Healthcare Systems and Challenges with Internet of Things

Healthcare Systems and Challenges with Internet of Things

Akhila Jagarlapudi

Department of Electrical Engineering VJTI, Mumbai

Sankalp Arora

Department of Electrical Engineering, VIT, Vellore

AbstractKeeping in mind the growing need of good remote healthcare facilitates, it always gives rise to a question that how well are these options and their performance. In this paper, we walk through the key challenges and propose a layout that can be implemented for higher success. During Pandemic times, health- care has become the prime importance and hence IoT plays a vital role in improving the present conditions.

KeywordsIoT; Health-care; Sensors; Big Data

1. INTRODUCTION

   As we know, Internet Of Things is essentially a network of various devices that are embedded with electronic devices and merged with some software enhance-ability to collect and exchange data [1]. As we know, the world population is increasing tremendously and urban areas are constantly facing a great deal of poor health challenges. Even though the medical resources and facilities in cities are expanded daily, still the sufficient level is not attained. The massive pressure towards the management of healthcare in cities has triggered the advancement in technologies to come out with the proper solutions to the booming problems [2]. The Internet of Things (IoT) is a rapidly evolving technology that can help health care services go to the next level. It enables seamless networking between patients, medical equipment, and clinicians by ensuring that economical, low-cost, dependable, and handy gadgets can be carried or integrated with patients. The sensors will continuously capture signals, which will then be associated with key physiological characteristics and transmitted through the wireless network. With the current health records, the generated data is stored, processed, and evaluated. The advancement of technology will have a profound impact on every human's existence and health monitoring; it will significantly reduce the number of health-caused deaths. Progressive technology will have a dramatic impact on every human's life and health monitoring; it will significantly reduce healthcare costs and advance disease prediction accuracy.

2. CLOUD-BASED INTERNET OF THINGS SYSTEM The physical layer, network layer, software layer, and application layer are the four protocol layers that make up this system. The physical layer is made up of devices having sensors and transmitters integrated in them. The network layer is in charge of sending signals from sensors to Cloudlets, while the software layer is in charge of storing data in the cloud and making it accessible to those who need it. Finally, analytics and diagnosis are carried out in the application layer [2]. Physiological data, consisting of many necessary physiological characteristics, is obtained from the small sensors inserted within the patient's body. Then there's some modest hardware for preparing the data and some communication software for

   sending it. Sensors must be compact, light, and not obstruct the patient's mobility or movements. Those sensors must be powered by compact, low-energy batteries. The batteries are expected to work continuously without needing to be charged or replaced.

   The system components in charge of data transfer must be able to accurately and securely convert patient recordings from any location to the health centre.

   Short-range low-power digital radio Zigbee or Bluetooth can be used for transmission. Additionally, the data collected can be transferred to a health centre through the Internet for storage. The concentrator, which could even be a smart phone, can control the sensors in the IoT system over the Internet.

   Smart phones nowadays are equipped with far more complex features, allowing them to function as both LTE and WiFi. In this method, smart phones can operate as concentrators. The concentrator's data will be transferred to the cloud for storage. If this information is saved, it will be very useful for clinicians to access on demand or for analytics.

   When local resources are insufficient to meet the requirements, a small processing unit called cloudlet is employed for both storing and processing locally.

   It also aids in the execution of time-sensitive tasks on the medical data of patients. When data is kept in a cloudlet, it is accessible at all times, allowing data analytics to produce more accurate diagnostic information.

   Maintaining the security of a patient's electronic medical records while storing them on the cloud has become critical. When transferring offline data to the cloud, proper privacy- preserving procedures should be used to avoid unauthorised access. As a result, secure cloud storage frameworks have been developed to deal with sensitive medical data, but it remains a concern [3].

   Because medical databases are so large, data analytics is a major undertaking. This process of linking sensor characteristics and clinical data is done by machine learning algorithms. Pattern recognition and machine learning techniques will be used to analyse data from wearable sensors. Machine learning must be improved in order to manage more heterogeneous and constantly changing sensor data. Those algorithms must also be able to deal with missing data values, flowing data, and information of varying dimensions, which is unavoidable.

3. CHALLENGES IN THE CLOUD-BASED SYSTEM While doing the analytics process in the application of IoT

   in medical fields, we face three major problems. To begin with, new measurement devices and equipment are introduced virtually every day in the field of medicine. As a result, IoT

   devices must be updated on a regular basis, and sensor data will change. Obviously, this will have a significant impact on database design, and IoT devices must be able to manage all of this.

   Machine learning techniques are likely to be improved in the future to cope with the continuously changing sensory data. Secondly, the data to be collected will vary based on the condition of the patient, as advised by the physician. As a result, extra input changing over time is rather infeasible. Although it is possible to link past sensor data with clinical records, it is difficult due to the rarity of patient situations. The concept of classification and regression methods can be useful in preparing common training data for machine learning algorithms, but it

   will add to the physician's workload.

   Finally, as we gather information from various sources, the sensory data will result in a variety of modalities. As machine learning algorithms deal with homogeneous data, heterogeneity remains a barrier. Graphical models can be useful for combining various input data in a centralised framework with a lot of customisations.

   The medical data is plotted graphically to continuously monitor the patient's health, even though the sensor data is numerical. In health monitoring, the concept of visualisation is crucial. For effective prediction, data from IoT warble sensors is spread using several graphical approaches. In emergency situations, visualisation tools must constantly be ready to interface with heterogeneous data in order to make timely and accurate predictions. The visualisation must be able to handle static images in order to compare patient medical records.

4. CHALLENGES FOR BIG DATA IN HEALTHCARE In a fee-for-service environment, the only way for healthcare practitioners to be compensated is for them to interact with patients face-to-face. As a result, there is a strong ias against pushing technology that makes non-face-to-face encounters more efficient. However, as we move away from that model and toward value-based care, where delivery organizations (hospitals, patient-centered medical homes, accountable care organizations, and so on) are paid on a global risk-based basis, there will be more incentive to use new technologies that reduce unnecessary in-office encounters [4]. Face-to-face interactions are treated as a cost center rather than a profit center in such an environment, and excellent population health outcomes are rewarded. The condition of the health data is the most significant technical impediment to realizing this vision. Health data is largely split into institution-centered silos as a result of older EHR systems. Those silos can be quite huge at times, but they are still silos. Much of the present effort is focused on exchanging individual records between silos using more standardized vocabularies (code sets) and message formats (ADT messages, C-CDAs, and even FHIR objects). But that does not solve the problem of data

   fragmentation.

5. A PROPOSED MODEL

   We examine an automatic system for monitoring a patient's body temperature, heart rate, bodily movements, and blood pressure in this study. Using the numerous health parameters and various additional symptoms gathered by the system, we extend the existing method to forecast if the patient is suffering from any chronic ailment or disease [5].

   Fig. 1. Layout of the proposed model

   The diagram above shows how to extract information about a patient's health status by monitoring several metrics and then using that information to determine whether the patient has a chronic problem or another disease.

   • Unprocessed data from various IoT devices is collected and stored on the server at level 1. Temperature, vibration, blood pressure, and pulse sensors are among the sensors included in these devices. Because some sensors produce analogue output that cannot be read by the Raspberry Pi, we must first convert the analogue values to digital using a convertor IC. We next build python code that receives the values from the sensors and updates them into the database at regular intervals using the Raspberry Pi with Linux OS loaded.

   • In level 2, the important information is acquired by filtering, classifying, and categorising the data recorded. This data consists of the patient's current health statistics as well as any symptoms he or she may be experiencing. This information will be used in the next level to determine whether or not the patient is suffering from any ailment. This contributes to the system's intelligence and efficiency.

   • We use data mining techniques in level 3's analysis & prediction phase to predict the type and nature of the sickness or disorders for which the system was created. Artificial intelligence can improve the system even further by making it smarter. Using the existing information base, we can infer the disease or ailment

   and categorise the outcome into multiple categories such as Ideal, Normal, and With Symptoms, for example.

6. SYSTEM MODULE

   1. Health monitoring Section

      This module contains the physical components of the system that enable it to be IoT enabled, and it is used to track the patient's health parameters via various sensors. Because Raspberry Pi only works with digital signals, it serves as a central server to which all of the sensors are linked through GPIO pins or the MCP3008 analog-to-digital converter if their output is analogue. The pi reads the real-time information and writes them to a mySQL database, which is then shown on the web interface.

   2. Emergency Alert Section

      This module focuses on the procedures to take after a patient's health has been observed to show abnormality, such as contacting his or her family members and the hospital. In our programme, we can set up specific threshold values that, if exceeded, will send an email or SMS to the patient's family or doctor.

   3. Health Status Prediction System

   In this module, we use the patients' health data as recorded by our system, as well as any symptoms they may be experiencing, to compare it to the existing knowledge base and predict if the patient has any disease or disorder, resulting in an efficient Expert System with proper data mining techniques.

7. CONCLUSION

   We discovered the importance and potential benefits of implementing IoT in remote health monitoring systems in this paper.

   The IoT-enabled small sensors will have a significant impact on every patient's life, allowing them to minimise their

   fear of danger even when they are away from home and their physician.

   The sensory data might be collected at home or at work. Also mentioned are the problems in disease detection, analytics, and prediction, all of which can be overcome to create a smooth integration into the medical industry.

8. REFERENCES

1. A. Zanella , N. Bui, A. Castellani, L. Vangelista and M. Zorzi, "Internet of things for smart cities," IEEE Internet of Things journal, vol. 1, no. 1,

   pp. 22-32, 2014.

2. M. Sathya, S. Madhan and K. Jayanthi , "Internet of things (IoT) based health monitoring system and challenges," International Journal of

   Engineering and Technology(UAE), vol. 7, pp. 175-178, 2018.

3. M. Li, S. Yu, Y. Zheng, K. Ren and W. Lou, "Scalable and secure sharing of personal health records in cloud computing using attribute

   based encryption," IEEE Trans. Parallel Distrib. Syst., vol. 24, no. 1, pp. 131-143, 2013.

4. D. V. Dimitrov, "Medical Internet of Things and Big Data in Healthcare," Healthcare informatics research, vol. 22, no. 3, pp. 156-

   63, 2016.

5. S. Banka, I. Madan and S. S. Saranya, "Smart Healthcare Monitoring using IoT," International Journal of Applied Engineering Research, vol.

13, no. 15, pp. 11984-11989, 2018.