Survey On Machine Learning Based Patient Monitoring Algorithm Using Oxygen Saturation

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 09 Issue: 06 | June 2022 www.irjet.net p-ISSN: 2395-0072
© 2022, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page 3335
Survey On Machine Learning Based Patient Monitoring Algorithm
Using Oxygen Saturation
Kavya M1, Harshavardhana Doddamani2
1Student, S J C Institute of Technology, Chickballapur, Karnataka, India
2 Assistant Professor, S J C Institute of Technology, Chickballapur, Karnataka, India
---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - For patients with cardiac and pulmonary
problems, as well as those undergoing surgery, continuous
monitoring of blood oxygen saturation level (SpO 2) iscritical.
In a clinical environment, SpO 2 monitoring is commonly used
to assess the effectiveness of lung medicines and ventilator
support. A pulse oximeter is a piece of medical equipment that
measures the amount of oxygen in a person's bloodstream, or
how much of the oxygen-carrying molecules known as
haemoglobin) are simply transporting oxygen around the
body. Pulse oximetry is based on the concept that two
wavelengths can be used to choose between arterial blood
oxygen levels, given that the measurements are made on the
pulsatile component of the signal. Here, the system
continuously monitors the temperature and Spo2 level in
order to construct a machine learning-based algorithm for
patient severity prediction and upload it to a private server at
regular intervals. So, the doctor or carer can keep an eye on
the situation from a far.
Key Words: Oximeter, spO2, Senso, Raspberry Pi,Pulse
1. INTRODUCTION
The technology business was worth more than $22.5 billion
in 2018, according to Global Data,andisexpectedtoincrease
to $55 billion by 2023. (Global Data, 2019). Wearable
gadgets have been employed in a variety of industries, but
healthcare is the most potential use in terms of tackling
rising healthcareexpenditures,anageingpopulation,and the
prevalence of chronic disease. Wearable technologies are
self-contained, non-invasive devices that record, analyse,
and collect biological data in thehealthcareindustryinorder
to improve human health and welfare. By focusing on
diagnosis, treatment, monitoring, and prevention, recent
innovations in healthcare have added value. Customization,
early detection, remote patient monitoring (RPM),
medication adherence, knowledge libraries, and improved
decision-making are just a few of the advantages found
throughout the healthcare value chain, all while cutting
healthcare costs. Corporations that provide health
technology to clients for a variety of reasons have taken
notice of the growing demand and use of the technology
(GlobalData, 2019). Experts predict that COVID-19 will
remain for months, if not years, before being eradicated. Its
cure depends on early detection and screening, which will
define how it is treated. Because of the virus's quick
contactless transmission, screening it is difficult.
Furthermore, COVID-19 take a look at outcomes can take in
to forty-eight hours to arrive. The medical community
requires inexperienced techniques for detecting the virus
with inside the least amount of time. (Mukhtar, Rubaiee,
Krichen, & Alroobaea, 2021) In this study, we gift a
technique for figuring out and tracking a patient's blood
oxygen saturation tiers to decide the severity of the illness.
We gather facts from them the use of IoT gadgets and
remotely tracking their fitness situations. The proposed
technique assists docs and healthcare specialists to reveal
sufferers constantly and decide the severity of the
circumstance via way of means of enforcing this detecting
and tracking system. As a result, hazard on a patient's
existence and the hazard of contamination to scientific
employees may be decreased via way of means of the use of
this early detection system.
Types of Patient Monitoring system :
Remote patient monitoring:
Remote patient monitoring (RPM) devices are expected to
grow at a rate of 12.5 percent per year over the next decade.
Given the cumulative effects of the ageing population, the
high expense of in-patient care, andtheimmensedemand on
hospitals imposed by COVID-19, the trend is entirely
predicted. More equipment for remote monitoring is being
used by healthcare providers than ever before. In order to
free up beds, many hospitals began remotely monitoringthe
vital signs and symptoms of both coronavirus and non-
coronavirus patients. Authorities were in favour of RPM
methods. The Centers for Medicare and Medicaid Services
(CMS) in the United States has increased thelistoftelehealth
services that can be reimbursed. And the US Food and Drug
Administration (FDA) authorized the use of non-invasive
devices designed for hospitals in home settings. Remote
patient monitoring Also called telemonitoring and in-home
monitoring, remote patient monitoring is the set of
technologies and practices enabling healthcare providers to
track real-time changes in a patient’s health data from a
distance and use it in a treatment plan. It’s an integral
component of the broader telehealth industry and e-health
domain.

In addition, the US Food and Drug Administration (FDA) has
approved the use of non-invasive hospital devices in home
settings. Patient monitoring via the internet Remote patient
monitoring, also known as telemonitoring or in-home
monitoring, is a combination oftechnologiesand procedures
that allow healthcare providers to follow real-time changes
in a patient's health data and use it in a treatment plan from
a distance. It's an important part of the broader telehealth
and e-health industries. An IoT system can be thoughtof asa
remote patient monitoring system. It consists of four
important components in the most common scenario: a
personal medical device with a Bluetooth module, a patient
mobile app, a cloud repository, and hospital-side software.
'Activity monitoring' is anothertermforprocessmonitoring.
The only objective of process monitoring is to track the
usage of inputs and resources, as well as to examine how
activities and outputs are provided, and it is applied at the
early stages of a project. It is frequently carried out in
conjunction with compliance monitoring and feeds into the
impact assessment.
Compliance monitoring: As the name implies, compliance
monitoring's goal is to assure adherence to donor
restrictions, grant requirements, contract requirements,
local governmental regulations and laws, ethical standards,
and, most significantly, adherence to the project's desired
outcomes. Compliance monitoring may be required at any
point during the project life cycle.
Situation monitoring is a term used to describe context
monitoring. It keeps track of the project's general
environment. At any stage during the project cycle, context
monitoring allows us to detect and measure risks,
assumptions, or other unforeseen scenarios that may
develop within the institutional, political, financial, and
policy environment. These assumptions and hazards are
external elements that are outside the project's control;
nonetheless, context monitoringallowsustorecognisethem
in time to influence the project's success or failure.
Beneficiary monitoring, also known as 'Beneficiary Contact
Monitoring (BCM)', is a sort of monitoring that can be
required at any point during the project cycle. Its main goal
is to track the overall perceptions of project beneficiaries,
both direct and indirect. Itcomprisesbeneficiarysatisfaction
or complaints about the project and its components, such as
their involvement, treatment, resource access, and whether
or not those resources are equal, as well as their overall
transformation experience.
Stakeholder concerns and feedback mechanisms are also
tracked through beneficial monitoring.
Financial Oversight: The mainpurposeoffinancial oversight
is to measure the financial efficiency within a project. It
tracks the actual spending and allocated budget of the
project and helps the project team develop strategies to
maximize results with minimal input. This is often done in
combination with process and compliancemonitoringandis
important for accountability and reporting purposes.
Organizational Monitoring: As the name implies,
organizational monitoring tracks organizational
development, communication ,collaboration,sustainability,
and capacity building with partners andstakeholders within
the organization and related to project implementation.
Result monitoring: Here, monitoring is intertwined with
evaluation. Gather data to show the overall impact and
impact of your project on target groups. This helps the
project team determine if the project is doing well with the
intended results and if there are any unintended
consequences
Risk-based monitoring:
A risk-based approach to studying quality and monitoring
clinical trials has a simple focus. Identify potential threats
Plan to monitor these activities Adjust the monitoring
method as needed However, RBM is a very subtle approach.
Monitoring techniques should be tailored to the risks
identified by the clinician, CRO, and / or sponsor. Significant
risk can be monitored in a variety of ways, including the
above methods (onsite, remote, centralized), but the
ultimate effort is often a combination of these methods. All
three clinical trial monitoring methods can be used together
for effective monitoring, data integrity and patient safety
monitoring and protection.
Centralized monitoring:
Central monitoring includes analytical assessments
performed by a sponsor representative or representative at
a central location other than where the trial is being
conducted.
Onsite monitoring:
Onsite monitoring includes face-to-face assessments by the
sponsor's representative or the representative of the study
facility
Vital Monitoring:
Monitoring patient vital signs such as heart rate, blood
pressure, and temperature is one of the most basic and
important types of patient monitoring. As the patient's
condition worsens, the first signs appear in his vital organs.
The devices that monitor vital signs may vary depending on
the information displayed. Most devices display at least the
three vital signs mentioned above, but more advanced
devices can also display respiratory or blood oxidation
levels.

2. OBJECTIVE
The foremost goals of this have a look at are as follows: 1. To
Collect oxygen saturation of sufferers 2. To layout and
expand Machine getting to know primarily based totally set
of rules for Severity Prediction of Patient 3. To validate the
overall performance of evolved Machine getting to know
primarily based totally set of rules for Severity Prediction
3. PROBLEM STATEMENT
Healthcare sensorsembeddedinwearabledevicesarepretty
not unusual place those days. But using oximeters in
wearable gadgets and far off tracking to stumble on signs of
COVID-19 is fairly new. The present answers are both
transportable oximeters which want to be person initiated
however can’t be placed on their palms fora longtermor the
oxygen video display units in hospitals which aren’t
transportable. If a medical doctor wishes to peer the SpO2
readings constantly over a protracted length of time, they
should admit the patient. There is a gift scarcity of wearable,
compact, wi-fi system that may transmit the SpO2 and pulse
charge over IoT to the far off location, wherein the medical
doctors can display the sufferers with out requiring them to
be admitted to the hospital.
4. LITERATURE SURVEY
[1].A look at (Chen & Tang, 2020) mentionedhowdrastically
movement disturbances should affect blood oxygen
saturation tracking in a converting environment. Their look
at offered a brand new adaptive set of rules primarily based
totally on adaptive filtering to deal with the issue. They
extensively utilized the envelope statistics of the PPG
(Photoplethysmography) to retrieve the aspect of the
photoelectric extent pulse charge wave, construct actions as
reference signals, and perform adaptivefilteringtosuppress
motion intervention primarily based totally at the
interference evaluation of the photoelectric extent pulse
wave sign in a converting environment. The adaptive
cancellation method offered of their look at examines the
calculation consequences after appearing virtual sign
processing at the amassed unique data.
[2] In some other study, researchers (Mukhtar et al., 2021)
invented a scientific tool made of collapsible sensors that
may be used remotely and in real-time to reveal the fitness
of folks who show off signs of the coronavirus or are
inflamed with it. Wearable scientific sensors are coupled
with an Arduino hardware interface and a telephone utility
of their tool. An IoT framework changed into installed,
permitting a couple of gadgets to attach in real-time. By
utilising the algorithm, the scientific tool is used to come
across the patient's crucial repute of the affects of the
coronavirus or its signs with a pulse, cough, temperature,
and oxygen concentration (SpO2).
[3] Recently, in a studies project, authors (Tham, Markom,
Bakar, Tan, & Markom, 2020) evolved a tracking tool for the
aged to evaluate coronary heart or pulse fee and outside
capillary oxygen saturation(SpO2).MAX30100 changedinto
used because the front-cease sensor and the node MCU
(ESP8266) to accumulate and transmit information to the
cloud. Both are merged withinside the tracking gadgettoget
a median and bizarre classification. The tracking gadget for
SpO2 and coronary heart fee is prepared to apply primarily
based totally at the outcomes obtained. In addition, the IoT
gadget permits a big wide variety oftestedcustomers tohold
song of the patient's repute.
[4] While Kadarina and Priambodo (Kadarina & Priambodo,
2018b) have designed the entire gadget of their studies that
specializes in growing transportable clinical gadgets and
statistics visualizationona pc-primarilybasedtotallyclinical
pc the usage of an open-supply IoT platform. The cellular
clinical tool may have the capacity to degree coronary heart
charge and SpO2, abnormality reputation and ship them to
the server wherein caregivers may have get right of entry to
to view the statistics visualization. Recently a few
researchers (Si et al., 2021) designed a helmet-hooked up
wi-fi oxygen saturation tracking at the head (WORTH) to
evaluate nearby cerebral oxygen saturation the usage of
NIRS. A particularly included middle block, which
incorporates an optical module, a CPU unit, a wi-fi verbal
exchange module, and an strength control module, is
included with inside the band. A regulated hypoxia check
and a squat-to-stand check are used to evaluate the WORTH
band's performance. The findings display that the WORTH
band can correctly degree cerebral oxygen saturation, same
to that of a scientific screen and that it may additionally be
used throughout movement applications.
[5] In the early stages, researchers (Kanva, Sharma & Deb,
2014) used cell phone cameras to estimate pulse oximeter
(SpO2) levels and heart rate. Chao Lietal. (Li, Hu & Zhang,
2017) has developed an IoT-based heart diseasemonitoring
system. The system continuously analyzes patientvital signs
such as blood pressure, ECG, SpO2, and associated
environmental data. Finally, Adiputra et al. (Adiputra,
Hadiyoso & Sun Hariyani, 2018) has developed a low cost
SpO2 device for health monitoring using finger sensors and
node MCUs. Data sent to the server, stored in the database,
and displayed on the website.
[6] Another study by researchers (Ekawita,Nasution,Yuliza,
Suardi, and Suwarsono, 2020) proposed a non-invasive
glucose monitoring method based on a wireless data
transmission system based on the Android application of
mobile phones. With this method, the measurement data is
sent automatically. In the recorded Android memory. After
that, the patient can monitor his blood glucose level on a
regular basis. In addition, automatic data storage makes
measurements easier, faster, cheaper, and allows you to

access the data stored in your phone's memory anytime,
anywhere.
4. PROPOSED SYSTEM
As specified, a wearable oximeter tool might be used to
accumulate the oxygen saturation statistics the use of its
sensor and ship it to the server thru a gateway. In this
research, the device tool is composed of:
1. Heart rate and pulse oximetry sensor to study parameters
required to calculate coronary heart price and SpO2.
2. The Raspberry Pi will interpret the sensor's uncooked
statistics, degree coronary heart price and SpO2,after which
switch it to the Raspberry Pi through USB serial wire.
3. Cloud platform may have a database device to preserve
coronary heart price and SpO2 statistics in tables. Once the
ordinary or atypical end result is given along side the
uncooked statistics, a application will evaluate the cost
received from the sensor to decide if it's far beneath or over
the ordinary range, coronary heart price, and SpO2values to
the server. Then, the computation will show up to offer facts
for a caregiver to determine at the early detection.
5. CONCLUSIONS
In this study, we gift a method for figuring out and trackinga
patient's blood oxygen saturation stages to decide the
severity of the illness. We acquireinformationfromthem the
usage of IoT gadgets and remotely tracking their fitness
situations. The proposed technique assists docs and
healthcare specialists to screen sufferers constantly and
decide the severity of the circumstance via way of means of
enforcing this detecting and tracking system. As a result,
threat on a patient's lifestyles and the threat of
contamination to scientific employees may be decreased via
way of means of the usage of this early detection system.
REFERENCES
[1] Adiputra, R. R., Hadiyoso, S., & Sun Hariyani, Y. (2018).
Internet of things: Low fee and wearable SpO2 tool for
fitness tracking. International Journal of Electrical and
Computer Engineering, 8(2), 939–945.
https://doi.org/10.11591/ijece.v8i2.pp939-945
[2] Ahmed, M. F., Hasan, M. K., Shahjalal, M., Alam, M. M., &
Jang, Y. M. (2020). Design and implementation of an OCC-
primarily based totally real-time coronary heart charge and
pulse-oxygen saturation tracking gadget.
https://doi.org/10.1109/ACCESS.2020.3034366
[3] Boudargham, N., Abdo, J. B., Demerjian, J., & Guyeux, C.
(2017). Exhaustive Study on Medical Sensors. International
Conference on Sensor Technologies and Applications,
(September 2017),1–2.Retrievedfromhttps://hal.archives-
ouvertes.fr/hal-03221928/
[4] Chen, Q., & Tang, L. (2020). A wearable blood oxygen
saturation tracking gadget primarily based totally on
bluetooth low electricity era. Computer Communications,
160(May), 101– 110.
https://doi.org/10.1016/j.comcom.2020.05.041
[5] Dias, D., & Cunha, J. P. S. (2018). Wearable fitness
gadgets-important signal tracking, structures and
technologies. https://doi.org/10.3390/s18082414
[6] Ekawita, R., Nasution, A. A., Yuliza, E., Suardi, N., &
Suwarsono, S. (2020). Development of Non-Invasive Blood
Glucose Level Monitoring System the usage of Phone as a
Patient Data Storage. Jurnal Penelitian Fisika Dan
Aplikasinya (JPFA), 10(2), 103.
https://doi.org/10.26740/jpfa.v10n2.p103-113

Survey On Machine Learning Based Patient Monitoring Algorithm Using Oxygen Saturation

More Related Content

Similar to Survey On Machine Learning Based Patient Monitoring Algorithm Using Oxygen Saturation (20)

More from IRJET Journal (20)

Recently uploaded (20)

Survey On Machine Learning Based Patient Monitoring Algorithm Using Oxygen Saturation