WO2007110625A2

WO2007110625A2 - Device for prediction of human or other mammalian labour onset

Info

Publication number: WO2007110625A2
Application number: PCT/GB2007/001091
Authority: WO
Inventors: Patricia Connolly; Paul Couchman
Original assignee: Jopejo Limited
Priority date: 2006-03-24
Filing date: 2007-03-26
Publication date: 2007-10-04
Also published as: EP2004046A2; CA2663420A1; AU2007231122A1; WO2007110625A3; CN101448455A; GB0605950D0

Abstract

A device for prediction of imminence of onset of labour and a method for using the device are described. The device is a portable assembly of: a plurality of recording electrodes and optionally a reference electrode, each electrode comprising a contact surface for application to a subject; a means for data acquisition; a data analysing means to process the data to determine a result indicative of the electromyographic readings of the subject and a display adapted to display this result. It is characterised in that the data analysing means comprises a means adapted to produce a result indicative an electromyographic parameter known to vary cyclically in the period preceding labour, a data register to store previous results, and a data comparator to compare a currently generated result with previous results and thereby to identify a peak in the said parameter.

Description

DEVICE FOR PREDICTION OF HUMAN OR OTHER MAMMALIAN LABOUR ONSET

The present invention relates to a device for the prediction of the onset of labour in humans or non-human mammals, a method for predicting labour using the device and the use thereof. More specifically the invention relates to a device for predicting the onset of labour which may be premature, on time or delayed relative to an initial estimate, a method of using the device, to facilitate operation by the subject or another individual who is not a clinician, and the use thereof.

Measurement of uterine contractions to determine the state of pregnancy and diagnose real labour is well documented. For example US 4,967,761 (Nathanielsz) discloses a method comprising sensing myometrial contractility, producing signals representing the sensed contractility, storing the signals and determining characteristic representative frequencies which are then classified as term or preterm labour. WO 94/13202 (Rosenberg) discloses a method and device as a diagnostic tool by means of assessing muscular contraction. Such clinical assessment by means of evaluating the mechanical activity in terms of strength, duration and frequency of palpated contractions do not however correlate with onset or progress of labour, and are liable to considerable interobserver variation.

Measurement of electrical activity, electrohysterography, has been proposed as an alternative approach to the detection of labour. For example WO 94/14373 (Garfield) records signals from embedded electrodes, WO 95/31932 and WO 96/39931 (Garfield) disclose a method which stores data and compares activity and US 5,373,852 (Harrison) uses radiotelemetric transmission for sensing pressure, temperature and electrical activity. WO 97/25922 (Rosenberg) discloses a further method for analysis of electromyographic data.

However such publications relate exclusively to clinical, usually intrusive techniques for data collection. Moreover, data analysis and/or diagnosis is generally also by complex methods requiring expert operation. These are therefore of limited application, inconvenient since they require the intervention of a clinician, and hence usually require the subject to attend a hospital, clinic or the like and are therefore limited in application, for example to instances where a prior indication of possible clinical abnormality exists. They are unlikely to be practical as a routine tool for home or other nonclinical use to improve accuracy of the predicted delivery date in a routine pregnancy.

More recently, focus in research for labour prediction has centred around specific biomarkers present in body fluid which may indicate the uterus 's preparation or readiness for labour. The use of such biomarkers will present the patient with the challenge of body fluids sampling and test interpretation at home if they are to come into widespread use.

Increasingly there is a need for a device which non-intrusively predicts the onset of labour and/or which indicates the imminence of labour and which is suited for non clinical, i. e., home use. In particular, there is a desire for a device which can both be operated in a domiciliary environment and produce results which can be interpreted without the need for a clinician.

WOO 1/45555 (Simpson & Walker) describes a device for prediction of onset of labour comprising a portable assembly of: a plurality of recording electrodes and optionally a reference electrode, each electrode comprising a contact surface for application to a subject, and in particular to an abdominal and/or vaginal surface of a subject; a means for data acquisition; a means for conversion of analogue to digitised data; a data analysing means to process the data to determine a result indicative of uterine preparedness for labour; and a display adapted to display this result. Although the device can be operated by a non-expert, it relies upon complex data analysis of frequency shifts and advanced frequency spectral analysis to determine the state of uterine preparedness.

The present invention is directed at the provision of an alternative device and method, and in particular a device and method based on less complex data processing principles, for home or other non-clinical use to improve accuracy of the predicted delivery date in any pregnancy. The invention is in particular directed at provision of a non-expert system which will provide an indication of imminence of onset of labour during any pregnancy and for any purpose, and not merely where a clinical need or abnormality is identified.

Accordingly in a first aspect of the invention there is provided a device for prediction of imminence of onset of labour comprising a portable assembly of: a plurality of recording electrodes and optionally a reference electrode, each electrode comprising a contact surface for application to a subject, and in particular to an abdominal and/or vaginal surface of a subject; a means for data acquisition; a data analysing means to process the data to determine a result indicative of the electromyographic readings of the subject and a display adapted to display this result.

The device is characterised in that the data analysing means comprises a means adapted to produce a result indicative an electromyographic parameter known to vary cyclically in the period preceding labour, a data register to store previous results, and a data comparator to compare a currently generated result with previous results and thereby to identify a peak in the said parameter.

Readings are taken consistently periodically, for example once a day, preferably at a consistent time. The device identifies and indicates a parameter peak (for example, it identifies and indicates this when a given measured data reading is determined by the comparator to be lower than the most recent stored equivalent data reading in the data register).

The principles of the invention can be applied to any parameter derivable from successive readings of electromyographic potentials of a subject which can be shown to vary cyclically in a consistent manner to the extent at least that it delivers a characteristic peak at a sufficiently consistent time period in advance of the onset of labour to make the peak serve as a marker from which a more accurate indication of the imminence of onset of labour can be derived. In a simple mode of operation, such a characteristic peak is used to refine a previously obtained estimate of labour onset, which may for example have been derived as a delivery estimate in an entirely conventional manner, or by some other data analysis technique. Such a previous estimated delivery date may similarly be stored within the system on a suitable data register. Once a peak identified as such a characteristic marker is resolved, the result can be processed and passed to the display means, which displays a result indicative of the imminence of labour derived from the said marker.

Two data parameters in particular are advanced below as being suitable for use with the device in accordance with the principles of the invention. Both of these (have been measured to) produce characteristic peaks a known interval in advance of the onset of labour, and can be used as markers to give an indication of imminence of labour in the manner described. In accordance with the first embodiment, the data analysing means comprises a means adapted to produce a result indicative of the total energy level within at least one specific narrow band of the overall frequency spectrum for electromyographic data. It has surprisingly been found that if a narrow band is resolved from the spectrum such energy levels are cyclical during pregnancy and that a final peak occurs 7 to 14 days before labour onset. Thus an indicated peak can provide a simple way of indicating labour onset and improving the accuracy of an initial prediction, in that, in a simple mode of operation, a peak within 15 days of predicted delivery can be taken as such an indicator. On identifying such a peak, the comparator causes a suitable indication to be made via the display.

It has accordingly been established that a frequency based analysis of electrophysiological signals obtained by a relatively simple device suitable for home recording can effectively predict labour onset. Surprisingly it has been found that the clearest predictive method lies not in frequency shifts or advanced frequency spectral analysis of the type described in the prior art. Instead the tracking of energy levels within specific bands of frequency from day to day shows that there is a characteristics rise and fall of the relative energies within these bands prior to labour. This implies that specific electrical signals from nerve or muscle groups have some sort of cyclic activity as the womb prepares for labour. Although the invention is not limited by any theory, it is suggested that this may be linked to the hormonal cycles, gap junction formation or muscle toning. In any case the result is perfectly clear in when Fourier transform techniques are used to group frequencies and energy levels. Observation of the rise and fall of energy distribution within the frequency bands allows the prediction of labour onset from anywhere between a few weeks to a few days according to the mathematical rules applied in the analysis.

The device works best when separate recordings are obtained for consecutive days or sessions during gestation in the third trimester whereby the energy patterns in the frequency bands of electrophysiological activity from the patient can be ascertained.

In accordance with the principles of the invention at least one narrow frequency band is used, and data indicative of the energy within the band is produced by the acquisition and analysis means. Data across a plurality of such narrow bands within the overall spectrum may be used. Frequencies in the range 0 -5 Hz recorded from the mother have been found to be particularly useful in the determination of cyclic energy patterns and have been grouped in short intervals. A suitable frequency interval for such a narrow band to perform the analysis in accordance with the invention is of 0.2 to 0.5 Hz, and most preferably around 0.25 Hz. Suitable narrow frequency bands across which total energy may be determined in accordance with the invention might be 0 to 0.25 Hz, 0.25 to 0.50 Hz etc. The frequency bands may be narrowed or extended as required.

In a preferred embodiment the energy in each frequency band relative to the overall energy in the total signal is produced as the result indicative of the energy within the band. This minimises day to day fluctuations such as skin impedance during electrical recording etc. However in another embodiment the actual total energy in the signal for the recording session in individual frequency bands could be derived and used as the result.

The device records uterine electrical activity, in particular by application to the abdominal or vaginal surface, in generally similar manner to that known in the art. That is to say, a plurality of recording electrodes for measuring action potential are provided which are applicable to an abdominal or vaginal surface of a patient under analysis, conveniently an analogue to digital converter connected to the electrodes acts to convert electromyographic signals produced by the electrodes into digitised data indicative of the electromyographic signals and action potentials. The optionally digitised data are stored and analysed with reference to previously recorded results and optionally predetermined reference parameters, the results being used to actuate a display adapted to display this data as indicating imminence of labour.

The invention differs from most of the prior art by exploiting the surprising result that analysis of uterine electrical activity can be used not merely to monitor the progress of, and/or diagnose active labour as in the prior art, but can be used to obtain an indication of uterine preparedness for labour in the initial phase of parturition before onset of active labour. The ability to utilise such analysis in a predictive manner was first described in WOO 145555 but this current invention develops a simple but effective data analysis technique to produce a device that significantly extends the abilities of analysis and prediction.

It does this in the embodiment by exploiting the surprising result that if data are analysed across narrow bands within the overall spectrum conventionally used for the analysis of electromyographic signals, a plot of total energy across one or more of such narrow bands shows a resolvable cyclical pattern which can be used in conjunction with an initial conventionally estimated delivery date to improve the accuracy of the latter. A final peak in such a cycle of energy across a resolved narrow band can be shown to occur within 7 to 14 days of onset of labour. It follows that it can reasonably be concluded that the detection of such a peak which falls within the final 15 days of the estimated period can be used to given a more accurate indication that 7 to 14 days remain to onset of labour. The display may give a suitable alert of this date.

The data analysis means performs at least one such an analysis to determine a cyclical peak in at least one such narrow frequency range. It is not intended that the device should be limited only to such an analysis. A device in accordance with the invention may further refine any resultant estimate by being adapted to carry out further data analysis.

In a second embodiment the data analysis means includes a burst counter that allows for the identification of classical electrical burst activity over a predetermined recording period, containing a range of frequencies, the comparator being adapted to identify a peak in such data (again, for example, by indicating when a measured burst count is below the most recent recorded burst count).

Such burst activity has been noted in the past to be of greater strength and frequency during labour. However the current device is able to measure the occurrence of such activity which is present before and during labour onset and the pattern of such activity is discerned by the device.

Observations on consecutive days of women from 32 weeks of labour to labour onset have allowed the pattern of 'burst' electrical activity typically found when recordings are made from the uterus or areas surrounding it to further be analysed and monitored in consecutive sessions. This burst activity has been found to display a certain pattern in the approach to labour onset and birth. Surprisingly, it does not rise continually to birth, but peaks 3 to 5 days before labour. Observation and analysis of such burst activity allows determination of a further marker of the imminence of onset of labour.

Preferably, the first and second markers of the first and second principal embodiments are combined. The combination device thus identifies two markers which successively enhance the accuracy of a prediction of likely labour onset. First, the device identifies a cyclical peak in the total energy (and in particular in the relative total energy) of at least one narrow frequency band across the usual spectrum in the manner above described. Suitable narrow frequency bands across which total energy may be determined in accordance with the invention might be 0 to 0.25 Hz, 0.25 to 0.50 Hz or like width bands up to a typical maximum frequency of 5 Hz. Of course other higher or lower frequency bands may be examined in similar manner

The incidence of such a peak that occurs within 15 days of an original estimated date is treated as a first marker, and delivers a result that onset of labour is expected within a first period, typically 7 to 14 days.

Second, the device identifies a second marker comprising a peak in the burst activity count over a consistent recorded period, and this peak is referred as a second marker indicating that onset of labour is even more imminent, for example, within 3 to 5 days.

Since the device and method of the invention is predictive rather than merely diagnostic of labour it enables subjects to determine the expected onset of their labour with accuracy. Whilst such a tool might have value in giving advance warning of abnormal premature labour allowing timely measures to be taken to provide medical assistance where necessary in the instance of premature labour in advance of the event itself, thereby allowing for example preventative action, simplifying any remedial action and reducing premature birth problems, it is also of use in monitoring the progress of routine pregnancy.

In fact, as a simple unit for non-expert use it is in this regard, as a means to improve the accuracy of prediction of labour during normal pregnancy where no disorder is present, for reasons that have nothing to do with clinical diagnosis or prediction for any medical or veterinary purpose, that the invention makes a particularly useful contribution. Onset of labour is still hard to predict, both because of the approximate nature of the original calculation and because of the statistical spread between early and late delivery that is a feature of a perfectly healthy pregnancy. The device and method help to avoid false alarms which are distressful and an undue cost burden on the medical services. They improve resource allocation in providing support services during normal pregnancy. They allow a human mother to make a prediction for social purposes, for example to allow for the prospective mother to organise her support some days prior to the birth of her child. They allow a livestock manager to make a prediction for economic purposes, to organise stocking practice more efficiently.

Preferably, the apparatus includes signal digitisation means to digitise the collected analogue signal before processing by the analysis means. The digitisation of the signal is for current computational convenience but in other embodiments of such a device analogue means of manipulation could be envisaged.

Readings are taken periodically, for example once or twice a day, preferably at a consistent time. Data are preferably collected for a consistent period. To that end, the apparatus conveniently includes timer means to time the data collection period, acting with the data retrieval means to ensure each data collection takes place over a fixed period. Conveniently this automatically starts data retrieval and ends data retrieval at the end of the period. Alternatively the data retrieval period is user controlled, and the timer serves to advise the user, for example via the display means or visual or audio alarm, that the collection period is complete. Additionally or alternatively the device includes a reminder system, for example acting via the display means or a visual or audio alarm, to remind a user when each successive data collection action is due.

The display is adapted to work in conjunction with the comparator to provide a simple display indicating a data peak, and thus indicating imminence of onset of labour. For example a two-state model indicating a positive or negative result for a peak in the or each narrow band cycle or a three-state model indicating this first peak marker and a second marker based on the burst analysis may be provided.

As the analysis is based at least in part on, and seeks to increase the accuracy of, a conventionally generated estimated delivery date (EDD), the apparatus preferably further includes means to input and store an initial estimated delivery date for example in the form of a delivery date register. The analysis means may then reference this stored date in assessing whether an identified cyclical peak in narrow band total energy is close enough to the predicted date to be representative of imminence of onset of labour as above. The display is adapted to display a positive result only if the comparator detects a peak and the analysis means shows this criterion of proximity to the delivery date estimate to be met. The present invention thus enables a prediction of the onset of labour and/or a monitoring of the pre-labour phase of parturition, as the uterus prepares for the contractile or labour phase of parturition. Moreover, the invention does this in a compact and portable manner, and by displaying only a small number of discrete results, provides a result which can be readily interpreted by a subject without the need for intervention by a clinician, and is thus particularly suited to home use. For example, given the peak analysis on the two markers discussed above, a reading can be given at 7 to 14 days from onset of labour, and a further reading at 3 to 5 days.

Without being limited by theory, the present invention in considered to provide highly reliable results based on the finding that certain parameters govern the production of uterine electrical signals.

Specifically there is an increase in the number of intercell connections (gap junctions) which then facilitate the spread of electrical impulses from one part of the womb (uterus) to another, allowing for progressively synchronous and effective mechanical activity (contractions). These changes occur progressively throughout the pre-labour phase of parturition. We judge that the observed change in the rhythms governing electrical activity reflect these cellular changes. The changes can be resolved by analysis of relatively narrow bands only and determination of total energy within the bands to show a cyclical pattern in the later stages of pregnancy, with a final peak being temporally relatable to the onset of labour itself.

This finding is highly significant in the ability to provide capable of displaying results which are readily understood by a non-specialised user, and which, in contrast to many prior art devices, do not require the analysis of a clinician to obtain a meaningful prognosis or diagnosis. In particular, the device makes it possible to pre-record reference data and/or to pre-programme particular reference parameters which are characteristic of a small number of pre-defined pre-labour states indicating the progress of preparedness for contractile labour. For example, data are pre-recorded and/or parameters are pre-programmed which correspond to uterine electrical activity indicative of the pre-parturition state and uterine activity indicative of a state of full or near full uterine preparedness for active contractile labour, and the display means are configured to display one or other state accordingly as a positive/negative result. Optionally, further data and/or parameters are stored representative of one or more intermediate states as the preparation phase of parturition has progressed to a greater or lesser extent. A comparison of changes in electrical activity against these referenced parameters can offer an indication of the imminence of full preparedness for labour.

The device thus offers a tool which is not diagnostic of active labour but is predictive of the onset of labour by enabling monitoring to be made during the preparation phase of parturition. Moreover, the device achieves this objective in a compact and portable manner which is particularly suited for home use.

The display of the invention is adapted to display the collected and analysed data in terms of one of a small number of discrete results indicating the imminence of onset of labour. The digitised data indicative of uterine electrical activity is compared with suitable pre-recorded reference data and/or predetermined reference parameters by the data analyser and is characterised as being indicative of a state of preparedness corresponding to one of the said discrete results, and the display is then actuated to make such an indication.

In its simplest embodiment, the display may be configured to indicate a positive or negative result as to whether the uterus is substantially prepared for contractile labour and onset of labour is imminent. However, it is known that the first phase of parturition in which the uterus prepares for labour extends over a period of time, and WO0145555 describes that detectable electrical changes occur within the uterus over this time.

It is found that detectable changes can be obtained for many subjects at least eight weeks before the onset of labour, and these detectable changes can therefore be used to effect an intermediate reading where labour is approaching but not yet imminent. The device therefore displays negative/intermediate/positive results, which for practical purposes from the perspective of the subject could be thought to correspond to: labour distant (retest in a week or so)/ labour likely in a week or two (retest daily, remain prepared to initiate deliver procedures)/ labour imminent (prepare for onset of labour, treat any symptoms as probable labour).

Display means may be auditory, visual or both and conveniently indicates a result without the need for the subject to interpret levels or the frequencies of activity, i. e., in the form of a set of illuminating lights, tactile patterns such as vibration or discrete auditory signal or alarm or an alphanumeric display to display simple messages. In a multi-state embodiment such as the preferred three-state embodiment, multi-coloured lights may be used, each colour indicating a different state of preparedness.

The display means preferably includes at least an alphanumeric display to display simple instructional or informational messages, in particular messages to this effect. The display may now also indicate with some certainty after a predictive session that has identified the first above marker (peak in energy across a narrow band) 'Labour onset is expected within 7 to 14 days' and count down daily from this prediction in the screen messages to the expectant mother. The device will be able to further predict as the countdown continues and identifies the second predictive marker (peak in burst count) 'Labour onset now likely in 3 to 5 days' or some such equivalent message.

As is illustrated by this embodiment, the particular advantage of present invention is that the display means can give simply understandable results which do not require the intervention of a clinician and which offer the potential for significant reduction in both false alarms and unexpected early labour producing complications because of inaccessibility of medical assistance.

Preferably the device detects spread of electrical impulse throughout the uterus, and may classify as a range, in the scale of 0-100% intercell connection within the uterus.

The device of the invention suitably comprises any means for operatively associating the assembly of components such as a casing, mounting, cassette, card, frame or the like. A mounting may be rigid or flexible, for example a flexible mounting may comprise an article of clothing such as a belt, patch or the like which may be conveniently worn by the subject for prolonged periods without inconvenience.

The device may be manually operated or may comprise means for intermittent operation allowing periodic inspection on a regular basis.

Activation may be manual or automated for example as a periodic activation using a timer control means, and/or as a manual activation such as through a power switch. The device preferably incorporates a portable power supply such as a rechargeable or non-rechargeable battery. Additionally or alternatively means are provided to connect the device to a mains power supply.

The comparator compares previous responses stored in a data register, in particular to determine a cyclical peak in energy across at least one resolved narrow band and/ or a peak in burst count. The data register may also be provided for input of pre-recorded data indicative of known pre-parturition electrical responses. The comparator may be adapted further to reference this data to refine and confirm any result analysis. Comparison with pre-recorded data maybe with data programmed by a medical supervisor, for example enabling detection of a particular condition or pre-labour period or combinations thereof as alternative or co-operative settings. Alternatively the pre-recorded data may be acquired by the device in a programming mode by the subject in a non-labour and in particular a preparturition condition in simple manner.

Preferably the device operates by analysing several consecutive sessions from a patient to identify cycles in electrical energy levels at different frequencies from the recorded signal and to record burst activity from the uterus on a daily or frequent basis. Ideally, for best predictive value the patient should start daily recordings 25 to 30 days before the estimated due date (EDD). A session of recording may vary from a few minutes to longer. With an EDD and the frequency energy analysis from session to session a predictive value of labour onset with a 7 to 14 day period from a characteristic peak in the energy cycles is possible. By further adding the count and pattern of burst activity before labour we can further predict a 3 to 5 day period to labour onset. It has been identified importantly that the frequency of burst activity (number of bursts in a fixed time recording session) rises to a peak in the 3 to 5 days before labour and then falls off. Identification of this peak provides and additional prediction of labour onset

Data acquisition and digitisation is suitably carried out by means of a microprocessor or other computing means associated co-operatively with the electrodes using known techniques. Data analysis is by any suitable means and method associated with parameters related to change in electrical signals. In particular, the signals are analysed by performing a spectral analysis of power density of electromyographic potentials to determine total energy across one or more narrow frequency bands, and the analyser is adapted to perform such an analysis, for example by suitable programming.

The preferred frequency analysis method for analysing the signal comprises producing a power spectrum, and thereafter analysing the total energy in specific narrow frequency bands with respect to all energy in the total recorded signal. The analyser is adapted to perform such an analysis (e. g. by suitably programming). A suitable overall frequency range for measurement, as will be known in the art, is up to 5 Hz. A suitable narrow frequency band is 0.2 to 0.5 Hz in width, and for example around 0.25 Hz. A particularly useful indicative frequency band is found to be 0.25-0.5 Hz, and in a preferred embodiment the data collection and analysis means is adapted to determine total energy at least across this band.

The energy contained in the specific frequency bands is recorded by the analyser and compared on a day to day or session to session basis by the comparator at least to identify a cyclical peak. The emergence of peaks in total energy contained in any one frequency band when compared with EDD analysed alone leads to improved predictions of labour onset. For example, if the total energy in the frequency band 0.25-0.5 Hz of the signal recorded is found to be reaching a peak with 15 days of the EDD then labour onset can reasonably be predicted to follow within 7 to 14 days and the device and analyser can be adapted to give such a refined indication of imminence of labour via a suitable message on the display. Alternatively a pattern of the energy totals or ratio in the 0.25-0.5 Hz band may be observed over a number of days. Regular occurrence in rise and fall of energy in the days before labour is indicative of the approach of labour onset. Other frequency bands may be used as may other means of analysing the rise and fall of the energy within the band.

The total power, measured as the area under the above within the range of frequencies is also significant.

In a further embodiment the device will also count or otherwise monitor how frequently bursts of electrical activity are found in the recorded signal. The pattern of frequency of burst activity change before labour, rising to a peak and then falling off and the device will analyse this and display a message to the user about labour onset. Therefore for example a peak in the amount of bursts occurring in a session followed by a sharp fall of in burst activity in the subsequent sessions indicates that labour onset will follow between 3 and 5 days later.

Thus the complete embodiment of the ideas can result in a device that first predicts labour onset by the peaks in energy in certain frequency bands and then further confirms the imminence of labour onset by monitoring how many electrical bursts occur in each recoding session. Clearly either of these methods could also be used alone in a device.

Data analyser and display means operating this preferred embodiment of the invention thus comprise an analyser such as a microprocessor or filter adapted, for example suitably programmed, to split data into frequency bands, to perform an analysis of energies within bands and to record and compare these from session to session, to detect and record burst activity and note its number of occurrences in any one session, to compare such burst activity on a session to session basis, and to have the means to display in software appropriate indicative messages of the imminence or not of labour onset.

The present invention thus provides means for displaying frequency-analysed electrohysterographic data in form of a small number of discrete results such as a positive or negative result and optionally further one or more intermediate results, in particular results which indicate not merely a diagnosis of labour, but a prediction of its likely onset.

Additionally in the current invention it is possible to monitor, in subsequent recording sessions in the third trimester, 'burst activity' from the uterine electrical signals and this leads us to be able to identify a peak in the number of burst recordings that occurs just 3 to 5 days before labour onset. This peak of activity has never been observed or recorded in any prior published studies. In the research work that underpins this application, undertaken one of the most comprehensive studies of electrophysiological recording throughout all of the third trimester of labour has been undertaken where signals have been measured on a daily or twice daily basis from a patient every day from 32 weeks of gestation. A total of thirty patients were used in our stud and some 140 Mbytes of data analysed to reveal the patterns of burst activity and energy changes in frequency bands.

Burst activity is obvious to the eye of the trained observer looking at recordings from the abdomen or uterus of a pregnant woman and is most simply recorded in this manner but can equally be embodied in a device by using mathematical algorithms that detect, shape, amplitude or frequency components of the incoming signal.

In a further aspect of the invention there is provided a method for predicting onset of labour comprising applying the principles of a device as hereinbefore defined to the skin of a subject in the uterine region, activating the device for a sufficient period to record electrical activity and obtaining a displayed result indicating imminence of labour.

In particular, the method comprises applying potential measuring electrodes to an abdominal or vaginal surface of a patient in the uterine region, activating data acquisition means for a sufficient period to record electrical activity, acquiring data corresponding to the electrical activity, and resolving at least one data parameter known to vary cyclically in the period prior to onset of labour; repeating the method successively over a period of time, in particular at consistent intervals and for a consistent recording period, comparing the result for the resolved parameter with previous results for the same to identify a peak value; deriving from the known relationship of such peak value to the likely onset of labour a result indicative of the imminence of onset of labour.

Any parameter can be used which has a characteristic peak displayed a consistent time period before the onset of labour. The method is used to identify a peak in the parameter, and suitable reference made, either automatically through suitable analysis means in the manner above described, or by user reference or any other suitable means, to the consequent indication which can be derived regarding the imminence of labour. Preferably, the method involves analysing the data by a power frequency analysis technique, and in particular the analysis comprises a tracking of distinct cyclical peaks of energy within discrete frequency bands as above described. Session to session or day to day comparisons of the energies over the period of use of the device are preferable.

That is to say, the method comprises the step successively over a period of time of repeatedly analysing the data by a power frequency analysis technique to produce a result indicative of the total energy level within at least one specific narrow band of the overall frequency spectrum for electromyographic and identifying a peak in the successive results.

Preferably the method additionally or alternatively allows the electrical bursts of activity in an session to be discerned, counted and further analysed. Session to session or day to day mapping of the burst activity over the period of use of the device is preferable. A peak is identified as above described.

That is to say, the method comprises the step successively over a period of time of repeatedly analysing a count of electrical bursts of activity in a session and identifying a peak in the successive results.

Preferably, both analysis methods above are used, the method of this aspect of the invention being modified to comprise the steps of: determining an initial estimate of labour onset/ delivery date; successively at regular intervals analysing the data by a power frequency analysis technique to produce a result indicative of the total energy level within at least one specific narrow band of the overall frequency spectrum for electromyographic data, identifying a peak in the successive results within 15 days of predicted delivery and outputting this identified peak as a first marker of the imminence of onset of labour; successively at regular intervals analysing a count of electrical bursts of activity in a session, identifying a peak in the successive results and outputting this identified peak as a second marker of the imminence of onset of labour.

Each marker result is preferably displayed, for example via an alphanumeric display means to output the first marker by means of a display 'Labour onset is expected within 7 to 14 days' and to output the second marker by means of a display 'Labour onset now likely in 3 to 5 days' .

Application of the device may be temporary or permanent, i. e., the device may be applied periodically by the subject or may be worn as part of thick clothing.

The method may comprise a preliminary non-labour programming stage to acquire non-labour data. Alternatively the method may be operated with a preprogrammed device.

Further features of the method will be apparent from the foregoing description of the device and its use.

In a further aspect of the invention there is provided the use of a device or method as hereinbefore defined for predicting premature, on time or delayed labour in advance of physically detectable or observable contractual activity.

The use may be human or animal use. In each case the device may be used for humans and animals in remote regions enabling them to remain in their natural environment throughout the duration of pre-labour pregnancy and to detect prediction of labour onset for a suitable period enabling travel to assistance.

Use of the device may be use within the home or in transit, during a subject's normal daily activities in public, in hospitals, clinics and the like. It is a particular advantage that the use is simple, convenient and the device may be operated by skilled or unskilled personnel, by the subject or any other person.

In a further aspect of the invention there is provided a device programmed to predict the onset of labour as hereinbefore defined.

Further aspects of the invention will be apparent from the following examples, to be read in conjunction with the figures wherein:

Figure 1 illustrates an embodiment of the invention in the form of a belt/corset worn device;

Figure 2 is an example of variation in energy in the frequency band 0.25-0.50

Hz as a fraction of total energy in the signal on a day to day basis in pregnant

Patient A; Figure 3 shows burst activity recorded from the abdomen of a pregnant woman across a pair of electrodes;

Figure 4 illustrates the pattern of such burst activity in successive recordings during the days approaching onset of labour.

Example 1

A device of four electrodes and reference electrode has been utilised to record electrophysiological signals from the abdomen of a pregnant women. The four electrodes are paired such that two electrodes form a channel of input to the device. The electrodes may be arranged as to detect transversely or longitudinally the electrical signals spreading across the abdomen of a patient. Data from either or both channels may be analysed independently or in combination.

A possible arrangement, by way of example only, is that illustrated in Figure 1. A capture/analysis/comparator/display device (1) is mounted on belt/corset (2) connected to recording electrodes (3) and reference electrode (4).

The device (1) in the example comprises coloured lights (6a, 6b, 6c) indicative of discrete results states. In the preferred embodiment, three discrete results states are envisaged, generally representing a pre-labour state, an indication when the first marker (peak in narrow band energy level) is detected and an indication when the second marker (peak in burst count level) is detected. For example, green, amber and red lights may be used.

An LCD panel (5) is pre-programmed to display simple alphanumeric information. This may include, for example, approximate time to labour and/or instructions on recommended duration before a re-test is carried out. For example this may display a 7-14 day count when the first marker (peak in narrow band energy level) is detected and a 3-5 day count when the second marker (peak in burst count level) is detected. In this way, the device is adapted for simple home use by a subject without acquiring any special medical knowledge, to give both reassurance when labour is distant, early warning when labour is not so distant, and a full alert when labour is imminent.

The device also comprises an on/off switch (7) and a battery panel. Typical electrical data obtained after Fourier analysis and calculation of total energy in a frequency band is shown in Figure 2.

The first peaks in energy may be recorded several weeks prior to onset of mechanical contractions. Pre-programmed data and/or pre-programmed reference parameters ensure that this can be interpreted by the device, and displayed as an intermediate result.

AU results that indicate peaks are effected by the development of intercell connections facilitating spread of electrical impulses across the womb.

Figure 2 illustrates variation in energy in the frequency band 0.25-0.5 Hz as a fraction of total energy in the signal on a day to day basis in pregnant Patient A. Figure 2a illustrates data over the whole period, and Figure 2b over the immediate pre-labour period. Several clear peaks and troughs in the signal can be detected visually in Figure 2a. The presence of such peaks in the 15 days before EDD indicates that labour will start within 7 to 14 days. Labour onset occurs at the end of the x-axis.

Example 2

In addition to the detection of the peaks as shown which indicate the onset of labour in 7 to 14 days, the rate of occurrence in the electrical 'burst ' activity of the uterus is indicative of the approach of labour. Figure 3 shows the actual data recorded from the abdomen of a pregnant woman across a pair of electrodes. In the figure a 50 second trace of electrical signal obtained across a pair of electrodes on the abdomen of a pregnant patient is shown. Two strong bursts of electrical activity and two lesser bursts can be seen. Furthermore a count of the number of all electrical burst in a recording session, say ten minutes of recording, on subsequent days of the pregnancy shows that burst activity reaches a peak a few days before birth. This is illustrated in Figure 4, which shows a graph of a count of the electrical burst in each ten minute recording session for a pregnant patient. It can be seen that the number of burst per session (y-axis) reaches a peak 5 days before labour onset. X- axis is dbb = days before birth.

Thus, the analysis of example 1 may be combined with the analysis of example 2 to give two markers of imminence of labour onset, or either used separately as a single marker, to improve upon an initially conventionally derived EDD.

Claims

1. A device for prediction of imminence of onset of labour comprising a portable assembly of: a plurality of recording electrodes and optionally a reference electrode, each electrode comprising a contact surface for application to a subject; a means for data acquisition; a data analysing means to process the data to determine a result indicative of the electromyographic readings of the subject and a display adapted to display this result; characterised in that the data analysing means comprises a means adapted to produce a result indicative an electromyographic parameter known to vary cyclically in the period preceding labour, a data register to store previous results, and a data comparator to compare a currently generated result with previous results and thereby to identify a peak in the said parameter.

2. A device in accordance with claim 1 wherein the data analysing means comprises a means adapted to produce a result indicative of the total energy level within at least one specific narrow band of the overall frequency spectrum for electromyographic data.

3. A device in accordance with claim 2 wherein the data analysing means is adapted to analyse data across a plurality of narrow bands within the overall spectrum.

4. A device in accordance with claim 2 or claim 3 wherein the data analysing means is adapted to analyse data in one or more narrow bands having a frequency range of 0.2 to 0.5 Hz within an overall signal frequency range of 0 -5 Hz.

5. A device in accordance with any of claims 2 to 4 wherein the data analysing means is adapted to determine the energy in a narrow frequency band relative to the overall energy in the total signal and produced this as the result indicative of the energy within the band.

6. A device in accordance with any preceding claim wherein the data analysis means includes a burst counter that allows for the identification of classical electrical burst activity over a predetermined recording period, containing a range of frequencies, the comparator being adapted to identify a peak in such data.

7. A device in accordance with claim 1 wherein the data analysing means comprises a means adapted to produce a result indicative of the total energy level within at least one specific narrow band of the overall frequency spectrum in accordance with one of claims 2 to 5 and a burst counter in accordance with claim 6 and is adapted thereby to identifies as markers which successively enhance the accuracy of a prediction of likely labour onset first a cyclical peak in the energy of the at least one narrow frequency band and second a peak in the burst activity count.

8. A device in accordance with any preceding claim further comprising signal digitisation means to digitise the collected analogue signal before processing by the analysis means.

9. A device in accordance with any preceding claim further comprising timer means to time a data collection period, acting with the data retrieval means to ensure each data collection takes place over a fixed period.

10. A device in accordance with any preceding claim further comprising a delivery date register to input and store an initial estimated delivery date.

11. A device in accordance with claim 10 wherein the analysis means is in data communication with the delivery date register to reference the stored date in assessing whether an identified cyclical peak in narrow band total energy is close enough to predicted date to be representative of imminence of onset of labour.

12. A device in accordance with any preceding claim wherein the display is adapted to display the collected and analysed data in terms of one of a small number of discrete results indicating the imminence of onset of labour.

13. A device in accordance with any preceding claim wherein the data analysis means is adapted to produce a power spectrum, and thereafter analyse the total energy in specific narrow frequency bands with respect to all energy in the total recorded signal.

14. A method for predicting onset of labour comprising applying the electrodes of a device in accordance with any preceding claim to the skin of a subject in the uterine region, activating the device for a sufficient period to record electrical activity and obtaining a displayed result indicating imminence of labour.

15. A method for predicting onset of labour comprising applying potential measuring electrodes to an abdominal or vaginal surface of a patient in the uterine region, activating data acquisition means for a sufficient period to record electrical activity, acquiring data corresponding to the electrical activity, and resolving at least one data parameter known to vary cyclically in the period prior to onset of labour; repeating the method successively over a period of time; comparing the result for the resolved parameter with previous results for the same to identify a peak value; deriving from the known relationship of such peak value to the likely onset of labour a result indicative of the imminence of onset of labour.

16. The method of claim 15 comprising the step successively over a period of time of repeatedly analysing the data by a power frequency analysis technique to produce a result indicative of the total energy level within at least one specific narrow band of the overall frequency spectrum for electromyographic potentials and identifying a peak in the successive results.

17. The method of claim 16 wherein in one or more narrow bands having a frequency range of 0.2 to 0.5 Hz within an overall signal frequency range of 0 -5 Hz.

18. The method of one of claims 16 to 17 wherein the data is analysed by determining the energy in a narrow frequency band relative to the overall energy in the total signal and producing this as the result indicative of the energy within the band.

19. The method of one of claims 16 to 18 comprising the steps of: determining an initial estimate of labour onset/ delivery date; identifying a peak in the successive results occurring within 15 days of predicted delivery; and using this identified peak as a marker of the imminence of onset of labour.

20. The method of claim 19 wherein on identification of a peak within 15 days of predicted delivery a result is output via the display indicating labour onset is expected within 7 to 14 days.

21. The method of one of claims 15 to 20 comprising the step successively over a period of time of repeatedly analysing a count of electrical bursts of activity in a session and identifying a peak in the successive results.

22. The method of claim 21 comprising the steps of: determining an initial estimate of labour onset/ delivery date; identifying a peak in the successive results; and using this identified peak as a marker of the imminence of onset of labour.

23. The method of claim 22 wherein on identification of a peak a result is output via the display indicating labour onset is expected within 3 to 5 days.

24. A method in accordance with claim 15 and further including the steps of one of claims 16 to 20 and the steps of one of claims 21 to 22 applied in accordance with the following protocol:

25. determining an initial estimate of labour onset/ delivery date;

26. successively at regular intervals analysing the data by a power frequency analysis technique to produce a result indicative of the total energy level within at least one specific narrow band of the overall frequency spectrum for electromyographic data, identifying a peak in the successive results within 15 days of predicted delivery and outputting this identified peak as a first marker of the imminence of onset of labour;

27. successively at regular intervals analysing a count of electrical bursts of activity in a session, identifying a peak in the successive results and outputting this identified peak as a second marker of the imminence of onset of labour.

28. The use of a device in accordance with one of claims 1 to 13 or method accordance with one of claims 14 to 27 for predicting premature, on time or delayed labour in advance of physically detectable or observable contractual activity.