Patient Health Monitoring System Using IOT And AI ^†

Abstract

Our method is to keep a close eye on the patient’s health and inform the person responsible for their care on a regular basis. Since it is not possible for an individual to always watch over a single patient in a hospital setting, the primary objective is to monitor the patient around-the-clock. There is going to be aWe employed an apparatus that continuously observed the patient and sent frequent updates to the attending physician, regardless of the physician’s location. This device also has an extra feature that allows it to sound an alarm to notify hospital staff members of a patient emergency.

1. Introduction

There are rumours that deaths are happening in hospitals globally because of medical mismanagement and doctors’ negligence. AI is one of the fastest growing technologies throughout the world. AI has made the impossible possible, for example, robots that behave like humans (in some cases, better than humans). Expert systems have also been developed that can monitor and diagnose patients, in addition to performing tasks beyond human capabilities. This technology has been slow to improve; it has been in use since the 1990s, but only became famous in 2020 [1,2,3]. The invention of sensors has advanced every field. RPHM provides each and every detail of a patient’s health condition to their doctor and nurses, so that they can have complete knowledge to diagnose the patient. This helps to give the best treatment to the patient, so that they can recover as soon as possible. The proposed system helps to reduce deaths due to a lack of medical care. There will always be a doctor who is monitoring the patient through the sensors and updating the nurses on what medicines should be given through online mode. In this way, we can reduce the death rate due to a lack of monitoring and miscommunication between doctors.

2. Related Work

In recent years, in the Journal of Innovation & Knowledge, Ali et al. have said that RPHM facilitates decision making, early prediction and cure, performance efficiency, continuous monitoring, and time efficiency. But, it experiences errors in data digitization and consolidation, data availability, privacy issues, and decision making. It also has the potential for future enhancements; until now, monitoring of patients has occurred through video and providing in-person updates to doctors in hospital. In 2022, Mr. Thanveer’s paper “Remote Patient Monitoring Using Artificial Intelligence” advises to provide complete care before preparations for an emergency, as it is less expensive and there is less chance for miscommunication and supply demand mismatch. Greater utilization of telecommunication services increased provider costs, with arguments over rural and urban charges. In the paper “Smart Health Monitoring System of Patient Through IOT”, the patient’s BP and body temperature were monitored continuously but not accurately, resulting in decision errors and data collection errors. In “A Review of Uncertainty Quantification in Deep-Learning”, Abdar says implementation is easy, the training process is the same, complexity is reduced, it is fast, and it is suitable for big datasets. In this study, training process implantation is complex, more samples are needed for training, calculations are complicated, and it is very slow. In 2017, in “Smart Health Monitoring System”, Mr. Khan says that it helps some rural areas to connect to the clinics near the villages. However, there are some demerits like calculation errors, slow internet, and server crashes [4,5,6].

3. Methodology

The patient health monitoring system contains Raspberry Pi to control the sensors as well as other electronic parts; it is the most efficient platform and is more versatile. Using pulse sensors amped to measure patients’ heartbeat, it can visualize the heartbeat by connecting to the Arduino UNO. Pulse-rate-amped sensors have three primary links: red wire, dark wire, and purple wire. These should be connected to the Arduino board to reach the heartbeat. Now that we have the heartbeat, with the help of ECG measuring sensor, we will be able to display the ECG curves [7]. Pulse sensors measure the pulse of the patient. Consequently, the heartbeat pulse will be stored in the database. These stored records will be periodically delivered to doctor in charge, irrespective of location, place, etc. The additional and most important feature of this system is that the doctor will be provided with the specific range of the heartbeat and BP for every patient; if there are any abnormal changes or it drops below the expected range, there will be an alert for staff nearby the patient or in the hospital in the form of alarm with the bed number [8]. This ensures the patient receives immediate treatment, which can help them to survive. In this way, the patient can be monitored continuously by doctors, who can check and receive updates on the patient from any location. RPHM gives each detail of a patient’s health condition to the doctor and nurses, so that they can have complete knowledge to diagnose the patient. They should know where the emergency is, so the bed number is displayed with alarm sound. In this study, we used classification algorithms: Random Forest, Logistic, and Decision Tress. A classification algorithm is a Supervised Learning technique that is used to identify the category of new observations because of training data. By using IOT sensors, we can collect data from a patient and these data can be used to train the model to give an automatic alarm when a specific condition is met [9].

4. Results and Comparison

The RPHM system can be used to continuously monitor a patient’s condition (heartbeat, BP, etc.), record the information obtained, and give the information to the doctors and nurses who are in charge.

The additional feature is that it gives an alert alarm when there is an emergency [10].

RPHM gives each detail of the patient’s health condition to the doctor and nurses, so that they can have complete knowledge to diagnose the patient. This helps to give the best treatment to the patient, so that they can recover as soon as possible (Figure 1). We do not know when there will be an emergency, so the patient should be continuously monitored; when there is an emergency, this system will produce an alarm sound so that the staff and doctor can provide immediate treatment, which helps the patient to survive [11]. They should know where the emergency is, so the bed number is displayed with the alarm sound, allowing staff to reach the patient as fast as possible. Compared to the existing systems, this is more accurate in terms of measuring [12].

5. Conclusions and Future Enhancement

Continuous monitoring of patients in hospital is very important and it should be conducted very efficiently. The proposed system helps to reduce deaths due to a lack of medical care. There will always be a doctor who is monitoring the patient through the sensors and updating the nurses on what medicines should be given through online mode in

Table 1. Applications of sensors.

Sensors	Applications
Accelerometer, Motion	Patient motion activity and inactivity sensing detects the presence or lack of motion.
Optical AFE, ECG	Heart rate monitoring spo2, industrial monitoring
Temperature	Medical equipment, environment monitoring, HVAC
Bio Impedance	Photo plethysmography, photodiode measurements
Capacitive sensors	For touch screen sensors

.

As we know, every patient’s health condition is different from others. So, we utilize user input from the attending doctor to facilitate patient monitoring that is easy and effective. Specifically, the heartbeat range and BP (etc.) are different for each patient; if there are any abnormal changes or it drops below the minimum range given by the doctor, then it sends alerts to staff and doctors, which helps to provide immediate treatment. In this way, we can reduce the death rate due to a lack of monitoring and miscommunication between doctors. In future enhancements, we can improve the system to measure recovery in bones that are fractured and show the recovery percentage to the injured person and doctor. It would also tell the patient when they should visit the doctor.