WO2007093482A1

WO2007093482A1 - A device and a method for managing data relating to blood glucose level for a person

Info

Publication number: WO2007093482A1
Application number: PCT/EP2007/050584
Authority: WO
Inventors: Jette RANDLØV
Original assignee: Novo Nordisk A/S
Priority date: 2006-02-16
Filing date: 2007-01-22
Publication date: 2007-08-23

Abstract

A device and a method for managing data relating to blood glucose (BG) level for a person. The BG level is measured before and after an event to obtain data pairs, and the data pairs are processed, and based on the processed data pairs, graphics and/or text is displayed on a display means, the graphics/text reflecting a correspondence between data obtained before and after the event. Thereby the shown information relates the BG levels to the time of the event, rather than to a specific time of the day. It is possible to derive from the information what typically happens to the BG level in connection with the event. This is a valuable tool when deciding whether or not changes to the behaviour of the person should be made, and what kind of changes. Changes in behaviour may include changing insulin dose, doing more exercise, changing meal contents, etc. The event may be a meal, doing exercise, injecting fast acting insulin, etc. Very suitable for controlling BG level for diabetics or pre-diabetics.

Description

A DEVICE AND A METHOD FOR MANAGING DATA RELATING TO BLOOD GLUCOSE LEVEL FOR A PERSON

FIELD OF THE INVENTION

The present invention relates to a device and a method for managing data relating to blood glucose (BG) level for a person. More particularly, the present invention relates to a device and a method for managing such data obtained before and after an event, respectively, for the person. The event may, e.g., be a meal or doing exercise.

BACKGROUND OF THE INVENTION

Some chronic diseases require strict control, e.g. in the form of regular medication and/or life style management. An example of such a disease is diabetes which requires continuing medical care and self-management by the person having the disease in order to avoid complications. Persons with type 1 diabetes and many persons with type 2 diabetes administer insulin as part of their diabetes treatment plans.

In order for the person having diabetes to be able to perform strict control, it is necessary to measure the level of blood glucose (BG) in the bloodstream a number of times during the day. Thereby the BG level can be monitored in order to avoid that too low (hypoglycaemia) or too high (hyperglycaemia) BG levels occur, and in order to ensure that insulin intake is adjusted to match the present BG level. To this end persons having diabetes normally keep a logbook or diabetes diary containing information relating to measured BG levels, food intake (e.g. in the form of carbohydrates consumed), exercise, injection site, injection timing (i.e. before or during a meal), insulin intake including type of insulin, e.g. bolus injection or background injection, long acting, medium acting or short acting insulin, etc.

It is further desirable to be able to present the data normally kept in a logbook or diabetes diary in a manner which allows the person, or medical staff being responsible for the treatment or disease management of the person, to draw conclusions and make decisions, e.g. relating to insulin intake, food intake and/or exercise for the person. To this end various statistical means have been adopted. For instance, a report may be generated and displayed which shows the person's BG level at various times of the day and indicate any undesired highs or lows. Similarly, a report may be generated and displayed which shows the person's food intake, insulin intake, etc. These reports may be in the form of text or a suitable graphical representation, such as a bar graph, a pie chart, a histogram, etc. One such useful report is the so-called modal day report. In this kind of report, data relating to several days is displayed versus the time of the day. Thereby data relating to many different days is superimposed, and this allows the person to spot possible patterns in the data. This daily trend plot helps in glycaemic control vis a vis the daily activities of the person. In a modal day plot, a user (e.g. the person in question or health care personnel) can select the period range, e.g. day, week, month, quarter, year, etc., for data points to be analysed. A target/desirable range can be decided and the analysis of data points can be done keeping those points into consideration. Furthermore, it may be possible to generate a statistical summary report.

T. Deutsch et al., Time series analysis and control of blood glucose levels in diabetic patients', Comput. Methods Programs Biomed. 41 (1994), 167-182, describes features of a computer-based decision support system for assisting in the management of insulin- dependent diabetic patients. The two major features are time series analysis of blood glucose data, and their interpretation in relation to the provision of advice for controlling the patient's blood glucose level. The time series analysis and the interpretation are based on modal day reports.

A drawback of the modal day based representation is that it only shows the data versus the time of day. It is often more relevant to relate the data to specific events, such as meals or exercise, and since the time of such events may vary from one day to another, the modal day based representation is not capable of showing such a relation.

It is therefore desirable to be able to display data of the kind defined above relatively to the time of certain events. It is also desirable to be able to display such data in a processed form which further facilitates making decisions regarding lifestyle, insulin intake, etc.

SUMMARY OF THE INVENTION

It is, thus, an object of the invention to provide a device and a method which is capable of processing data relating to blood glucose level for a person and of displaying the processed data in a manner which relates to an event for the person.

It is a further object of the invention to provide a device and a method which facilitates diabetic control relatively to events.

According to a first aspect of the invention the above and other objects are fulfilled by providing a device comprising : display means for displaying graphics and/or text,

processing means being interfaced with said display means, the processing means being adapted to causing the display means to display graphics and/or text,

wherein the processing means is adapted to processing data pairs relating to blood glucose level for a person, said data pairs being obtained before and after an event, respectively, for said person, and wherein the processing means is adapted to causing the display means to display graphics and/or text reflecting a correspondence between data obtained before and after said event, based on said processed data pairs.

According to a second aspect of the invention, the above and other objects are fulfilled by providing a method for managing data, the method comprising the steps of:

obtaining data pairs relating to blood glucose level for a person, said data pairs being obtained before and after an event, respectively, for the person,

- processing the data pairs,

- displaying graphics and/or text reflecting a correspondence between data obtained before and after said event, based on the processed data pairs.

It should be noted that a person skilled in the art would readily recognise that any feature described in connection with the first aspect of the invention could also be combined with the second aspect of the invention, and vice versa.

The display means may be or comprise a screen, e.g. of the kind which may be used in cellular phones, personal digital assistants (PDA's), computers, etc. Graphics displayed on the display means may, e.g., be diagrams, charts, graphs, histograms, etc. presenting processed data in a graphical manner.

The processing means may be or comprise a processor, such as a central processing unit (CPU). The processing means is adapted to receive the data pairs, process the data pairs in a desired manner and to send a suitable output to the display means in order to cause the display means to display a graphical and/or textual representation of the processed data pairs. The data pairs relate to blood glucose level for a person, and the data pairs have been obtained before and after an event, respectively. Thus, each data pair represents the blood glucose level for the person before a certain event occurs and after the event has occurred. Thus, processing data pairs of this type provides information as to how the blood glucose level for the person changes in connection with the event, and whether or not it changes at all.

Thus, the obtained data relates the blood glucose level and variations in the blood glucose level to a specific event rather than to a specific time of the day. This is very advantageous since this relationship is most often much more relevant for diabetic control. Furthermore, the fact that the graphics and/or text which is displayed by the display means reflects a correspondence between data obtained before and after the event, respectively, ensures that the displayed graphics and/or text reflects what happens to the blood glucose level of the person when the event occurs. Thus, when data pairs relating to a specific event (e.g. a meal, such as breakfast, lunch or dinner, or doing exercise) occurring on a number of different days have been processed, and the result of this processing is displayed by the display means, it is possible to derive what typically happens with the blood glucose level during the corresponding event. Such information is very valuable in order to assess whether or not long term changes need to be made to the lifestyle, insulin intake, etc. for the person in order to improve the blood glucose level around that event. Furthermore, presenting these results in a manner which is easy to overview provides a very valuable tool for self- monitoring and self-management.

The data may be the result of blood glucose measurements, e.g. performed by the person himself or herself. Such data may have been obtained previously by means of another device, and the measured data may subsequently have been transferred to the device according to the invention for processing and possibly storage. Alternatively the device according to the invention may be a device suitable for performing the measurements, in which case it is not necessary to transfer the measured data to the device before the processing takes place.

The device and the method according to the invention are very suitable for helping diabetics in controlling their treatment and lifestyle. However, the device and the method may also be used by other persons who need to closely monitor their BG level. This is, e.g., the case for persons who are at risk of becoming diabetics, such as so-called ^λpre-diabetics' having a raised BG level and being in the process of developing type 2 diabetes. Such persons need to monitor their BG level carefully in order to control their lifestyle in a manner which prevents the development of type 2 diabetes, or at least postpones the occurrence of this to the maximum possible extent. The event may advantageously be a meal consumed by the person, such as a habitual meal. In the present context the term 'habitual meal' should preferably be interpreted as a meal which the person normally consumes, e.g. breakfast, lunch, or dinner. However, the data pairs may be obtained before and after any meal the person consumes, including snacks consumed between the main meals. In case the event is a meal, the time of the meal may either be marked manually by the user, or it may be automatically recorded as the time of a bolus injection.

Alternatively, the event may be the person doing exercise, injection of a correction injection, injection of a basal insulin dose, injection of fast acting insulin preceded or unpreceded by an air shot, any event marked as 'stressful', and/or any other event having influence on the glucose level of a person.

The display means may be adapted to display graphics, and the processing means may in this case be adapted to causing the display means to display a two-dimensional plot of corresponding data obtained before and after the event, respectively, said two-dimensional plot being based on the processed data pairs. Such a two-dimensional plot is a very useful tool for assessing what typically happens to the BG level of a person when a specific event occurs. Thus, such a two-dimensional plot will readily reveal whether the BG level before the event is typically high or low, whether the BG level after the event is typically high or low, and whether a high BG level before an event will typically be followed by a high or a low level after the event, etc. This is very advantageous.

The two-dimensional plot may, e.g., be of a kind where BG level before the event is plotted along a first axis and BG level after the event is plotted along a second axis being orthogonal to the first axis. Thereby corresponding values of BG levels measured before and after a specific event occurred, are represented by a single point in the plane defined by the two axes. Various points in the plane will represent corresponding measurements as described above obtained at various days, but all relating to the same event, e.g. a specific meal, such as breakfast. By looking at where in the two-dimensional plane the points are grouped, and whether or not they are grouped at all, the user is readily provided with information relating to how the BG level typically changes in connection with the event in question.

Alternatively, any other suitable two-dimensional representation of the processed data pairs may be used. One example is a plot showing four time periods corresponding to the three normal habitual meals (i.e. breakfast, lunch and dinner) and bedtime. Each of four orthogonal axes of the plot, extending from the origin of the plot, represents a meal (or bedtime) in such a manner that time advances in a clockwise direction. Thereby each quadrant defined by the axes represent the time between the meals represented by the two axes positioned adjacent to that specific quadrant. Within each quadrant a plot similar to the one described above may be plotted. Thereby a two-dimensional plot is obtained, which in a clear manner shows BG data relating to the whole day.

Furthermore, according to one embodiment of the invention it is possible to select the data which should be used for the processing. Such a selection may, e.g., relate to time, i.e. data measured during a specific time window, e.g. the preceding week or month, or a specific week last month, etc. In this case only measurements obtained during the specified time window will be used for processing, and the resulting displayed graphics and/or text will only reflect BG levels for the person during that time window. Alternatively, the selection may be made using other criteria. For instance, it may be desirable to investigate what happens to the BG level in connection to the event, e.g. how the BG level develops after the event, when the BG level before the event was relatively high. In this case it is possible to select all the data pairs where the BG level before the event was above a specific threshold value. Using a two-dimensional plot as described above, the selected and processed data will immediately reveal a typical pattern to the user. Selection of data pairs may alternatively be done based on other criteria, e.g. meals where more than two courses have been consumed or meals with wine.

The processing means may further be adapted to derive, from the two-dimensional plot, a main trend in the correspondence between data obtained before and after said event. In this case the processing means may be adapted to causing the display means to display said derived main trend. According to this embodiment, a main trend, i.e. a typical pattern, is derived and displayed or highlighted to the user. Thereby the main trend or typical pattern will be very clear to the user, even if the user is not used to or skilled in looking at a two- dimensional plot and deriving a conclusion directly based on that. Looking at the plotted main trend it will in most cases be possible to directly see if the present lifestyle and medication are acceptable, or whether improvements are possible. Furthermore, points in the two- dimensional plot which do not follow the main trend or typical pattern will clearly stand out, and it will therefore be possible to pay special attention to such points. This will be described further below.

The main trend is preferably a small area of the plot having the most points. This may be calculated or determined in a number of various ways. It may advantageously be done using a standard statistical method, e.g. finding the linear regression line of the two-dimensional plot, or determining the best second order polynomial fit. It may, e.g., be determined that if a set percentage of the points lie within ±2mM of the fitting curve, then the fitting curve is the main trend. The processing means may further be adapted to analysing points in the two-dimensional plot which do not follow the derived main trend. For instance, the main trend may be that the BG level before the event varies relatively much, and that the BG level after the event is relatively stable at a relatively low value. However, the two-dimensional plot may also reveal that a few data pairs do not follow this trend, e.g. a low BG level before the event is suddenly followed by a high BG level after the event. Such derivations from the main trend may represent dangerous exceptions which need attention, and it is therefore desirable to be able to process the data relating to such points separately, and to include the result in the conclusion regarding possible changes in the lifestyle, medication, etc. of the person, thereby avoiding possible adverse effects of the situations causing the exceptions.

Alternatively or additionally, the display means may be adapted to display text, and the processing means may be adapted to causing the display means to display a behavioural advice to the person, said behavioural advice being based on the processed data pairs. According to this embodiment the processing means analyses the data pairs, and based on this analysis a behavioural advice is generated and displayed on the display means. The term 'behavioural advice' should be interpreted to mean an advice to the person regarding how to behave in order to improve control of the BG level. Such an advice may, e.g., be 'increase/decrease the bolus dose of insulin before lunch', 'reduce the carbohydrates consumed at dinner', 'increase exercise', 'increase/decrease the background (long acting) insulin dose', 'advance the time for bolus injection', etc. Displaying a behavioural advice as described above has the advantage that the person will readily know exactly what to do in order to improve control of the BG level, even if the person does not possess special knowledge about the various mechanisms and parameters which influence the BG level. Thus, even if the person observes that a graphical representation of the processed data seems to indicate that improvements are desirable, or even necessary, he or she may not know which parameters to adjust and how, in order to improve control of the BG levels. A textual message as described above provides this information to the person, and by following the behavioural advice, the person will improve the control of the BG levels. This is particularly advantageous in case the device is used in connection with self-treatment, i.e. the person himself or herself performs the measurements and adjusts his or her behaviour without the intervention by medical staff.

The processing means may be adapted to processing data pairs relating to blood glucose level and being obtained before and after two or more events, respectively, for the person. The processing means may further be adapted to causing the display means to display graphics and/or text reflecting a correspondence between data obtained before and after each of said events, based on said processed data pairs. In this case data pairs relating to different events may be displayed by the display means, e.g. breakfast as well as lunch and/or dinner and/or exercise. Thereby the person may change his or her behaviour relatively to the events where it turn out to be necessary, but remain his or her behaviour relatively to the events where the control of the BG level seems to be optimal or close to optimal.

The processing means may further be adapted to processing data relating to blood glucose level obtained at bedtime for the person, and to causing the display means to display graphics and/or text reflecting said processed data. According to this embodiment, the display means is capable of displaying processed data relating to one or more events, e.g. one or more meals, as well as data relating to bedtime. Preferably, the display means is capable of displaying four different screens, one for each of the usual habitual meals, i.e. breakfast, lunch and dinner, and one relating to bedtime. For the habitual meals, the data is processed with reference to the time where the habitual meal is actually consumed, while the data relating to bedtime is processed with reference to the time of measurement.

The device may be or form part of a medical device, such as an injection device, a device for measuring BG values, etc. Alternatively, the device may be or form part of an electronic device, such as a hand held or portable electronic device, e.g. a personal digital assistant (PDA), a cellular phone, a laptop computer, etc. Alternatively, the electronic device may be or form part of a stationary personal computer (PC) or stationary medical equipment of the kind being permanently installed in a hospital.

BRIEF DESCRIPTION OF THE DRAWINGS

The invention will now be further described with reference to the accompanying drawings in which:

Fig. 1 is a table illustrating a possible distribution of BG levels before and after a habitual meal,

Fig. 2 shows measured BG values relating to breakfast, lunch, dinner and bedtime, as well as corresponding two-dimensional plots of corresponding processed data,

Figs. 3-8 show various possible main trends of the two-dimensional plots shown in Fig. 2,

Fig. 9 shows a two-dimensional plot of the kind shown in Fig. 2, displaying a main trend as well as deviations from the main trend, and Figs. 10-12 show possible main trends relating to injection of fast acting insulin.

DETAILED DESRIPTION OF THE DRAWINGS

Fig. 1 is a table illustrating a possible connection between BG levels measured before and after a habitual meal for a person. For this particular person, it is expected that in 10% of the data pairs, the BG level will be high before as well as after the habitual meal, in 4% of the data pairs, the BG level will be high before, but low after the habitual meal, in 80% of the data pairs, the BG level will low before, but high after the habitual meal, and in 6% of the data pairs, the BG level will be low before as well as after the habitual meal. The values shown in the table appear as a result of a processing of measured data pairs. In this case 'high' or 'low' may be defined by whether a measured value is above or below a suitable threshold value. The table of Fig. 1 thereby represents a very simple display of processed data pairs reflecting a correspondence between data obtained before and after the habitual meal, respectively.

The table shown in Fig. 1 reveals that in the majority of the data pair, the BG level is low before the habitual meal and high after the habitual meal. Thus, this may be regarded as a 'main trend', and this 'main trend' can easily be derived from the table.

The upper part of Fig. 2 shows so-called 'prandial plots' relating to breakfast 1, lunch 2, dinner 3 and bedtime 4, respectively, for a person. For breakfast 1, lunch 2 and dinner 3 measurements of BG level before and after each of the habitual meals is shown relatively to the time of the meal, '0' on the first axis defining the time of the meal in question, and the units on the first axis being hours before and after the meal.

The lower part of Fig. 2 shows two-dimensional plots relating to the prandial plot immediately above each of the two-dimensional plots. Thus, the two-dimensional plots relate to breakfast 5, lunch 6, dinner 7 and bedtime 8, respectively. In the two-dimensional plots 5, 6, 7, 8, the first axis represents BG level before the habitual meal, and the second axis represents BG level after the meal. The bedtime plot 8 shows the bedtime BG level, i.e. the latest measurement in a time interval and the first measurement the following morning. Thus, a point in one of the two-dimensional plots represents a data pair comprising a BG level before a particular meal and a BG level after that meal, and it can readily be derived what typically happens to the BG level when a habitual meal is consumed. Thus, the two-dimensional plots 5, 6, 7, 8 show how the BG levels before and after a meal are connected.

In the two-dimensional plots 5, 6, 7, 8 the frequency of a specific combination of BG levels before and after a meal may be shown by varying the colour of the corresponding point. In Fig. 2 this is done using greyscales, i.e. the darker a specific point, the more frequently that specific combination of BG levels occurs. In the example shown in Fig. 2, the scale far to the left shows that the darkest colour correlates that 10% of the total data-pair plots for the viewed mealtime lies in that one specific point, the grey scale ranging down to 0%.

Fig. 3 shows one example of a possible main trend occurring in a two-dimensional plot of the kind shown in Fig. 2. In Fig. 3 a pattern is centred around the diagonal of the displayed plot. Thus, a low BG level before the meal results in a low BG level after the meal, a medium BG level before the meal results in a medium BG level after the meal, and a high BG level before the meal results in a high BG level after the meal. In other words, consuming the meal has virtually no effect on the BG level. This indicates that the person is most likely not including a BG value measured immediately prior to the meal when the bolus insulin dose is calculated. However, the bolus insulin dose generally corresponds very well to the intake of carbohydrates during the meal. This person would probably benefit from taking the measured BG level into account when the bolus insulin dose is calculated. This should be done by taking a higher dose than the dose corresponding to the meal if the measured BG level before the meal is relatively high, and possibly a lower dose if the measured BG level before the meal is very low. If the measured BG level before the meal is low, the bolus insulin dose should still correspond to the meal.

Fig. 4 shows another example of a possible main trend occurring in a two-dimensional plot of the kind shown in Fig. 2. In Fig. 4 the main trend is a substantially vertical line in the left part of the plot. Thus, the BG level before the meal is typically in the target range 4-8 mM, but the BG level after the meal varies a lot from very low to very high. This indicates that the contents of carbohydrates in this habitual meal are often estimated incorrectly, and the bolus insulin dose injected before the meal therefore often does not correspond to the meal. The person could therefore most likely improve control of the BG level by estimating the contents of carbohydrates of this habitual meal more carefully, and by adjusting the bolus insulin dose accordingly.

Fig. 5 shows yet another example of a possible main trend occurring in a two-dimensional plot of the kind shown in Fig. 2. In Fig. 5 the main trend is a substantially horizontal line in the lower part of the plot. Thus, the BG level before the meal varies a lot from very low to very high. However, the BG level after the meal is maintained relatively stable at a relatively low level. This indicates that in connection with this habitual meal, the contents of carbohydrates of the meal are most often estimated correctly, and that the bolus insulin dose injected before the meal is set with due respect to the estimated contents of carbohydrates as well as to the measured BG level before the meal. Thus, the person already has a good control of the BG level relatively to this meal. However, the strongly varying BG level before the meal may constitute a problem, and it may be possible to improve the control of the BG level by adjusting the treatment relatively to the habitual meal preceding the habitual meal relating to the shown plot, or by adjusting the behaviour during the time interval between the previous meal and the meal illustrated in Fig. 5. Thus, the main trend shown in Fig. 5 indicates to the person, that he or she should look further into the preceding habitual meal and the time between the preceding meal and the present meal.

Fig. 6 shows yet another example of a possible main trend occurring in a two-dimensional plot of the kind shown in Fig. 2. In Fig. 6 the main trend is a concentrated 'lump' of points positioned in the upper right corner. Thus, the BG level is relatively high before as well as after the meal, and the person is not successful in controlling the BG level. This is a very serious situation, and the person should therefore consider consulting health personnel in order to initiate substantial behavioural changes. The situation illustrated in Fig. 6 may, e.g., occur if the person is out of control most of the time. This main trend indicates that treatment adjustment may be appropriate, and the person should therefore consider consulting health care personnel.

Fig. 7 shows yet another example of a possible main trend occurring in a two-dimensional plot of the kind shown in Fig. 2. In Fig. 7 the main trend is a slightly increasing line positioned in the lower part of the plot. Thus, the BG level before the meal is varying from relatively low to relatively high. The BG level after the meal is generally similar to or lower than the BG level before the meal. However, there is a tendency that a relatively high BG level before the meal results in a higher BG level after the meal than it is the case when the BG level before the meal is relatively low. This may be due to the insulin sensitivity factor being estimated incorrectly, i.e. a too large value is estimated. Thus, it may be possible to improve control of the BG level by correcting high BG levels before the meal more aggressively. Furthermore, it may be possible to even further improve control of the BG level by looking further into the preceding habitual meal and/or the time interval between the preceding meal and the present meal, as described above, in order to obtain a more constant and relatively low BG level before the meal.

Fig. 8 shows yet another example of a possible main trend occurring in a two-dimensional plot of the kind shown in Fig. 2. In Fig. 8 the main trend is a slightly decreasing line positioned in the lower part of the plot. Thus, the BG level before the meal is varying from relatively low to relatively high. The BG level after the meal is relatively low. However, there is a tendency that a relatively high BG level before the meal results in a very low BG level after the meal. Thus, it seems that the higher BG levels before the meal are Overcompensated'. This may be due to the insulin sensitivity factor being estimated incorrectly, i.e. a too small value is estimated. A very low BG level may be dangerous, and an adjustment of the BG level after the meal when the BG level before the meal was relatively high may therefore be desirable. Thus, is may be possible to improve control of the BG level by correcting high BG levels before the meal less aggressively. Furthermore, it may be desirable to look into the preceding habitual meal and/or the time between the preceding and the present habitual meal, as described above.

Fig. 9 shows a two-dimensional plot of the kind shown in Fig. 2, displaying a main trend as well as deviations from the main trend. In the plot of Fig. 9, most of the points are located along a substantially horizontal line positioned in the lower part of the plot. Thus, the main trend corresponds to the main trend shown in Fig. 5. However, a few points 9 do not follow the main trend. These points 9 are positioned in the upper left corner of the plot. Thus, they represent a situation where a relatively low BG level before the meal results in a relatively high BG level after the meal. If a behavioural advice is given based on the main trend, such deviations may represent dangerous exceptions, especially if they are of a kind which results in lower BG levels than expected. In the example shown in Fig. 9, the exceptions 9 are in the hypo-area, suggesting that the person overcompensated when the bolus insulin dose was calculated.

Thus, it may be desirable to take exceptions from the main trend into account when the results are analysed. An analysis of exceptions may comprise the following points.

1. Determining how many exception there are.

2. Determining to which extent the exceptions deviate from the main trend pattern.

3. Determining to which extent the exceptions deviate from the diagonal of the plot.

4. Determining the BG level before the meal giving rise to the exceptions.

5. Assigning a weight to the exceptions and taking the weighted exceptions into account when deciding whether or not and how the person should adjust his or her behaviour in order to improve control of the BG level.

The weight of the exceptions may advantageously be assigned on the basis of a combination of the number of exceptions and their distance to the main trend and/or the diagonal. The obtained number may be multiplied by a fixed number if the exceptions result in a lower BG level than expected. Fig. 10 shows an example of a possible main trend relating to injection of fast acting insulin. In the case illustrated in Fig. 10 each injection was preceded by an air shot. The first axis represents BG level before the injection and the second axis represents BG level after the injection. It appears from the Figure that the injection has very little influence on the BG level in that a relatively low BG level before the injection results in a relatively low BG level after the injection, a relatively high BG level before the injection results in a relatively high BG level after the injection, and a medium BG level before the injection results in a medium BG level after the injection.

Fig. 11 also shows an example of a possible main trend relating to injection of fast acting insulin. However, in this case the injections were not preceded by an air shot. It appears that when the injection of fast acting insulin is not preceded by an air shot, as illustrated in Fig. 11, it is more difficult to control the BG level than when the injection is preceded by an air shot, as illustrated in Fig. 10. Thus, it seems that the range of BG values, in particular after the injection, is somewhat larger in Fig. 11 than in Fig. 10.

This is more clearly illustrated in Fig. 12 showing the result of a subtraction of the main trend illustrated in Fig. 11 and the main trend illustrated in Fig. 10. The remaining main trend, i.e. the marked area of Fig. 12, represents data pairs which are present in Fig. 11, but not in Fig.

10. It is clear that these data pairs represent relatively high BG levels after the injection, i.e. injecting fast acting insulin without a preceding air shot tends to result in a higher after-event BG level. Accordingly, the main trend of Fig. 12 may result in an advice to the user to always use air shots prior to injection of fast acting insulin.

Claims

1. A device comprising:

- display means for displaying graphics and/or text,

wherein the processing means process data pairs relating to blood glucose level for a person, said data pairs being obtained before and after an event, respectively, the processing means cause the display to display a two dimensional plot reflecting the correspondence between data obtained before and after said event,

characterised in that the processing means is further adapted to derive, from said two- dimensional plot, a main trend in the correspondence between data obtained before and after said event, and wherein the processing means is adapted to cause the display means to display said derived main trend.

2. A device according to claim 1, wherein the event is a meal consumed by the person or a person doing exercise.

3. A device according to claim 1 or 2, wherein the processing means is further adapted to analysing points in the two-dimensional plot which do not follow the derived main trend.

4. A device according to any of the preceding claims, wherein said main trend is that a certain blood glucose level before an event corresponds to a certain blood glucose level after said event.

5. A device according to a claim 4, wherein said main trend is that a low blood glucose level before an event corresponds to a low blood glucose level after said event, a medium blood glucose level before an event corresponds to a medium blood glucose level after said event and a high blood glucose level before an event corresponds to a high blood glucose level after an event.

6. A device according to a claim 4, wherein said main trend is that a stable blood glucose level before an event corresponds to a wide range of blood glucose levels after said event.

7. A device according to a claim 4, wherein said main trend is that a wide range of blood glucose levels before an event corresponds to a stable blood glucose level after said event.

8. A device according to a claim 4, wherein said main trend is that a high blood glucose level before an event corresponds to a high blood glucose level after said event.

9. A device according to any of the preceding claims, wherein the display means is adapted to display text, and wherein the processing means is adapted to causing the display means to display a behavioural advice to the person, said behavioural advice being based on the processed data pairs.

10. A device according to any of the preceding claims, wherein the processing means is adapted to processing data pairs relating to blood glucose level and being obtained before and after two or more events, respectively, for the person, the processing means further being adapted to causing the display means to display graphics and/or text reflecting a correspondence between data obtained before and after each of said events, based on said processed data pairs.

11. A device according to any of the preceding claims, wherein the processing means is further adapted to processing data relating to blood glucose level obtained at bedtime for the person, and to causing the display means to display graphics and/or text reflecting said processed data.

12. A device according to any of the preceding claims, wherein the device is or forms part of a medical device.

13. A device according to any of the preceding claims, wherein the device is or forms part of an electronic device.

14. A method for managing data, the method comprising the steps of:

- processing the data pairs,

- displaying graphics and/or text reflecting a correspondence between data obtained before and after said event, based on the processed data pairs, displaying a two dimensional plot of corresponding data based on the processed data pairs,

derive and display a main trend in the correspondence between data obtained before and after said event.