P E R S O N AL HEALTH MONI T O R I N G : A N I N F O R M A T I O N S Y S T E M I N A W I R E L E S S N E T W O R K

Transcription

1 P E R S O N AL HEALTH MONI T O R I N G : A N I N F O R M A T I O N S Y S T E M I N A W I R E L E S S N E T W O R K TDT4520 PR OGR AM - A ND INFORMA TION SYSTEM S, AN IN-DE PTH STUDY FALL 2007 SIGNE OVERÅ HÅKON HAUGROS SUPERVISOR: JOHN KROGSTIE, IDI NTNU CO-SUPERVISOR: ØYVIND STRØMME, ACCENTURE

2 ii

3 ABSTRACT The LifeShirt is a non-invasive continuous monitoring system that collects data on cardiac, pulmonary and other physiological data. The collected data is stored on a memory card or can be transmitted in real time over a wireless network. We wanted to research the possibility of using the LifeShirt in an integrated health information system, in order to create a health care program that would be more cost efficient or provide new treatment options. We have studied the LifeShirt in general to find its strengths and weaknesses. With this in mind we have outlined the requirements of an information system for various user illness groups. We decided on a specific scenario and outlined an information system for supporting the follow-up of this illness. While there are many difficulties to overcome, we believe that an implementation of the outlined system could greatly enhance the quality of life for the specified family. To investigating this further there is a need to implement the system for additional testing and evaluation of this approach. iii

4 iv

5 PREFACE This report is written for the course TDT4520 Program and Information Systems - in-depth study, which is part of the Masters program in Computer Science at the Norwegian University for Science and Technology (NTNU). This project was a co-operation between the Institute for Computer Science and Informatics (IDI) at NTNU and Accenture Innovation Lab Norway. Accenture has previously done some research with the LifeShirt system. They have made an application to use for demonstration purposes, which we will use for guidance and learning purposes. IDI has not done anything with LifeShirt until now, but they have a close co-operation with the Norwegian centre for electronic patient journal (NSEP) who has been involved with health monitoring for quite some time. One idea has been to integrate the LifeShirt in a metropolitan area network, which means that Wireless Trondheim was involved in this project as well. We will only use the existing network and services, thus their role in the project has been a passive one. We would like to thank our supervisors John Krogstie from IDI, NTNU and Øyvind Strømme from Accenture. They have given us invaluable feedback and council during this fall. We would also like to thank Øystein Nytrø from NSEP for valuable advice and interesting viewpoints. Trondheim, December 19, 2007 Håkon Haugros Signe Overå v

6 vi

7 TABLE OF CONTENTS 1 Introduction Motivation Problem definition Report outline Prestudy Health monitoring in general Wearable health monitoring Research within wearable health monitoring MobiHealth WEALTHY system Wearable shirts Existing solution LifeShirt The system Monitoring Configuration options Wireless network connection Wireless Trondheim Coverage Research around Wireless Trondheim Research Research questions Research approach Focus Problems with LifeShirt as is The look and feel The software Actual use of the LifeShirt in a monitoring home Overall impression Proof of concept in Wireless Trondheim Connection Authentication Security issues Problems encountered vii

8 6 User groups Chronic pain patients For the patient For relatives For medical personnel Sudden Infant Death Syndrome (SIDS) For relatives For medical personnel Rehabilitation after heart surgery For the patient For relatives For the medical personnel Elderly For the patient For relatives For medical personnel Irritable bowel syndrome (IBS) For the patient For relatives For medical personnel Chronic Obstructive Pulmonary Disease For the patient For relatives For medical personnel Functionally disabled children For relatives For medical personnel Detailed use case Patient background Today s monitoring Monitoring with the LifeShirt Diagnosing with the LifeShirt Physician support Technical requirements of application for this use case Solution Architecture viii

9 8.1 Overview Clients Server Hardware Network Infrastructure Network Server Local application Remote application The LifeShirt user The future: a central server Application architecture Web portal Local portal Dataflow LifeShirt data entry User logs into local application Physician uses remote web interface Data offline input Open issues The future: Extensibility Discussion and evaluation Discussion of the results overview On the practical level Evaluation of results Conclusions and further work Conclusions Further work Graduate thesis Other possible future work References Appendix A LifeShirt Report wake summary ix

10 LIST OF FIGURES Figure Vivologic screenshot Figure Wireless Trondheim coverage map Figure Chronic pain use case Figure 6-2 SIDS use case Figure Rehabilitation after heart surgery Figure Today's situation (Electronic Arts Inc., 2000) Figure Possible future situation (Electronic Arts Inc., 2000) Figure Architecture overview Figure Infrastructure overview Figure Web portal Figure Local portal Figure LifeShirt data entry Figure Local application user Figure Use of web portal Figure Offline storing of data LIST OF TABLES User groups Functional requirements x

11 INTR ODUC TION 1 INTRODUCTION Recently there has been a change in people s demands and expectations of health care systems (Park & Jayaraman, 2003). Information is easily available through the internet which leads to increased patient knowledge. Our demands are leading to health care that improve quality of life throughout the continuum of life. All this should be achievable at a reasonable prize. A person s increasing demands leads to focus on preventing instead of treating (Park & Jayaraman, 2003). Also, health care spending is increasing. In the US, health care spending increased with 0.8% from 2000 to This indicates the need for a process change in the health care system. Also, studies show that seven major diseases account for 80% of deaths in the US in These are all diseases that benefit from early detection and intervention, which is something ubiquitous health monitoring could provide. If some of those diseases could be prevented or treated at an early stage, it would both save health spending and lead to increased quality of life. To make process changes, the use of technology needs to be taken a step further. The focus on ubiquitous computing has been present for some time now, in the health sector as well. There exist many new solutions and ideas, but introducing process changes does not happen over night. Health monitoring is one such area where one can really make a difference. Constant monitoring of your vital signs could be the answer to people s expectation of more involving and continuous health care. We want to look further into one already existing health monitoring system and improve the system to make it easier, more affordable and quality-of-life improving to use in an actual situation. We want to show that such a system could actually make a difference. 1.1 MOTIVATION Health care costs are increasing in western societies because of more advanced treatments and equipment, as well as the increasing age gap in the populations. A larger age gap means that fewer than normal have to pay for more advanced and expensive health care. A way to decrease health care costs is to use technology in new ways. It is a constant effort to streamline hospital processes so that medical personnel can focus on medical problems, and not administrative ones. Technology can automate and make hospital processes more efficient. Technology can also introduce new health care methods that are more cost efficient. 1

12 INTR ODUC TION Introducing more technology isn t just about reducing the costs; improvements can also benefit the patient by increasing their quality of life. This ranges from reduced waiting time at the family doctor to a life saving new technology. New integrated information systems can reduce costs and provide better care for end users. The LifeShirt system 1, which is a monitoring system described in detail in chapter 2.4, has great potential, its sensors and readings are highly accurate and the vest is comfortable and unobtrusive. At the moment it is mostly used in laboratories for pre-clinical studies of drugs; however an augmentation or integration with other systems could quite rapidly produce a usable system for many different scenarios. 1.2 PROBLEM DEFINITION The main focus of this project is to test the LifeShirt and improve the existing solution with an integrated information system. The LifeShirt system today is intended for medical personnel and researchers. It has not been developed much further than for demonstration and research purposes. Although it is possible to use for medical personnel as is, it has not been published and reported information about use in Norway. To make such a system we first have to study the LifeShirt in general, and find its strengths and weaknesses. Also, other similar solutions have to be studied. It is not obvious that an integrated information system will lead to huge improvements and that it will be used by patients. To show the need for such a system, we will look into different scenarios where you might get a better quality of life and treatment using the shirt. This includes specific illnesses, categories of patients with similar health problems and even specific patient cases. General improvements for different actuators will be discussed. This includes the patient himself, relatives and of course medical personnel. The integrated information system will be planned and designed. This will be a high level design. Also, we will try to integrate the LifeShirt with Wireless Trondheim to find out how suited the LifeShirt is in a metropolitan area network. The LifeShirt system might turn out to be not suited for integration with other systems. 1.3 REPORT OUTLINE A short description of each chapter in this report follows below. 2. Prestudy This chapter does a prestudy into the state of the art of personal health monitoring. It also looks into the LifeShirt and similar solutions

13 INTR ODUC TION 3. Research This chapter describes the project s focus and approach. 4. Problems with LifeShirt as is 5. Proof-of-concept with wireless network This chapter describes an integration test with the LifeShirt. 6. User groups This chapter presents potential user groups that could benefit from using the LifeShirt. 7. Specific scenario An in depth study of an actual family who is interested in using an integrated information system with the LifeShirt. 8. Solution architecture An outline of a solution architecture for an information system for the specific scenario in chapter Discussion and evaluation This chapter discuss and evaluates our results. 10. Conclusion and further work This chapter contains our conclusion and outlines possible further work. 3

14 PRESTUDY 2 PRESTUDY This chapter describes health monitoring in detail. First a general introduction to health monitoring and the possibilities that comes along with new monitoring technology, and then narrowed down to wearable monitoring. Similar solutions to the LifeShirt system and research in these fields are discussed and in the end the LifeShirt system itself is described in detail. In the end we shortly describe what Wireless Trondheim is. 2.1 HEALTH MONITORING IN GENERAL Over the last 100 years health monitoring equipment has changed with the technology development (Korhonen, Pärkkä, & Van Gils, 2003). Health monitoring equipment has usually been found in hospitals due to their use and price. This situation has changed, and during the last couple of years there has been an increased focus on monitoring in your home. With the new technology it is possible to have monitoring equipment placed everywhere without much effort and to a reasonable prize (Park & Jayaraman, 2003). This makes it possible for nearly everyone to set up monitoring equipment in their house and monitor themselves. Self monitoring could be quite useful of different reasons. The most important goals are to avoid long hospital stays, to increase the quality of life of the patients and to give safer and quicker diagnoses. If patients could be monitored in their home in the same way they would have been in a hospital, this would lead to free hospital beds and money saved. This is of course for patients that only need monitoring, for example in recovering after a surgery. Fewer people hospitalized leads to the possibility to organize medical personnel in a different way. Patients that leave hospital after surgery or after recovering from a disease might be nervous and frightened to be ill again. Such people could be monitored at home, and an alarm could go off if something wrong happened to the patient. The triggered alarm could be received my medical personnel and an ambulance could be sent within seconds. This safety and comfort that help will arrive quickly if something happens, is definitely a factor to enhancing quality of life. Monitoring yourself leads to more information about your vital signs (Park & Jayaraman, 2003). This information can be available to medical personnel which gives them a lot more information than a 20 minutes doctor visit. Medical personnel are then able to make safer and quicker diagnoses. This in turn, might lead to lives saved and fewer errors from health personnel. A lot of research done within this field has focused on monitoring in the patients home, with monitoring equipment placed on objects and furniture. Also, the focus has been on making it 4

15 PRESTUDY work. It has not been taken in use and there is therefore no user experience reported (Korhonen, Pärkkä, & Van Gils, 2003). Another field that can achieve the same results as home monitoring is wearable monitoring. This is discussed in the next chapter. 2.2 WEARABLE HEALTH MONITORING The technology for wearable health monitoring has been around for some time. However for many reasons it hasn t seen a wide scale adoption. There are several issues to be overcome when it comes to wireless and wearable health monitoring. Wearable sensors have to be un-obtrusive if they are to be worn in ubiquitous environments (Lukowicz, Kirstein, & Tröster, 2004). Usually the sensors will be embedded into clothing garments, or the garment is adjusted to hide the sensors. The sensors will have to be selfsustained with a minimal of repairs and attention required. A trade-off for the sensors is between energy consumption and its functionality. Energy will normally have to be used scarcely, so sensors have to be low-power and energy efficient. Some systems generate their own energy, by scavenging kinetic or thermal energy. For instance a hearing aid is implanted internally and it is expected to last for several decades. It has to self-replenish its energy and function fault-free for a long time. The wearable sensors are connected to wearable computers (Lukowicz, Kirstein, & Tröster, 2004). These computers are often located together with the sensor, thus making the sensor autonomous. Another paradigm is a central computer carried on the body. It integrates with all the sensors, and processes the data. There are multiple trade-offs for wearable computers, they have to be energy efficient, small in size so as to be un-obtrusive and they have to provide the necessary input and output possibilities. There are several ways to make peripherals so that wearable computers can be minimized. Wearable systems usually require wireless communication. Either from the sensors to the central worn computer or from the body network as a whole to a server located elsewhere. Protocols for the communication have to take into consideration that while a small protocol such as ZigBee is energy efficient, it is unable to cope when networks merge, for instance when all the patients in a hospital use wireless sensors. On the other side, protocols such as IPv6 provide everything you ll ever need, but they are heavy weight and energy consuming. At the moment, wearable health monitoring is mainly found in personal devices intended for fitness minded individuals. They range from the very simple such as a step counter to the more advanced heart rate monitors. Users of these devices may come to expect more of a clinical examination, and be willing to learn how to use more advanced monitoring solutions. 5

16 PRESTUDY 2.3 RESEARCH WITHIN WEARABLE HEALTH MONITORING A lot of research has been done within this field. This is a wide area, from huge and expensive monitoring machines located at hospitals, to cheap and simple step counters that everyone can use. This chapter will look into solutions similar to the LifeShirt. That is both shirts and complete systems. All these systems and shirts have one thing in common, they have never been taken much longer than to demonstration and research. No feedback on actual use from patients or medical personnel is to be found. Still, demonstrations and research with different kinds of systems gives useful information of what is likely to be a good useful solution MOBIHEALTH MobiHealth 2 is a project with the idea of a body area network platform that should make it possible for different health functions to integrate with a plug-and-play approach. This would also enable transmission of signals to a distant healthcare; remote monitoring. A body area network is defined as a collection of communicating devices which are worn on the body, providing an integrated set of personalized services to the user. The MobiHealth system consists of the body area network, the back end system and the front end system presented to the user (Jones, 2006). The communication is done with Bluetooth. The body area network can consist of a lot of different monitoring equipment. The end user system displays different monitoring data as curves. This project was finished in 2004 and shares similar motivation as the ones described in chapter 1 (Bults, Wac, & van Halteren, 2004). It ended up with a demonstration for use in a home environment, but was not taken any further. Especially the communication via Bluetooth leads to limitation of use WEALTHY SYSTEM The WEALTHY system consists of a wearable fabric with integrated sensors and a PPU Portable Patient Unit (Paradiso, Loriga, & Taccini, 2005). In the PPU analog signal conditioning, signal processing and GPRS transmission is done. It has a simple user interface with the possibility to click a button to set off an alarm. All data are sent to a Central Monitoring System which includes servers and modules for patient and medical personnel. This system has focused on the patients comfort and the wearable fabric with integrated sensors. The sensors and monitoring equipment should not be disturbing to the user and the user should be able to function in the everyday life as normal. If not, the probability for use decreases considerable. This is an important point worth having in mind

17 PRESTUDY The wearable fabric is based on conductive and piezoresistive materials in form of fibers and yarns. The sensors are woven and knitted into the wearable fabric. WEALTHY has made an effort in sending the signals from the sensors to the PPU in the best way possible (Paradiso, Loriga, & Taccini, 2005). Their focus has mostly been on making the shirt and monitoring system work together and making the signals work together in an intelligent way. The research done with this system has shown that the technology available today can make the shirt very comfortable and that the possibilities in collecting monitoring data are endless (Paradiso R., 2003). But, to make the shirt attractive to patients you also need to focus on the patient s needs and of course what should happen after the monitoring data is captured. In other words, you must have the whole picture in mind and not only the technology itself WEARABLE SHIRTS Shirts have been developed for different purposes and not only within health perspective. This includes use in for example exercising and working. Within the health area, in addition to the LifeShirt system which we are going to focus on, a few of the most known shirts are mentioned here. LifeShirt will be presented in the next section SMARTSHIRT The SmartShirt, or The Wearable Motherboard as it may also be called, was developed by Georgia Tech. The project was funded by the US Department of Navy in October 1996 (Jayaraman). SmartShirt was one of the world s first intelligent garments and was originally intended for the army. The SmartShirt detected bullet wounds and vital signs during combat (Park & Jayaraman, 2003). After a while it became clear to the developers that this shirt could be used for other purposes, including the health area. This lead to the idea of a shirt where only the shirt itself is constant, and all the other factors are changeable (Jayaraman). Although Georgia Tech still focuses on use in the army, thoughts around the SmartShirt and its opportunities has led to research and more focus on the future of health monitoring BIOSHIRT Some Korean researchers at the Electronics and Telecommunications Research Institute (ETRI) have been developing a shirt they have called BioShirt (Shin, 2004). BioShirt was originally intended for use in athletics and monitors temperature, heart rate and speed. Data is sent to a wearable monitor through Bluetooth. This solution is also possible to use in a health perspective, and some research has been done in this area with BioShirt. The goal in this research is to perceive emergency situations (Shin, 2004). The idea is to have an algorithm that makes an alarm-decision based on the data from sensors. If the input data is 7

18 PRE STUDY in a dangerous range for the person who wears BioShirt, the algorithm generates an alarm to contact an operator. The operator then tries to contact the patient to get his health status. In this system, the only persons that can read data are health personnel and there is only interaction when the patient is in a potential danger. This means that some medical personnel needs to be on standby all the time, ready to contact a patient if the alarm goes off ECG SHIRT ECG stands for electrocardiography and ECG Shirt is a shirt that only monitors the heart. Several actors have made research with a shirt like this; among them is The Tampere University of Technology (TUT) Institute of Electronics (Hahto & Halme). Their ECG shirt consists of a detachable unit that is the control unit, and the shirt itself. This unit contains room for memory card to store data, and does also have a button the patient can press if he feels a health change. Data from a memory card are loaded into MatLab-based software and evaluated. If the button was pressed, this is shown together with the data. This shirt has all the sensors incorporated and you don t need to connect with wires. The obvious limitation here is that this shirt only monitors the heart. In many cases, the patients need monitoring of other vital signs as well. If you already are sick and want to use monitoring to prevent illness or seizures, you often need to check other signs than the heart beat. This could be your body temperature or maybe even your brain activity. This is not possible with this shirt. 2.4 EXISTING SOLUTION LIFESHIRT LifeShirt is a wearable monitoring system developed by VivoMetrics (VivoMetrics, 2007). VivoMetrics was created in 1999 when they got patents for wearable sensors that monitor respiration and cardiac function. The VivoMetrics team consists of experts in both the medical and engineering field THE SYSTEM LifeShirt is a monitoring system that can monitor several different areas of your body. The basic system consists of the wearable shirt itself, a wire with connectors to different sensors, a PDA and a software program. The wearable shirt comes in all sizes, both for children and adults. This basic system monitors your respiration, heart and position (whether or not you re in an upright position). Respiration is monitored by sensors woven into the shirt around the chest and the abdomen. In addition to this, it is possible to monitor as much as 30 different values of your body (VivoMetrics, 2004). Among these are blood pressure, blood oxygen saturation and tidal CO2. You can choose to buy whatever additional equipment you like, which means that each shirt can be especially adjusted for each patient. 8

19 PRESTUDY LifeShirt has many advantages compared to commercial devices like step counters or thermometers. Some of them are: LifeShirt can monitor several functions at the same time. Monitoring data is saved on a memory card and can also be transmitted through a wireless connection. You can configure the PDA in many ways through a configuration file. You can mark events that happen to you, both feelings and influence from the outside world. To wear the shirt you need to place three ECG-electrodes on your body and connect the data cable to each one of them. The cable also connects to the respiratory sensors. The shirt is constructed so that the different cables are hidden inside the garment so that it won t bother you. Place a configured memory card into the PDA and connect the PDA to the cable. Then you are ready to start monitoring MONITORING To start monitoring, you power on the PDA. Some initial calibration will appear and when all this is done, the monitoring starts. The display will tell you that a monitoring session is active. You can abort the monitoring whenever you want. Also, you have several opportunities to log different things that are happening to you. This can be both things you feel and things you do. For example if you are feeling chest pain, you can log this. If you are exercising or taking medications, you can log this as well. One of the first questions you need to answer before a monitoring session is whether it is a sleep study or a day study. This means that you can do different kinds of monitoring. Vital signs are different when you are awake and when you sleep. By logging this, the experts evaluating the monitoring later on, knows when you slept and not. When you have finished a monitoring session, the memory card contains a file with all the monitoring data. This file can be imported into the software program mentioned earlier, VivoLogic. VivoLogic presents data with graphs through time. There is one graph for each different variable. Figure 2-1 shows an example of a screenshot where the patient is coughing. In addition to the normal curves, it is also possible to see different variations of this curve. This can be the derivative of the normal curve or even double derivative. This is useful for medical personnel and perhaps researchers, but not likely to be used by others. 9

20 PRESTUDY FIGURE VIVOLOGIC SCREENSHOT When the monitored patient logged different events, this is shown in the timeline with a mark stating which event occurred. This makes it easier for medical personnel to concentrate on the time around specific events. VivoLogic also gives the opportunity to generate reports. This report is a summation of the monitoring session, with minimum, maximum, mean and median for the different values. For an example report see chapter 12. This report is mostly understandable for non-medical personnel as well CONFIGURATION OPTIONS The PDA is controlled via configuration files placed on the memory card. These are generated with a program supplied in the software package. They contain all the information showed on the PDA screen and all information about what to monitor. In the configuration files you can: Specify name of the patient Specify which calibrations are to be performed before monitoring begins Enable additional sensors Specify sensor parameters such as the sampling rate Enable PIN-code to start/stop sessions The PDA software supports internationalization. Although translations aren t supplied, it should be relatively easy to change the language. 10

21 PRESTUDY WIRELESS NETWORK CONNECTION VivoMetrics has also developed a Wi-Fi solution for LifeShirt. A wireless network card can be attached to the PDA. A configuration file copied to the memory card tells how and where to connect. The wireless option supports three modes of operation: Ad-hoc. Direct connection between the LifeShirt and a computer with a wireless network card Static IP connected to a wireless network Dynamic IP connected to a wireless network When delivered, the LifeShirt is calibrated for a static IP address to the router supplied from VivoMetrics. The monitoring data is stored in the memory card as for the usual monitoring, but is sent over the wireless network as well. The software presenting these data real time is called VivoMonitor and comes along with the Wi-Fi solution. VivoMonitor is a lot like VivoLogic, but with fewer options. The graphs are shown in real time, which means that they are continuously moving across the screen. You can choose the time interval you want to see at once. Marks for different events are shown in the same way as in VivoLogic. 2.5 WIRELESS TRONDHEIM Wireless Trondheim is a research project by NTNU in cooperation with local and regional officials. Its aim is to build a real-time, city sized wireless network where students and residents get access to a wireless service. At the same time providing a unique laboratory in which researchers, company and technology suppliers can test new ideas and equipment. The network is free of charge for all students and employees at the university as well as employees in the municipal government. For everyone else it s available for a fee of 10 NOK (about 1.2 ) for 12 hours COVERAGE At the moment Wireless Trondheim covers most of the downtown area of Trondheim, shown in the map below. Also marked on the map are places where there s guaranteed indoor coverage. 11

22 PRESTUDY FIGURE WIRELESS TRONDHEIM COVERAGE MAP RESEARCH AROUND WIRELESS TRONDHEIM TRIMAKS Information Systems perform research and development of new information services used over a wireless broadband network. TRIMAKS looks among other things into: Context-aware services, for instance services based on your location Services for tourists Personalized shopping oriented services Issues of trust in mobile services For more information see (Andresen, Krogstie, & Jelle, 2007) and (TRIMAKS). 12

23 RESE ARCH 3 RESEARCH A lot of research has been done within health monitoring in the last couple of years, due to new technology and the fact that it has been an increased focus on quality of life (Park & Jayaraman, 2003). Also, healthcare spending has increased in the last decade and it is of great interest to stop this trend. We were already in the possession of the LifeShirt system, which is described in detail in chapter 2.4. We would try to go further with the LifeShirt technology already developed, and make it even easier to use, easily accessible and increase quality of life for the patient. This chapter describes the focus and goals of this project. That is, what this project was going to find an answer to, and how this was done. 3.1 RESEARCH QUESTIONS These are the main questions this project was going to focus on and try to answer. How could an information system for the use of personal health monitoring improve the quality of life for several patient groups? Is it possible to integrate a monitoring information system with an already existing metropolitan area network in an easy way? 3.2 RESEARCH APPROACH A literature study of existing solutions and of health monitoring in general was performed to get to know the subject better, and also find out results and thoughts from previous work. It was of interest to find out if any had already made an integrated information system with a wearable monitoring system and if such a system had been accepted by patients. Also, we wanted to find out as much as possible about which user groups other similar system had focused on and why. The goal is to make an integrated information system with a central server and several users. The thought is that users wearing the shirt can be monitored everywhere. To reach this goal the shirt must be able to operate in a wireless network, preferably in a wide area network. We tried to make a proof of concept with the LifeShirt connected wirelessly through Wireless Trondheim. To find out if the idea of an integrated information system was something that could be useful to patients, we then tried to find user groups that could benefit from such a solution. We tried to find out what requirements each group would have to the system. This also helped us to find out how a system could improve their quality of life. 13

24 RESE ARCH An information system that covers all possible scenarios and user groups will be complex and difficult to create. Already developed systems are on the demonstration and prototype stage, so creating a collectively exhaustive system for very different uses will have a high degree of risk. The risk is both in functional mismatch as well as the actual completion and deployment of the system. We therefore decided to focus on a specific scenario, creating a system based on requirements from involved stakeholders. We tried to model the system so that it could be extended and reused later on. In the end we made an architecture approach to the integrated solution. This was based on feedback from users in the specific scenario, the already existing software from the LifeShirt system and earlier research done within this field. This solution is planned to be developed further and implemented during the spring of FOCUS The main focus of this project was, as already mentioned, to take already existing technology a step further by making a whole solution with an integrated information system. This means trying to make the already existing hardware more useful through new software. We focused on user groups that would make use of such a system, what they needed and how it could improve their quality of life. That is, for both themselves and their surroundings. Our focus was on increased quality of life for patients, and also to make a better solution than the already existing LifeShirt system. We did not focus on other issues like for example security, maintenance or economy. 14

25 PR OBLEM S WITH L IFESHIR T AS IS 4 PROBLEMS WITH LIFESHIRT AS IS This chapter describes problems and weaknesses we have found when testing the LifeShirt system. 4.1 THE LOOK AND FEEL The shirt is relatively comfortable to wear, perhaps as comfortable as you can expect of a wearable monitoring system. But where the shirt is non-obtrusive, the PDA is bulky and heavy. It has to be carried in a pouch in the front or on the side due to the length of the data cable. This could be a nuisance to the user, both for practical and fashion reasons, thus decreasing the usability of the LifeShirt. VivoMetrics states that you can exercise when wearing LifeShirt and you can also log this with the PDA. It would be difficult to do anything but light exercise, such as walking, due to the pouch and the PDA attached to your shirt. To avoid problems with the PDA when sleeping, there is a special pillow that hides the PDA inside. The data cable connecting the shirt to the PDA is not particularly long, which means that you have reduced mobility in bed. For some this might be a problem. Another problem with the PDA is the cards, especially the network card, which sticks out and makes the PDA bigger and not as streamlined. Of course this is a small problem compared to the whole picture, but still it is unnecessary to make the PDA feel even bigger. The data cable is of a certain length and is so long that it suits the larger sizes of the shirt. Again, this could potentially be a problem for children which can wear smaller sizes of the shirt. There is a fixed length from the upper sensor to the lower sensor. This means that the wire in between would stick out and be a disturbing moment for the smaller sizes of the shirt. 4.2 THE SOFTWARE Both VivoLogic and VivoMonitor present quite complex graphs with a lot of short medical terms as explanation. This means that a layman will understand only a fraction of the program, or maybe even nothing. The documentation that comes along with these programs contains only information about how to install and start the program, and not how to use it. It is safe to say that VivoMonitor and VivoLogic are primarily meant for health personnel and researchers. This leads to the fact that a patient will be dependent on having an expert nearby, especially if it is necessary to monitor in real time. If it is sufficient to monitor first and check results later, it 15

26 PR OBLEM S WITH L IFESHIR T AS IS may be enough to send the memory card to a specialist after the monitoring is finished. Then the specialist can read all data through VivoLogic. Use of the LifeShirt system has to be part of a larger treatment program. A normal patient cannot buy the system himself and use the software supplied for self-monitoring. This is due to the cost of the system and the advanced software. These programs only present graphs from vital signs monitored. There is no complex calculation with several vital signs at once, like for example if one vital sign goes high and another low, this might indicate that something is wrong to people with a special illness. This also leads to another point worth mentioning, personalization. If patients with different diseases and needs use the LifeShirt, they probably have different wishes for what they want from the program. Today, there is limited possibility to personalize the software to fit the patient s needs. 4.3 ACTUAL USE OF THE LIFESHIRT IN A MONITORING HOME In a real monitoring situation there are some unfortunate situations that can occur. If you are using the wireless connection, you get no warning when you are out of range. The PDA acts normal. The monitoring is still recorded in the memory card, but if you are being monitored by someone else, this might be a problem. Currently, there is no support for an alarm system. This means that monitoring of patients in a certain danger, are out of the question. As mentioned you can mark events like taking meds or exercising. If medical personnel should get something useful out of hours of monitoring data, you are obligated to mark such events. This requires the user to interact with the PDA which might be irritating and easy to forget. As discussed in the section 4.2, the software is quite complex and difficult to understand. This means that it would be very difficult for the patient himself to use the software. This again means that medical personnel need to have a certain role in a monitoring situation. The more people that needs to be involved, the more difficult it is to take in use. It is not enough to be motivated yourself; your physician needs to know and use the system as well. 4.4 OVERALL IMPRESSION The hardware and LifeShirt equipment have minor problems, but they are all manageable. The software is also informative and useful to medical personnel. But the software has limitations that an integrated information system could solve. The main focus here is self-monitoring, and it is in fact quite difficult to monitor yourself with this software. Of course, that is unless you are medical personnel or have gotten advanced 16

27 PR OBLEM S WITH L IFESHIR T AS IS training in how to use the software. You need to know medical terms, and you need to know how to read complex graphs. If medical personnel were to monitor all patients wearing LifeShirt it would require a lot of personnel, training, time and money. If you use the LifeShirt to monitor yourself and bring the memory card the next time you see a doctor, this requires the doctor to have VivoLogic. It will be time-consuming for a doctor to go through monitoring data. It would be nice to retrieve only relevant time intervals. If you use the LifeShirt in real time monitoring, this requires a specialist to sit and watch the screen somewhere. To find such personnel and resources is at the moment difficult. We also miss the possibility to personalize the software, and maybe show different user interfaces dependent on who you are. Medical personnel, relatives and the patient himself might have different needs and would benefit from having different options in the application. Centralization might be of interest, and is missing in today s solution. If the software was available through internet, and monitoring data was saved in a database on a central server, it would be much easier to all persons involved and the options you have with the data would also increase. This also leads to the fact that different people can use the system whenever they like, and not only when a memory card is present or when you are watching the screen real time. 17

28 PR OOF OF C ONCE PT IN WIR ELESS TR ONDHEIM 5 PROOF OF CONCEPT IN WIRELESS TRONDHEIM At the moment we want to connect the LifeShirt to our wireless city wide network and use it with VivoMonitor. Later on we plan to implement our own information system to replace and augment VivoMonitor s functionality. This chapter describes how the LifeShirt could be integrated in a Wide Area Network (WAN). 5.1 CONNECTION The LifeShirt supports connection to a wireless zone via static or dynamic configuration. By default the LifeShirt is configured to use a static IP address, this is practical in controlled environments such as laboratories. Practically all of the LifeShirt current users fit into this category. If we are to use this configuration we have to reserve an IP range for LifeShirt and do more initial configuration. A more preferable option is to use the dynamic configuration where the LifeShirt will automatically retrieve an IP address. We don t need to be able to contact the LifeShirt, so as long as the LifeShirt is able to connect to the network and send data to our server, we re happy. Via the configuration files we set the LifeShirt to connect to wireless Trondheim, use either static ip or dhcp and set the ip of our server. 5.2 AUTHENTICATION Normally a user will have to authenticate via a web portal or via VPN to gain access to the wireless network of Wireless Trondheim. The LifeShirt supports neither of these methods of authentication. We have therefore gotten the MAC address authenticated so that the LifeShirt has unlimited access to the network. 5.3 SECURITY ISSUES VivoMetrics never intended the LifeShirt to be connected to the internet; therefore there are many security concerns. In our case this concerns aren t a focus of research, however in further research or implementations they should be looked into. The embedded Linux in the PDA has not been configured for security. It has open ports and could be manipulated in this way. However according to VivoMetrics the PDA itself wouldn t get a virus, the only thing that could get damaged is the compact flash card. The LifeShirt supports Wired Equivalent Privacy (WEP) encryption; however this is not used in Wireless Trondheim. All data is therefore sent unencrypted. For a technical proof of concept this is valid, for a more permanent solution steps have to be taken to secure the transmission of 18

29 PR OOF OF C ONCE PT IN WIR ELESS TR ONDHEIM vital signs data. The Norwegian Data Inspectorate is quite strict on the privacy of personal data; the current solution would not pass an inspection. 5.4 PROBLEMS ENCOUNTERED While everything is supposed to work in theory, we have been unable to get the LifeShirt to function correctly. We have had difficulties connecting the LifeShirt to any other wireless network than the one supplied from VivoMetrics, we are unsure of the reason for this. The LifeShirt connects to the supplied network, even though the configuration files point it to another network. Together with VivoMetrics we have analyzed this problem without finding a solution. We expect to solve this problem once we get access to the LifeShirt s internal systems. VivoMetrics uses only the static configuration in their installs. Due to this the dynamic configuration is not well tested, and not supported that well. This may be one of the causes of our problem mentioned above; however a field test with static configuration was unsuccessful as well. When we tested the LifeShirt with the supplied wireless access point, the LifeShirt had problems reconnecting to VivoMonitor after it had been outside of network range. We are unsure of whether this robustness problem is in the hardware or software. In further studies we have to look more into this, since it s highly likely that the LifeShirt will lose the network connection inside many buildings. 19

30 USER GR OUPS 6 USER GROUPS To show the need for an integrated monitoring system, table 6-1 show patient groups where this could be applied. For each group, requirements and possible improvements from today s situation are described shortly for three roles; the patient, relatives and medical personnel. The rest of the chapter contains a detailed description of each patient group together with key facts about the group. 6.1 CHRONIC PAIN PATIENTS These patients suffer from pain they probably never will get rid of. Many such patients suffer from cancer, and have got the death diagnose. In addition to the cancer, they suffer from severe pain and have difficulty living a normal life. Pain can be difficult to describe, and it is often difficult for medical personnel to understand the pain, both where it is located and what the cause is. Treatment of chronic pain is simply to try to decrease the pain as much as possible (Jordhøy, 2003). This is done through medication and through the way the patient is living. Exercise and activity level can affect the patient s pain. The goal is to make the patient maintain a normal life as long as possible. Since the pain varies from patient to patient and is difficult to describe and locate, it is usually very difficult to compensate or treat. It is estimated that in Europe chronic pain causes 500 million lost working days a year (Nytrø, 2007). Many sufferers feel that their treatment is not effective or sufficient. Figure 6-1 shows a use case of this scenario. 20

31 USER GR OUPS Chronic pain treatment Sudden infant death syndrome Rehabilitation after heart surgery Elderly Irritable bowel syndrome Functionally disabled children Chronic obstructive pulmonary disease USER GROUPS Patient Relatives Medical Personnel Simple interface. Logging capabilities (start/stop). Search with different parameters. Monitoring depends on ailment. Does not apply. Have to store exercise. Easier to follow prescribed treatment. Increased safety. Simple interface. Better care. Increased sense of independence. Increased safety. Alarms when time for medication. Monitoring at night. Alarms very important. Monitor respiration and cardiac state. Store data around the alarms for further research. Store position and movement of the child. Easier follow-up on children. Simple interface. Monitor EKG, activity, respiration Alarm at nursery. Falls and other accidents can be detected. Self motivated. Simple interface. Extended interactive diary. Searching to see connections with food. Does not apply. Alarms when vital signs are out of range. Similar to medical personnel, but reports easier written. Logging of medication intake. Find own threshold for physical exercise. Alarm if breathing problems occur. System generates report. Can help understand ailment, thus able to prescribe proper treatment. Use monitoring data for research. Correlate multiple cases. Alarms. Control that patient performs prescribed exercise. SpO2? Reactions to medications, adjust strength etc. Generates reports. Sets parameters for different individuals. Logging of alarms in backend. Search for threshold values. Control on whether medications work. Research. 21

32 USER GR OUPS FIGURE CHRONIC PAIN USE CASE FOR THE PATIENT A simple interface that only presents the most necessary data in an easy way is important to make patients use it. If the interface contains a lot of options, presents difficult curves and uses medical language, the motivation from the patient would most likely decrease. The interface has to be self-explanatory and with help functions that make it easy to navigate through menu options. The patient needs logging capabilities. The patient should be able to log start and stop of various events, like exercising or resting, and this should be stored together with the monitoring data between start and stop. Also, a simple text explaining what event the patient performed is necessary. This makes it possible for the user to keep track of reactions to different events in a more concrete way. Data logging also makes it possible to present data to medical personnel this could increase quality of diagnosing and treatment. There should be a possibility to search for your own stored monitoring data. The search should include time search, event search, monitoring data search and so on. Results should be presented in a way that makes it easy for the patient to read and compare data. This makes it easier for the patient to go back and see how he has reacted to the same event several times, and see patterns in his pain. This could make the patient avoid certain events that obviously makes the pain worse and it could help create an effective exercise plan. 22

33 USER GR OUPS Chronic pain varies from patient to patient, both in intensity and where in the body the pain is located. As mentioned, it is often cancer patients that belong to this patient group. The vital signs that would be useful to monitor will therefore vary and needs to be set up independently. The use of a monitoring system needs to be self-motivated. If used actively you can see connections between the pain and things you do in the everyday life. It could also make the exercise program more suited for each patient. This could make the patient more familiar with his own body and reactions. This again could lead to increased quality of life. Our approach to this system is one-to-one communication but in the future it might be possible to share events and monitoring data that comes along with it over internet. This could be like an internet forum, where you can share data and discuss them with others in the same situation, or perhaps medical personnel. To meet other people in the same situation as you could be very useful. You could share information and data, and get to know what conclusions other has made based on their monitoring. This could increase your quality of life. Such an extension raises several issues like for instance security, maintenance and error discovery. For these reasons this is something far ahead from today s situation but could be very useful in a future version FOR RELATIVES Relatives will usually have little or no use for such an application. Due to the fact that the use needs to be self-motivated, the patient himself is doing all the monitoring. Of course, relatives can be of help and support, but will have no additional demands to the system. In cases where the patient is so ill that he cannot take care of himself, relatives could use the same system as described above to get information and learn more about the patient s pain. If the patient is unable to communicate, it could ease the treatment a lot to see the patient s reaction to different events FOR MEDICAL PERSONNEL As mentioned the use of a monitoring system in these cases, are mostly for the patient himself. But monitoring means more data, which means more information about the patient. More information about the patient leads to a safer and quicker diagnoses (Park & Jayaraman, 2003). This again leads to the right treatment. Medical personnel should have reports about patients generated from the system to see data and different events connected. This could also help them see reactions to different medication and correct them if necessary. It would be easier for the patient to describe the pain to medical personnel when they have access to the monitoring data as well. Hopefully this leads to a better understanding and right medication and support. 23

34 USER GR OUPS 6.2 SUDDEN INFANT DEATH SYNDROME (SIDS) When apparently healthy infants suddenly and unexpectedly die, it is referred to as sudden infant death syndrome (SIDS). The causes of this syndrome is not known, however it is widely believed to be caused by respiratory failure, especially a quiet apnea during sleep (Southall, 1988). SIDS is most probable during the first 43 weeks after birth (Skadberg, Morild, & Markestad, 1988). A change in the sleeping position of an infant from a prone to a supine position will greatly reduce the chance of SIDS. An industry has sprung up around the monitoring of infants to avoid SIDS. However the basis and effect of this monitoring has come under doubt (Hunt, Corwin, Lister, & al, 1999). Nevertheless, infants with respiratory problems may benefit from monitoring these first 43 weeks. The current monitoring is quite costly and the introduction of a monitoring system like LifeShirt might be easier and cheaper to use. Early born infants or infants with apnea problems may profit of short term monitoring given that the monitoring discovers respiration problems. Especially parents that have already lost one or more children in SIDS might wish for such a monitoring system. This could make them feel more secure and have the possibility to monitor their new baby in a better way. Figure 6-2 shows a use case for this scenario. 24

35 USER GR OUPS FIGURE 6-2 SIDS USE CASE FOR RELATIVES Monitoring is only interesting at night when the kid sleeps. Alarm is essential for relatives in such a case. If abnormalities are detected, an alarm should wake relatives immediately. Monitoring data around events where alarm goes off should be stored. This data should be possible to see afterwards and compare with other events. Sleeping position and movements should also be stored. This could help relatives to see sleeping patterns and prevent the child from sleeping in wrong positions if necessary. Relatives could also bring monitoring data in discussion with medical personnel. If relatives are worried that something is wrong, needs advice or facts, such data would increase understanding for all parts FOR MEDICAL PERSONNEL As mentioned there is many unanswered question around SIDS and monitoring data from several kids could be useful information in future research. Storing monitoring data makes it possible to compare data from both healthy kids and kids with abnormalities. 25

36 USER GR OUPS 6.3 REHABILITATION AFTER HEART SURGERY After surgery a patient has to go through a cardiac rehabilitation. The first part of the convalescence is carried out in a hospital, where the patient is continuously monitored. After a while the need for monitoring decreases and the patient is moved to a rehab centre or goes through home based rehabilitation. Monitoring the patient at the hospital is costly, if the patient can change to home based rehabilitation earlier, the same hospital bed may be used for other patients. Figure 6-3 shows a use case for this scenario. FIGURE REHABILITATION AFTER HEART SURGERY FOR THE PATIENT Long hospital stays can be taxing on a patient. Home based rehabilitation will most often increase the patient s quality of life. The patient may be monitored with an automated system. 26

37 USER GR OUPS The system should have audible alarms, so that if the patient falls or starts having an irregular hearth rhythm, an alarm will go off. The system should be connected to a communication device so that an alarm can be sent to a monitoring central where a proper action can be taken. The system can have two alarm types, one for alerting the patient that he is over exerting himself and should stop current activity. The other could be for more severe alarms. Communication with someone else is then required. This could be relatives or medical personnel that can react instantly and check up on the patient. A cardiac patient has usually made lifestyle choices that have led to the cardiac problems. After surgery a patient should increase physical activity to reduce further problems. This increase in activity has to be conducted safely and an exercise program is usually devised together with a physician. The patient needs logging capabilities. Especially the logging of exercise is important. If the user has to log his activities he will be more motivated to follow the program. The system should also provide a simple interface for the patient to monitor himself FOR RELATIVES The relatives are interested in the fast recovery of the patient. They can provide assistance and aid with the exercise program. Changes in lifestyle are difficult to do alone, engaging the whole family can make it easier for the patient to do the necessary changes in diet and physical activities. The relatives can provide assistance upon alarms from the system. The system itself can sound alarms, but it can also send off signals to other devices in order to sound an alarm. For instance it can be hooked up to the home stereo system, or send a text message to cell phones. When the patient is a child or a handicapped person, the relatives may need to perform all the operations for the patient. For instance the logging of exercise. It is important that the system has a simple interface FOR THE MEDICAL PERSONNEL The physician will want to know how the exercise program is progressing. Recovery is heavily dependent on the physical activities after the surgery. If the patient is not following the prescribed exercise program, the program may need to be revised. The system should be able to generate reports for all the patients a physician is monitoring. These reports should include whether the patients are following the exercise program, anomalies in hearth rhythm and events that has transpired. This information can be used in a consultation with the patient to better adapt a program for the patient. 27

38 USER GR OUPS 6.4 ELDERLY The western world will see an increase of elderly people, in proportion to the general population, in the time to come. Health care costs will increase, especially in countries with complex public health care systems such as Norway. Governments have to devise ways to reduce spending without reducing the quality of health care. One way of doing this is to have people retire later, thus providing a larger work force to pay for the health care. Governments who try to revoke early-retirement privileges face staunch opposition. The population expects a good retirement with well provided health care. There is however a consensus in the population that health care expense has to be decreased. At a certain point elderly are not able to take care of themselves anymore. They might get proper help from family, but if this is not possible they might need to stay in hospitals or care centers. If these elderly could be monitored continuously at home, it would free up hospital or care center beds. The patient could live a normal life at home and know that help would arrive if something happened. This leads to better health care at a lower cost, and also improved quality of life for the patient. One example of use is a care center with elderly living in separate apartments. Each person will wear a LifeShirt, and the person on call can monitor all the elderly at once. Should an elderly have a problem, for instance shortness of breath or a fall, it can be detected and proper aid can be dispatched. Thus the elderly will have a large degree of independence while at the same time have the safety of knowing that they are taken care of. This scenario is quite similar to the use case in Figure 6-3. A system for the monitoring of elderly has been envisaged before in a Japanese study; see (Tamura & al, 1998) FOR THE PATIENT There is a certain risk that elderly who lives alone get ill and unable to call for help. This fact might lead to people feeling anxious or living in fear. Knowing that an alarm will go off, and aid will be dispatched quickly should they have a problem will greatly increase their sense of safety. This is important for situations such as shortness of breath or irregular cardiac rhythm. Alarms can be programmed so that if the elderly falls, a nurse or another assistant can be dispatched to help the elderly. The patient will profit from a better health care and added security. If the patient is on medication the system has to support the logging in an extended diary. This diary has to have a simple interface. If the medication program is static, the system can provide assistance in remembering to take prescribed medicine. For instance the system can make a 28

39 USER GR OUPS sound and ask if the medication is taken at a given time. The user then has to confirm that the medication has been consumed FOR RELATIVES As mentioned before, automated audible alarms for various settings are important. If the elderly is living in the same house, or nearby the relatives, the system can dispatch alarms to the relatives instead of a care center. For instance to the cell phone of the closest relative. In some cases an elderly may require the assistance of his relatives. The relatives will then have to do the logging and interact with the system. The system interface should be simple. If the elderly is unable to communicate his symptoms, and the symptoms aren t severe enough to cause an alarm, the relatives can better understand how the elderly feels FOR MEDICAL PERSONNEL Physicians will receive a better data basis for prescribing medications. When the effects of medication can be monitored, they may better adjust the dosage according to the effect they have on the patient. At a care center or a retirement centre medical personnel may do rounds to check up on elderly at least once a day. If this monitoring can be replaced by automated alarms only when something happens, the elderly will feel more safe at the same time as medical personnel are freed up to do other tasks. 6.5 IRRITABLE BOWEL SYNDROME (IBS) Irritable bowel syndrome is the most common digestion disease in the western world (Blomhoff, Diseth, & Jacobsen, 2002). This is a syndrome that leads to stomach pain, severe production of gas and often diarrhea. The symptoms vary from person to person, and each person has ups and downs. In some periods they feel no symptoms, and in other periods they nearly can t go to work. There is no medication for this disease, and no treatment other than helping the patient find out how he could have few or no symptoms and pain. There is a connection between this disease and the food the patient is eating (Blomhoff, Diseth, & Jacobsen, 2002). The problem is that the food patients are reacting on is individual. Some might react on something that makes other people feel better. Fiber is a good example. Some patients need a lot of fiber to ease their symptoms, but some people get worse after eating fiber. This diagnose is set when everything else of possible illnesses is ruled out. There exist no approved treatment today, and many patients need to find out by themselves how they could live without pain. 29

40 USER GR OUPS This monitoring system could also be used on patients with several food allergies or food intolerance. In these cases it might also be difficult to learn what to eat without getting ill, and they would have the same problems as people with irritable bowel syndrome. This scenario is quite similar to the use case in Figure FOR THE PATIENT As mentioned, the patient often is encouraged to write a food diary. This could be done in an extended way through the system. It should be possible to mark whenever you eat, and what you are eating. This should be stored together with monitoring data between a specific time before and after the meal. This time interval should be possible to choose individually. It should also be possible to write notes together with the meal, describing how the patient feels. This could be useful when looking back in time. Presentation of these monitoring data should be made simple and easy to understand. A help function describing parameters in an easy way should also be available. A search function is also useful to see connections between meals and monitoring data. It should be possible to search for meals, special monitoring parameter values and time. This could help the patient see connections between different food, and also what food to stay away from. In a bigger meal like supper it might be difficult to know which part you are reacting to. When searching for a special food type you can see whether you have reacted in the same way every time you have eaten it. This is much more time consuming with an ordinary diary. The use of such a system for these patients needs to be self-motivated. It is time consuming to report everything you eat, and also too look for connections and reactions. But it is definitely worth it if it leads to a better understanding of yourself and the food that you can eat. This leads to fewer symptoms and you can live a normal life without pain or concerns. You could get a lot more information out of such a system in a shorter time, than searching yourself in a written diary FOR RELATIVES This diagnose is based on no results from other medical examinations and on the patient describing his symptoms. In children this could be difficult, since they often have problems describing symptoms. Therefore, this diagnose is not present on children below 4 year (Blomhoff, Diseth, & Jacobsen, 2002). It could be relevant for relatives to older children to use the system, until they can use it themselves. In these cases, the use of the system is the same as for the patient. Patients being so sick that they can t take care of themselves are not an issue with this syndrome. But relatives might support and help the patient make decisions based on the system. This requires no extra requirements for the system. 30

41 USER GR OUPS FOR MEDICAL PERSONNEL As for other self-motivated patient groups, the interesting aspect for medical personnel is to get more information about the patient. Reports generated about food eaten and monitoring data should be available for medical personnel to analyze. This could help them support and discuss the symptoms more detailed with the patient. Researchers believe that the reason for this disease is complex and is still unknown. There has been a lot of research over the last ten years, but there is still many questions left unanswered (Blomhoff, Diseth, & Jacobsen, 2002). The data you can get from monitoring patient could be very useful in such research. 6.6 CHRONIC OBSTRUCTIVE PULMONARY DISEASE Chronic Obstructive Pulmonary Disease (COPD) is a disease of the airways (U.S Department of health and human services, 2003). It is slowly progressing and over time causes a gradual loss of lung function. It is caused by smoking and in some cases exposure to occupational dust and chemicals. Symptoms of COPD range from severe shortage of breath to chronic cough. Since the main cause of COPD is caused by smoking over a long time frame, patients with COPD are usually mid-age to elderly people. There is no known cure to COPD at the moment, treatment is only to relieve symptoms and ease pain (U.S Department of health and human services, 2003). Medications used are mainly bronchodilators and steroids, both are used to make breathing easier. The most important action if you are diagnosed with COPD is to stop smoking, and then your COPD will not progress further. This scenario is quite similar to the use case in Figure FOR THE PATIENT The system has to support medications logging. When the user indicates that he uses a medication, the application have to save vital signs for a time period prior and after the event. The system should also store how the patient felt after a while, whether the medication is providing a relief, or if the patient feels no change. Patients with COPD may suffer from severe shortage of breath and may need to go to the emergency room. An automated alarm system integrated a care central can ensure that aid is dispatched to the patient upon an emergency. System should work as an aid to the patient in order to find his thresholds. When the patient is doing physical activities and starting to over exert himself, the system should also sound an audible alarm. Thus the system may prevent a more severe situation arising. The patient may then adapt his activities in order to not cause extra pain, thus increasing his quality of life. 31

42 USER GR OUPS FOR RELATIVES The system should provide audible alarms so relatives or other people in the vicinity may come to aid the patient. These alarms are for severe situations, as well as the situations where the patient is simply over exerting himself FOR MEDICAL PERSONNEL The system should be able to generate reports about a physician s patients. These reports should include anomalies in respiration or heart rate, level of activity and progress in prescribed exercise program. The report should highlight those patients that have problems. The physician will also want to monitor the effect of prescribed medications. This is done with the data from the patient s logs. The logs indicate vital signs around the time of medication use as well as the subjective feelings of the patient. This can be used in generated reports, or in a patient-physician consultation. The medication logs can be used in research around the effects of a given medication. Information from many different patients can be correlated automatically by the system. 6.7 FUNCTIONALLY DISABLED CHILDREN Functionally disabled persons often require intensive care. When you give birth to a functionally disabled child, your life might be turned upside down. If the child have disabilities requiring help 24 hours a day, caring for them is time consuming and might be difficult to combine with work. Often this leads to children staying in hospital or a caring centre, even though their parents want them at home. Children with disabilities that make them unable to communicate their pain and feelings could be very difficult to help. If multi-handicapped children are unable to communicate, they are unable to warn their surroundings if they get for instance a hypoglycemia (low blood sugar). This makes it very difficult for relatives to take care of them. Often children who are unable to communicate and can have seizures from different reasons, needs to be hospitalized. This is not an ideal situation for anyone. They take up hospital beds even though they only need extra help once in a while, and it is difficult for relatives to have their child in a hospital all the time. The patient s quality of life is not optimal either, since he has to spend most of his time in a hospital without actually needing it. Parameters that could be worth monitoring will of course vary from patient to patient. This depends on the handicap and diseases the patient has. This scenario is quite similar to the use case in Figure

43 USER GR OUPS FOR RELATIVES A monitoring system should help the relatives do the job needed to keep their child out of hospital. This should be possible through an alarm system. If a vital sign goes out of legal interval (which should be possible to set individually) an alarm goes off and relatives can do the desired action to prevent damage to the child. This could be very useful for children without the ability to talk, and relatives could have their child at home in a safe way. A search function should be available. There should be possible to search for values of vital signs and time periods. This should return monitoring data above or below the desired value. This could help relatives see patterns in their child s behavior and vital signs, which in turn can help them understand and perhaps prevent future events to happen. To ease the use of the system, a helping function should be available FOR MEDICAL PERSONNEL The alarms should be saved together with monitoring data in a number of seconds before and after the alarm. This interval should be possible to set individually. The search function mentioned for relatives should also be available to medical personnel. This could give useful information about what actually causing the abnormality in vital signs. Also, it could help make the treatment better, and find right medication if that is of interest. Generation of reports should be possible in the system. Reports should present monitoring data around significant events in an easy readable way. The reports should be presented in a way that medical personnel are familiar with, and with all information useful to them. Reports could be printed and used together with the patient journal. If relatives want reports, it should be possible to change the report outline to make them understandable to others as well. 33

44 DETAILED USE CA SE 7 DETAILED USE CASE We have selected to explore a concrete case about a functionally disabled child. This chapter describes this particular patient and the requirements needed for relatives and medical personnel to take LifeShirt and the information system in use. 7.1 PATIENT BACKGROUND Linda is a 7 year old functionally disabled child. She has respiration problems of unknown origin. These respiration problems occur with varying frequency and in different situations. Since Linda can t talk, diagnosing the cause of the respiration problems is difficult. In addition it s very difficult to detect any warnings of an apnea or another problem arising. Linda has been diagnosed with Arnold Chiari syndrome. Arnold Chiari syndrome leads to structural defects in the brain (National Institute of Neurological Disorders and Stroke, 2007). She has gone through several operations, and she has currently an intracranial shunt. The shunt is there to extract excess fluid from her brain. Linda has also been diagnosed with bradycardia; this is when one has a heart beat below 60 beats per minute. This limit is relative, and a heart beat below 60 can be normal for athletes who are in a good shape, however for other people, such a low heart beat signalizes that the heart is unable to pump enough blood to the body. Due to Linda s bradycardia and respiration problems, she has been resuscitated multiple times. 7.2 TODAY S MONITORING Linda is under constant surveillance by her parents. Her mother is a nurse and can therefore provide resuscitation should it be necessary. At night Linda is connected to a SpO2 monitor that monitors the saturation and pulse. A wireless video camera is installed in Linda s bedroom, it transfers video and audio to a black and white screen so that Linda can be monitored from other places in the home. When the parents are unavailable to take care of Linda, whether it s because of work or as a means of load removal, Linda is put into a hospital. There she is under constant surveillance by a selected group of nurses. Figure 7-1 shows today s situation where Linda s parents are sleeping with the TV right next to the bed. The camera is monitoring Linda when she is in bed. 34

45 DETAILED USE CA SE FIGURE TODAY'S SITUATION (ELECTRONIC ARTS INC., 2000) 7.3 MONITORING WITH THE LIFESHIRT Since Linda can have an apnea at any time, it is vital that an alarm will go off as soon as possible. This can be done by analyzing the respiration pattern monitored with the LifeShirt. At nighttime Linda can be hooked up to the SpO 2, for additional readings of the saturation. It is assumed that a SpO 2 sensor will be obtrusive for normal activities during the day, and thus the system has to rely on the other sensor readings. Information gathered from the LifeShirt can be transferred to a PC and will together with the video surveillance form an automated system for alarms when there is a problem. Figure 7-2 shows a possible future situation. A sound system is installed in each room, and when something is wrong with Linda, an alarm goes off everywhere in the house. Computers and laptops are all connected with the system through a wireless network and these can be used to check monitoring data and stored data at any time. 35

46 DETAILED USE CA SE FIGURE POSSIBLE FUTURE SITUATION (ELECTRONIC ARTS INC., 2000) 7.4 DIAGNOSING WITH THE LIFESHIRT Linda is unable to communicate, and her symptoms occur with no apparent pattern, it is therefore very difficult to diagnose her. Monitoring data of her vital signs leading up to an apnea can help to understand why it happened. At the moment it is difficult to understand what happens before an apnea. By utilizing the constant monitoring with the LifeShirt of several vital signs at once, there s a possibility for finding the cause of her apneas. 7.5 PHYSICIAN SUPPORT Medical personnel from the pediatric department at Linda s hospital are positive to a new monitoring solution for Linda. Linda s physician and the chief of the pediatrics department have provided this project with valuable input. They have been shown VivoLogic and the capabilities of the LifeShirt system, and given requirements for what they will want in a monitoring application for this use case. 7.6 TECHNICAL REQUIREMENTS OF APPLICATION FOR THIS USE CASE The application will have a different target audience than VivoLogic. It is to be used by physicians, parents or other people. Thus it will need to generate different user interfaces for different users. 36

47 DETAILED USE CA SE Since users can be interested in various items, graphs have to be more dynamic. They should support resizing of scale. Thresholds for different measurements should be possible to input for a subject. When a threshold is breached, an alarm should be set off. E.g. heart rate lower limit: 40 and upper limit 150. If the threshold is breached, sound audible alarm and contact monitor personnel on duty. The system should support extensive logging of alarms and events in the back-end system. This data will be mainly used for research purposes and data stored should be sufficient for this use. Since the system will be used by the parents, more detailed explanation of different parameters must be available. It should also be easy to see if the values are within normal range. Increased search functionality for both thresholds and time periods should be available, and presented in an easy way to the user. Medical personnel and the parents might have different focus on this search result. The system has to be able to generate reports, primarily for the physician. This generation should be dynamic and let the user specify which attributes the report should contain. Significant events should be listed with the monitoring data of the five seconds before and after the event. The application has to support generation of reports for parents or other users as well. Reports should be formulated and adapted to the user group in question. Layman terms will replace medical terms for the parents et cetera. The system has to support different methods of data capture and integration. The data can be captured on a memory card in the LifeShirt and then later uploaded to the main system, or it can be transmitted continuously over wireless internet. The system has to support different behavior depending on what mode is chosen. In this case the valuable readings are: Heart beat (ECG) Oxygen saturation, SpO2 Respiration Table 7-1 summarizes these requirements. 37

48 DETAILED USE CA SE FR1 FR2 FR3 FR4 FR5 FR6 FR7 FR8 FR9 FR10 FR11 FR12 Relatives and medical personnel should be met with different user interfaces User interface for relatives should contain only the most necessary information in a simple and easy understandable presentation Graphs should support scale resizing It should be possible to specify threshold values for different parameters When a parameter reaches the threshold, an alarm should go off Monitoring data around an alarm should be stored A help function describing different parameters should be available to relatives Monitoring data should be presented to relatives in a way that makes it easy to see if parameters are within normal range It should be possible to search for monitoring data, based on threshold values, alarms and time period It should be possible to generate reports based on input parameters The system should support real time monitoring The system should support uploading data from a memory card FUNCTIONAL REQUIREMENTS 38

49 SOLUTION ARCHITEC TUR E 8 SOLUTION ARCHITECTURE In this chapter we outline an information system for our specific scenario. The main point of our solution is to provide support for the parents of the patient, and it is therefore planned for home use. We assume that the home has its own wireless network. 8.1 OVERVIEW This section presents an overview of the architecture and describes each of the main components. The solution is divided into four separate areas; Clients, Server, Hardware and Network. The clients provide either a web based presentation layer or a local application presentation layer. The server layer receives data from hardware components. These data are processed and stored. This layer exposes web services to clients. Figure 8-1 shows the high level architecture. 39

50 SOLUTION ARCHITEC TUR E FIGURE ARCHITECTURE OVERVIEW CLIENTS The clients have two responsibilities; to provide a graphical user interface to the user and to store user input and retrieve data. We have two clients, one web client and one local client. The local client will be available at computers in the local area network where the LifeShirt is used. 40

51 SOLUTION ARCHITEC TUR E There are two different clients for two different usage areas. Making a clear distinction between them makes it easier to control who have access to configuration settings. This makes it safer and the user has increased control. On the other hand, it could be very useful to have the application available over internet, so that medical personnel and others can easily access monitoring data SERVER The server contains a SQL database containing all relevant data. Data is retrieved and stored through the business logic. Business logic processes data from both the user and the LifeShirt hardware. Web services provide an interface for the clients to access and store data. The server will continuously monitor the vital signs based on the current configuration. In the configuration one can specify threshold values and what action should be taken when they are breached. For instance a heart rate below 40 should trigger an audible alarm HARDWARE The server should have adequate memory and storage. A normal computer system should be sufficient. The server will be connected to an audio system in the user s home NETWORK The server will be deployed in a LAN located in the user s home. The LAN will have a wireless access point and be connected to the internet. 8.2 INFRASTRUCTURE This section provides an overview of the physical infrastructure demonstrated in the figure below. 41

52 SOLUTION ARCHITEC TUR E FIGURE INFRASTRUCTURE OVERVIEW NETWORK An Ethernet LAN will be connected to a wireless access point, providing both wired and wireless transfer. If one wireless access point does not provide sufficient coverage, multiple wireless access points will be installed. The LAN will be connected to the internet, providing remote access SERVER The home server is the main hub of the system. It will receive the vital signs from the LifeShirt, process and store them. It is connected via a wire to a home audio system so that an audible alarm can be sounded throughout the house. The server will expose web services to applications wishing to connect to live monitoring. It will also act as a web server providing users to connect to it remotely through the internet LOCAL APPLICATION The local application is a normal windows application; it will connect to the server s web services. 42

53 SOLUTION ARCHITEC TUR E REMOTE APPLICATION A user can connect to the home server via a simple web interface. This connection is over HTTP and should be secured with SSL THE LIFESHIRT USER The LifeShirt will be configured to automatically connect to the wireless network. This is done in the initial set-up. No further input from the LifeShirt should be required when the LifeShirt is started for a monitoring session THE FUTURE: A CENTRAL SERVER In the future the home server may be no more than a shell forwarding the data to a server in a remote location where all the main processing and storage happens. Such a central could be manned constantly, providing support for monitored patients. 8.3 APPLICATION ARCHITECTURE This chapter presents the application architecture. It is decomposed into two components; a web portal and a local portal WEB PORTAL Figure 8-3 shows an overview of the web portal components. The portal is the interface the user always is presented with, independent of the user s role. After a log in, the system retrieves information about what site should be viewed. The front end logic contains all logic necessary to present data on the screen. The physician site is adjusted to meet their requirements for a user interface. This will typically be presentation of several patients in one screen and detailed reports of monitoring data with medical terms. The patient site is adjusted to meet the everyday man s requirements. This means a simple interface with help functions where this is required. FIGURE WEB PORTAL 43

54 SOLUTION ARCHITEC TURE LOCAL PORTAL Figure 8-4 shows an overview of the local portal components. The graphical user interface is in many ways similar to the web interface, but contains more choices and user input. For instance input that changes configuration in the application, like thresholds. You can also choose to upload monitoring data from a memory card, and this data is processed in the offline data processing module. The data is transformed in a way that makes it readable to the business logic at the server. FIGURE LOCAL PORTAL 8.4 DATAFLOW This section provides a high level description of some of the important data flows. These are LifeShirt data entry, user login in local application, physician use of the web interface and data offline input LIFESHIRT DATA ENTRY The LifeShirt will automatically send its data stream to the home server; this is configured manually at installation. The business logic processes the data to find relevant data to store. If all monitoring data is stored, there would be too much data and only a small part of it would be usable later on. FIGURE LIFESHIRT DATA ENTRY 1. LifeShirt sends data to server 2. Business logic processes data 44

55 SOLUTION ARCHITEC TUR E 3. Relevant data is stored in the SQL server USER LOGS INTO LOCAL APPLICATION The user needs a username and password to get the right data and settings from the system. These are presented on the screen as graphs and menus. FIGURE LOCAL APPLICATION USER 1. User inputs username and password 2. System asks for user s settings 3. Application calls for data from a web service 4. Business logic asks for data from database 5. Data is returned from database to business logic 6. Data is sent to front end logic 7. Data is sent to GUI 8. Graphs and menus are presented to the user 45

56 SOLUTION ARCHITEC TUR E PHYSICIAN USES REMOTE WEB INTERFACE Physicians can log in to the system through the web interface. This makes the system accessible to medical personnel everywhere. When someone logs in from the web, the system detects whether this is a physician or patient/relative, and presents different user interfaces based on this. This is because physicians and patients have different needs. When physicians log in, the system detects what data that is relevant to the particular physician and generates an overview report. For instance a list of all the patients the physician wants to follow. FIGURE USE OF WEB PORTAL 1. Physician logs into site 2. System asks for user data in front end logic 3. System asks for user data through a web service at the server 4. Business logic asks for user data from database 5. User data is returned to business logic 6. Business logic sends relevant user data to front end logic 7. Front end logic detects what site should be used and sends data 46

57 SOLUTION ARCHITEC TUR E 8. Physician are presented with a overview report DATA OFFLINE INPUT If the user wants to use the LifeShirt somewhere without wireless connection, he could upload the monitoring data from a memory card. The business logic finds interesting monitoring data and stores this. Again, there is no use in storing all data, whereas data before and after abnormal vital signs is valuable. Between point 1 and 2 in this dataflow we assume the same dataflow as in Figure 8-6. FIGURE OFFLINE STORING OF DATA 1. The user logs in 2. User choose to import data from memory card 3. Data is uploaded from memory card 4. Data is sent to front end logic 5. Data is sent to business logic at server via a web service 6. Data is processed in business logic 7. Relevant data is stored in database 47