Method for oscillometric blood pressure determination employing curve fitting

• US 5,704,362 A
• Filed: 08/13/1993
• Issued: 01/06/1998
• Est. Priority Date: 08/13/1993
• Status: Expired due to Term

• Current Assignee: Critikon Co. LLC
• Original Assignee: Johnson & Johnson Medical Incorporated (Johnson & Johnson)
• Inventors: Hersh, Lawrence T., Medero, Richard, Friedman, Bruce
• Primary Examiner(s): NASSER, ROBERT L
• Application Number:
  US08/106,280
• Time in Patent Office: 1,607 Days
• Field of Search: 128/677-686, 364/413.03
• US Class Current: 600/486
• CPC Class Codes: G06F 18/00 Pattern recognition
• CPC Class Codes: G06F 2218/12 Classification; Matching

Abstract

A method of measuring blood pressure where a plurality of oscillometric data values are measured from a subject at a plurality pressure levels, a function curve having the shape of an oscillometric envelope is selected and computational model fit to the data, using the Gauss-Marquardt method of model fitting. Once the curve is model fit, blood pressure is measured using the oscillometric method. Included in the step of model fitting is the step of computing the parameter of the function curve that minimizes the error between the function curve and the data values at the respective pressure levels thereby approximating the oscillometric envelope based upon data values.

Claims

1. A method for measuring the blood pressure of a subject, comprising the steps of:

obtaining from the subject a plurality of oscillometric data values from at least one complex amplitude taken at a plurality of pressure levels;

selecting a function curve having a physiologically-expected shape of an oscillometric envelope, the function curve being defined by at least one parameter;

computationally model fitting the function curve to the plurality of data values by computing the parameter of the function curve that minimizes the error between the function curve and the data values at the respective pressure levels thereby approximating the oscillometric envelope based upon data values; and

computing a systolic pressure and a diastolic pressure based upon the curve.

2. The method as set forth in claim 1, wherein the step of computationally model fitting a function curve to the plurality of data values further comprises the step of storing computations of at least one previous determination for use in a subsequent determination.

3. The method as set forth in claim 1, wherein the step of computationally model fitting a function curve to the plurality of data values comprises the step of computationally model fitting a Gaussian curve to the plurality of data values.

4. The method as set forth in claim 3, wherein the step of computationally model fitting a Gaussian curve to the plurality of data values employs the Marquardt method.

5. The method as set forth in claim 4, wherein the step of computationally model fitting a Gaussian curve to the plurality of data values by the Marquardt method is over Gaussian parameters of amplitude, mean, and deviation.

6. The method as set forth in claim 5, wherein the step of computationally model fitting a Gaussian curve to the plurality of data values by the Marquardt method over the Gaussian parameters of amplitude, mean and deviation further comprises the step of storing the Gaussian parameters of at least one previous determination for use in a subsequent determination.

7. The method as set forth in claim 6, wherein the step of computationally model fitting a Gaussian curve to the plurality of data values by the Marquardt method over the Gaussian parameters of amplitude, mean and deviation and storing the Gaussian parameters of at least one previous determination for use in a subsequent determination comprises the steps of storing the Gaussian parameters of a plurality of previous determinations and computing a weighted average of the Gaussian parameters of the previous determinations.

8. The method as set forth in claim 7, wherein the step of computing a weighted average of the Gaussian parameters of the previous determinations includes the step of weighing the Gaussian parameters of the more recent determinations heavier than older determinations.

9. The method as set forth in claim 6, wherein the step of computationally model fitting a Gaussian curve to the plurality of data values by the Marquardt method over the Gaussian parameters of amplitude, mean and deviation and storing the Gaussian parameters of at least one previous determination for use in a subsequent determination comprises the step of predicting an expected new Gaussian parameters and then performing another blood pressure determination utilizing the expected new Gaussian parameters.

10. The method as set forth in claim 6, wherein the step of computationally model fitting a Gaussian curve to the plurality of data values by the Marquardt method over the Gaussian parameters of amplitude, mean and deviation and storing the Gaussian parameters of at least one previous determination for use in a subsequent determination comprises the steps of storing the Gaussian parameters of at least one previous determination, generating data values based upon such previous determination, and using the generated data values in a subsequent determination.

11. The method as set forth in claim 1, wherein the step of computationally model fitting a function curve to the plurality of data values such that the function curve approximates an oscillometric envelope based upon the data values comprises the step of scaling an expected envelope.

12. The method as set forth in claim 1, wherein the step of computationally model fitting a function curve to the plurality of data values such that the function curve approximates an oscillometric envelope based upon the data values comprises the step of shifting an expected envelope along the pressure axis so that the amplitudes of the expected and measured data agree, thereby accounting for a change in blood pressure level.

13. The method as set forth in claim 1, wherein the step of computationally model fitting a function curve to the plurality of data values such that the function curve approximates an oscillometric envelope based upon the data values further includes the steps of filtering the data values to reject the data values that are not close to the fitted function curve and then computationally fitting another curve to the remaining data values, from which the systolic pressure and diastolic pressure are computed.

14. The method as set forth in claim 1, wherein the step of computationally model fitting a function curve to the plurality of data values such that the function curve approximates an oscillometric envelope based upon the data values comprises the step of identifying the data values relevant to the computation of systolic pressure and the data values relevant to the computation of diastolic pressure and then computing a systolic curve to the systolic data values and computing a diastolic curve to the diastolic data values and wherein the steps of computing a systolic pressure and a diastolic pressure based upon the curve is performed relative to the systolic curve and the diastolic curve, respectively.