Method and apparatus for controlling a ventilator

[0001] This application claims the benefit of U.S. provisional patent application No. 60/481,693, filed Nov. 21, 2003, the entire contents of which are incorporated herein by reference. This application is related to U.S. Pat. No. 4,986,268 entitled “Method and Apparatus for Controlling an Artificial Respirator,” the disclosure of which is incorporated by reference.

[0002] 1. Field of the Invention

[0003] The present invention relates to a method and apparatus for controlling a ventilator such as a mechanical ventilator (i.e. an artificial respirator) or a respiratory assist device. In particular, the present invention relates to a method and apparatus for controlling a ventilator based on the measured levels of oxygen of the patient on the ventilator, as well as other physical conditions of the patient.

[0004] 2. Background of the Invention

[0005] Mechanical ventilators and other respiratory assist devices are extensively used to treat and manage all patient populations. In the past few decades, there have been significant changes in the features offered by the ventilators and they have become increasingly responsive to individual patient needs. However, despite much advancement in these devices, most ventilators used today are still mainly open-loop controlled devices and their added features have to some extent contributed to their complexity. The clinicians are required to make many important selections among the wide range of options available in advanced mechanical ventilators. Optimal adjustment of these machines oftentimes requires in depth knowledge about the ventilator along-with thorough review of the patient's status and his/her underlying illness. These adjustments are particularly cumbersome and frequent in more fragile and less medically stable patients.

[0006] There have been many attempts in the past to automatically control some of the main outputs of mechanical ventilators. See Y. Mitamura et al., “A dual control system for assisting respiration,” Medical and Biological Engineering, vol. 13, no. 6, pages 846-854, 1975, Yu et al., “Improvement in arterial oxygen control using multiple model adaptive control procedures,” IEEE Transactions on Biomedical Engineering, BME-34(8), pages 567-574, 1987, and U.S. Pat. No. 4,986,268 to F. T. Tehrani, issued Jan. 22, 1991, entitled “Method and apparatus for controlling an artificial respirator.”

[0007] Also, see U.S. Pat. No. 5,103,814 to T. Maher, issued Apr. 14, 1992, entitled “Self-compensating patient respirator,” Morozoff P. E., and Evans R. W., “Closed-loop control of S_aO2 in the neonate,” Biomedical Instrumentation and Technology, vol. 26, pages 117-123, 1992, U.S. Pat. No. 5,365,922 to D. B. Raemer issued Nov. 22, 1994 entitled “Closed-loop non-invasive oxygen saturation control system,” Tehrani et al. “Closed-loop control of the inspired fraction of oxygen in mechanical ventilation,” Journal of Clinical Monitoring and Computing, vol. 17, No. 6, pages 367-376, 2002, and U.S. Pat. No. 6,671,529 to N. R. Claure et al., issued Dec. 30, 2003, entitled “System and method for closed-loop controlled inspired oxygen concentration.”

[0008] Some of the prior art on this subject is focused on controlling the patient's oxygenation, and some is intended to automatically control the breathing frequency and tidal volume. The systems intended for controlling only the oxygen level of the patient on the ventilator, either do not provide the automation of all factors that affect oxygenation and/or they do not provide a reliable and sufficiently robust response against oxygen disturbances.

[0009] In addition to advancement in mechanical ventilators, there have been many attempts in recent years to prevent the collapse of the airways and apnea in spontaneously breathing patients specially during sleep, by using less elaborate machines than mechanical ventilators, generally known as CPAP machines (CPAP stands for Continuous Positive Airway Pressure). In these machines, either a constant pressure is applied to the patient's airways throughout respiration (i.e. CPAP), or a combination of CPAP and pressure support in inspiration is used to ventilate the patient (e.g. bilevel CPAP machines). See U.S. Pat. No. 4,773,411 to J. B. Downs issued Sep. 27, 1988, entitled “Method and apparatus for ventilatory therapy,” International Patent Publication No. WO 99/61088 to Resmed Limited, issued Dec. 2, 1999, entitled “Ventilatory assistance for treatment of cardiac failure and Cheyne-Stokes breathing,” U.S. Pat. No. 6,539,940 to R. J. Zdrojkowski et al., issued Apr. 1, 2003, entitled “Breathing gas delivery method and apparatus,” and U.S. Pat. No. 6,752,151 to P. D. Hill, issued Jun. 22, 2004, entitled “Method and apparatus for providing variable positive airway pressure.”

[0010] In one embodiment, the present invention describes a method and apparatus that can reliably and robustly control PEEP (or CPAP), and F_IO2. These are novel features which significantly improve the oxygenation of patients during ventilatory therapy provided by mechanical ventilators as well as respiratory devices such as CPAP machines.

[0011] Furthermore, in a more elaborate embodiment of the invention, in addition to PEEP (or CPAP) and F_IO2, the I:E ratio of the patient can be automatically adjusted and by further inclusion of the features of U.S. Pat. No. 4,986,268, the breathing frequency, and tidal volume can be automatically controlled in mechanical ventilation. Application of these features results in a significantly more effective and optimal treatment to the patient based on his/her conditions and requirements, in total or assist ventilatory therapy.

[0012] A method and apparatus for controlling a ventilator includes first means receiving at least input data indicative of the patient's measured oxygen levels, and in a more elaborate embodiment of the invention, the first means also receives respiratory mechanics and/or pressure-volume data, as well as data indicative of measured carbon dioxide levels of the patient. The first means which preferably comprises a programmable microprocessor, is controlled by a software algorithm to operate on the input data, and to provide digital output data to control the ventilator and the gas mixer of the ventilator. The software algorithm is divided into two control programs. One control program which can either be used by itself or along with the other program, is designed to automatically adjust F_IO2 and PEEP (or CPAP), based on at least the measured oxygen levels of the patient. The control program also operates on data from a pressure volume (PV) monitor/analyzer to set the initial PEEP value in certain groups of respiratory patients. The processing means detects hazardous conditions based on the input data and/or artifacts, replaces and/or corrects the measurement artifacts, and instructs generation of appropriate warning signals. The other control program, most of which is described in U.S. Pat. No. 4,986,268, is designed to control the frequency and ventilation for a next breath of the patient on the ventilator based on at least data indicative of measured CO₂ and O₂ levels of the patient, barometric pressure (as a reference pressure), and respiratory elastance and airway resistance (respiratory mechanics) data; and to make necessary adjustments in the I:E ratio based on the patient's respiratory mechanics data. The output data from the 1^st means indicative of PEEP (or CPAP), F_IO2, the adjustment in the I:E ratio, breathing frequency, and ventilation, and status of alarms are transmitted to a Signal Generator which is equipped with converters and/or other electronic components to generate the control and appropriate warning signals. The control signals for the breathing frequency, ventilation, PEEP (or CPAP), and the adjustment in the I:E ratio are supplied to the ventilator. The control signal for F_IO2 is supplied to a mixer regulator unit which adjusts the concentration of oxygen added to the inhalation gas in the gas mixer of the ventilator. Based on the instructions from the 1^st means, the alarm circuit generates appropriate warning signals when needed.

[0013] FIGS. 1-4 illustrate a preferred embodiment of the present invention. However, it is understood that this invention is not limited to the precise arrangements shown in the figures and can be embodied in other arrangements without deviating from the scope of the invention.

[0014] FIG. 1 is a block diagram of a mechanical ventilator and the control apparatus according to an alternative embodiment of the invention.

[0015] FIGS. 2a-2c show the flow chart of a software algorithm that also incorporates the control technique described in U.S. Pat. No. 4,986,268, to automatically control breathing frequency, tidal volume, and the adjustment in the I:E ratio of the patient on the ventilator, according to a preferred method of the present invention.

[0016] FIGS. 3a-3i show the flow chart of a software algorithm to automatically control PEEP (or CPAP) and F_IO2 according to a preferred method of the present invention.

[0017] FIG. 4 shows a preferred detailed schematic diagram of a Signal Generator and an Alarm Circuit, for use in a preferred practice of the present invention.

Definitions

[0018] In the specification and claims:

[0019] 1—The term “ventilator” refers to a device which is used to provide total or assist ventilatory treatment to patients, and includes mechanical ventilators (i.e. artificial respirators) or CPAP (Continuous Positive Airway Pressure) machines.

[0020] 2—The term “PEEP” represents “Positive End-Expiratory Pressure” and is interchangeable with the term “CPAP,” which represents “Continuous Positive Airway Pressure,” for example, when assist ventilation is provided to spontaneously breathing subjects.

[0021] 3—The term “F_IO2” represents “concentration of oxygen in a patient's inspiratory gas” which is the same as “fraction of inspired oxygen.”

[0022] 4—The term I:E represents the “ratio of inspiration time to expiration time.”

[0023] FIG. 1 shows a block diagram according to an alternative practice of the present invention. The digital processor 10 includes a programmable controller coupled to receive the outputs of 8 bit A/D converters 12, 14 and 16 as shown. The A/D converters 18 and 20 are each a single 8 bit A/D converter. The A/D converter unit 22 is an A/D board containing three 8 bit A/D converters. The inputs 24, 26, and 28 of the A/Ds are from an oxygen sensor, preferably a pulse oximeter, 30, a CO₂ sensor, such as a transcutaneous monitor or preferably a capnograph, 32, and a lung mechanics calculator and PV monitor, 34. The outputs 24, and 26 are each a single analog signal while the output 28 represents 3 analog signals; 1—representing respiratory elastance, 2—representing respiratory airway resistance (air viscosity factor in the lungs), and 3—representing the lower inflection point on the inspiratory or expiratory PV curve of the patient, or alternatively, the measured intrinsic PEEP (PEEPi) of the patient on the ventilator. The inputs to the oxygen sensor and the carbon dioxide sensor are respectively shown at 40 and 42 coming from the patient. The input 40 is preferably the arterial hemoglobin oxygen saturation data and the input to the CO₂ sensor shown at 42 is preferably the exhaled gas from the patient from which the end-tidal CO₂ concentration or the end-tidal partial pressure of CO₂ is determined by the sensor. The lung mechanics calculator and PV monitor, 34, receives data from the mechanical ventilator shown at 56, or from the patient through the ventilator circuit, on the line illustrated at 36 and communicates back to the ventilator as shown at 38. The digital processor's outputs shown at 44 are applied to a Signal Generator Circuit, illustrated at 46. The Signal Generator Circuit sends alarm instruction signals 52 to the alarm circuit 54.

[0024] The mechanical ventilator 56 receives the control signals 48 from the Signal Generator Circuit 46. These consist of signals to control PEEP, breathing frequency, tidal volume, and the adjustment in the I:E ratio of the patient. A Mixer Regulator circuit 58, receives control signals to adjust F_IO2, 50, from the Signal Generator Circuit 46. An oxygen air mixer 62 receives the adjusted output signal 60 from the Mixer Regulator 58. The concentration of oxygen in the mixer is thereby adjusted by mixing the determined concentration of oxygen 66 coming from the oxygen supply 70 and that of air 64 coming from the air compressor 68. The enriched oxygenated air 72 from the mixer is provided to the ventilator 56 which delivers it to the patient at 74.

[0025] Referring to FIG. 2a-2c, there is illustrated a flow chart of the algorithm to control the breathing frequency, ventilation, and the adjustment in the I:E ratio in an alternative embodiment of the invention. As seen at the start of the flow chart, the initial values of breathing frequency and tidal volume are transmitted to the output ports in step 100. Then the main loop at A is started and in the next step at 102, based on data indicative of CO₂ and O₂ levels of the patient which are preferably provided by a capnograph and a pulse oximeter respectively, the arterial partial pressures of CO₂ and O₂ are calculated by using the following equations:
P_aCO2=P_etCO2+K₁ $P_{{\mathrm{aO}}_2}=\frac{-\ln \left[1-{\left(S_{{\mathrm{pO}}_2}\right)}^{0.5}\right]}{0.046}+\mathrm{CP}$

[0026] Where P_aCO2 and P_aO2 are arterial partial pressures of CO₂ and O₂ respectively, P_etCO2 is the end-tidal partial pressure of CO₂ measured by the CO₂ sensor, and K₁ is the difference between the arterial partial pressure of CO₂ and the end-tidal partial pressure of CO₂. K₁ can be measured in advance and depending on the patient's conditions, it can be adjusted to set the desired P_aCO2 of the patient. S_pO2 is the arterial hemoglobin oxygen saturation of the patient measured by a pulse oximeter and CP is an added correction factor which is used to correct and shift P_aO2 based on the patient's measured blood pH level. If the patient's blood pH level is in the 7.45-7.55 range, CP is set to zero. Otherwise, CP needs to be adjusted by +3.5 mm Hg per every −0.1 deviation in pH from the above range. After the calculation of P_aCO2 and P_aO2, their values are compared to defined minimum acceptable levels to determine whether there has been any measurement artifact in step 104. If any artifact is detected, the calculated value is discarded and the previous calculated value is resumed. In the next step at 106, if P_aCO2 and/or P_aO2 are not within certain defined ranges, alarms are transmitted to the output ports. In the step that follows at 108, if the calculated P_aCO2 and P_aO2 values are both lower than their minimum threshold limits (which are different from the minimum acceptable values used in step 104), the possibility of pulmonary embolism is assumed, predefined levels of ventilation and breathing frequency are provided, and an alarm is generated in steps 110 and 112, and the program returns to A. However, if the calculated P_aCO2 and P_aO2 values are not found to be simultaneously lower than their minimum threshold levels in 108, then the effect of CO₂ on the required ventilation is calculated in step 114:
V_C=C₁.P_aCO2−C₂

[0027] Where V_C is the ratio of alveolar ventilation as the net effect of CO₂ to the resting value of ventilation, C₁ is the sensitivity factor of the controller to CO₂ (e.g., C₁=0.405) and C₂ is a constant (e.g., C₂=14.88).

[0028] Next, in step 116, the P_aO2 value is compared to a high threshold limit of 104 mm Hg. If P_aO2 is greater than or equal to this threshold value, the effect of oxygen on ventilation is set to zero in 118, and the next step at 122 is followed. Otherwise, if P_aO2 is found to be less than the threshold value in step 116, then control is passed to step 120 in which the effect of oxygen on the required ventilation is calculated by using the following equation:
V_O=(4.72×10⁻⁹)(104−P_aO2)^4.9

[0029] Where V_O is the ratio of alveolar ventilation as the net effect of oxygen to the resting value of ventilation. It is recognized that the above equations are based on the use of a capnograph and a pulse oximeter to measure the carbon dioxide and oxygen levels of the patient respectively. If other measurement techniques are utilized to provide data indicative of said levels, then other alternative equations may be used to determine the required ventilation for the patient, without deviating from the scope and the essential attributes of the invention.

[0030] In the next step at 122, the effect of increase in the metabolic rate ratio, MRR, (i.e. rate of metabolism/basal rate of metabolism), on ventilation is calculated by using the following equation:
V_M=0.988(MRR−1)

[0031] Where V_M is the ratio of alveolar ventilation as the net effect of increase in the metabolic rate ratio, MRR, to the resting value of ventilation, and MRR is an input to the algorithm. In the next step at 124, total alveolar ventilation for the next breath is calculated:
V_A=(V_A at rest)(V_C+V_O+V_M)

[0032] Where V_A is alveolar ventilation in liters/minute and V_A at rest is the alveolar ventilation at rest which is input and stored in the software. In the next step at 126, the physiological dead space of the patient, and the total dead space including that of the equipment are calculated, if not provided in advance, as follows:
V_D=(0.1698V_A/60)+0.1587
V_Dt=V_D+V_ED

[0033] In these equations, V_D is the patient's dead space in liters, V_ED is the equipment dead space due to the tubes and connections, and V_Dt is the total dead space. It should be noted that the constant factors in these equations are based on measured experimental values for adults and can therefore be different for individual patients. Also, for other patient populations, they need to be adjusted. For example the constant factor of 0.1587 should change to a much smaller value for infants (e.g., 2.28×10⁻³). In the next step at 128, data indicative of barometric pressure and the patient's airway resistance (or the air viscosity factor in the lungs) and respiratory elastance are read from the input ports. The barometric pressure data which is affected mostly by the altitude, is used as a reference pressure (for the purpose of calibration) in the invention.

[0034] In the next step at 130, the optimal frequency for the next breath is computed. This calculation is based on minimization of the respiratory work rate and is done in order to stimulate natural breathing, provide a more comfortable breathing pattern to the patient, and thereby, expedite the weaning process in assisted ventilation. The following equation, which is a modified version of an equation derived in 1950 by Otis et al. to describe the control of breathing frequency in mammals, is used to calculate the optimal breathing frequency in the invention: $f=\frac{-K^{\prime }{V}_D+\sqrt{{\left(K^{\prime }{V}_D\right)}^2+4K^{\prime }{K}^{″}{\Pi }^2{V}_{\mathrm{AR}}{V}_D}}{2K^{″}{\Pi }^2{V}_D}$

• • [0035] where f is the optimum breathing frequency in breaths/second, V_AR is the alveolar ventilation in liters/second and is equal to V_A/60, K′ is the respiratory elastance (reciprocal of respiratory compliance) in cm H₂O/liter and K″ is the airway resistance in cm H₂O/liter/second. Next in step 132, the required minute ventilation and tidal volume are calculated:
    V_E=V_A+60fV_Dt
    V_T=V_A/60f+V_Dt

[0036] Where V_E represents total minute ventilation in liters/minute and V_T is tidal volume in liters. In the next step at 134, additional safety rules are applied. If breathing frequency, f, tidal volume, V_T, or minute ventilation are not within prescribed safe ranges, their values are limited and adjusted.

[0037] In the next step 136 which follows, the breathing frequency is compared with an upper limit value F_max. This upper limit frequency is defined as:
F_max=1/5τ

[0038] Where τ is the respiratory time constant and is equal to K″/K′. If in step 136, the breathing frequency is found to be higher than F_max, then in the next step at 138, its value is reduced to F_max (in which case V_T is also adjusted according to procedures in steps 132 and 134), and step 140 is followed. Otherwise, if the computed breathing frequency is less than or equal to F_max, it does not need further adjustment and the program is transferred to step 140. In step 140, the expiration time, T_E, is compared to 2.5 times τ. If it is found to be less than 2.5 τ, then step 142 is followed and the I:E ratio (the ratio of the inspiratory time to the expiratory time) is adjusted, so that T_E becomes at least equal to 2.5 τ. Otherwise, if T_E is found to be greater than or equal to 2.5 τ in step 140, it does not need to be adjusted (i.e. the adjustment value is zero) and the program is transferred to step 144. The reason for the adjustments in the breathing frequency and T_E in steps 138 and 142 mentioned above, is to provide sufficient time for exhalation based on the patient's respiratory time constant and to avoid build up of intrinsic positive end-expiratory pressure (PEEPi).

[0039] In step 144 that follows, the calculated values for ventilation, breathing frequency, and the adjustment in the I:E ratio for the next breath are provided to the output ports. At this point, if the ventilator is in the pressure control/assist mode, the inspiratory pressure is calculated by using the following equation:
P_m=K′V_T+PEEP

• • [0040] where P_m is the inspiratory pressure in cm H₂O. Thereafter, the control data indicative of P_m is also provided to an output port and the routine is held for the duration of the next breathing cycle. After the delay is passed, the program returns to the beginning of the loop at A.

[0041] It should be noted that the major portion of the procedure depicted in FIG. 2 to calculate the optimal breathing frequency and tidal volume of the breaths of a patient and controlling them automatically, has been described in U.S. Pat. No. 4,986,268. In the present invention, the necessary adjustments in the I:E ratio are controlled automatically as described above, and the levels of F_IO2 and PEEP are automatically controlled by another algorithm which is described next.

[0042] Referring to FIGS. 3a-3i, there is illustrated a flow chart of a control algorithm which is operated upon by the digital processor. This algorithm is either run by itself, or in an alternative embodiment of the invention, it is run in parallel to the algorithm of FIGS. 2a-2c described above. The purpose of this algorithm is to automatically control the levels of F_IO2 and PEEP provided to the patient on the ventilator and thereby improve the patient's oxygenation. The method depicted in FIGS. 3a-3i can be used for patients on mechanical ventilation or those on respiratory assist devices receiving CPAP treatment. Depending on the type of the ventilatory treatment, the term PEEP in the flow chart is meant to be interchangeable with CPAP.

[0043] As is seen, at the start of the flow chart, the desired set point for arterial partial pressure of oxygen of the patient is defined in step 200. This is done on the basis of the patient's conditions and his/her underlying illness. Then in the next step at 202, the initial value of F_IO2 is set and transmitted to the output port.

[0044] In step 204 that follows next, the initial value of PEEP is set and transmitted to an output port. The initial value of PEEP can be set by using different options. For certain patient groups such as COPD patients, the initial PEEP can be chosen to be 80% to 85% of the intrinsic PEEP (PEEPi) which needs to be measured in advance. For some other patient groups such as ARDS patients, the initial PEEP setting can be chosen to be 3-4 cm H₂O above the lower inflection pressure point of the inspiratory (or the expiratory) pressure volume curve of the patient. This value can either be calculated by the lung mechanics calculator and PV monitor unit and provided automatically to the digital processor via an input port, or the calculated value of the pressure can be provided manually by the clinician either through one of the input ports or via software. The third option is that the clinician arbitrarily decides an initial setting for PEEP and provides it to the digital processor, preferably via software. After setting the initial PEEP value in 204, the next step in 206 is followed. At this point, a time parameter (e.g., TP) for PEEP adjustment is defined and initially set to zero. The purpose of defining this parameter is to guarantee that PEEP adjustments are done only after a certain time has elapsed since the latest adjustment, thereby giving enough time to an adjustment in PEEP to make an impact on the patient's oxygenation.

[0045] In step 208 which follows next, another parameter, AP, for PEEP adjustment is defined. If this parameter is set to zero, then PEEP is controlled manually and only F_IO2 is automatically adjusted. If AP is set to one, then both F_IO2 and PEEP are automatically controlled.

[0046] In the next step 210, the threshold values for arterial hemoglobin oxygen saturation, S_pO2, (or alternatively for arterial partial pressure of oxygen) are defined. In a preferred practice of the invention, four threshold values are defined for S_pO2 and they are set at 90%, 93%, 95%, and 97% respectively. However, the threshold values may differ for different patients. They should be defined based on the patient's conditions and the desired levels of oxygenation.

[0047] Next, program control passes to step 212 in which a loop indicator (e.g., LI) is defined and is set to 1.5, and the main loop starts at A′.

[0048] In the next step in 214, the patient's S_pO2 data is read from one of the input ports, and in step 216, the arterial partial pressure of oxygen is calculated from the S_pO2 data as: $P_{{\mathrm{aO}}_2}=\frac{-\ln \left[1-{\left(S_{{\mathrm{pO}}_2}\right)}^{0.5}\right]}{0.046}+\mathrm{CP}$

[0049] Where P_aO2 is the arterial partial pressure of oxygen, and CP is an added correction factor which is used to shift P_aO2 based on the patient's measured blood pH level. If the patient's blood pH is within 7.45-7.55, then CP is set to zero. Otherwise, for every +0.1 deviation in pH from this range, CP is adjusted by −3.5 mm Hg as was also mentioned in the description of FIG. 2 earlier.

[0050] In step 218 that follows next, the calculated partial pressure of oxygen, P_aO2, is compared with a minimum acceptable value. This is done to detect artifacts in the measurement of S_pO2. If the calculated P_aO2 is found to be less than the minimum acceptable value, then control passes to step 220 in which an artifact is assumed and an alarm is generated. Then step 222 is performed in which the S_pO2 data is discarded and the previous value Of P_aO2 in the memory is resumed and step 224 is followed. However, if in 218, the calculated P_aO2 is found to be greater than or equal to the minimum acceptable value, its value is accepted and control passes to step 224.

[0051] In step 224, S_pO2 is compared to a minimum safe value, which is the first threshold value defined previously in step 210 (e.g., 90%). If S_pO2 is less than or equal to the minimum safe value, loop B is started in 226 and the loop indicator, LI, is set to 2.5. Then in step 228, F_IO2 is increased stepwise (i.e. in a step-like arrangement) to a high value, F1, (e.g., 60%), and an alarm is generated in 230. Control then passes to loop F at which the procedure of PEEP adjustment begins as will be described later. However, if S_pO2 is found to be higher than the minimum safe value in step 224, control passes to 232 where S_pO2 is compared to a second threshold value (e.g., 93%). If S_pO2 is less than the second threshold value, then steps 234 and 236 are followed in which the loop indicator, LI, is examined and compared to 2. If LI is less than 2, control passes to another loop E which will be described later. If LI is greater than or equal to 2, the next step in 238 is performed in which LI is compared to 3. If LI is less than 3, control passes to loop B (where F_IO2 was set high at F1, e.g., 60%), otherwise, the program transfers to step 240. In this step, LI is compared to 4. If it is less than 4, control passes to loop C; otherwise, the program transfers to loop D (loops C and D will be described later).

[0052] Back to step 232, if S_pO2 is found to be higher than or equal to the 2^nd threshold value (e.g., 93%), then steps 242 and 244 are followed in which LI is compared to 2. If it is less than 2, control passes to loop E. Otherwise, in the next step at 246, LI is compared to 3. If less than 3, loop C is defined and started at 248, and LI is set to 3.5. Then in step 250, F_IO2 is set stepwise at a moderately high value, F2 (e.g., 45%), and control transfers to loop F in which the procedure of PEEP adjustment is followed. However, if in step 246, LI is found to be greater than or equal to 3, control passes to step 252 in which LI is compared to 4. If LI is less than 4, then S_pO2 is compared to a third threshold value (e.g., 95%) in step 254. If S_pO2 is less than the third threshold value, control passes to loop C in which F_IO2 was set at a moderately high level, F2 (e.g., 45%). Otherwise, if S_pO2 is found to be higher than or equal to the third threshold value in 254, then the next step in 256 is followed in which loop D is defined and started and LI is set to 4.5. Next in step 258, F_IO2 is set stepwise at a slightly high level, F3 (e.g., 30%), and control passes to loop F.

[0053] Back to step 252, if LI is found to be greater than or equal to 4, then S_pO2 is compared to a 4^th threshold value (e.g., 97%) in step 260. If S_pO2 is less than the 4^th threshold value, control passes to loop D in which F_IO2 was set at a slightly high value, F3 (e.g., 30%). Otherwise, if S_pO2 is higher than or equal to the 4^th threshold value in 260, then loop E is started in 262 and LI is set to 1.5. In loop E, a proportional, integral, derivative (PID) control procedure is performed to adjust F_IO2 (PID control is a control technique comprising proportional, integral, and derivative terms). In the next step at 264, using the P_aO2 set point defined in step 200, the proportional, differential, and integral components of error are calculated as follows:
Y₁(k)=P_aO2(set-point)−P_aO2
Y₂(k)=[Y₁(k)−Y₁(k−1)]/T
Y₃(k)=Y₃(k−1)+TY₁(k)

[0054] In the above equations, Y₁(k), Y₂(k), and Y₃(k) represent the proportional, differential, and integral components of error in P_aO2 respectively, and T is a sampling interval.

[0055] In step 266 that follows, the required F_IO2 is calculated by using the following equations:
E(k)=αY₁(k)+βY₃(k)+γY₂(k)
G(k)=E(k)+0.21

[0056] Where E(k) is an error function, α, β, and γ are the PID coefficients, and G(k) is the required F_IO2. In a preferred practice of the invention, T is set to 0.75 seconds, and α, β, and γ are set to 6.45×10⁻⁵, 3.22×10⁻⁵, and 7.29×10⁻⁶ respectively. These parameters were tuned to minimize steady-state oscillations and to keep the overshoot/undershoot in the F_IO2 response of the PID controller below 25% of the total change. It is also recognized that other error correction schemes can be used to determine F_IO2. As long as those schemes reduce the error in the oxygen level of the patient in a similar way as described above, they will be within the scope of the present invention.

[0057] In the next step in 268, the calculated value of F_IO2 is compared with a minimum of 0.21 (i.e. 21%). If the F_IO2 value is less than 21%, in step 270 which follows, it is set to a minimum of 21% and control passes to loop F. However, if in 268, F_IO2 is found to be greater than or equal to 21%, control passes to step 272 in which F_IO2 is compared to a maximum allowed value (e.g., 80%). If F_IO2 is less than or equal to the maximum allowed value, the next step in 274 is followed where the calculated value of F_IO2 is sent to the output port and control passes to step 276. In this step F_IO2 is compared to 60%. If it is less than 60%, control passes to loop F. Otherwise, an alarm is generated in 278 and then control transfers to loop F.

[0058] Back to step 272, if the calculated value of F_IO2 is found to be higher than the maximum allowed value, it is reduced to the maximum value in step 280, an alarm is generated, and then control transfers to loop F.

[0059] Up to the beginning of loop F at step 282, the focus of control is on automatic control of F_IO2. As shown, two different mechanisms are incorporated in the control process of F_IO2 in a preferred practice of the invention. One, a rapid stepwise control scheme which responds instantly to fast declines in S_pO2, and the other, a more finely controlled PID algorithm that provides fine control of F_IO2 in the absence of sharp and hazardous declines in S_pO2. The stepwise controller in a preferred practice of the invention has three loops, each with its defined minimum and maximum S_pO2 threshold levels. These three loops were shown respectively at B, C, and D, and the PID control loop was shown at E in the flow chart of FIG. 3. The controller switches from the PID control to the rapid stepwise algorithm only if rapid declines in S_pO2 are detected. Once in the stepwise mode, the controller continuously checks S_pO2, and if it rises, the controller reduces F_IO2 to minimize the exposure of the patient to high and toxic levels of F_IO2. The controller is designed to correct hypoxemia within seconds and to avoid hyperoxemia. As shown, the controller detects artifacts in the measurement of S_pO2, discards the artifacts, and generates alarms when appropriate. The algorithm also enables clinicians to define the desired oxygenation levels for different patients. This is done by defining an appropriate P_aO2 set point, by setting the threshold values for S_pO2, and by adjusting the correction parameter, CP, in accordance with the measured pH levels in the patient's blood as described above.

[0060] After the determination of the required F_IO2, the procedure of adjusting the PEEP value is started at F in step 282. In this step, the ratio of PEEP/F_IO2 is calculated. Then in 284, the control parameter AP, which was defined in step 208, is examined. If it is less than 1, it means that PEEP is not adjusted automatically and it is instead adjusted manually by the operator. In this case, the controller merely watches the PEEP/F_IO2 ratio and generates warning signals, if the ratio is either too low or too high. In step 286, the ratio is compared to a minimum allowed value (e.g., 0.12). If it is less than the minimum value, an alarm is generated in 288 and control passes to I (which will be described later). However, if the PEEP/F_IO2 ratio is found to be equal to or greater than the minimum value in step 286, then the next step in 290 is performed where the ratio is compared to a maximum allowed value (e.g., 0.22). If the ratio is less than or equal to the maximum value, control passes to I. Otherwise, an alarm is generated in step 292 and then control is transferred to I.

[0061] Back to step 284, if AP is not less than 1, it means that PEEP should be calculated and automatically adjusted. Therefore, the automatic PEEP adjustment control loop is started next at G at step 294. In the step 296 that follows, the PEEP/F_IO2 ratio is compared to a minimum allowed value (e.g., 0.12). If it is less than the minimum, the procedure at H is started and it is examined how long ago the last adjustment in PEEP was made. In step 300 that follows, the time parameter, TP, is compared to a defined fixed interval, T1, for example 240 seconds. If TP is less than 240 seconds, it means that the last PEEP adjustment was made less than 4 minutes ago. Then the procedure at J is started. Control passes to step 302 in which no change is made in PEEP and the time parameter, TP, is increased by a fixed amount (e.g., 0.75 seconds):
TP_(new)=TP_(old)+0.75

[0062] Thereafter, control passes to I. However, if in step 300, it is found that TP is equal to or greater than 240 seconds, it means that the last adjustment in PEEP was made at least 4 minutes ago or longer. Therefore, control passes to step 304. In this step, TP is set back to zero. Then in 306 that follows, PEEP is increased by a fixed amount (e.g., 2 cm H₂O):
PEEP_(new)=PEEP_(old)+2 cm H₂O

• • [0063] Thereafter, control passes to I.

[0064] Back to step 296, if the PEEP/F_IO2 ratio is not found to be less than the minimum allowed value, control transfers to step 308. In this step the PEEP/F_IO2 ratio is compared to a maximum allowed value (e.g., 0.22). If the ratio is not less than the maximum value, step 310 is next performed. At this point, the PEEP/F_IO2 ratio is compared to a slightly higher value than the maximum, RG, (e.g., 0.24). If it is not greater than this value, control passes to J. Otherwise; step 312 is performed in which the time parameter, TP, is compared to the fixed interval of 240 seconds. If TP is less than 240 seconds, control passes to J. Otherwise; TP is set back to zero in step 314, and PEEP is reduced by a fixed amount (e.g., 2 cm H₂O) in step 316:
PEEP_(new)=PEEP_(old)−2 cm H₂O

[0065] Thereafter, control passes to I. In step 318 at I, the routine is held for a fixed interval (e.g., 0.75 seconds) and then control returns to the beginning of the main loop at A′.

[0066] Back to step 308, if the PEEP/F_IO2 ratio is found to be less than the maximum allowed limit (e.g., 0.22), the step 320 is next followed. In this step S_pO2 is compared to a predefined minimum allowed value (e.g., 92%). If it is higher than or at least equal to the predefined minimum value, the PEEP level is not changed and control passes to J. However, if in 320, S_pO2 is found to be less than the predefined minimum value, then control passes to H, where it is determined whether at least 4 minutes have passed since the last PEEP adjustment, and if so, PEEP is increased by a fixed amount (e.g., 2 cm H₂—O) as was shown earlier.

[0067] In performing the automatic PEEP adjustments, the PEEP/F_IO2 is kept within a clinically acceptable range. As shown above, if the PEEP/F_IO2 is too low, PEEP is increased by a fixed increment (e.g., 2 cm H₂O). Also, if the PEEP/F_IO2 ratio is within the acceptable range and S_pO2 is low, then PEEP is increased by a fixed increment (e.g., 2 cm H₂O) to improve patient's oxygenation. On the other hand, if the PEEP/F_IO2 ratio increases beyond a maximum defined value, the program reduces PEEP in fixed amounts (e.g., 2 cm H₂O). In any case, the interval between two successive PEEP adjustments is at least equal to a fixed period (e.g., 240 seconds), to allow for the changes in PEEP to have an observable and measurable impact on the patient's oxygenation.

[0068] It should be noted that the above examples for the incremental step size for PEEP adjustment (e.g. 2 cm H₂O) and the minimum and maximum values for the ratio of PEEP/F_IO2, are indicated for patients receiving ventilatory treatment in a more acute clinical setting such as the intensive care or a constant care unit of a hospital. Smaller incremental adjustments (e.g. 1 cm H₂O) and more conservative ranges for the ratio of PEEP (or CPAP)/F_IO2 may be adopted if the invention is used to improve the breathing and oxygenation of more stable, spontaneously breathing patients.

[0069] FIG. 4 illustrates in detail, a preferred circuit diagram of the Signal Generator Circuit, 46, and the alarm circuit 54. The preferred component types and values are shown in the chart below:

	Component	Type/Value
	IC1	DAC0802LCN
	IC2	LM741CN
	IC3	SN7400N
	C1	0.1	μF
	C2	0.03	μF
	D1	1N4148
	R1	5.1	kΩ
	R2	50	kΩ pot
	R3	10	kΩ pot
	R4	2.7	kΩ
	R5	330	Ω

[0070] There has been described a method and apparatus for controlling a ventilator. The invention utilizes data indicative of measured oxygen levels of the patient to automatically control F_IO2, and PEEP (or CPAP). In an alternative embodiment, the invention further uses the respiratory mechanics data (i.e. respiratory elastance and airway resistance) to automatically make the necessary adjustments in the I:E ratio of the patient on the ventilator. It further incorporates the features of U.S. Pat. No. 4,986,268 and uses data indicative of measured levels of oxygen and the respiratory mechanics data of the patient, along with data indicative of barometric pressure (as a reference calibrating pressure), and data indicative of measured carbon dioxide level of the patient to automatically control the breathing frequency and tidal volume of breaths of the patient on the ventilator. The invention also detects and corrects artifacts in the measured oxygen and carbon dioxide data and applies safety rules. In its different embodiments, the invention can improve total and/or assist ventilatory treatments provided to different patient groups.

[0071] The present invention may be embodied in other specific forms without departing from the scope and the essential attributes thereof. Therefore, reference should be made to the appended claims rather than to the foregoing specification, with regard to the scope of the invention.