+ All Categories
Home > Documents > My Smart Health: an integrated suite for Remote Self...

My Smart Health: an integrated suite for Remote Self...

Date post: 21-May-2020
Category:
Upload: others
View: 4 times
Download: 0 times
Share this document with a friend
7
My Smart Health: an integrated suite for Remote Self Monitoring of Diabetes Maria Teresa Baldassarre 1 , Giovanni Bruno 2 , Danilo Caivano 1,2 , Gennaro Del Campo 2 , Massimiliano Morga 2 , Giuseppe Visaggio 1,2 Department of Informatics, University of Bari, Italy 1 {baldassarre, caivano, visaggio}@di.uniba.it; SER&Practices Spin Off, Bari, Italy 2 {massimiliano.morga, giovanni.bruno, gennaro.delcampo}@serandpractices.com Abstract—Diabetes mellitus is a constantly increasing disease. Evidences prove how self monitoring of blood glucose (SMBG) is associated to with better metabolic control in Type I diabetes patients. Nonetheless, there are several issues that limit the effectiveness and reliability of self monitoring such as: bulky devices for blood glucose checks, large amounts of data to report on log books and to interpret in real time, limited communication between diabetologist and patient, and scarce availability of data for medical and scientific use in evidence based medicine. “My Smart Health”, proposed in this paper, is an integrated suite for remote SMBG. It aims to improve management of diabetes from several perspectives among which assure reliability of SMBG values collected, appropriateness of therapies identified, as well as timely communications with patients. The solution provides innovative software components and discrete hardware devices that enforce the traditional processes for blood glucose control and, at the same time, reduce the social impact of the disease, support the stakeholders involved (i.e. patients, diabetologists and GP) in defining and adopting the most appropriate therapy, readily analyze the data and assure real time feedback to patients after each blood glucose check. Index Terms—Diabetes, SaaS, Evidence Based Medicine, blood glucose self monitoring I. INTRODUCTION Diabetes mellitus is a chronic pathology largely diffused all over the world. According to the WHO, the last estimate of the number of diabetic subjects is about 270 million people [1, 2]. Nevertheless, this disease is increasing and the WHO forecasts that by 2015 the number of diabetics may even double. In Italy, approximately four million patients, with a recognized pathology, and about two million borderline ones, represent the diabetic population. Of this, 53.7% is male, and 59% is over 65 years of age, while about 33.7% is between 45 and 65 years, but over 7% is less than 35. The request for glycaemia strips, reimbursed by the National Health Service, has registered an increasing trend in the last years with a growth of 11% per year and a cost of 600 million euros. There is an increasing trend in the market of self-monitoring diabetes due to the impact of the pathology on the Italian population. In spite of the high cardiovascular risk of diabetic patients, it turns out that lipid profile monitoring (cholesterol and triglycerides) is carried out more systematically than blood glucose monitoring. There are many evidences that prove how self monitoring of blood glucose is associated to a better metabolic control in patients with Type I diabetes, allowing for an adequate adjustment of the insulin doses, on behalf of healthcare providers, and most of all allowing the patient to self adjust the insulin dose according to the value collected. Indeed, such controls (as for glycaemia and arterial pressure) and their joint management with a GP may motivate the patent to constantly conform to a specific lifestyle (diet and exercise) and pharmacologic therapy. In general, the aim of the Health Ministry in facing the diabetes emergency and the metabolism syndrome foresees structuring, for what is possible, local services that are able to reduce inconveniences for users, improve the quality of life of patients, allow for timely interventions of diagnosis and therapy and, in the perspective of reducing healthcare expenses, reduce the number of hospital accesses. Domiciliary self monitoring of blood glucose has provided a very important contribution and is considered as a means for diabetes therapy, that helps and is integrated with other classic instruments such as diet, exercise, medicine and health care. Nevertheless, the other side of the medal consists in the way self monitoring is often carried out and its related issues which can be summarized as follows: Methods and Techniques: there is no assurance and continuity in the control. Having to proceed autonomously on their own and without direct medical control, patients often forget to collect the vital parameters or they invert the values when reporting them on their blood glucose log book. In other cases, for example, when a patient is not at home and forgets something (glucose diary, value reader, pin-stick) he/she is not able to self monitor. Indeed this occurs for one third of the patients. Instruments: collecting a blood sample for the test, by using a lancing device, monitoring system, test strips with control solution, is often lived as a trauma especially by children. Furthermore, each test requires specific ad hoc test strips that are not interchangeable from one device to another. As so, the current devices present many hygienic (sterile lancets and controlled disposal) and logistic limitations (diabetics must always have enough lancets with them), other than the ones related to traumatic and pain aspects. Data Management and Knowledge Assets: doctors encounter objective difficulties in managing the large amount of the data. This is mostly due to the increasing number of diabetic patients and among them, the increasing number of domiciliary self monitoring patients. Each patient carries out from a minimum
Transcript
Page 1: My Smart Health: an integrated suite for Remote Self ...worldcomp-proceedings.com/proc/p2013/BIC3991.pdf · diabetes emergency and the metabolism syndrome foresees structuring, for

My Smart Health: an integrated suite for Remote Self Monitoring of Diabetes

Maria Teresa Baldassarre1, Giovanni Bruno2, Danilo Caivano1,2, Gennaro Del Campo2, Massimiliano Morga2, Giuseppe Visaggio1,2

Department of Informatics, University of Bari, Italy1 {baldassarre, caivano, visaggio}@di.uniba.it;

SER&Practices Spin Off, Bari, Italy2 {massimiliano.morga, giovanni.bruno,

gennaro.delcampo}@serandpractices.com

Abstract—Diabetes mellitus is a constantly increasing disease. Evidences prove how self monitoring of blood glucose (SMBG) is associated to with better metabolic control in Type I diabetes patients. Nonetheless, there are several issues that limit the effectiveness and reliability of self monitoring such as: bulky devices for blood glucose checks, large amounts of data to report on log books and to interpret in real time, limited communication between diabetologist and patient, and scarce availability of data for medical and scientific use in evidence based medicine.

“My Smart Health”, proposed in this paper, is an integrated suite for remote SMBG. It aims to improve management of diabetes from several perspectives among which assure reliability of SMBG values collected, appropriateness of therapies identified, as well as timely communications with patients. The solution provides innovative software components and discrete hardware devices that enforce the traditional processes for blood glucose control and, at the same time, reduce the social impact of the disease, support the stakeholders involved (i.e. patients, diabetologists and GP) in defining and adopting the most appropriate therapy, readily analyze the data and assure real time feedback to patients after each blood glucose check.

Index Terms—Diabetes, SaaS, Evidence Based Medicine, blood glucose self monitoring

I. INTRODUCTION Diabetes mellitus is a chronic pathology largely diffused all

over the world. According to the WHO, the last estimate of the number of diabetic subjects is about 270 million people [1, 2]. Nevertheless, this disease is increasing and the WHO forecasts that by 2015 the number of diabetics may even double. In Italy, approximately four million patients, with a recognized pathology, and about two million borderline ones, represent the diabetic population. Of this, 53.7% is male, and 59% is over 65 years of age, while about 33.7% is between 45 and 65 years, but over 7% is less than 35. The request for glycaemia strips, reimbursed by the National Health Service, has registered an increasing trend in the last years with a growth of 11% per year and a cost of 600 million euros. There is an increasing trend in the market of self-monitoring diabetes due to the impact of the pathology on the Italian population. In spite of the high cardiovascular risk of diabetic patients, it turns out that lipid profile monitoring (cholesterol and triglycerides) is carried out more systematically than blood glucose monitoring.

There are many evidences that prove how self monitoring of blood glucose is associated to a better metabolic control in

patients with Type I diabetes, allowing for an adequate adjustment of the insulin doses, on behalf of healthcare providers, and most of all allowing the patient to self adjust the insulin dose according to the value collected. Indeed, such controls (as for glycaemia and arterial pressure) and their joint management with a GP may motivate the patent to constantly conform to a specific lifestyle (diet and exercise) and pharmacologic therapy.

In general, the aim of the Health Ministry in facing the diabetes emergency and the metabolism syndrome foresees structuring, for what is possible, local services that are able to reduce inconveniences for users, improve the quality of life of patients, allow for timely interventions of diagnosis and therapy and, in the perspective of reducing healthcare expenses, reduce the number of hospital accesses.

Domiciliary self monitoring of blood glucose has provided a very important contribution and is considered as a means for diabetes therapy, that helps and is integrated with other classic instruments such as diet, exercise, medicine and health care. Nevertheless, the other side of the medal consists in the way self monitoring is often carried out and its related issues which can be summarized as follows: Methods and Techniques: there is no assurance and continuity in the control. Having to proceed autonomously on their own and without direct medical control, patients often forget to collect the vital parameters or they invert the values when reporting them on their blood glucose log book. In other cases, for example, when a patient is not at home and forgets something (glucose diary, value reader, pin-stick) he/she is not able to self monitor. Indeed this occurs for one third of the patients. Instruments: collecting a blood sample for the test, by using a lancing device, monitoring system, test strips with control solution, is often lived as a trauma especially by children. Furthermore, each test requires specific ad hoc test strips that are not interchangeable from one device to another. As so, the current devices present many hygienic (sterile lancets and controlled disposal) and logistic limitations (diabetics must always have enough lancets with them), other than the ones related to traumatic and pain aspects. Data Management and Knowledge Assets: doctors encounter objective difficulties in managing the large amount of the data. This is mostly due to the increasing number of diabetic patients and among them, the increasing number of domiciliary self monitoring patients. Each patient carries out from a minimum

Page 2: My Smart Health: an integrated suite for Remote Self ...worldcomp-proceedings.com/proc/p2013/BIC3991.pdf · diabetes emergency and the metabolism syndrome foresees structuring, for

of one, up to to 5-6 checks per day (insulin dependents). Consequently, at each clinical control (usually every 2-3 months) the number of values that the doctor must analyze is considerable and this requires time for consultation. At the same time, given the characteristics of the blood glucose log books (paper, local files), the data contained cannot be shared by the scientific community and allow for large scale studies on the pathologies and the effectiveness of the treatments or therapies adopted.

In this sense, the “My Smart Health (MSH)” solution, illustrated in this paper consists in an integrated system for collecting diabetic patient parameters in mobility. This translates in practice to a better acknowledgement of the disease allowing patients to maintain a high quality of life, reducing at the same time healthcare costs for managing diabetes. MSH aims to improve self-monitoring of glycemic levels with easy to use devices. The benefits in terms of technological innovation are twofold. On one hand it improves patient’s relation with the disease thanks to a discrete and reliable device that provides real time feedback on the measurements carried out; on the other, it allows diabetologists, to rapidly obtain information from the data collected and closely monitor their patients identifying critical situations and improvements on the course of the disease with attention to each single patient.

The rest of the paper is organized as follows: section II sets the state of art on currently available self monitoring systems, points out the benefits and advantages of the solution proposed with respect to the state of art on what literature and market offers; section III describes the MSH solution framework; section IV points out the innovations of the solution proposed and how it overcomes the issues raised in literature; section V illustrates how the concepts of evidence based medicine are implemented in MSH; finally, conclusions are drawn.

II. STATE OF ART ON SELF-MONITORING SYSTEMS Traditional self monitoring devices have a set of problems

that can be classified into three categories: size and low appeal; interpretation of the collected data delegated to the patient; difficulties in the analysis of historical data and its reliability. In the following we provide an analysis of the literature for what concerns these problems related to existing solutions.

A. Size and low appeal Glucometers currently present on the market, although

differing by manufacturer, reliability, speed of measurement and another, often have a non negligible dimension that does not allow for a discrete use of the device. As proof of how this aspect is a known problem, producers of diagnostic systems have attempted to solve it by proposing devices with reduced size. Bayer's CONTOUR® USB [3], is a pen drive that allows to save data via USB with an integrated meter and graphical display; OneTouch Verio® IQ [4] is also an alternative to traditional meters in terms of color, dimensions, and a layout similar to a MP3 player. Although these solutions are more discrete than traditional ones, the reduction of size is minimal and yet to be overcome. This issue should not be underestimated as it directly impacts on the social aspect of the

disease. Often diabetic patients are embarrassed to carry out self monitoring when in public and not at home. This is especially true for young patients and may lead to a non measurement with enormous impact on disease management and prevention of hypo and hyperglycemia events. A recent product characterized by its small dimensions is the GMATE-IRIS meter [5] that transfers data through an audio jack to a last generation smartphone. The counterpart of this solution is that it is only compatible with two smartphones (IPhone 5 and Samsung Galaxy SIII), both quite pricy, making the solution is unfeasible for all users. Another limitation is the fact that the IRIS model recovers the energy it needs for its operations directly from the smartphone battery. It cannot be used connected to any other smartphone.

B. Interpretation of the collected data delegated to the patient As stated previously, measurement of glucose values is a

primary aspect for coping with diabetes. However, it must be supported by an effective interpretation of the collected data, as a mere detection of a value has no meaning and does not produce any effect on diabetes management if it is not contextualized and correctly interpreted. The meters currently on the market show the value to the patient who, if appropriately instructed by their healthcare professional, will be more or like able to adopt corrective actions with respect to the measured data. In each case, a correct interpretation cannot be context independent and requires consideration of other parameters such as type if meal, time from previous or next meal, exercise carried out, mood (stress surely influences glucose levels), other physical parameters such as sex, age, height, correlation with previous measures, trend of the last period, just to mention a few. Many of these parameters are not directly collectable by the patients who, consequently, will most likely not know the relation between measures collected and the appropriate action to undertake. On the other end, even though a diabetic specialist is aware of such correlations and of the patient’s specific case, he is not able to readily analyze all of the measures collected and provide feedback in real time.

C. Difficulties in analyzing historical data and its reliability The two issues discussed above not only impact on the

single measures of self monitoring but also have consequences on the global management of the disease and of its course for a patient. A set of measures not carried out, not appropriately characterized with respect to the parameters mentioned, or falsified by the patient himself represent an information deficit for the diabetologist who, in turn, must analyze the patient’s clinical situation, identify the most appropriate therapy and monitor the evolution of the disease. None of the meters commercially available are able to avoid the risk of falsification of the data. Indeed, a patient can manually report the (falsified) value on the blood glucose log book, or can change the content of the file produced after the measurement in case of electronic diary (falsifying the data).

Furthermore, even correct and timely collected measures are not able to express the entire information potential due to the restrictions of the meters: measures are reported manually in paper log books or extracted by specific software integrated

Page 3: My Smart Health: an integrated suite for Remote Self ...worldcomp-proceedings.com/proc/p2013/BIC3991.pdf · diabetes emergency and the metabolism syndrome foresees structuring, for

with the glucometer and reported in an electronic diary. All of this data is analyzed by the diabetologist during the patient’s medical visit. It is clear that the large amount of data, along with the short time available for the visit represent a limit for a complete and detailed analysis of the data collected over a long time span that goes at least from one visit to another.

Given these considerations, MSH aims to improve self monitoring measures of blood glucose levels, increase the ease of use of the devices involved and improve the general management of the data collected in time. In the next section a detailed description of the solution is provided.

III. MY SMART HEALTH SOLUTION MSH is a solution for self monitoring of blood glucose

levels integrated with hardware and software tools for: synchronous collection and analysis, anti-falsification and proactive verification in mobility of physiological parameters of the metabolic syndrome aligned with various diagnostic therapeutic protocols.

Figure 1 illustrates the logical architecture and deploy of the MSH solution.

Fig. 1 Logical architecture of the MSH Solution

It is made up of the following components:

A. MSH-CENTER This component is a software system provided as Software

as a Service (SaaS) [6, 7] and accessible through a web portal by authorized users (diabetologist, patient, specialized operator, etc.) containing the following information: personal and physiological information of each patient; electronic log book of blood glucose measures; advanced report tools that allow to select and effectively present data included in the electronic log book; etc. Once logged into the system the diabetologist can select a patient from a list of patients and view information such as the blood glucose diary (Fig.2).

MSH-CENTER also includes a Feedback&Communication module used for detailed analysis of collected data, production of statistics and trends that support evidence based medicine [8, 9, 10]. This module will be further detailed in section IV.

B. MSH-MOBILE This part of the system is made up of a software component

installed on a certified and compatible smartphone, and a compatible audio jack able to visualize the physiological data collected by MSH-METER, save it locally in the electronic log book, elaborate and send it to the MSH-CENTER via

GPRS/UMTS. MSH-MOBILE has a visual and vocal interface that guides the patient step by step in carrying out the self monitoring process (Fig.4). It also receives messages from the MSH-CENTER and therefore represents a means for communicating information to the patient (Fig.5).

Fig.3. Blood Glucose diary of a selected patient

Fig.4 Guided interface Fig.5 Message from MSH-CENTER

C. MSH-METER This component is a self powered glucometer connected to

the MSH-MOBILE through audio jack (Fig.6).

Fig.6 MSH METER

The component is able to collect the blood glucose values from a traditional test strip and communicate it to the MSH-MOBILE through the audio jack. A self control method assures

Page 4: My Smart Health: an integrated suite for Remote Self ...worldcomp-proceedings.com/proc/p2013/BIC3991.pdf · diabetes emergency and the metabolism syndrome foresees structuring, for

interaction between MSH-METER and MSH-MOBILE which is installed on the smartphone.

The MSH architecture is based on two technological paradigms: Cloud Computing [11] and Future Internet [12] with particular attention to “internet of things”. The first paradigm simplifies the administration and evolution of the system for doctors, patients and healthcare facility as it is managed by the MSH-CENTER. Furthermore, adoption of Future Internet maintains the communication between MSH-MOBILE, MSH-CENTER, diabetologists and service center in order to better control assistance to patients. This control is very reliable being based on an implementation of “internet of things” between MSH-METER and MSH-CENTER.

IV. INNOVATIONS OF MSH In this section we provide further insight as to how MSH

overcomes the three categories of problems highlighted in section II.

A. Problem I: Size and low appeal MSH-METER is a meter that combines ergonomic design

with reduced dimensions, approximately half the size with respect to other commercially available devices, making it a suitable and discreet solution that preserves the social status of any diabetic patient. It assists self monitoring measurements and sends the data collected to the MSH-MOBILE component. Measurement is completed by simply connecting these two components through the audio jack of any smartphone. The connection allows for data transfer which consequently is: reliable and secure, in that it uses modulation frequency for transmission of signals, a stable and well known technology used for communication. Furthermore, the connection between MSH-METER and smartphone is mechanical, so it is not affected by poor reception or interferences as for wireless or infrared connections; anonymous, as transmission does not use radio frequencies or other types of signals (Bluetooth or WiFi) that may be intercepted and decoded. The data is collected from the MSH-METER and transferred AS-IS to the smartphone through audio signal.

Furthermore, unlike its competitor GMATE-IRIS, MSH-METER does not use the power of the smartphone to function, rather it is self powered by an internal lithium battery. Consequently it can be used on any smartphone with an audio jack that satisfies the OS requirements. The battery begins to supply power only after the self monitoring procedure has been activated through the MSH-MOBILE installed on the smartphone. Tests have shown that a battery lasts for about nine months of therapy of an average patient.

B. Problem II: Interpretation of the collected data delegated to the patient MSH supports patients in interpreting the measurement

values. After having completed the self measurement, MSH-MOBILE contextualizes the data and indicates if it is pre or post prandial as well as the elapsed time from the last meal (breakfast, lunch, dinner, snack). The data is locally saved in MSH-MOBILE and, at the same time, sent to MSH-CENTER and saved. Next, MSH-MOBILE will query MSH-CENTER

with particular reference to the Feedback&Communication component, in order to receive feedback in real time on the measurement sent.

The Feedback&Communication component (Fig.7) is an analytical engine based on rules that allow to analyze the data received in real time and validate it with respect to one or more interpretation protocols defined by the diabetologist.

Fig.7 Feedback&Communication conceptual module

In practical terms, the MSH-CENTER is accessible through a webpage by authorized users, in this case the diabetologist. A protocol is a set of rules that the doctor creates with the assistance of a wizard that guides him in defining the decision rule(s) using a set of conditional operators (<, >, <=, >=) together with additional conditions that are taken into account for interpreting the data. The method and technology used to implement the decision support system is that of decision tables [13]. Here, the conditions are represented by indicators such as blood glucose measurements and eventually also other indicators such as sex, age, previous data values, pharmaceutical treatment undertaken etc.; the conditional states are the possible values for each condition; a rule or action entry is the combination of conditional states, i.e. the measurement values collected for each condition; the actions identified, represent the feedback information that is communicated to the patient and/or doctor after each measurement. The actions are classified according to a chromatic legend from most to less severe code: RED, ORANGE, YELLOW, GREEN. Figure 8 is the graphical interface that the diabetologist uses for defining a protocol in th MSH-CENTER. In this particular case the protocol expresses actions (rules) to undertake depending on the condition “blood glucose value” (<60=RED; between <60-100> = YELLOW; >200=RED; between <130-200>=YELLOW). Each time a protocol is defined, the diabetologist also specifies the recipient of the message among the possible user profiles (patient, doctor, service center, local healthcare unit, etc.) as well as where to send the message (sms, email, smartphone connected to the meter). As last step, the diabetologist assigns a protocol to each patient. To summarize, MSH-MOBILE collects the data from the meter, sends it to MSH-CENTER and queries the decision support system. Depending on the interpretation of the data, a message is sent to a specific recipient through a specific communication channel. So for e.g. if “Blood glucose value = 57”, a RED CODE is activated and both diabetologist and patient are alerted with a message.

All of the communication messages sent out following to interpretations of the therapeutic protocols are saved in the Feedback&Communication module. Each time a measurement is carried out by a patient and transferred by the MSH-MOBILE, it activates a decision rule of the protocol, the

Page 5: My Smart Health: an integrated suite for Remote Self ...worldcomp-proceedings.com/proc/p2013/BIC3991.pdf · diabetes emergency and the metabolism syndrome foresees structuring, for

system saves it in the archive of actions endorsed (made up of a temporal reference, the message sent and the communication channel used for sending it). All of the actions saved can be consulted in a specific section of MSH-CENTER which offers the following functionalities:

• Filter on actions. Actions per patient can be ordered and filtered according to date of issue, chromatic code of alert, etc;

• Detailed consultation of the action undertaken for each measurement collected;

• Mark notifications. The actions to carry out, that have been consulted by a patient can be marked as “already read”.

The “Risks” section of the MSH-CENTER platform provides a summary of all the follow-up initiatives and feedback provided by the system based on the measurements received, according to the protocols defined by the diabetologists and applied to each patient. So, this section allows to monitor patients on a daily basis and readily take action in case of alert cases. Figure 9 summarizes the alert situations that have occurred for patient Mr. Mario Rossi specifying for each case, the date and alert code. It is also possible to have detailed information for each alert by clicking on it.

Fig. 8 Protocol Definition Interface

Fig. 9 Risks module for each patient

In this sense, this module not only provides a reference for the diabetologist who can control the general situation of any patient, but also and especially for a service center that is able to constantly monitor all registered patients in real time and promptly take action in case of emergency.

C. Problem III: Difficulties in analyzing historical data The entire process of self monitoring, cataloging,

collection, analysis, interpretation, and data management that characterizes the MSH-CENTER is a relevant contribution for the analysis of historical data and, consequently, plays an important role in validating the therapeutic treatment of each diabetic patient. Indeed, the MSH platform provides diabetologists a precise electronic log book of each patient without “falsifications” (every measurement reports the exact time it is taken at); furthermore, it also provides a consultation section dedicated to synthesis reports, graphical representations, trends of the data collected from the MSH-MOBILE unit, classified and stored in the MSH data bases of the MSH-CENTER component.

In particular, the module related to “Statistics” allows to define two types of graphs and is divided into two distinct logical areas: one dedicated to the representation of data related to single patients (Fig10), and one for elaborating complex statistics on a wider range of patients (Fig.11) that belong to a specific cohort constantly monitored by the local healthcare center and by a diabetologist.

For what concerns the graphs on single patients, data points refer to a period of thirty days prior to the current one. It is also possible to create other graphs based on longer periods of observation such as monthly, three months, up to a six month period. For example, the graph in Fig.10 provides textual information on the visualization date, reference period, number of days with at least one self monitoring data collected, number of self monitoring glucose level collected over the period, the risk baselines (60-200) in accordance to the OMS/NDDG criteria [14] and to the criteria for diagnosis of diabetes mellitus proposed by the ADA Expert Committee [15]. The graph shows the trend of the measures related to the patient (Mr. Mario Rossi) over at three month observation period. Other trends can be produced with respect to pre-post prandial values, mean, mean +/- standard deviation, or graphs that point out the distribution of the data points within reference intervals of glycemic values (for example: <60; <60-100>; <100-130>…).

It is evident that a graphical visualization of a patient’s clinical situation is an added value for the diabetologist as he/she has a general picture of the trend of measurements in a unique interface that summarizes all the relevant information necessary for decision making, allowing to focus on a single data point if the case. Consequently, these functionalities reduce the effort of the diabetologist for analyzing the clinical data, and the effects of the therapy and eventually make any changes. Although such operations are also possible without the MSH suite, they obviously would request much more work, effort and data elaboration between various devices (paper/electronic log books, spreadsheets with data, statistical tool for elaborations etc.) with high risk of falsifications and errors.

Page 6: My Smart Health: an integrated suite for Remote Self ...worldcomp-proceedings.com/proc/p2013/BIC3991.pdf · diabetes emergency and the metabolism syndrome foresees structuring, for

Fig. 10 Graph on trend of patient’s self monitoring measurements

Fig. 11 Comparison of glucose values during two observation periods

MSH reduces such risks and allows for real time data elaboration and decision making based on the decision support system. Results are accessible via web and therefore consultable by various stakeholders (specialist, local healthcare center, patient, etc.) anywhere.

The graphs related to more complex elaborations on cohorts of patients can be seen as a Business Intelligence [16] subsystem, used by authorized personnel for epistemological analysis and wider investigations on entire populations of diabetic patients monitored by the healthcare center that uses MSH. This component provides various criteria for grouping and filtering data related to the population of patients. Possible filters are for example: age, sex, weight, type of diabetes, etc. So a type of graphical analysis that can be requested ad hoc may be to compare the averages of blood glucose levels of patients classified according to the intervals of the protocol during two different observation periods (Fig.11).

Availability of such information and data is important for the scientific community as it provides evidence on the effectiveness and validity of therapies defined and adopted. This concept is described in more detail in the next section.

V. EVIDENCE BASED MEDICINE IN MSH The MSH solution represents a valid support system for the

diabetologist, as he/she can elaborate complex statistics on an entire cohort of patients and undertake common preventive initiatives and validate their effectiveness and evolution from both an epistemological and scientific perspective. All the information collected, the protocols defined and evolved in time, the decision support system created with decision models, the actions carried out, monitored and controlled in time make up the knowledge base of therapies and contribute to create the Evidence Based Medicine (EBM) [8, 9, 10] component. Moreover, EBM in this specific case constitutes an approach to the clinical practice, where the clinical decisions derive from the integration between the doctor’s experience and the adoption of the best available evidence on therapies or protocols, mediated by the patient’s experience.

This said, the EBM unit inside the proposed solution makes a cluster of statistical instruments available for doctors, diabetoligists, associations and authorized personnel, which are useful for analyzing the enormous quantity of clinical data collected by the system in order to determine, based on founded evidences, the best therapies to adopt. A general idea of the EBM concept is illustrated in Fig.12.

Fig. 12 Knowledge base of consultable therapies /protocols

Page 7: My Smart Health: an integrated suite for Remote Self ...worldcomp-proceedings.com/proc/p2013/BIC3991.pdf · diabetes emergency and the metabolism syndrome foresees structuring, for

As stated in the previous section the therapies/protocols associated to each patient are formalized through a decision table and are part of a decision support system. Consequently, by monitoring diabetic patients, it is possible to compare the effect of the different therapies and determine the best one; by studying a single patient, it is possible to check the effect that variations to a therapy/protocol (formally consisting in the change of a decision table) have on the patient’s health condition. In this way, the doctor can immediately check the cause-effect relations between the therapy changes (for e.g. a different diet or a higher or lower injection of insulin) and the response of the patient. Furthermore, by studying the trend of the examined data, it is possible to go back to the therapy adopted and make the appropriate changes. For example, a trend showing a significant improvement of the glucose level pre and post-lunch, could encourage the doctor to revise the prescribed therapy and analyze its main features. The doctor could then decide to adopt it and experiment it on other patients as well. Finally, the EBM unit provides instruments for generating evidence and critical experience about the current therapies being adopted, and so, it can be seen as a valuable means for supporting continuous improvement of the clinical practice.

VI. CONCLUSIONS “My Smart Health” is a solution made up of hardware

devices and software components that allow data collection, synchronous and asynchronous analysis, proactive checking in mobility of physiological parameters involved in the monitoring process of glycemic values. It is a novel and innovative solution compared to competitor products currently available on the market. It reduces the social impact of diabetes and improves the support provided to patients during the definition and adoption of a diagnostic and therapeutic solution on behalf of all stakeholders such as diabetologists, general med doctors, pharmacists, and all the healthcare centers involved.

The advanced data analysis features along with the decision support system and therapies knowledge base provide considerable support towards scientific research for medical purposes on relevant amounts of data related to an entire cohort of patients. Moreover, the Feedback&Communication module allows to create and manage protocols, feedback and communications that can either be the same for all users or personalized per single patient, depending on the needs of the diabetologist and the clinical situation of the patient. Overall, the solution represents an effective means of communication towards patients allowing for a more effective action of therapeutic follow-up.

MSH has already been subjected to different cycles of verification and validation. First of all the software components have been tested through 3 types of tests: unit test, integration test, and system test. The unit test was conducted with a white box strategy while the remaining two in a black box mode.

Then we carried out an on field investigation that involved a small group of 10 volunteers for a period of 6 months. They provided important feedbacks especially useful for improving the human-machine interface. We have recently started setting up a wider and more complex experiment involving 100 subjects. After this phase, we plan on starting a large-scale production and distribution of the product over national and international markets. Also the MSH-Meter was subject to multiple cycles of testing and prototyping that have allowed its validation and, above all, its evolution in order to be compatible with the largest number of smartphones on the market.

REFERENCES [1] T.Kuzuya, et al., “Report of the committee on the classification

and diagnostic criteria of diabetes mellitus”, Diabetes Research and Clinical Practice, 55(1), 2002, pp.65-85

[2] Report of a WHO Consultation, “Definition, diagnosis and classification of diabetes mellitus and its complications. Part 1”, World Health Organization, Department of Noncommunicable Disease Surveillance, Geneva, 1999

[3] ContourUSB. www.byercontourusb.com [4] OneTouch Verio. http://www.lifescan.it [5] GMATE-IRIS, www.gmate.biz [6] M.Turner, D.Budgen, P.Brereton, “Turning software into a

service”, IEEE Computer, 36 (10), 2003, pp.38-44. [7] V.Choudhary, “Software as a service: implications for

investment in software development”, 40th Hawaii International Conference on System Sciences, 2007. HICSS 2007

[8] GH. Guyatt “Evidence-based medicine”.ACP 1991; 114(2):A-16 [9] DL. Sackett, WMC Rosenberg, JAM Gray, et al. “Evidence-

Based Medicine: What it is and what it isn’t”, BMJ 1996; 312:71-2 url: http://www.bmj.com/cgi/content/full/312/7023/71

[10] GH Guyatt, MO Meade, RZ Jaeschke, et al., “Practitioners of evidence based care”. BMJ 2000; 320:954-955, url:http://www.bmj.com/cgi/content/full/320/7240/954

[11] T.Velte, A.Velte, R.Elsenpeter, Cloud Computing, a practical approach, McGraw Hill Inc., New York, 2010

[12] Future Internet. "Shaping Policies for a Digital World: The Seoul Declaration for the Future of the Internet Economy". OECD 2008, (http://www.oecd.org/FutureInternet/)

[13] J. Huysmans, K. Dejaeger, C.Mues, J.Vanthienen, B.Baesens, “An empirical evaluation of the comprehensibility of decision table, tree and rule based predictive models”, Decision Support Systems, Volume 51, Issue 1, April 2011, Pages 141-154

[14] AA Motala, MA Omar, “Evaluation of WHO and NDDG criteria for impaired glucose tolerance”, Diabetes Res Clin Practice, 1994; 23(2), pp.103-109.

[15] American Diabetes Association, “Diagnosis and classification of diabetes mellitus”, Diabetes Care January 2010 vol. 33 no. Supplement 1 S62-S69, doi: 10.2337/dc10-S062

[16] L.T. Moss, S.Atre, Business Intelligence Roadmap: the complete project lifecycle for decision support applications, Addison Wesley, 2003, ISBN: 0201784203.


Recommended