Method and apparatus for determining potential onset of an acute medical condition

1 . A method of managing risk, of a subject, of suffering an adverse respiratory condition relating to a physiological system of the subject, the method using a system, the system comprising a wearable acoustic sensor, the method comprising steps of:

a) providing the subject with the wearable acoustic sensor, operatively configured to provide an output relating to sensed cyclical events in said physiological system;

b) providing a data processor in data communication with the wearable acoustic sensor; and

c) causing the data processor to:

i) monitor the output from the wearable acoustic sensor relating to said sensed cyclical events in the physiological system, wherein the data processor receives an electronic data signal representative of the output from the wearable acoustic sensor;

ii) apply a predictive numerical model comprising Extended Kalman Filtering (EKF) to generate, from output from the wearable acoustic sensor relating to an event in a most recently observed event cycle in the physiological system, a prediction of the output from the wearable acoustic sensor in relation to a corresponding event in a future cycle in the physiological system to be observed,

iii) compare the prediction with the output from the wearable acoustic sensor relating to the corresponding event in the future cycle in the physiological system, when observed, and

iv) execute an output response of the system based on the comparison, the output response comprising an operation of a device that provides a stimulation.

2 . The method according to claim 1 , comprising performing principal component analysis on the output from the wearable acoustic sensor to identify a principal component on which to perform the EKF, and wherein EKF is performed on the identified principal component only.

3 . The method according to claim 1 , wherein comparing the prediction with the output from the wearable acoustic sensor comprises comparing the prediction and the output from the wearable acoustic sensor to data in a database of historical output and corresponding predictions, and determining whether the comparison with the database indicates the adverse respiratory condition is imminent, the method including allocating increased processing resources to allow for running a machine learning model when the comparison with the database yields a disparity between the prediction and the corresponding event in said future cycle in the physiological system.

4 . The method according to claim 1 , wherein the physiological system is the respiratory system and the adverse respiratory condition is asthma.

5 . The device of claim 4 , wherein the processor is configured to allocate increased processing resources to perform steps i) to iv) when the comparison yields a disparity between the prediction and the corresponding event in said future cycle in the physiological system.

6 . A wearable condition-management device wearable by a subject at risk of suffering an adverse respiratory condition relating to a physiological system of the subject, the device comprising:

a) an acoustic sensor operatively configured to provide output relating to observed cyclical events in said physiological system, and

b) a computer processor in data communication with the acoustic sensor, the computer processor being programmed to execute instructions causing it to:

i) monitor output from the acoustic sensor relating to said observed cyclical events in the physiological system, wherein the computer processor receives an electronic data signal representative of the output from the wearable acoustic sensor,

ii) apply a predictive numerical model comprising Extended Kalman Filtering (EKF) to generate, from output from the acoustic sensor relating to an event in a most recently observed event cycle in the physiological system, a prediction of output from the acoustic sensor in relation to a corresponding event in a future cycle in the physiological system to be observed,

iii) compare the prediction with the output from the acoustic sensor relating to the corresponding event in the future cycle in the physiological system when observed, and

iv) execute an output response of the device based on the comparison, the output response comprising a controlled operation that provides a stimulation.

7 . The device of claim 6 , wherein the computer processor is programmed to perform principal component analysis on the output from the acoustic sensor to identify a principal component on which to perform the EKF, and wherein the computer processor is programmed to perform EKF on the identified principal component only.

8 . The device according to claim 6 , wherein the response comprises executable instructions causing the system to consult a database of historical output and predictions for determining whether the comparison indicates the adverse respiratory condition is imminent.

9 . The device according to claim 6 , wherein the physiological system is a respiratory system and the adverse respiratory condition is asthma.

10 . The device of claim 6 , wherein the sensor comprises piezo-acoustic sensing means configured for receiving frequencies characteristic of inflammation signs within an upper respiratory tract of the subject.

11 . A computing system comprising the wearable condition-management device of claim 6 , wherein the computing system is configured to perform the comparison by comparing the prediction and the output from the wearable acoustic sensor to data in a database of historical output and corresponding predictions, and wherein the system is configured to determine whether the comparison with the database indicates the adverse respiratory condition is imminent, and wherein the computing system is further configured to allocate increased processing resources to allow for running a machine learning model when the comparison with the database yields a disparity between the prediction and the corresponding event in said future cycle in the physiological system.