Method of blood pressure estimation using trend analysis

A method of blood pressure trend determination of a subject comprising:

• providing a device comprising at least one of a photoplethysmographic (PPG) sensor or a magnetic sensor to the subject, wherein the PPG/magnetic sensor uses a pulse oximeter to measure changes in skin light absorption;
• transmitting results of the PPG/magnetic sensor measurements to a processor, wherein the processor comprises a non-transitory computer readable medium having instruction stored thereon, wherein the instructions, when executed by the processor, cause the processor to:
• receive the PPG/magnetic sensor measurements over a first time interval;
• identify a second time interval within the first time interval with statistically stable measurements, wherein the statistically stable measurement is obtained by computing a z-score value on a subinterval that is compared to a prior z-score value computed from a previous subinterval, wherein the z-score:
• is a statistical measure in which an absolute value represents a distance between a raw score of a random variable, x, and a population mean, μ, in units of standard deviation, σ:
• x−μ
• z=σ
• and in order to compute a stabilized point in time when PPG measurements have stabilized, a distance between a population mean of a previous subinterval and a population mean of current subintervals is calculated as follows:
• μ
• 2
• -
• μ
• 1
• σ
• 1
• ⁢
• and
• ⁢
• μ
• 1
• -
• μ
• 2
• σ
• 2
• wherein the PPG measurements have stabilized when a distance between two successive z-score computations becomes less than a preset noise tolerance level;
• identify within the second time interval a sequence of time intervals containing a number of both systolic and diastolic cardiac periods (SCP, DCP) suitable for blood pressure trend computation, wherein the SCP are identified by calculating the time interval between PPG peaks and the DCP are identified by calculating the time interval between PPG troughs;
• apply a statistical analysis to determining the sequence of time intervals;
• eliminate a noise in the sequence of time intervals by applying an adaptive or fixed low-pass Fourier filter to each element of the sequence;
• compute a trend of systolic blood pressure by finding a slope of a best-fit line to systolic cardiac period values over time;
• compute a trend of diastolic blood pressure by finding a slope of a best-fit line to diastolic cardiac period values over time and inverting the slope sign;
• and
• combine the trend analysis results to determine the blood pressure trend over a time period;
• and
• providing an input/output device that at least one of stores, displays, or transmits blood pressure trend of the subject.

A computer implemented method for obtaining a blood pressure trend from a subject, comprising:

• providing a device comprising at least one of a photoplethismographic (PPG) sensor or a magnetic sensor to the subject, wherein the PPG/magnetic sensor uses a pulse oximeter to measure changes in skin light absorption;
• transmitting results of the PPG magnetic sensor measurements to a processor, wherein the processor comprises A non-transitory computer readable medium having instruction stored thereon, wherein the instructions, when executed by the processor, cause the processor to:
• receive the PPG/magnetic sensor measurements over a first time interval;
• identify a second time interval within the first time interval with statistically stable measurements, wherein the statistically stable measurement is obtained by computing a z-score value on a subinterval that is compared to a prior z-score value computed from a previous subinterval, wherein the z-score:
• is a statistical measure in which an absolute value represents a distance between a raw score of a random variable, x, and a population mean, μ, in units of standard deviation, σ:
• x−μ
• z=σ
• and in order to compute a stabilized point in time when PPG measurements have stabilized, a distance between a population mean of a previous subinterval and a population mean of current subintervals is calculated as follows:
• μ
• 2
• -
• μ
• 1
• σ
• 1
• ⁢
• and
• ⁢
• μ
• 1
• -
• μ
• 2
• σ
• 2
• wherein the PPG measurements have stabilized when a distance between two successive z-score computations becomes less than a preset noise tolerance level;
• identify within the second time interval a sequence of time intervals containing a number of both systolic and diastolic cardiac periods (SCP, DCP) suitable for blood pressure trend computation, wherein the SCP are identified by calculating the time interval between PPG peaks and the DCP are identified by calculating the time interval between PPG troughs;
• apply a statistical analysis to determining the sequence of time intervals;
• eliminate a noise in the sequence of time intervals by applying an adaptive or fixed low-pass Fourier filter to each element of the sequence;
• compute a trend of systolic blood pressure by finding a slope of a best-fit line to systolic cardiac period values over time traversing;
• compute a trend of diastolic blood pressure by finding a slope of a best-fit line to diastolic cardiac period values over time and inverting the slope sign;
• and
• combine the trend analysis results to determine the blood pressure trend over a time period.

A non-transitory computer readable medium for obtaining a blood pressure trend from a subject having instruction stored thereon, wherein the instructions, when executed by the processor, cause the processor to:

• receive the PPG/magnetic sensor measurements over a first time interval;
• identify a second time interval within the first time interval with statistically stable measurements, wherein the statistically stable measurement is obtained by computing a z-score value on a subinterval that is compared to a prior z-score value computed from a previous subinterval, wherein the z-score:
• is a statistical measure in which an absolute value represents a distance between a raw score of a random variable, x, and a population mean μ, in units of standard deviation, σ:
• x−μ
• z=σ
• and in order to compute a stabilized point in time when PPG measurements have stabilized, a distance between a population mean of a previous subinterval and a population mean of current subintervals is calculated as follows:
• μ
• 2
• -
• μ
• 1
• σ
• 1
• ⁢
• and
• ⁢
• μ
• 1
• -
• μ
• 2
• σ
• 2
• wherein the PPG measurements have stabilized when a distance between two successive z-score computations becomes less than a preset noise tolerance level;
• identify within the second time interval a sequence of time intervals containing a number of both systolic and diastolic cardiac periods (SCP, DCP) suitable for blood pressure trend computation, wherein the SCP are identified by calculating the time interval between PPG peaks and the DCP are identified by calculating the time interval between PPG troughs;
• apply a statistical analysis to determining the sequence of time intervals;
• eliminate a noise in the sequence of time intervals by applying a low-pass Fourier filter to each element of the sequence;
• compute a trend of systolic blood pressure by finding a slope of a best-fit line to systolic and diastolic cardiac period values over time;
• compute a trend of diastolic blood pressure by finding a slope of a best-fit line to diastolic cardiac period values over time and inverting the slope sign;
• and
• combine the trend analysis results to determine the blood pressure trend over a time period.