventional type which is intended to be wrapped
around the upper arm of the subject so as selectively to block blood flow
through the arm. The cuff 34 is inflated by an inflation pump 36
controlled by the microcomputer 12, and the cuff 34 is deflated by a
deflation valve 38 also controlled by the microcomputer 12. The pressure
sensor 24 monitors the air pressure within the blood pressure cuff 34, and
produces a pressure signal which is digitized in an A/D converter 40 and
applied as an input to the microcomputer 12. In addition, the pressure
signal generated by the pressure sensor 24 is applied to a pulse detector
42, which generates a series of pulses that are also applied as an input
to the microcomputer 12.
Those skilled in the art will recognize that a wide variety of conventional
components can be used for each of the devices 12-42 described above, and
the details of construction of these devices form no part of the present
invention. Suitable ECG sensors, oximetric finger probes, impedance
sensors, respiration strain gauges, blood pressure cuffs, and pressure
sensors are all well known to those skilled in the art. Furthermore, a
wide variety of pulse detectors can be used to perform the functions
described above. Simply by way of example, the electronic circuit shown in
FIG. 19 can be used for the pulse detector 26; the electronic circuit
shown in FIG. 18 can be used for the pulse detector 28, and can be
modified by tuning resistor values for the pulse detectors 32 and 42; the
circuit shown in FIG. 17 can be used to implement the impedance sensor 20;
and the circuit shown in FIG. 20 can be used to implement the respiration
strain gauges 22. The circuits of FIGS. 17-20 have been provided merely by
way of illustration, and they are in no way intended to limit the scope of
this invention. It is anticipated that many applications of the inventions
described herein will utilize circuitry which differs significantly from
that of FIGS. 17-20.
The microcomputer 12 is programmed to process the input signals described
above to generate reliable measures of heart rate, respiration rate, blood
pressure, and blood pressure waveform. The following detailed discussion
will take up these four aspects of the monitor 10 in sequence.
HEART RATE MONITORING
FIGS. 2a-7b provide pairs of signal waveforms that will be used to explain
the principle of operation used in heart rate monitoring. Within each pair
of waveforms the time base is identical; thus, FIGS. 2a and 2b represent
the same time interval, and vertically aligned portions of the two
waveforms correspond to the same instant in time. The signals labeled "ECG
Waveform" were taken from conductor 44 of FIG. 1 and the signals labeled
"ECG Pulses" were taken from conductor 46. Similarly, the signals labeled,
"Finger Probe Waveform" were taken from conductor 48 and the signals
labeled "Finger Probe Pulses" were taken from conductor 50.
FIGS. 2a and 2b show a typical ECG waveform and the ECG pulses derived
therefrom. In the absence of artifacts such as those associated with chest
motion, the ECG pulses are regular and periodic. However, as shown in
FIGS. 3a and 3b, the ECG waveform can be distored by chest motion. Such
distortions can cause the pulse detector 26 to fail to detect pulses in
the ECG waveform associated with cardiac pulses of the subject. In the
waveform of FIG. 3b, two chest motion artifacts are noted.
FIGS. 4a and 5a show the waveform generated by the oximetric finger probe
18. This waveform is normally regular and periodic, but finger motion can
distort the waveform. FIGS. 4b and 5b show pulse trains generated by the
pulse detector 28. These pulse trains can be interrupted by finger motion,
and Figures 4b and 5b have been marked to show such finger motion
artifacts.
FIGS. 6a, 6b and 7a, 7b show the ECG pulses and the finger probe pulses for
two time intervals. In the time interval of FIGS. 6a and 6b the finger
probe pulses are regular and periodic throughout the time interval, while
the ECG pulses are interrupted by a chest motion artifact. Conversely,
during the time interval shown in FIGS. 7a and 7b, the ECG pulses are
regular and periodic while the finger probe pulses are interrupted by
finger motion artifacts. These waveforms show that the accuracy of the
information in either the ECG channel or the finger probe channel may be
compromised, by chest motion and coughing on the one hand and by finger
motion on the other hand. However, in many cases the pulse information
derived from the oximetric finger probe at the fingertip is not disturbed
at the same time as is the pulse information derived from the ECG sensor.
According to this invention the microprocessor monitors the pulse trains
generated by both of the pulse detectors 26 and 28 and selects the pulse
train with the more accurate information as the better indicator of heart
rate. The general approach is to chose the channel having the signal with
the more regularly repeating pulse intervals. In this way, the information
provided by two separate sensors in interrelated and combined to provide
an improved heart rate measurement.
FIGS. 8a-8c provide a flow chart of a first preferred embodiment of a
program for interrelating the ECG and finger probe pulses.
The program of FIGS. 8a, 8b begins by initializing internal variables and
setting the variables Skip Count and Half Count to zero. A pulse search
routine is then executed which searches for pulses generated by either of
the pulse detectors 26, 28. Once a pulse is found, the program branches at
decision diamond 52, depending upon whether the pulse found is an ECG
pulse or a finger pulse.
Assuming the pulse is an ECG pulse, the program then searches in block 54
for a next pulse. Once the next pulse is found the program branches at
decision diamond 56, depending upon whether this next pulse is a finger
pulse or an ECG pulse.
During normal operation, ECG pulses will alternate with finger pulses.
Assuming this to be the case the variable W.sub.EF (equal to the time
window from an ECG pulse to a next adjacent finger pulse) is set equal to
the measured interval and Skip Count is set equal to zero. Then the
variable Half Count is incremented and checked at decision diamond 58. In
the first pass through the program Half Count would be equal to 1 and
control would then branch to a next pulse search as indicated at block 60.
Once a pulse is found in block 60, control branches at decision diamond
62, depending upon whether this pulse is an ECG pulse or a finger pulse.
If both the ECG channel and the finger probe channel are functioning
properly, and if the previous pulse was a finger pulse, this pulse should
be an ECG pulse. If so, the variable W.sub.FE (equal to the time window
from a finger pulse to the next adjacent ECG pulse) is set in block 64 and
the variable Skip Count is set to zero. Then the variable Half Count is
incremented and is compared with the constant 2. If Half Count equals 2,
indicating that both W.sub.EF and W.sub.FE have been set, then the
variable W.sub.T is set equal to W.sub.EF plus W.sub.FE in block 66.
W.sub.T is indicative of the total time interval for a heartbeat, and in
this branch of the program is equal to the sum of the time from an ECG
pulse to the next adjacent finger pulse, and from that finger pulse to the
next adjacent ECG pulse.
Once W.sub.T has been set in block 66 the subroutine Update is called. As
shown in FIG. 8c, this subroutine sets a new pulse rate equal to the
inverse of W.sub.T, resets Half Count to zero, and then returns. Control
is then returned to block 54.
In the event that no finger pulse is detected between two adjacent ECG
pulses, then the program branches at decision diamond 56 to block 68, at
which the variable W.sub.EE is set equal to the time window or interval
between two adjacent ECG pulses. The variable W.sub.EE is then compared in
decision diamond 70 with a value equal to 75% of the most recent value of
W.sub.T. If W.sub.EE is less than this value, indicating that the current
ECG pulse has occurred too soon and is therefore probably an artifact,
control branches at decision diamond 70 to block 72, which skips the
current ECG pulse and increments Skip Count. Control is then returned to
block 54.
In the event that the comparison in the decision diamond 70 indicates that
the most recent ECG pulse has not occurred earlier than the allowed
window, then W.sub.EE is then compared for reasonableness in decision
diamond 74. If W.sub.EE is less than 125% of the most recent value of
W.sub.T, then Skip Count is checked in decision diamond 76. If Skip Count
is equal to zero then W.sub.T is set equal to W.sub.EE and the subroutine
Update is called. On the other hand, if Skip Count is not equal to zero
(indicating one or more skipped pulses), then W.sub.T is reduced by 5% and
control is returned to block 78. If W.sub.EE is greater than 125% of
W.sub.T (and therefore outside of the expected range), then W.sub.T is
either incremented or decremented by 5%, depending upon the state of the
variable Skip Count, and control is returned to block 78.
In general terms, the portion of the program associated with blocks 68
through 76 is only executed when two adjacent pulses are ECG pulses. In
this case W.sub.EE is set, compared with upper and lower values, and then
used to update W.sub.T if Skip Count is zero. If W.sub.EE is either too
small or too large, W.sub.T is modified without calling the routine Update
in order to bring W.sub.T more nearly equal to W.sub.EE.
Similarly, if two adjacent pulses are finger pulses, then control branches
at decision diamond 62 to block 80, which sets the variable W.sub.FF (the
time window or interval between two adjacent finger pulses), compares
W.sub.FF with upper and lower limits as described above, checks the
variable Skip Count, and then sets W.sub.T equal to W.sub.FF in the event
W.sub.FF is reasonable and Skip Count is zero. The portion of the program
associated with block 80 operates identically to that associated with
block 68, except that block 80 relates to the use of two adjacent finger
pulses to set the variable W.sub.FF.
From the foregoing description it should be apparent that the program of
FIGS. 8a, 8b automatically selects the pulse train with the more regular
waveform for use in updating the variable W.sub.T and therefore the
measured pulse rate. In the event both the ECG channel and the finger
probe channel are functioning properly, the ECG and finger pulses
alternate in time and the variables W.sub.EF and W.sub.FE are summed to
obtain W.sub.T, the total time between pulses and the reciprocal of the
new pulse rate. On the other hand, if either the ECG pulses or the finger
pulses drop out (due to a motion artifact, for example), the program of
FIGS. 8a, 8b selects the remaining pulse train and then measures the
interval between adjacent pulses in the remaining pulse train to set to
set W.sub.T and therefore the new pulse rate.
The program of FIGS. 8a, 8b operates on a pulse by pulse basis to determine
the more regular pulse train. This is not a requirement for all
embodiments of this invention, and FIG. 9 shows an alternative program
which operates with groups of pulses.
The program of FIG. 9 first counts the number of ECG pulses and the number
of finger pulses detected during a preset measurement time interval. The
measured ECG pulse count is then checked to determine whether it is close
to the measured finger pulse count in decision diamond 82. If not, the
count is repeated until the ECG pulse count is within the desired
tolerance of the finger pulse count. Once this is the case, the variable
Previous Count is set equal to the average of the ECG and finger pulse
counts in block 84. The program then counts in block 86 the number of ECG
and finger pulses detected during a next measurement time interval, which
is equal in duration to the previous measurement time interval. Control
then branches depending upon whether the variable Previous Count is closer
to the ECG count or to the finger count of block 86. If Previous Count is
closer to the ECG count then Previous Count is set equal to the ECG count.
Conversely, if Previous Count is closer to the finger count, then Previous
Count is set equal to the finger count. In either case, the new pulse rate
is set equal to the updated Previous Count divided by the measurement time
interval, and control is returned to block 86.
The program of FIG. 9 monitors the ECG pulses and the finger pulses
occurring during the measurement time interval and automatically chooses
the pulse train having a pulse count which more nearly corresponds to the
previous pulse count. In this way, the more periodic and regular pulse
train is selected for use in determining the new pulse rate.
In both the programs of FIGS. 8a, 8b and the program of FIG. 9 the ECG and
the finger probe pulse train are interrelated and the pulse train with the
more regular pattern is selected to provide an improved, more reliable,
more artifact-free measure of heart rate. Of course, this invention is not
restricted to use with finger probes and ECG sensors. Rather, any two
measures of pulse rate can be used, such as pulse sensors of the type
which include an occluding cuff or which monitor ECG, chest impedance, or
some other parameter from which pulse information may be derived. In
addition the heart rate monitoring system described above can be used in
monitors which do not measure respiration rate or blood pressure.
RESPIRATION MONITORING
The vital signs monitor 10 includes an impedance sensor 20 and a strain
gauge 22, as described above. The outputs of these two sensors are
combined to produce a particularly reliable measure of respiration rate.
The impedance sensor 20 measures the fluctuating electrical impedance
associated with the changing volume of the chest of the subject. The
strain gauge 22 mechanically monitors the physical changes in chest
dimension. Both of these measuring techniques are subject to artifacts
associated with chest motion such as coughing. FIGS. 10a and 10b show the
signal outputs of the impedance sensor 20 and the strain gauge 22,
respectively, for a preset measuring interval. The impedance waveform of
FIG. 10a includes regular peaks associated with respiration, as well as
additional peaks associated with coughs. The peaks associated with coughs
are only slightly greater in amplitude than the peaks associated with
normal respiration, and it is difficult to distinguish reliably between
these two types of peaks in the impedance waveform. However, as shown in
FIG. 10b, the artifacts associated with chest motion stand out clearly in
the strain gauge waveform. The microcomputer 12 is programmed to use the
strain gauge waveform as a measure of chest motion artifacts in order to
correct the breathing rate as determined by the impedance waveform.
This program is flowcharted in FIG. 11a. As shown in FIG. 11a the first
step is to count the number of respiration peaks in the impedance channel
and the number of large peaks associated with chest motion artifacts in
the strain gauge channel during a preset measuring interval. Then the
variable Breath Count is set equal to the impedance channel count minus
the strain gauge channel count, and finally the respiration rate is set
equal to the Breath Count divided by the measuring interval. By combining
the impedance waveform with the strain gauge waveform as described above,
a more accurate measure of respiration rate is obtained.
Another algorithm which can be used to obtain an accurate measure of
respiration rate is flowcharted in FIG. 11b. With this algorithm, the
variable BREATH COUNT is used to count peaks on the impedance channel
waveform during a measuring period having a time duration equal to
MEASUREMENT INTERVAL. In order to suppress artifacts associated with chest
motion, BREATH COUNT is not incremented during periods of unusual chest
motion, as indicated by the presence of a large peak on the strain gauge
channel. Furthermore, the total duration of the periods of unusual chest
motion is measured and stored in the variable MOTION INTERVAL. The
respiration rate is then set equal to BREATH COUNT divided by the actual
time period of measurement (MEASUREMENT INTERVAL--MOTION INTERVAL).
As before, this invention is not restricted to use with the particular
sensors described above. For example, piezoelectric films can be used to
measure changes in chest dimension and electromyograms can be used to
measure muscle potentials of the lower chest in place of the strain gauges
described above. In addition, air flow sensors such as microphones on the
trachea, flow meters on an airway, and thermistors; CO.sub.2 sensors in
expired air; and other types of sensors can be substituted for the
impedance channel described above. Furthermore, the respiration rate
monitoring system described above can be used in monitors which do not
measure heart rate or blood pressure.
BLOOD PRESSURE MONITORING
The vital signs monitor 10 also determines blood pressure of the subject,
this time by combining pressure information from the blood pressure cuff
34 with pulse information from the oximetric finger probe 18.
In general terms, the blood pressure cuff 34 is inflated to a level higher
than the systolic blood pressure, until arterial pulsations in the
fingertip (as measured by the oximetric finger probe 18) cease. The
occluding cuff 34 on the arm is then gradually deflated. Once the cuff
pressure drops below the systolic blood pressure, blood flow resumes
beneath the cuff. This immediately causes a pulsation detected by the
oximetric finger probe 18 at the fingertip, and it is the appearance of
this first pulsation that signals that the cuff has dropped to systolic
pressure. The microcomputer 12 reads the systolic pressure from the
pressure sensor 24. As the pressure in the cuff 34 continues to decrease
below the systolic pressure, pulsations are received by the microcomputer
12 from both the pressure sensor 24 associated with the cuff 34 and from
the oximetric finger probe 18. The pulses arriving at the fingertip occur
shortly after the pulses appearing at the cuff. There are two factors
which contribute to this delay. First, it takes a period of time for the
pulse to propagate from the upper arm to the fingertip. Second, the
partial occlusion of the artery, due to the pressure in the cuff 34 having
a value below systolic and above diastolic, causes a delay in the pulse
below the cuff. The second factor is dependent upon the difference between
cuff pressure and diastolic pressure. As the cuff pressure drops closer to
the diastolic pressure, the second type of delay decreases, and when the
cuff pressure falls below diastolic pressure, the second type of delay is
no longer a factor. Therefore, once the cuff pressure falls below the
diastolic blood pressure, the delay between the cuff pulse and the finger
probe pulse becomes substantially constant.
FIG. 13 is a graph showing the time delay between (1) the arm pulse as
detected by the pressure sensor 24 and the pulse detector 42 and (2) the
finger pulse as detected by the oximetric finger probe 18 and the pulse
detector 28, as a function of the pressure of the blood pressure cuff 34.
As shown in FIG. 13, when the cuff pressure is just below systolic
pressure there is a delay of about 260 milliseconds between the cuff pulse
and the finger pulse. As the cuff pressure decreases to the diastolic
value, the delay decreases to a value of about 65 milliseconds. As the
cuff pressure further decreases below diastolic, the delay is not
decreased, but simply oscillates around the 65 millisecond value. As the
cuff deflates, the transition from a decreasing delay to a fixed delay
indicates the diastolic pressure value.
The oscillations in the delay value (which are most easily seen when the
cuff pressure is below diastolic) are synchronous with the subject's
respiratory cycle. Previous work has shown that these oscillations are due
to variations in cardiac stroke volume which occur as a result of
fluctuations in intrathoracic pressure during the respiratory cycle.
FIGS. 14a and 14b show waveforms associated with oscillations in pressure
of the cuff 34 and oscillations in the signals generated by the finger
probe 18, respectively. The waveform of FIG. 14a is similar to that of
FIG. 12b, except that the time scale of FIG. 14a has been expanded to
enable more careful analysis of pulse timing. The left hand portions of
the waveforms of FIGS. 14a and 14b illustrate the suppression of pulses at
the fingertip when the cuff pressure is higher than the systolic pressure.
As the pressure drops to the systolic level, the trace of FIG. 14b begins
to indicate arterial pulsations at the fingertip. The arrival of this
first pulse is an indication that the instantaneous pressure indicated by
the pressure sensor 24 corresponds to systolic blood pressure of the
subject. In the center portion of FIGS. 14a and 14b both waveforms
indicate the presence of pulsations as the cuff pressure drops from the
systolic to the diastolic. Careful examination of the timing differences
between pulses in FIGS. 14a and 14b show the progressive decrease in the
delay time as described above in conjunction with FIG. 13. After the cuff
pressure reaches diastolic, the right hand portion of the waveforms of
FIGS. 14a and 14b illustrate the substantially constant delay between the
respective pulses. Thus, the data shown in FIGS. 14a and 14b provide the
necessary information for determining both the systolic and the diastolic
blood pressure through the use of interrelationships between cuff
pulsations detected by the pressure sensor 24 and fingertip pulsations as
detected by the oximetric finger probe 18.
FIG. 15 shows a flowchart of a suitable program for implementing the blood
pressure monitoring technique described above. As shown in FIG. 15 the
first step is to inflate the blood pressure cuff 34 until the amplitude of
the pulses detected by the oximetric finger probe 18 is at very low level.
This is used to determine that the pressure of the cuff 34 is above
systolic. Then the cuff 34 is deflated at a rate such as 3 millimeters of
mercury per second.
The microcomputer 12 then monitors the pulse detector 42 for cuff pulses.
If no cuff pulses are found within an allowed time, the routine aborts.
Otherwise, the routine then monitors the pulse detector 28 for finger
probe pulses. Once a finger probe pulse is found, the microcomputer sets
the systolic pressure equal to the DC cuff pressure at the time when the
previous cuff pulse occurred. The routine then monitors the pulse detector
42 for additional cuff pulses and the pulse detector 28 for additional
finger pulses. The delay time between each cuff pulse and the subsequent
finger pulse is then calculated and compared with the previous delay time.
As long as the newly calculated delay time is less than the previous delay
time, the routine sets the previous delay time to the newly calculated
delay time and returns to search for more cuff and finger pulses. However,
in the event the newly calculated delay time is equal to or greater than
the previous delay time, then the diastolic pressure is set equal to the
currently prevailing DC value of the cuff pressure as measured by the
pressure sensor 24 and the A/D converter 40.
Once systolic and diastolic pressures have both been measured, the cuff is
deflated completely and the routine terminates.
It is important to note that the foregoing technique for blood pressure
measurement does not increase the hardware complexity of the vital signs
monitor 10 in any way. The oximetric finger probe 18 and the pulse
detector 28 are present in the vital signs monitor 10 for other purposes,
and therefore the present invention provides important improvements in
accuracy over the prior art oscillometric technique without increasing the
cost or hardware complexity of the vital signs monitor 10.
Once again, this system is not restricted to use with the particular
sensors described above. Other pulse detectors can be substituted for the
oximetric finger probe to detect pulses distal to the cuff, and other
pulse detectors can be used to detect pulses at or near the cuff. In
addition, the blood pressure monitoring system described above can be used
in monitors which do not measure heart rate or respiration rate.
BLOOD PRESSURE WAVEFORM MONITOR AND VASOMOTOR ACTIVITY DETECTOR
It is well known that oximetric sensors such as the oximetric finger probe
18 provide excellent relative measures of the blood pressure waveform.
However, standard oximetric waveforms are not calibrated as to core blood
pressure, and this lack of calibration is a significant limitation in many
settings. The monitor 10 overcomes this problem by automatically combining
core blood pressure information derived from an indirect blood pressure
measuring system such as that described in the preceding section with
waveform information derived from a waveform detector such as the
oximetric finger probe 18. By providing an electronic system such as the
microcomputer 12 with signals from both sources, it is possible to
generate a continuous waveform which is calibrated as t | | |
