WO2001093762A2

WO2001093762A2 - A fail-sure computer aided method for anticoagulant treatment

Info

Publication number: WO2001093762A2
Application number: PCT/DK2001/000397
Authority: WO
Inventors: Peter Michael Nielsen
Original assignee: Medimage Aps
Priority date: 2000-06-09
Filing date: 2001-06-08
Publication date: 2001-12-13
Also published as: WO2001093762A9; WO2001093762A3; AU6378501A

Abstract

The present invention relates to a method for determining a set of corresponding dose of a drug and time to take the drug. The method comprises the step of measuring a blood coagulation value at a patient and by use a first computer system determining the set of corresponding dose and time to take the drug. The determination by use of the computer system is also preferably performed at the patient. Furthermore, the present invention relates to a validation method in which data relating to the determination of the set of corresponding dose and time to take the drug is transmitted to a second computer system. By use of this second computer system a set of corresponding dose and time to take the drug is determined and this set may be validated against the set determined by the first computer system. The present invention also relates to a personal apparatus at the patient having means for taking a blood sample, means adapted to determine a set of corresponding dose and time to take the drug and means for transceiving data to and from a validation apparatus. The validation apparatus comprises means for transceiving data to and from the personal apparatus, means adapted to determining a set of corresponding dose and time to take the dose based on data received from the personal apparatus and means adapted to validate this set to the set determined by the personal apparatus.

Description

TITLE:

A fail-sure computer aided method for Anticoagulant Treatment

Background of the invention:

Anticoagulant Therapy (ACT) has gained widespread use in conditions wherever reduction of risk of thrombosis is mandatory. The indications for ACT are numerous, and a vast number of patients are receiving the treatment - in Denmark 20.000 and world-wide 5.000.000 patients. Oral anticoagulants have a narrow therapeutic range and monitoring of treatment through measurement of the blood coagulation capabilities, in example the Prothrombine Time (PT) or the INR is critical. Adverse effects of the drugs are life threatening, major bleeding is associated with overdosage, especially in case of uncontrolled management of the therapy.

Several devices equipped with standard kit solutions for measurements of INR are commercially available today. The patients receiving ACT keep the device at home and procedures for testing are developed towards simplicity, as the patients are supposed to perform the tests themselves.

The result of the test is displayed, but the patient will have to call the doctor to get the prescription for further treatment. The doctor of course will have to keep records to ensure treatment quality.

Several solutions have been developed by doctors for the calculation of medication dosage on a computer, as the doctor receives information about INR test results from patients or laboratories . These calculation methods are all extrapolative and do not include knowledge enhancement. No such application is commercially available.

Also, the human factor involved in the known systems involving a doctor for prescription for further treatment could in some situation turn out to be disadvantageous. For instance the instantaneous effect of a drug will in many situation be widely influenced by the drug taken as long as several days before and as the medicine in many situation has the possibility of harming as well as treating the patient a wrong dosage could lead to a very dangerous situation in case the doctor is not specialised and experienced in his field. Even in such a situation it is common that not one doctor alone takes care of the same patient during the hole treatment and as differences are present in the way doctors treat patients a less than optimal treatment of the patient may occur.

Furthermore, the involvement of specialised and experienced doctors may be uneconomical as these doctors skills may be used more efficient for other purposes in case the dosage is performed by a computer system which function is supervised and validated.

Thus, it is a purpose of the present invention to provide a method and an apparatus being able to

determine a blood coagulation value, preferably, at a patient and determine at least one dose of a drug and at which time(s) the dose(s) is(are) to be taken based upon the determined blood coagulation value, wherein the method and apparatus is simple to use for the patient, the treatment has been optimised and the determination of the dose of the drug and at which time the drug is to be taken is done preferably without the assistance of a doctor.

This is in a broad aspect of the present invention provided by a method for determining at least one dose of a drug and at which time the dose is to be taken by use of a first computer system/pattern determination system, the method comprises the steps of

providing a measured blood coagulation value measured by use of a blood analysis means, preferably located at a patient determining an expected blood coagulation value, comparing, such as determining the difference between, the measured blood coagulation value and the expected measured blood coagulation value, determining at least one dose of a drug and at which time the drug is to be taken by use of a first computer system; said determination being based at least on the measure blood coagulation value.

The phrase "computer system" is used to emphasise that in some preferred includes a computer system having a pattern determination system

At least one dose of a drug and at which time the drug is to be taken denotes a set of data comprising the amount of a drug and the specific point in time at which the dose is to be taken. The set may in a preferred embodiment of the invention comprise only one dose and at which time this dose is to be taken, but it is also preferred to determine a series of doses with connected times to take the doses.

In the broad aspect of the present invention a treatment of a patient will typically be initiated by prescribing by a doctor a specific amount of drug and the time the dose is to be taken. This initiation process may in a preferred embodiment of the present invention be applied to start the use of the method according to the present invention and the first dose of the drug the patient have to take may be prescribed by the method when a blood coagulation value has been determined for the first time.

This first time determined - and also the succeeding - blood coagulation value will preferably be stored in a memory for further processing and use. After a blood coagulation value has been measured, a blood coagulation value which would be expected at the time of the measurement will be determined and compared to the measured blood coagulation value. These steps may very advantageously serve as validation steps of the means used for measuring the blood coagulation value as a too high difference between the measured and expected values may be the result of malfunctioning of the measuring means. Furthermore, this step may preferably also serve the purpose of cancelling the method and initiating an alerting procedure, which could be needed in a situation where the drug may have a side effect on the patient.

The expected blood coagulation value may in some preferred embodiments be a value chosen within the therapeutic interval. In other preferred embodiments this values is determined by use of the computer system it self.

Based on the measured blood coagulation value a new dose of a drug at which time the dose is to be taken may be determined by a first computer system. In a preferred embodiment of the present invention this computer system is a Bayesian network. Such a network has the ability of being able to, after a "training process", to determine the dose of a drug and the time at which the dose is to be taken on the basis of the measured blood coagulation value.

The "training" of the network is based on many observations of corresponding dose of a drug and at which time the dose is to be taken observed by doctors. Based on these observation the connections inside the network are established and the network may be ready to use. Another important feature of the present invention is the ability of "self- learning". As the determined sets are stored together with the measured and optionally also the expected blood coagulation values, a database can be generated containing the history of the treatment. This database may then be applied for updating the Bayesian network, which then can be applied for treatment of patients.

As the history effect of the drug often plays an important role in treatment - the effect of a dose of drug may be delayed - a situation where an over-dosage could be the result if the history effect is not encountered in the method. This may in a preferred embodiment of the present invention be taken into account by only executing the determination of the at least one dose of a drug and at which time the dose is to be taken if the difference between the measured and expected blood coagulation exceeds a predetermined value. By this way of evaluating the difference between the expected and the measured blood coagulation value, a change of dose that would lead to large changes in blood coagulation value on behalf of small variations in actual measured blood coagulation values may be inhibited.

Furthermore, the determination of the at least one dose of a drug and at which time the dose is to be taken may in preferred embodiments of the present invention be inhibited by evaluating the value of the measured blood coagulation value. In this situation the method according to the present invention only determine the expected blood coagulation value if the measured blood coagulation value is within a certain pre-set range (preferably set by a medical adviser to assure that no life threatening situations occur), thereby ensuring that the measured blood coagulation value for instance will not lead to the determination of a dose and the time at which the dose is to be taken which exceed potentially harmful (lethal) values. Also, this evaluation of the measured blood coagulation value may be used for generating a signal to be sent to the patient and/or a doctor signalling a potentially harmful (lethal) situation.

After the at least one dose of a drug and at which time the dose is to be taken has been determined the patient is preferably informed. This may preferably be performed by outputting, preferably at the patient, a first set of data related to the method and/or of the at least dose of the drug and at which time the dose is to be taken if determined.

The first set of data may comprise the at least one dose and time and information relating to the performance of the method and the means used such as the performance of the blood analysis means. The first set of data may preferably be presented to the patient by readable information such as information presented on a display such as a monitor.

As the at least one dose and at which time the dose is to be taken is determined by the first computer system/pattern determination system so may also the expected blood coagulation value be determined by use of the same system. Alternatively, the expected blood coagulation value may be determined by reading a value stored for instance in a memory of the first computer system/pattern determination system. In the first situation, the first computer system/pattern determination system may be able to derive a value that corresponds quite close to the actual treatment situation, which may include the history of the treatment. In the second situation which may be applied in the case control of the specification of the expected blood coagulation value, a database is applied for providing an expected blood coagulation value.

As the method according to the present invention is able to determine at least one dose of a drug and at which time the dose is to be taken without assistance of a doctor or other personal, it is in particular embodiments preferred that the method is evaluated by for instance a second computer system/pattern determination system. In order to be able to evaluate the method it is preferred that the method according to the present invention further comprises the steps of - outputting to a first transmitter means the measured blood coagulation value and if determined also the at least one dose of a drug and at which time the dose is to be taken and/or transmitting the at least one dose and at which time the dose is to be taken and the data relating to the performance of the method to a second computer system/pattern determination system.

The data relating to the performance of the method is data enabling the second computer system/pattern determination system to determine, on the same basis as the first computer system/pattern determination system, the at least one dose and at which time the dose is to be taken.

The transmitter means used for transmitting may be provided by a standard modem being able to transmit data through the public telephone network or via the internet or the transmitting means may be a wireless telephone or it may be a means specially designed to transmit data to the second computer system.

In this preferred embodiment of the invention it is also preferred to transmit data relating to the performance of the system including the means used for taking the blood sample.

As one of the purposes of transmitting data to a second computer system is to evaluate the performance of the method, also the basic parameter, the measured blood coagulation value, is transmitted to the second computer system. Based on this measured blood coagulation value the second computer system may then be able to reproduce or repeat the steps performed by the first computer system by the second computer system and a comparison of such two sets (doses and connected times to take the doses) may be carried out by the second computer system.

Accordingly, the method preferably further comprises the step of verifying, by use of the second computer system/pattern determination system, the at least one dose and at which time the dose is to be take determined by and removed from the first computer system/pattern determination system, and if the dose and time to take the dose is verified the transmitting acceptance data to first computer system/pattern determination system, or - if the dose and time to take the dose is not verified the transmitting rejection data to the first computer system/pattern determination system.

Furthermore, the first computer system/pattern determination system is preferably adapted to receive the acceptance and the rejection data and in such embodiments the method further comprises the steps of, if the acceptance data has not been received, such as received successfully, by the first computer system/pattern determination system from the second computer system/pattern determination system, then transmitting criticality data to the second computer system/pattern determination system, and - outputting a criticality response, preferably at the patient.

As the received second set of data has to be transmitted successfully it may be ensured that performance of the method for determining corresponding dose and time (at the patient) may only be accepted based on an amount of information which has been found to be suitable in order to validate the method according to the present invention. Otherwise, a situation could occur wherein the determined dose and time would be accepted upon an insufficient basis, which in turn could lead to a potentially lethal dosage.

Again, a crucial feature of the method according to the present invention is safety and therefore the second computer system which may be able to contact a doctor or other suitable personal. The doctor is preferably contacted by transmitting the transmission criticality data to the second computer system. Furthermore, an output is provided to the patient in order to inform the patient of any potential error in the method for providing the at least one dose and at which time the dose is to be taken whereby the patient will be able to contact his doctor for further advice.

One of the situations in which the transmission does not perform the way it was designed for could be where the first transmitter and/or the first receiver means is(are) faulty. In such a situation, the transmission criticality data is preferably outputted to a second transmitter means for transmitting the transmission criticality data to the second computer system, whereby a fail safe system is provided being able to contact a doctor or the like in case the first transmitter and/or the first receiver means is faulty.

An important aspect of the present invention is its ability to being able to suggest or determining a new point in time to determine a next dose of a drug and at which time the drug is to be taken. In such embodiments of the method according to the present invention the method may further comprise the steps of

determining a new point in time for determining at least one set of corresponding dose of a drug and at which time the dose is to be taken, outputting this new point in time at the patient, and - outputting this new point in time to the first transmitter means.

By the possibility of being able to supply the patient with the time for determining a next dose and time at which the dose is to be taken the method may also be able to watch for a missing or forgotten determination of a set of dose and time. In such embodiments the method may output a signal at the patient signalling time for determining a new dose and time, and if this determination is not carried out within a reasonable time then transmitting data to the second computer system signalling this event.

In a very important embodiment of the method according to the present invention the first set of data and/or the at least one dose of a drug and at which time the dose is to be taken is outputted at the patient only if acceptance data is received by the receiver means. The acceptance data may preferably be generated by the second computer system and is preferably generated in case the at least one dose and at which time the dose is to be taken determined by the first computer system is validated satisfactorily i.e. fulfils a predetermined acceptance criteria. By only outputting the first set of data and/or the at least one dose and time determined in case the set determined has been validated satisfactorily, situations where the patient is informed of a wrong or even a lethal dose may be inhibited.

A further purpose of the present this aspect of the present invention is to provide a method being able to

evaluate the determined dose and at which time the dose is to be taken, and - alert a doctor and the patient, in case the determined dose and at which time the dose is to be taken meet a predetermined alerting criteria,

This purpose ensures a high quality determination of the dosage of the drug and prevents the personal apparatus from determining and/or outputting the determined at least one dose of a drug and the time at which the dose is to be taken at a patient in case of determination of potential dangerous values.

This is in a broad aspect of the present invention provided by a method for verification of a method for determining a first at least one dose and at which time the dose is to be taken the method comprises the steps of receiving a measured blood coagulation value, preferably measured at a patient, and determining at least one dose of a drug and at which time the dose is to be taken by use of a second computer system/pattern determination system; said determination being based on at least the received blood coagulation value.

Preferably, the method for verification performs the same steps as the method which has determined the first at least one dose and at which time the dose is to be taken, but the verification method according to the present invention is not limited to that situation. The verification method may also with success be used as a second method for determining a second set of at lest one dose of a drug and at which time the dose is to be taken, where after this second set may be evaluated against the first set.

In the first situation, the verification method has the ability of tracking the method applied by the first computer system, whereby a certain step applied by the first and the second computer system giving a different result may be identified which in turn may be used for making decisions related to the difference.

In a very important embodiment of the verification method according to the present invention the method further comprises the steps of receiving at least one dose of a drug and at which time the drug is to be taken preferably determined at the patient, determining the differences between the determined and received doses and at which time the doses are to be taken, and if one or both of said differences are below predefined values, preferably defined by a medical adviser, then transmitting acceptance data to a receiver means of the first computer system/pattern determination system or if one or both of said differences are above predetermined values, preferably determined by a medical adviser, then transmitting rejecting data to the receiver means of the first computer system/pattern determination system.

Whether one or both differences are to be below or above certain values in order to either accept or reject the method is preferably judged by a medical adviser in advantage of utilising the method - such a judgement is also needed in other embodiments of the present invention, and in similar situations the judgement is also preferably made by the medical adviser. The rejection data may be any data signalling rejection of the dose and at which time the dose is to be taken determined by the first computer system. The rejection data may also preferably comprise information to the patient informing him of how to proceed the treatment for instance by taking a dose of the drug for security reasons, contacting a doctor and/or for instance changing a battery or like charging the first computer system and repeating the step relating to the determination.

In order to further verify the method for determining a dose of a drug and at which time the drug is to be taken the verification method may preferably further comprise the steps of receiving an expected measured blood coagulation value, determining and expected measure blood coagulation value, - determining the difference between the received and the determined expected blood coagulation value and, if the difference is above a predefined value, preferably defined by a medical adviser, then transmitting rejection data, signalling that the received expected blood coagulation value is erroneous, to the receiver means of the first computer system/pattern determination system.

As the verification further comprises the comparison of the expected measured blood coagulation value, also the determination of the expected blood coagulation utilised by the first computer system/pattern determination system is checked for instance for errors. Furthermore, as the expected measured blood coagulation value may be applied in supervising the results of the treatment when compared with the measured blood coagulation value, the possibility of alerting a doctor in case this comparison turns out the result that a too large difference exists.

Also in this preferred embodiment the rejection data may comprise information to the patient informing him of how to proceed the treatment for instance by taking a dose of the drug for security reasons, contacting a doctor and/or for instance changing a battery or like charging the first computer system and repeating the step relating to determining the first set of dose and time.

By this way of transmitting either acceptance data or rejection data to the receiver means of the first computer system/pattern determination system (at the patient), it is possible to inform the patient of the present status of the method for determining the at least one dose of a drug and the time at which the dose is to be taken. The acceptance data preferably comprises data signalling to the patient that he is allowed to take the drug according to the determined first dose of a drug and at which time the dose is to be taken and the rejection data preferably comprises data signalling to the patient that he is not allowed to take the drug according to first set of dose time.

In the situation where the criticality data is outputted to the transmitter means it is also preferred to output this data to a first rejection means of the first computer system/pattern determination system (at a patient) and to a second rejection means (distant from the patient), whereby contact to a doctor and/or the patient can be achieved in case the first transmission means is faulty. This further outputting may also preferably serve the purpose of a backup alerting system. In another aspect of the present invention the invention relates to an apparatus for determining a least one dose of a drug and at which time the dose is to be taken, the apparatus comprises measuring means for measuring a blood coagulation value, means adapted to determining an expected blood coagulation value, means for determining the difference between the expected blood coagulation valueand the determined blood coagulation, and means adapted to determining the at least one dose of a drug and at which time the dose is to be taken.

The measuring means is preferably an ordinary blood coagulation tester and the means adapted to determine an expected blood coagulation value and the means adapted to determined the first at least one dose of a drug and at which time the dose is to be taken are preferably a computer system such as an ordinary personal computer or and specific designed chip having a program designed to perform the steps needed. The means for determining the difference may also be such a computer system but it may also be a set of electronic logical gates designed provide an output based on the input.

In yet another important aspect of the present invention the invention relates to an apparatus for verification of at least one dose of a drug and at which time the dose is to be taken, preferably being determined at a patient, the apparatus comprises receiver means for receiving a measured blood coagulation value, preferably measured at a patient, and at least dose of a drug and at which time the dose is to be taken, means adapted to determine an expected measured blood coagulation value, means adapted to determine the difference between the measured blood coagulation value and the expected blood coagulation value, and means adapted to determining an at least one dose of a drug and at which time the dose is to be taken.

The receiver means may preferably be a modem or the like enabling data transmission via the public telephone network or it may be a portable telephone being able to transmit data.

Also in this aspect of the present invention the means adapted to determine an expected blood coagulation value and the means adapted to determined the at least one dose of a drug and at which time the dose is to be taken is(are) preferably a computer system such as an ordinary personal computer or and specific designed chip having a program designed to perform the steps needed. The means for determining the difference may also be such a computer system but it may also be a set of electronic logical gates designed provide an output based on the input.

DETAILED DESCRIPTION OF THE INVENTION

A detailed description of the invention, in particular of a preferred embodiment thereof, now follows in conjunction with the appended figures, in which:

Fig. 1 is a 3-D drawing of the structure of a preferred embodiment of the invention.

Technical and medical basis for the invention

This section aims at giving a detailed description of the technical and medical basis behind and utilised in the present invention - for instance the medical practice may very advantageously be implemented in this aspect of the present invention as a dose calculation application, as data stored in a database.

At mentioned in the introduction to the invention, several standard device solutions for measurement of blood coagulation capabilities are available today. Optimising a standard device set up through application of a decisive system - an intelligent facility - introduces great flexibility, reduction of time consumption without loss of treatment quality and security.

The intelligent facility is preferably based on Bayesian network technology providing the necessary robustness for working in fail secure environment.

As mentioned earlier the device must be able to transmit data to a central server facility for storage and monitoring of performance or malfunction, either through simple modem card for telephone line transmission or through built in GSM mobile phone unit. The latter facilitates easy installation and daily use. The transmission of each measurement and/or calculation results to server is carried out automatically.

The server software is designed for identification of incoming information, storage and display. Furthermore the central server system must respond to malfunctions and to results suggesting calibrations or treatments out of order.

Blood Coagulation: In the reaction sequences of the coagulation cascade twelve clotting factors are involved. The reactions are represented as distinct pathways (intrinsic, triggered by »foreign« substances not normally found in the circulation and extrinsic, triggered by tissue thromboplastin) which collectively lead to the conversion of prothrombin to thrombin, the enzyme responsible for converting circulating fibrinogen to fibrin. Fibrin strands mesh together to form a fibrin clot.

Blood coagulation tests measure the time required for the formation of a fibrin clot following the addition of a coagulation activating reagent. Laboratory assays typically use plasma recovered from anticoagulated (citrated) blood samples. The clotting time determined is a measure of the functionality of the patient's hemostatic system. The extrinsic pathway is evaluated by the PT time test (PT: Prothrombine Time). The test is sensitive to the coagulation factors II, VII, X and VII as well as fibrinogen (I) and is performed by adding thromboplastin reagent to the patient sample. The test is widely used to monitor oral anticoagulant therapy which suppresses the synthesis of vitamin K- dependent clotting factors. The intrinsic pathway is evaluated by the APPT test (Activated Partial Thromboplastine Time) which measures all of the coagulation proteins except Factor VI I and XI I.

Oral anticoagulant are prescribed to reduce the risk of thrombosis associated with various cardio- and cerebrovascular conditions and follow certain invasive procedures. Indications for oral anticoagulants include:

* Deep Vein Thrombosis and Pulmonary Embolism

* Atrial Fibrillation

* Myocardial Infarction, preventing complications * Mitral Valve Disease

* Stroke-in-progression/ repeted T.I. A.

* Other vascular malformations, i.e. AV-shunts.

* Cardiac Valve Replacement

Physicians as well as patients on oral anticoagulants are acutely aware of the need for and benefits of regular testing of prothrombine time. Oral anticoagulants have a narrow therapeutic range. If the drug level is too high, the patient is at risk for bleeding and if it is too low, dangerous clotting may occur. Either situation is sometimes life-threatening and always expensive. Certain dietary changes, alcohol consumption or other medications can affect the level of effect of anticoagulant therapy, further complicating the clinical picture and management of the patient.

Dosing is managed by measuring the amount of time it takes for the patient's blood to clot. Historically, results were reported in seconds, but results varied, depending of the potency of the thromboplastin used. PT results could not be compared from one lab to another. Laboratories and physicians all over the world are now using the INR system (Intemational Normalised Ratio) which is designed to standardise results from all laboratories regardless of the potency of the thromboplastine reagent utilised. The recommended INR is 2.0-3.0 for most clinical indications. More aggressive treatment is recommended for patients with mechanical heart valves (3.0-4.5)

When oral anticoagulant therapy (ACT) is initiated i.e. with a coumarine derivate (in Denmark Marevan usually is preferred) initial dosage is calculated to 10-15 mg depending on body weight. INR is measured and dosage is calculated on a daily basis until INR is stable and therapeutic range is achieved.The follow-up monitoring is then planned with measuring of INR: Three times in week 1 and Two times in week 2. Hereafter the interval is doubled until a 6 week interval is reached, provided that INR is within therapeutic range. Otherwise the intervals between tests are reduced.

Current PT testing protocol is useable but not optimal. The frequency of testing for "stable" patients is determined largely by logistics and patient convenience factors rather than optimal clinical considerations. Several studies have demonstrated that therapy within therapeutic range was achieved in only 50% of the time. Many factors influence drug levels and more frequent and/or customised testing schedules is to be preferred. Recently self-testing equipment have been introduced enabling patients to measure INR in a home lab. These apparatuses prove to be reliable and is expected to improve overall management of ACT.

Still time consumption is only slightly decreased regarding doctor's resources and quality and assurance control might even be dwindling, though current verification of machine performance is well developed.

For further development of dosing on a quantitative basis, fail secure management and quality and assurance control new technological approaches are necessary. Clinical practice

The corresponding sets of values: : The prescribed dose and the resulting INR measurement are obtained in the following procedure:

Initially an extensive investigation of the coagulation status is performed on a blood sample. One of the tests is an INR calculation. If all the parameters are normal and the patient has a normal body weight a dose of 15 mg of Marevan® is prescribed (if Marevan is preferred out of ACT-drugs). The following day a new INR test is performed. This should be raised significantly and depending on the slope either 10 mg or 7,5 mg of Marevan is prescribed. If, of course, the value is raised dramatically, an even lower dose (or nothing) of the drug is prescribed. The third day of treatment a new INR test is performed. The value should now be close to the chosen optimal value for this specific patient and the next dose is prescribed to maintain the value within the chosen INR interval for the patient and obtain and keep a zero slope of the curve of the measurement values. A dose of 5 mg Marevan a day will typically be sufficient to maintain a value within the therapeutic interval (INR 2,0-3,0) and from day four the doctor may be able to prescribe the dose 3 days ahead and a few days later one week ahead (and a few weeks later two weeks ahead).

If prediction problems is experienced and the measurement values tend to oscillate more, daily sets of values are necessary to predict the pattern of the daily dose of Marevan for longer periods of time.

The doctor will use his/her experience to treat "slow responders" and "quick responders" to the drug trying to hit the therapeutical interval as soon as possible. This predicting experience is utilised in the procedure of training an intelligent application, e.g. a Bayes'ian network. The training sets are identified among schemed examples of ACT from real life comprising both optimal treatment courses from normal, slow and quick responders. The patient scheme preferably comprises a co-ordinate system, wherein the INR measurement values are displayed on the y-axis and time is displayed on the x-axis. Each point put into the scheme represents an INR measurement at a certain time and will elicit a decision of a dose to be taken and a time for the next INR measurement. These decisions are written into the scheme in fields exclusively relating to the point leading to the decisions. Detailed description of preferred embodiments of Int ACT

The system comprises of two arrangements 1 and 20, one being a Portable Apparatus 1 placed at the Patients site the other being a Server 30 to be placed at a hospitals site.

The Portable Apparatus

The Portable Apparatus 1 in Fig. 1 comprises 4 modules: a Standard PT-tester 2, a Dose Calculation Module 3, a Display Module 4 and a Tele Module 5.

The Standard PT-tester 2 comprises a part taking a blood sample from the finger of a person and a part analysing the blood sample. The result is a PT-value. When the tester is being used, the person inserts a finger 6 into the opening 7 and activates the button 8. The needle 9 penetrates the skin on the finger and the blood is transferred to the analysing part 10 through a cannula 11. The result - a PT-value - is determined by the Standard PT-tester and transmitted through wire 12 to the Dose Calculation Module 3 in Fig. 1.

The Dose Calculation Module 3 comprises means for electronic storing and means for electronic calculation. In a preferred embodiment of the invention the means for electronic storing is a ram, and the means for electronic calculation is a microchip . Other means for electronic storing and calculation can be used.

The means for electronic storing is being used for storing the PT-value received from the Standard PT-tester 2, a Reference Value, a Limit Value, a Calculation Formula, a Time Between Blood Samples and a Log of events.

The Reference Value and the Limit Values (upper and lower) are initially being decided upon and stored in the means for electronic storing by the Medical Adviser. When the Dose Calculation Module 3 receives a PT-value from the Standard PT-tester 2, the value is being stored. The difference between the Reference Value (the expected blood coagulation value) and the PT-value (the measured blood coagulation value) is being calculated and compared to the Limit Values. If the difference between the Reference Value and the PT-value is above (considered in absolute sense) the Limit Value a new Data Set is being calculated by the Calculation Formula and stored in the means for electronic storing. The Data Set comprises at least a dose of drug and at which time the dose is to be taken and optionally a new Reference Value replacing the old Reference Value, and a new Time Between Blood Samples. If a new Reference value and a new Time Between Blood Samples are not determined, these values are set by the medical adviser respectively as a value within the therapeutic interval and as a suitable time interval.

The Calculation formula is preferably the Bayesian network, but the formula may also be based on the clinical practice described above implemented in a decisive application. This formula (algorithm) may in this case very suitable be based upon fitting a 3-dimensional curve to connected values of PT-value (PT_measur_ed), time (t) at determination of each PT- value and the dose (D) determined at t, i.e.

PTmeasured⁼f(t, D).

This curve is fitted based on actual determined values and future desired values of PT, the latter being based on for instance a curve fitted to empirically determined PT values and the connected time and dose.

According to this aspect of the present invention, a new point in time for repeating the dose determination procedure may be determined instead of being prescribed by a medical adviser utilising the clinical practice described herein. In this case the new point in time may be determined by the Bayesian network or it may be determined by a mathematical relation ship function reflecting a clinical practice, for instance reflecting time intervals between tests as a function of the level of fluctuations of PT-values. The decisions of the network must follow basic rules of determining the interval between measurements: Initially the INR is measured every day untill the INR value has entered into the therapeutic interval (criterium 1). The intervals may then be expanded if the gradient coefficient of the curve of measurement values decreases below a predetermined value (criteria 2). If both criteria are met the plan for expanding intervals as described in the "clinical practice" part above may be followed. The intervals are reduced if, for some reason, the criteria are no longer met. The intervals are reduced according to either the discrepancy of expected/ measured INR values or if the gradient coefficient of the curve of measurement values increases above a predetermined value. If none of the criteria are met the interval is reduced to one day (reset). This situation may very well coincide with an alert elicited.

The Log File comprises for each blood sample a Time Stamp, the PT-value with the corresponding Reference Value and a pointer to the Data Set being calculated using the Calculation Formula.

For each blood sample a Data Set is transmitted through wire 16 to the Display Module 4.

The Display Module 4 comprises a Screen 17 and User Interacting Buttons 19-22. When the Data Set is being received from the Dose Calculation Module, it is being visualised to the user on the Screen 17. At the same time the user is being prompted to press either the "yes" button 19 or the "no" button 20, either to confirm the result before the Log File is being stored for the blood sample or to take another blood sample. By activation of the data button 21 the user can at any time see the latest Data Set on the Screen including the Time Between Blood Samples, which tells the user when to take the next blood sample. The send button 22 is being used to activate the Tele Module, which is done after each blood sample in order to transfer the Data Set to the Server.

The Tele Module 5 comprises a Modem connecting to the Server 30 preferably through a wireless phone or similar wireless connection for data transfer. When data is received from the Display Module 4 the Tele Module 5 establishes a telephone connection to the Server 30. When the connection has been established, the Tele Module 5 reads the Log File from the Dose Calculation Module 3 and transmits its full content to the Server 30. From the Server 30 the Tele Module 5 receives a Status Signal indicating either if the Log File was accepted or not. The Status Signal is being transferred to the Display Module 4 for user visualisation on the Screen 17.

The Portable Apparatus 1 has a unique identification number enabling the Server 30 to recognise each Patient. The identification number is stored in the means for electronic storing. The Server

The Server 30 comprises 3 modules: a Display Module 31 , a Server Module 32 and an Interface Module 33.

The Server Module 32 comprises a Calculation Unit, means for electronic storing, and a Modem. The Log File is being received from the Portable Apparatus 1 through the Modem and is then being stored in the means for electronic storing. Based on the PT-values, Data Sets are being calculated by the Calculation Unit and compared to the corresponding calculation results from the received Log File. Any divergence between the calculated and the received values will result in an error signal being sent back to the Portable Apparatus 1 and being sent to the Interface Module. If no errors are found The Server Module 32 stores its own Log File comprising a Time Stamp, a pointer to the received Log File and the corresponding calculated values.

The Interface Module 33 receives in case of divergence between the received values and the calculated values an error signal from the Server Module 32. The purpose of the Interface Module 33 is to alert the Medical Adviser in case any error occurs. The Interface Module 33 may call the alert by connecting to the hospital communication system ("beeping" the doctor on call). At the same time the Patient is being alerted by means of the screen 17 on the display module 4 in the Portable Apparatus 1.

The Display Module 31 comprises a Screen 34 and a Keyboard 35. The error messages are sent to the Screen 34 together with the information in the Log File, to which the error is related. The Medical Adviser updates the Calculation formula and initially sets both the Reference Value and the Limit Value by means of the Keyboard 35. The Calculation formula and the Reference and Limit Values are being transferred to the Portable Apparatus 1 through the Server Module 32.

Finally the Medical Adviser is able to send messages to the Portable Apparatus 1 , using the Keyboard 35. This could e.g. be an appointment for medical examination etc. Furthermore the Medical Adviser is being warned, if a Patient has not connected to the Server 30 for a period of time exceeding the Time Between Blood Samples. Before the Patient uses the Portable Apparatus 1 for the first time, it is being initialised by the Medical Adviser. The initialisation is done through the Display module 31 at the Server 30. The Medical Adviser simply key-in the Reference Value in a pre filed form, Limit Value, A first value for Time Between Blood Samples and some values characterising the Patient, age, weight and sex. These values are being used to tune the Calculation Formula for the specific Patient. When the Medical Adviser has entered all relevant data these are transferred to the Portable Apparatus 1 by means of the Server Module 32 in the Server 30 and the Tele Module 5 in the Portable Apparatus 1.

The first time the Portable Apparatus 1 is being turned on the Patient is asked to confirm, using the yes button 19 the new reference data is being uploaded. The Patient can see who has entered these data, and the date the data was entered. By confirming, the Data are being transferred to the Dose Calculation Module 3. By rejecting the Medical Adviser gets a message saying that the Patient has rejected the data. This procedure can be repeated as many times as needed. E.g. if the Medical Adviser finds that the situation is for a change in the medical treatment.

In order to assure that the Portable Apparatus 1 and the Server 30 are always working correctly, the system has a self test function. The self test function is being activated automatically and periodically and checks for all kind of standard faults. If any Faults are found, the Medical Adviser and the Patient are being noticed by the Display Module (31 ,4) of respectively the Server 30 and the Portable Apparatus 1. At the same time all Blood testing and calculation functions are being disabled.

Both the Server 30 and the Portable Apparatus 1 can have outputting interfaces e.g. for connection of printing means or for connecting a PC.

The Portable Apparatus 1 are supplied with power from either batteries or from regular electrical circuitry.

For daily maintenance and cleaning, the PT-tester 2 can be disassembled and both the needle 9 and the cannula 11 can be replaced by the patient without use of tools. In order for the patient to keep track of the maintenance, a disc shaped plastic indicator can be set to the date of the day when the needle 9 has been replaced and the PT-tester 2 has been cleaned. The PT-tester 2 is being delivered in a special box, suitable for storing the PT-tester 2 and additional equipment e.g. needles, cannula, fluids for cleaning etc.

Neural Network Technology.

As mentioned above the determination of dose and at which time the dose is to be taken is in a preferred embodiment based on utilisation of neural network technology - artificial intelligence.

In the application considered here we utilise an approach in which neural networks will be used to control model based interpretation procedures. This approach allows us to build interactive procedures into early deliverables and to progressively remove these procedures as more automated and intelligent procedures are developed. The neural networks are used to identify appropriate interpretations based on descriptive models. The use of descriptive models is intended to provide a basis for implementing interpretation procedures that are robust and verifiable. It is essential to provide forward pathways for updating evidence and backward pathways to control the interpretation process.

The training sets for the model are easily identified. It is possible to run samples of well known patterns of reactions and dosage prescription charts in the Neural Network Computer Model.

Bayesian networks.

Graphical models cover a range of probability based inference models especially suited for handling problems with inherent uncertainty. The approach to intelligent systems taken by graphical is model oriented. We aim at models of the domain in question that is in direct correspondance with reality, such that real concepts have counterparts in the model. The most widespread type of graphical models is Bayesian Networks. Bayesian Networks consist of a qualitative and a quantitative part. The qualitative part consist of a graph whose nodes are stochastic variables, and whose links mirror dependencies between variables. This part can be drawn as a picture in which indirect (in)dependicies can be extracted relatively easy. Experience have shown that this eases knowledge elicitation from experts that are not neccesarily familar with the technicalities of the formalism. This part of the model describes the relevant concept in the domain and the way they interact. The quantitative part of the model consist of a collection of conditional probability tables that detail the interaction. Each variable in the model have an associated probability table that describes the distribution of that variable for each possible configuration of its parent variables, that is, the variables that have direct influence on the variable in question.

Bayesian networks have a solid coherent theoretical foundation that primarily is based on probability theory and graph theory. This enables the generation of a runtime model based on the specification of the network outlined above. This model is capable of very general reasoning as any variable can act as both input and output. This is in contrast to e.g. neural networks, where only one-way reasoning is possible, and is due to the model based approach where model updating consist of a global coherent calculation of probabilities for all variables given the evidence that have been entered into the system.

Thus, the runtime model is able to perform the various tasks most often involved in intelligent decision making: test planning, diagnosis, therapy planning, prognosis, and surveillance. Another feature is the ability to dynamically update the model based on experience, as it is able to catch variations over time, but also in order to tailor a general model to the specific characteristics of individuals.

In another preferred embodiment the Bayesian networks are extended with decision and utility nodes. Thus a formalism known as influence diagrams is obtained. Influence diagrams are designed for sequential decision problems, where the order of decisions , such as dosage of medicine, are predetermined.

In another preferred embodiment, Markov chains that are especially suited for repeated patterns of identical structure are utilised. When programming strategy has been selected and defined and training of the intelligent facility has been completed and application programme interface solution is verified the bulk of software is downloaded to EPROM to be installed in the Dose Calculation Module 3.

Claims

1. A method for determining at least one dose of a drug and at which time the dose is to be taken by use of a first computer system/pattern determination system, the method comprises the steps of providing a measured blood coagulation value measured by use of a blood analysis means determining an expected blood coagulation value, comparing, such as determining the difference between, the measured blood coagulation value and the expected measured blood coagulation value, determining at least one dose and at which time the dose is to be taken by use of a first computer system/pattern determination system; said determination being based at least on the measured blood coagulation value.

2. A method according to claim 1.wherein the determination of the at least one dose and at which time the dose is to be taken is only executed if the difference between the measured and expected blood coagulation exceeds a predetermined value.

3. A method according to claim 1 or 2, wherein the determination of the expected blood coagulation value is only executed if the measured blood coagulation value is within a certain preset range; said range being preferably set by a medical advice so as to assure that no life threatening situations will occur.

4. A method according to any of the preceding claims, wherein the expected blood coagulation value is determined by use of the first computer system/first pattern determination system.

5. A method according to any of the preceding claims, wherein the expected blood coagulation value is determined by reading a value stored in for instance a memory of the first computer system/pattern determination system.

6. A method according to any of the preceding claims, wherein the method further comprises the step of outputting to a first transmitter means the measured blood coagulation value and if determined also the at least one dose and at which time the dose is to be taken and/or transmitting the at least one dose and at and which time the dose is to be taken and, optionally, the data relating to the performance of the method to a second computer system/pattern determination system.

7. A method according to claim 6, which method further comprises the steps of verifying, by use of the second computer system/pattern determination system the at least one dose and at which time the dose is to be taken determined by and received from the first computer system/pattern determination system, and if the dose and time to take the dose is verified then transmitting acceptance data to the first computer system/pattern determination system, or if if the dose and time to take the dose is not verified then transmitting rejection data to the first computer system/pattern determination system.

8. A method according to claim 7, wherein the method further comprises the steps of, if the acceptance data has not been received, such as received successfully, by the first computer system/pattern determination system from the second computer system/pattern determination system, then transmitting criticality data to the second computer/pattern determination system and outputting a criticality response .

9. A method according to claim 8, wherein the criticality data is outputted to a second transmitter means for transmitting the criticality data to the second computer system.

10. A method according to any of the preceding claims, wherein the method further comprises the steps of determining a new point in time for determining at least one dose of a drug and at which time the dose is to be taken , outputting this new point in time at the patient, and outputting this new point in time to the first transmitter means.

11. A method according to any of the claims 3-10, wherein the at least one dose of a drug and at which time the drug is to be taken is outputted at the patient only if acceptance data is received by the receiver means.

12. A method of verification of the method according to any of the preceding claims for determining a first at least one dose of a drug and at which time the dose is to be taken, the method comprises the steps of

receiving a measured bloodcoagulation value - determining at least one dose of a drug and at which time the dose is to be taken by use of a second computer system/pattern determination system; said determination being based on at least the received blood coagulation value receiving at least one dose of the drug and at which time the dose is to be taken, determining the differences between the determined and the received doses and at which time the doses are to be taken and if one or both of said differences are below predefined values, preferably being predefined by a medical adviser, then transmitting acceptance data to a receiver means of the first computer system/pattern determination system or if one or both of said differences are above predetermined values then transmitting rejecting data to the receiver means of the first computer system/pattern determination system.

13. A method according to claim 12 further comprising the steps of - receiving an expected measured blood coagulation value, determining an expected measured blood coagulation value, determining the difference between the received and the determined expected blood coagulation value and if the difference is above a predefined value, preferably predefined by a medical adviser, then transmitting rejection data, signalling that the received expected blood coagulation value is erroneous, to the receiver means of the first computer system/pattern determination system.

14. A method according to claim 13, wherein the rejection data is transmitted to a first rejection means, preferably at a patient, and to a second rejection means, preferably distant from the patient.

15. An apparatus for determining a least one dose of a drug and at which time the dose is to be taken, the apparatus comprises measuring means for measuring a blood coagulation value, means adapted to determining an expected blood coagulation value, means for determining the difference between the expected blood coagulation value and the determined blood coagulation value, and means adapted to determining the at least one dose of a drug and at which time the dose is to be taken

16. An apparatus according to claim 15 further comprising means adapted to inhibit determination of the at least one dose of a drug and at which time the dose is to be taken if the difference between the measured and expected blood coagulation value exceeds a pre-set value, preferably pre-set by a medical adviser.

17. An apparatus according to claim 15 or 16, further comprising means adapted to inhibit determination of the expected blood coagulation value if the measured blood coagulation value is not within a certain predetermined range, preferably predetermined by a medical adviser.

18. An apparatus according to any of the claims 15-17, said apparatus further comprises outputting means for outputting at a patient, if determined, the at least one dose of a drug and at which time the dose is to be taken.

19. An apparatus according to any of the claims 15-18, said apparatus further comprises first transmitter means and first receiver means, said first transmitter means being adapted to transmit data to a second means adapted to determine at least one dose of a drug and at which time the dose is to be taken, and said first receiver means being adapted to receive data from the second means adapted determine at least one dose of a drug and at which time the dose is to be taken.

20. An apparatus according to claim 19, said apparatus further comprises second transmitter means for transmitting data to the second means adapted to determine the at least one dose of a drug and at which time the dose is to be taken so that data may be transmitted to said second means independently of the first transmitter means.

21. An apparatus for verification of at least one dose of a drug and at which time the dose is to be taken, the apparatus comprises receiver means for receiving a measured blood coagulation value, preferably measured at a patient, and at least one dose of a drug and at which time the dose is to be taken, means adapted to determine an expected measured blood coagulation value.means adapted to determine the difference between the measured blood coagulation value and the expected blood coagulation value, and means adapted to determining at least one dose of a drug and at which time the dose is to be taken.

22. An apparatus according to claim 21 , said apparatus further comprises means adapted to determine the difference between the received and determined doses of a drug and at which time the doses are to be taken and adapted to generate acceptance data if one or both of these differences are below pre-set limits, preferably pre-set by a medical adviser and adapted to generate criticality data if one or both of the difference(s) is(are) above pre-set limits.

23. An apparatus according to claim 22, said apparatus further comprises transmitter means adapted to transmit the acceptance data and/or the criticality data so that these data may be transmitted to means, preferably at a patient, adapted to handle these data in a predetermined manner and/or to means distant from the patient adapted to handle these data in a predetermined manner.