Health Monitoring of People with Diabetes usingIoT and 5G Wireless Network InfrastructuresVasileios Tsoulchas, Nikolaos Tsolis, Eirini Zoumi, Emmanouil Skondras, Dimitrios D. VergadosDepartment of Informatics, University of Piraeus, Piraeus, Greece,Email: ftsoulchas, tsolis, skondras, zoumi, [email protected]During the last years, Internet of Things (IoT) isevolving rapidly, providing a vast range of equipment that fitsin all aspects of modern life. Indicatively, wearable devices canbe used for monitoring peoples health, since several people needcontinuous medical supervision due to health problems they face.Meanwhile the continuous monitoring of their health status withtraditional ways, such as with daily visits at hospitals, bringsseveral limitations to their daily lives while at the same timeit does not provide supervision on a 24-hour basis. This paperproposes a model to monitor the health of people with diabetesmelitus, a disease with high incident rates mainly at the elderlybut also in younger people. Specifically, a study about the existingmedically approved technologies for continuous measurement ofdiabetes is described. Subsequently, the model for monitoringpatients blood glucose levels is descibed. Whenever a patientsblood glucose levels are Low or High, the model triggers an alarmto a Cloud infrastructure in order remote medical staff to provideimmediate cure to the patient. Furthermore, in order to assurethe immediate response of the remote medical staff, the proposedmodel is deployed upon a 5G wireless network architecture.Index TermsDiabetes Melitus, Health Monitor, Patient Consciousness, Internet of Things (IoT), 5G, CloudI. INTRODUCTIONMedical technology has managed to increase life expectancy. The elderly have increased also [1] mostly in thedeveloped world. According to Eurostat, 19.2% of Europespopulation corresponds to people aged over 65. After the ageof 65, the life expectancy is 11.73-16.35 years for men and18.18-20.21 years for women [2], usually with increased healthproblems. This means that the 47.5% of the expected life over65 for men and the 55.6% of the expected life over 65 forwomen, will have health problems. Also an increasing numberof the elderly in Europe live alone.Diabetes is a very common disease among the elderlyand not only. Worldwide, people suffering from diabetes isestimated to be increased from 285 million in 2010 to 440million in 2030 [3]. It is the main cause for blindness, kidneyfailure, heart attack, stroke and leg amputation. It has beencalculated that in the year 2015 there were 1,6 million deathscaused directly from diabetes and another 2,2 million deathsthat are caused due to high blood sugar (hyperglycemia)concentrations. In general, people suffering from diabetes areat least 56 million in Europe [4]. Furthermore, patterns ofdisease differ by ethnic group: type 2 diabetes is up to sixtimes more common in people in south Asian descent and upto three times more common among people with African andAfrican-Caribbean origin [5].If diabetes is not addressed immediately and correctly, overtime can damage the heart, blood vessels, kidneys, eyes andnerves. 50% of people with diabetes die of cardiovasculardisease (primarily heart disease and stroke), while at the sametime approximately 10-20% of people suffering from diabetes,die due to kidney failure [6]. There are many papers andresearch work for monitoring peoples health and activities[7][11]. Most of them are focused on monitoring people whenthey are at home [12][16], monitoring their activities andhealth using a variety of medical equipment, depending whichmedical parameters must be monitored [17][19].In this paper, an Internet of Things (IoT) device for monitoring patients blood glucose levels is used. This deviceperiodically measures the patients blood glucose levels andtransmits them to a smart device (e.g. a smartphone) of thepatient. In case where the patient faces an increase or decreasein the glucose measurement, the smart device communicateswith remote medical infrastructure using the appropriate communication equipment, and triggers an alarm. For assuringthe immediate response of remote medical staff in emergencysituations, the proposed model is deployed upon a dense 5G[20] wireless network architecture. In general, Ultra DenseNetworks (UDN) [21] aim at the support of high data ratesproduced by an increased number of users. Accordingly, alarge number of small cells, such as Femtocells, is deployedin-side the network coverage area in order to increase theoverall capacity of the access network [22]. In addition,heterogeneous network access technologies, such as 3GPPLong Term Evolution Advanced (LTE-A) [23] or IEEE 802.16Worldwide Interoperability for Microwave Access (WiMAX)[24] Macrocells and Femtocells, as well as IEEE 802.11pWireless Access for Vehicular Environment (WAVE) [25]RSUs, are used for the interaction between the medical IoTdevices and Cloud infrastructures. Through this infrastructure,the medical staff can immediately provide the required careor advice remotely each patient.The remainder of the paper is organized as follows: SectionII describes the existing techniques available for measuringdiabetes. Subsequently, Section III describes the proposed978-0-7381-2346-2/20/$31.00 ©2020 IEEEAuthorized licensed use limited to: Monash University. Downloaded on March 31,2021 at 07:17:45 UTC from IEEE Xplore. Restrictions apply.model and Section IV presents a case study. Finally, SectionV concludes the discussed work.II. MEASURING DIABETESThis section describes the available techniques for measuring diabetes. Firstly, regarding the Self Blood GlucoseMonitoring (SBGM) technique [26], a person who suffers fromdiabetes has to measure the glucose concentration in his blood,several times a day, depending on the severity of diabetes, bypuncturing a finger and get one drop of blood. This drop ofblood is inserted in a special sensor attached on a device, andafter a period of approximately one minute, the patient gets theresult. The device stores the value with the corresponding dateand time for monitoring the course of the disease. This methodis accurate but painful and not very comfortable in case thepatient needs to carry out many measurements during the day.Also, the patient must have his consciousness to complete themeasuring procedure.During the last years a new method has introduced, namelythe Continuous Glucose Monitoring (CGM) [27]. A bodysensor attached to the patients body for approximately 15 daysis used. After this period the sensor must be replaced for thepatients safety as well as for battery replacement. The sensormeasures the glucose from the interstitial fluid, which is notdifferent from the glucose measurement from blood, except a10 minutes delay. The sensor communicates wirelessly witha mobile device or a smartphone using an application. Theapplication receives the measurements from the sensor andnotifies the patient if the glycose measurement is out of range.The application stores the measurement, giving the opportunityto the patient and the doctor to examine the glucose variationsfor a selected period of time.Internet of Things (IoT) is mainly a network of interconnected devices that communicate one another by sending andreceiving data. Every day devices convert into IoT devices withthe addition of various sensors. Indicatively, the CGM sensorfor the continuous glucose measurement is an IoT devicethat measures glucose and through a communication network,send the measured values to a smart device or smartphone.The communication protocol could be any available in themarket of the IoT. Bluetooth, NFC, Wi-Fi, ZigBee, 6LoWPan,Thread, Sub1Ghz, WirelessHart. But for the patients comfort(not to carry many devices with him), the best solution isthe Bluetooth, since all the smartphones support this kindof communication. Bluetooth protocol can cover the rangethat is needed for the connection of the body sensor and thesmartphone which will be near the patient.For t
he moment, two companies have CE certification forthe medical use of their CGM based systems and theseresults can be used for medical evaluation. The models arethe Abbott Freestyle Libre [28] and the Dexcom G6 [29].The first one uses Near Field Communication (NFC) [30]with a few centimeters distance to achieve communication,and the patient must be conscious to get the measurement.The second system uses Bluetooth [31] communication whichgives approximately 8 meters distance for communication andthere is no need for the patient to be conscious to get ameasurement.The sensor that the Abbott Freestyle Libre [32] is using, communicates with a proprietary device using the NFCtechnology. The user must bring the device near the sensor(less than 4 cm) to get the measurement. The device thennotifies the user about the measurement value (Low, Highor Normal) and stores the value for monitoring the variationof the glucose for a period. On the other hand, the sensorthat the Dexcom G6 [33] is using, communicates throughBluetooth technology, either with a proprietary device or asmartphone that uses an application designed from Dexcom.The device needs to be near the sensor (6,5 meters) and thesensor sends the measurement values in predefined periods.The proprietary device or the smartphone, notifies the userabout the measurement value (Low, High or Normal) andstores the value for monitoring the variation of the glucoseover time. For both Abbott Freestyle Libre and Dexcom G6systems, the user/patient must either have his consciousnessto act according to the measurement value, or in case ofunconsciousness he must get immediate medical attention.III. THE PROPOSED MODEL FOR BLOOD GLUCOSEMONITORINGThe system architecture of the proposed model is presentedin figure 2. For each patient, a Dexcom G6 CGM deviceis used, whose design is presented is figure 1. As it canbe observed, the CGM device is equipped with a filamentfor performing the glucose measurements, which is less than0.4mm thick, as well as with a Bluetooth transmitter whichis used for sending the measured values to nearby smartdevices, such as smartphones, tablets or vehicular On-BoardUnits (OBUs) [34]. Furthermore, a mobile application whichis deployed to patients smart device is also used. Both theCGM device and the patients smart device are interconnectedusing the Bluetooth technology.Dexcom G6 IoT deviceSkinCellsInterstitial fluidGlucoseCapillaryThe thickness ofthe sensor filamentis less than 0.4 mmBluetooth transmitterFig. 1: The design of the Dexcom G6 IoT device.Algorithm 1 presents the measurement procedure, wherethe CGM device periodically monitors the patients bloodglucose levels. Specifically, during the initialization of theproposed model, the patient (or a person who oversees thepatient) sets a time interval for the measurement of his bloodglucose levels, depending on the severity of his health status.Furthermore, since not all patient cases are the same and not allpeople respond similarly to the same medicine, personalizedthresholds for Low (T Low) and High (T High) blood glucoselevel should be set for each patient, according to medical staffconclusion about each specific case.Authorized licensed use limited to: Monash University. Downloaded on March 31,2021 at 07:17:45 UTC from IEEE Xplore. Restrictions apply.When the time interval becomes equal to 0, the CGM deviceperforms a measurement about the blood glucose levels of thepatient. Subsequently, the CGM device transmits the measuredvalues to the smart device of the patient, while at the sametime the time interval obtains its initial value and start againthe countdown for the next measurement.Cloud InfrastructureWAVE RSULTE FemtocellWiMAX FemtocellLTE MacrocellWiMAX MacrocellLTE MacrocellLTE MacrocellWiMAX FemtocellLTE FemtocellWAVE RSUSDNController Virtual MachineData aboutpatients bloodglucose levelsVirtual MachineData aboutpatients bloodglucose levelsVirtual MachineData aboutpatients bloodglucose levelsVirtual MachineData aboutpatients bloodglucose levels WAVE RSULTE FemtocellLTE FemtocellLTE MacrocellWiMAX Macrocell5G Access Network InfrastructureLTE FemtocellUser EquipmentCGM IoTdeviceSmart device(e.g. smartphone)CGM IoTdeviceSmart device(e.g. smartphone)CGM IoTdeviceSmart device(e.g. smartphone) Fig. 2: The network topology where the proposed model is deployed.The patients smart device stores the received measurement data and compares them with the T Low and T Highthresholds. When the smart device of the patient recognizeseither High or Low glucose levels, it transmits the collectedinformation to a Cloud infrastructure which is used fromremote medical staff. Thus, an alarm about the health statusof the patient is triggered. Also, the smart device displaysa questionnaire to the user with questions about his currenthealth status in order additional useful information to becollected. If the user answer in the displayed questions, theadditional information is also transmitted to the Cloud, tofurther assist the remote medical staff to perform a diagnosis.Specifically, if the measured glucose level is Low, thenthe patients smart device displays the questionnaire for LowGlucose to the patient (Algorithm 2). The patient has to answerall questions in Tco?ntdo?n = 60 seconds. If all answers arenegative and within the time frame of 60 seconds, the resultof the procedure will be Negative. Accordingly, if at leastone answer is answered positively or the time frame of oneminute expires, the result of the procedure will be Positive.Algorithm 1 The pseudocode of the Measurement Procedure.PROCEDURE MEASUREMENTREPEAT;Ttime = TmeasureREAD Measurement;if T Lo? 0 doWAIT;end whileIf the result of the questionnaire is Negative, the patient isconsidered as conscious. In this case, the patients smart devicetransmits this information to the Cloud. Also, the measurementstatus is set to Low Glucose and the smart device storesthe measurement value, along with date/time information andthe measurement status. On the other hand, if the result ofthe questionnaire is Positive, the patient is considered asunconscious. In this case, the patients smart device obtainsthe patients geographical position using its GPS and transmitsthe coordinates to the Cloud.Algorithm 2 The pseudocode of the Low Glucose Questionnaire.PROCEDURE LOW GLUCOSE QUESTIONNAIREINTEGER Ccount = 0;BOOLEAN Answer;PRINT (Do you feel sweating and/or nausea?);READ Answer;if Ans?er = TRUE thenCcount = Ccount + 1;elsePRINT (Do you feel tachycardia and/or shivering?);READ Answer;if Ans?er = TRUE thenCcount = Ccount + 1;elsePRINT (Do you feel drowsiness and/or diplopia?);READ Answer;if Ans?er = TRUE thenCcount = Ccount + 1;elsePRINT (Do you feel spasms?);READ Answer;if Ans?er = TRUE thenCcount = Ccount + 1;end ifend ifend ifend ifif Cco?nt = 0 thenTRANSMIT TO CLOUD (Measurement && Measurement Status);elseGET GPS Coordinates; TRANSMIT TO CLOUDGPS Coordinates);(Measurement&&Measurement Status end ifAccordingly, if the measured glucose level is High, thenthe patients smart device displays the questionnaire for HighGlucose to the patient (Algorithm 3). Similar to the LowGlucose Procedure, the patient has to answer all questionsAuthorized licensed use limited to: Monash University. Downloaded on March 31,2021 at 07:17:45 UTC from IEEE Xplore. Restrictions apply.CGM IoT device User Smart device(e.g. smartphone)5G Access NetworkMeasureBloodGlucoseIf ( Measurement 108mg/dL ):Measurement(Date-time, Value, Status)Save_Measurement(Date-time, Value, Status)Low_Glucose_Questionnaire()CloudTransmit_Measurement(UserId, Date-time, Value, Status)Trigger_Alarm(UserId,Low_Glucose)Transmit_Additional_Information(UserId, Date-time, Answers)Transmit_Additional_Information(UserId, Date-time, Answers, GPS_Coordinates)If ( Questionnaire_result = Negative Tcountdown>=0):If ( Questionnaire_result = Positive Tcountdown=0):If ( Questionnaire_result = Positive Tcountdown