Method for predicting apnea-hypopnea index from overnight pulse oximetry readings
First Claim
Patent Images
1. A method for predicting respiratory disturbances in an individual comprising the steps of:
- a) identifying patients having obstructive sleep apnea;
b) determining actual Apopnea Hypopnea Index (AHI) in the patients;
c) obtaining pulse oximetry recordings from the patients;
d) determining patient data comprising a set of indices from the pulse oximetry readings comprising;
delta index, time spent at selected oxygen saturation levels and oxygen desaturation events;
e) performing random sampling with replacement of patient data from (d) to generate a plurality of data sets;
f) performing multivariate non-parametric analysis in the plurality of data sets to generate a plurality of prediction models;
g) obtaining a provisional predictive AHI value for each patient corresponding to each prediction model;
h) comparing the provisional predictive AHI value to actual AHI value for each patient to provide weight to each prediction model; and
i) providing each prediction model to a health care professional,wherein the plurality of prediction models and weight of the prediction models provides a predictive tool for predicting respiratory disturbances in an individual.
2 Assignments
0 Petitions
Accused Products
Abstract
The present invention provides a predictive tool for predicting Apopnea Hypopnea Index (AHI) in the diagnosis of Obstructive sleep apnea (OSA). The predictive tool is developed by recording pulse oximetry readings, obtaining delta index, oxygen saturation times and oxygen desaturation events from the oximetry readings. A multivariate non-parametric analysis and bootstrap aggregation is performed to obtain predictive models which can be used to predictive AHI in an individual and to classify an individual as having or not having OSA.
-
Citations
24 Claims
-
1. A method for predicting respiratory disturbances in an individual comprising the steps of:
-
a) identifying patients having obstructive sleep apnea; b) determining actual Apopnea Hypopnea Index (AHI) in the patients; c) obtaining pulse oximetry recordings from the patients; d) determining patient data comprising a set of indices from the pulse oximetry readings comprising;
delta index, time spent at selected oxygen saturation levels and oxygen desaturation events;e) performing random sampling with replacement of patient data from (d) to generate a plurality of data sets; f) performing multivariate non-parametric analysis in the plurality of data sets to generate a plurality of prediction models; g) obtaining a provisional predictive AHI value for each patient corresponding to each prediction model; h) comparing the provisional predictive AHI value to actual AHI value for each patient to provide weight to each prediction model; and i) providing each prediction model to a health care professional, wherein the plurality of prediction models and weight of the prediction models provides a predictive tool for predicting respiratory disturbances in an individual. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 16)
-
-
10. A method for predicting respiratory disturbances in an individual comprising the steps of:
-
a) identifying patients having obstructive sleep apnea; b) determining actual Apopnea Hypopnea Index (AHI) in the patients; c) obtaining pulse oximetry recordings from the patients; d) determining patient data comprising a set of indices from the pulse oximetry readings comprising;
delta index, time spent at selected oxygen saturation levels and oxygen desaturation events;e) performing random sampling with replacement of patient data from (d) to generate a plurality of data sets; f) performing multivariate non-parametric analysis in the plurality of data sets to generate a plurality of prediction models; g) obtaining a provisional predictive AHI value for each patient corresponding to each prediction model; h) comparing the provisional predictive AHI value to actual AHI value for each patient to provide weight to each prediction model; i) obtaining pulse oximetry recordings from the individual; j) determining a set of indices from the oximetry readings comprising, delta index, time spent at selected oxygen saturation levels and oxygen desaturation events; k) subjecting the set of indices to the plurality of prediction models to obtain a plurality of provisional predicted AHI values corresponding to each model; l) obtaining a final predicted AHI value by averaging the weighted provisional predicted AHI values; and m) providing each prediction model to a health care professional; wherein the plurality of prediction models and weight of the prediction models provides a predictive tool for predicting respiratory disturbances in an individual. - View Dependent Claims (11, 12, 13, 14, 15, 17, 18)
-
-
19. A computational system for predicting respiratory disturbances, the system having software thereon to execute the steps comprising:
-
a) receiving pulse oximetry readings; b) calculating indices comprising;
delta index, time spent at selected oxygen saturation levels and oxygen desaturation events;c) subjecting the indices to a plurality of Apopnea Hypopnea Index (AHI) prediction models to generate provisional AHI values corresponding to each prediction model wherein each prediction model has specific weight assigned thereto; d) calculating a final AHI value based on a weighted average of the provisional AHI value; and (e) providing the final AHI value to a health care professional. - View Dependent Claims (20, 21)
-
-
22. A computational system for predicting respiratory disturbances, the system having software thereon to execute the steps comprising:
-
a) receiving indices comprising;
delta index, time spent at selected oxygen saturation levels and oxygen desaturation events;b) subjecting the indices to a plurality of AHI prediction models to generate provisional AHI values corresponding to each prediction model wherein each prediction model has specific weight assigned thereto; c) calculating a final AHI value based on a weighted average of the provisional AHI value; and (d) providing the final AHI value to a health care professional. - View Dependent Claims (23, 24)
-
Specification