Abstract: A method of system monitoring or, more particularly, novelty detection, based on extreme value theory in particular a points-over-threshold POT method which is applicable to multimodal multivariate data. Multimodal multivariate data points collected by continuously monitoring a system are transformed into probability space by obtaining their probability density function (pdf) values from a statistical model of normality, such as a pdf fitted to a training data set of normal data. Extremal data is defined as that whose pdf value is below a predetermined threshold and a new analytic function, in particular the Generalised Pareto Distribution (GPD) is fitted to that extremal data only. The fitted GPD can be compared to a GPD fitted to the extremal datapoints of the training data set of normal data to determine if the monitored system is in a normal state.

Abstract: A method of analysing biomedical signals, for example electrocardiograms, by using a Hidden Markov Model for subsections of the signal. In the case of an electrocardiogram two Hidden Markov Models are used to detect respectively the start and end of the QT interval. The relationship between the QT interval and heart rate can be computed and a contemporaneous value for the slope of this relationship can be obtained by calculating the QT/RR relationship for all of the beats in a sliding time window based on the current beat. Portions of electrocardiograms taken on different days can efficiently and accurately be compared by selecting time windows of the ECGs at the same time of day, and looking for similar beats in those time windows.

Abstract: A way of quantifying the shape of an ECG waveform is disclosed by detecting the JT segment using two Hidden Markov Models and calculating the analytic signal of the JT segment. Parameters calculated from the analytic signal are used as shape descriptors for the JT segment. The shape descriptors may be displayed in a dimensionality-reduced mapping. Templates representing characteristic shapes can be produced by finding cluster centers in the shape descriptor space, and the novelty of new waveforms can be quantified by comparing the position in shape descriptor space of new shape descriptors to a predefined normal training set or to previously encountered waveforms. Novel shape descriptors can be used to retrieve the corresponding waveforms, and templates of such novel shapes can be created by averaging such waveforms, using dynamic time warping to allow for variations in heart rate.

Abstract: Vibration amplitudes are recorded as a function of rotation speed and of frequency and the data is analyzed to estimate a noise floor amplitude threshold for each of a plurality of different speed and frequency sub-ranges. On the basis of training data known to be normal speed-frequency areas which contain significant spectral content in normal operation are deemed “known significant spectral content”, so that during monitoring of new data points which correspond to significant vibration energy at speeds and frequencies different from the known significant spectral content can be deemed “novel significant spectral content” and form the basis for an alert. The estimation of the noise floor is based on a probabilistic analysis of the data in each speed-frequency area and from this analysis an extreme value distribution expressing the probability that any given sample is noise is obtained.

Abstract: A method of obtaining a consistent evaluation of the state of the system which has been monitored by measurement of multiple parameters of that system. The multiple parameters are used to calculate a single dimensional value based on the distance between the current state and normal states of the system using a Parzen Windows probability function. Consistent single dimensional values regardless of the dimensionality of the original data set can be obtained by finding a relationship between the single dimensional value and the probability of status of the system. Different relationships are obtained for different dimensionalities of data sets. Sensor malfunction can also be detected by testing the probability of the state implied by measuring all of the available parameters against the probability of the state implied by ignoring different individual ones of the parameters. A significant disparity in the two probabilities indicate possible sensor malfunction.

Abstract: Improvement in the reliability of segmentation of a signal, such as an ECG signal, is achieved through the use of duration constraints. The signal is analysed using a hidden Markov model. The duration constraints specify minimum allowed durations for specific states of the model. The duration constraints can be incorporated either in the model itself or in a Viterbi algorithm used to compute the most probable state sequence given a conventional model. The derivation of a confidence measure from the model can be used to assess the quality and robustness of the segmentation and to identify any signals for which the segmentation is unreliable, for example due to the presence of noise or abnormality in the signal.

Abstract: A method and apparatus for detecting an abnormality in e.g. in operating characteristics or function of a machine, apparatus or system, the method including providing a data sample set comprising n values of a measured physical parameter associated with the apparatus or system generated by repeating a measurement of the physical parameter n times. An extremal measured parameter value is selected from amongst the data sample set, determining a probability of observing the selected parameter value (e.g. of observing a value not exceeding the selected parameter value) by applying the selected parameter value to an extreme value probability distribution function having a location parameter and a scale parameter. The value of the location parameter and the value of the scale parameter are each constructed using an integer value m (e.g. notionally representing the size of a sub-sample data set comprising m of said measured parameter values) in which m is less than n (i.e. m<n).

Abstract: A system for assisting in dose optimisation for patients administering medication to themselves at home consists of a mobile telephone or PDA to which the patient is prompted to enter data indicative of their condition and also the dosage of medication they are taking. Such data is stored on the phone and transmitted to a remote server and is processed against predetermined criteria to decide whether the medication dose should be maintained or increased or decreased based on the patient's current condition. A message instructing the patient to maintain or adjust the dosage is displayed based on this determination. The data is made available both to the patient and clinicians by means of web pages provided by the remote server.

Abstract: An example method and apparatus for measuring the breathing rate from the photoplethysmogram signal (PPG) uses auto-regressive modelling of the signal. The PPG signal is windowed in overlapping windows of typically 30 seconds' length, overlapping by 25 seconds, to obtain discrete sections of the signal and each section is low-pass filtered, downsampled and detrended and then AR modelled using an all-pole auto-regressive (AR) model. The AR model allows identification of the dominant frequencies in the signal and the pole corresponding to the breathing rate is identified by considering its magnitude and the breathing rate it represents. Each 30 second window gives a breathing rate estimate and use of successive windows displaced by 5 seconds results in a breathing rate estimate every 5 seconds. The time series of breathing rate estimates can be Kalman filtered to reject measurements which have a large change in magnitude or represent a large change in breathing rate.

Abstract: A method of displaying medical data, particularly data representative of the condition of patients suffering from chronic medical conditions such as asthma, diabetes and hypertension. The display consists of two graphical elements, one of which indicates the current value of a parameter indicative of the patient's condition, this being displayed against another graphical element which represents a model of normality for that patient. The graphical element indicating the current condition may be, for example, a needle, against a scale which is constructed according to the patient-specific model of normality. This is particularly advantageous in the case of displays which have a small display area, such as mobile telephones and PDAs. Other forms of display are disclosed, such as histograms with the display being dynamically colour-coded and auto-scaled, or displays including limits which may vary.

Abstract: A system adapted to monitor adverse drug reactions in cancer patients undergoing chemotherapy which comprises a patient-based data terminal 3 such as a mobile telephone adapted to provide for periodic entry by the patient of data on a plurality of predetermined adverse reactions. The data is processed for display to the patient and for transmission to a remove server 7. The data is also processed to generate red or amber alerts which may be displayed to the patient and sent by the server 7 to a pager 13 held by the healthcare professional. The server 7 processes the received data for display and review via a secure web page. Alerts may be of different levels of urgency with urgent alerts being sent immediately to the pager 13, but less urgent alerts being sent in batches at regular intervals.

Abstract: A method of analysing biomedical signals, for example electrocardiograms, by using a Hidden Markov Model for subsections of the signal. In the case of an electrocardiogram two Hidden Markov Models are used to detect respectively the start and end of the QT interval. The relationship between the QT interval and heart rate can be computed and a contemporaneous value for the slope of this relationship can be obtained by calculating the QT/RR relationship for all of the beats in a sliding time window based on the current beat. Portions of electrocardiograms taken on different days can efficiently and accurately be compared by selecting time windows of the ECGs at the same time of day, and looking for similar beats in those time windows.

Abstract: A way of quantifying the shape of an ECG waveform is disclosed by detecting the JT segment using two Hidden Markov Models and calculating the analytic signal of the JT segment. Parameters calculated from the analytic signal are used as shape descriptors for the JT segment. The shape descriptors may be displayed in a dimensionality-reduced mapping. Templates representing characteristic shapes can be produced by finding cluster centres in the shape descriptor space, and the novelty of new waveforms can be quantified by comparing the position in shape descriptor space of new shape descriptors to a predefined normal training set or to previously encountered waveforms. Novel shape descriptors can be used to retrieve the corresponding waveforms, and templates of such novel shapes can be created by averaging such waveforms, using dynamic time warping to allow for variations in heart rate.

Abstract: A method for combining measurements from two or more independent measurement channels, particularly physiological measurements such as heart rate. Independent measurements of heart rate, for instance by ECG and pulse oximetry, can be combined to derive an improved measurement eliminating artefacts on one channel. A model of the process generating the physiological parameter, e.g., the heart rate, is constructed and is run independently for each channel to generate predictions of the parameter. The measured values are compared with the predicted values and the differences are used as an indication of the confidence in the measurement. The measurements from the two channels are ombined using weights calculated from the respective differences.

Abstract: A method of analysing tachometer and vibration response data from an apparatus having one or more rotary components is provided. The method comprises the steps of: providing vibration response data and corresponding tachometer data from the apparatus for a period over which a rotary component of the apparatus varies in rotational speed, the tachometer data being for that component; repeatedly performing at intervals throughout the period the sub-steps of: determining a forcing frequency of the component from the tachometer data and a corresponding vibration response frequency of the apparatus from the vibration response data, comparing the forcing and vibration response frequencies to determine the relative phase difference between the frequencies, and determining the corresponding amplitude of the vibration response from the vibration response data; and plotting the relative phase differences and vibration amplitudes on a polar diagram.

Abstract: A method and system for delivering advice to patients suffering from respiratory conditions based on changes in environment, such as weather or air quality changes. This system includes a patient specific model which uses input environmental data to predict changes in the patient's condition. The model is developed from an analysis of patient's responses to a variety of environmental triggers, and can be refined with time. The model can include only those specific triggers appropriate to a patient, can include the delay between the change in environment and change in condition of the patient, and can be run with data which is geographically localized to the patient's location. Conveniently the models can be run on personal devices held by the patients, such as mobile telephones, which are in communication with a server and/or a provider of environmental data.

Abstract: A method and apparatus for detecting an abnormality in e.g. in operating characteristics or function of a machine, apparatus or system, the method including providing a data sample set comprising n values of a measured physical parameter associated with the apparatus or system generated by repeating a measurement of the physical parameter n times. An extremal measured parameter value is selected from amongst the data sample set, determining a probability of observing the selected parameter value (e.g. of observing a value not exceeding the selected parameter value) by applying the selected parameter value to an extreme value probability distribution function having a location parameter and a scale parameter. The value of the location parameter and the value of the scale parameter are each constructed using an integer value m (e.g. notionally representing the size of a sub-sample data set comprising m of said measured parameter values) in which m is less than n (i.e. m<n).

Abstract: Improvement in the reliability of segmentation of a signal, such as an ECG signal, is disclosed through the use of duration constraints. The signal is analysed using a hidden Markov model. The duration constraints specify minimum allowed durations for specific states of the model. The duration constraints can be incorporated either in the model itself or in a Viterbi algorithm used to compute the most probable state sequence given a conventional model. Also disclosed is the derivation of a confidence measure from the model which can be used to assess the quality and robustness of the segmentation and to identify any signals for which the segmentation is unreliable, for example due to the presence of noise or abnormality in the signal.

Abstract: A method for combining breathing rate measurements from two or more independent measurement channels. Independent measurements of breathing rate, for instance by impedance pneumography and photoplethysmography, can be combined to derive an improved measurement eliminating artefacts on one channel. A model of the process generating the breathing rate, is constructed and is run independently for each channel to generate predictions of the breathing rate. The model may be a Kalman filter. The measured values are compared with the predicted values and the difference is used as an indication of the confidence in the measurement, the higher the difference the lower the confidence. The measurements from the two channels are combined using weights calculated from the respected differences.

Abstract: A method of displaying medical data, particularly data representative of the condition of patients suffering from chronic medical conditions such as asthma, diabetes and hypertension. The display consists of two graphical elements, one of which indicates the current value of a parameter indicative of the patient's condition, this being displayed against another graphical element which represents a model of normality for that patient. The graphical element indicating the current condition may be, for example, a needle, against a scale which is constructed according to the patient-specific model of normality. This is particularly advantageous in the case of displays which have a small display area, such as mobile telephones and PDAs. Other forms of display are disclosed, such as histograms with the display being dynamically colour-coded and auto-scaled, or displays including limits which may vary.