Table of Contents

I.0 Introduction................................................................................................................................3

1.1 Background to the problem...................................................................................................3

1.2 Research Questions................................................................................................................4

1.3 Aims and Objectives..............................................................................................................4

2.0 Literature review........................................................................................................................4

2.1 Existing applications of AI in healthcare...............................................................................5

2.2 Possible future applications of AI in healthcare...............................................................8

2.3 Challenges in Application of AI in healthcare....................................................................14

3.0 Proposed Methodology and Methods......................................................................................15

3.1 Research Methodology........................................................................................................15

3.2 Research Method.................................................................................................................15

3.3 Data Triangulation...............................................................................................................16

3.4 Ethical Issues.......................................................................................................................16

4.0 Conclusion..........................................................................................................................17

5.0 Proposed Timeline...................................................................................................................17

6.0 References................................................................................................................................19

1|Page
I.0 Introduction

1.1 Background to the problem

Healthcare is the most sensitive part of human life, and having a healthcare system that can

predict and thus prevent the occurrence of diseases, is key. While a few diseases that have a

major impact on human life have been prioritised in the research field to devise ways of

predicting their progression as well as their overall outcome for the patient within a given

timeline, there is still a long list of diseases that's yet to be addressed the same way or in a

similarly close approach (Tsui, 2020). This is the reason this research proposal is key in the

health industry. Artificial intelligence is a technology that has been applied in various aspects of

life and has, as a result, had a major impact on each of these aspects, such as aviation,

manufacturing, and even some aspects of the healthcare industry. Some of the instances where it

has been employed include diagnosing some serious diseases such as neural functionality and

cancer (Badnjević, Avdihodžić, & Gurbeta Pokvić, 2021, p. 230). However, these are not the

only major diseases in the health industry. They only cover a very tiny fraction of the problem.

Other diseases need prediction and diagnostic inception and designing the various healing

mechanisms that will guide the industry into faster disease prevention and treatment

management. This report's focus is to present a research proposal that guides the application of

Artificial intelligence in the healthcare industry.

The Torrens University Ethos is to be good. This means that as an institution, the goal is to

ensure that the education approaches employed and the projects and research approaches the

students and Alumni partake in are focused on the greater good of the society. In this regard, the

research proposal focuses on ensuring the healthcare industry is better equipped to save more

2|Page
lives and extend the prevention capabilities in the current healthcare industry. The proposal

offers the ideas and approaches on how they employ and assimilate AI, ensures the world

connects, in the process of offering not just good services, but better ones than the ones currently

in existence.

1.2 Research Questions

There are two key research questions to be answered in this proposal. These are:

1. How has AI been employed and used in the healthcare industry to diagnose, treat or

prevent different diseases?

2. In what other ways can AI technology be exploited to provide better care in the health

industry?

1.3 Aims and Objectives

Aims

This research aims to come up with a detailed list of other ways and techniques that AI can be

employed in healthcare. With this aim in mind, the proposal will develop a detailed analysis of

various requirements necessary for AI applications in healthcare.

Objectives

To achieve the aim of this project, they are objectives that have to be achieved in this proposal.

These are:

1. Address why Ai is necessary for healthcare

2. Explain ways in which AI can be employed in healthcare

3. Define several ways that AI can be applied in healthcare

4. Come up with a list of ways AI can be employed and implemented in the future

5. Explain how Ai with these applications will impact the healthcare industry

3|Page
2.0 Literature review

There are various applications of AI in the healthcare system already. Some of these applications

have impacted the industry positively, while others have been a major challenge in their

utilisation. The two aspects, understanding the past, present, and future of the healthcare system

concerning AI, are key. This is best carried out by analysing various articles written in this

regard.

2.1 Existing applications of AI in healthcare

Jiang et al. (2017, p. 239) discuss one key application of AI in healthcare that has a major impact

on the health industry. According to this article, AI has been employed in the early prediction of

stroke occurrence, the diagnosis of the disease, its treatment, and prediction of the outcome, as

well as prognosis. This has been achieved by applying machine learning-based algorithms to

detect the symptoms that cause and result in stroke earlier on. This process has two steps; the

first focuses on identifying the use of human recognition, and the second is the detection of

stroke onset symptoms. If any of these aspects seem to have deviated from what is considered

normal, a patient is under observation and even in line for treatment. Considering the impacts

stroke has had across the globe, this has helped reduce the mortality rate concerning stroke. Once

the disease has been detected, the diagnosis commences, which involves using neuroimaging

approaches such as the common healthcare technique CT and MRI, which help with the process

(Jiang et al., 2017, p. 240). In terms of stroke treatment, machine learning algorithms have been

used to predict the best approach for each patient based on past health history.

4|Page
Figure 1. Stroke detection, diagnosis, treatment, using AI(Source:

https://www.ahajournals.org/doi/10.1161/STROKEAHA.119.027479)

Another existing application of AI in healthcare is the statistical analysis of data to ensure that

the data is well analysed. No assumptions are made when developing and concluding on various

hypotheses (Trishan & Szolovits, 2018, p. 3). In healthcare, various hypotheses are formulated to

help analyse various diseases based on the data collected across several patients. In the analysis

and testing of such data, the accuracy of the outcome could have a significant impact on the

health and life of a patient and various discoveries in health that are considered revolutionary.

The application comes in this stage, where machine learning enhances the accuracy of the

various statistical data without applying any assumptions. The only focus is to look for patterns

and sequences that match the analogy/hypothesis under testing (Trishan & Szolovits, 2018, p. 3).

Apart from this, machine learning also comes in handy when analysing complex data that cannot

be easily analysed with simple statistical techniques without making assumptions that affect the

quality of data and the accuracy.

5|Page
Figure 2. .Data analysis process using machine learning in healthcare (source:

https://svn.bmj.com/content/svnbmj/2/4/230.full.pdf)

Lee and Yoon (2021, p. 7) describe another application of AI in the real world. This application

is the implementation of AI to carry out robotic surgeries in hospitals. The medical world is a

very dynamic one, with changes happening every few years. With this, so is the state of

infections and the level of treatment as well. Some of the surgeries that patients need to undergo

to be treated require surgery. The complexity and inflexible nature of these surgeries sometimes

make these impossible to acquire from health organisations. The complexity of these procedures

requires the doctors to be precise and have very convectional skills in the whole process. This is

why they employ AI to make up robots that have enough human intelligence to be part of the

process, without the likelihood of making a mistake, and under a well-controlled environment.

Also, this creates room for the surges to incorporate any new information arising even during

6|Page
surgery on a real-time basis and work with it effectively to better the patient's living conditions

(Lee & Yoon, 2021, p. 7).

Figure 3. AI techniques as they are applied in robotic surgery (Source:

https://www.imperial.ac.uk/news/200673/application-artificial-intelligence-ai-surgery/)

2.2 Possible future applications of AI in healthcare

Pesapane, Codari, and Sardanelli (2018, p. 4) stated that technologies such as MRI, tomographic

images, CT, position emission tomography, and tomosynthesis had been seen to increase the

amount of data that is processed in radiology. This influences the interpretation of images by

radiologists from only inference to merely description and detection. However, with the

increased number of images taken, the time for analysis is squeezed since too much time is taken

to analyse them. This makes radiologists lack the time for making clinical judgements. With AI,

in the future, it is possible to make use of the algorithms to examine and analyse them and

7|Page
identify patterns after training through numerous examinations of such images. These systems

will give information and characterisation of the abnormal findings in terms of conditional

probabilities. Such structured AI-aided reporting will help improve the analysis of images by

radiologists and also help them present the findings in a concise format (Pesapane et al., 2018, p.

5). The figure below shows how AI will benefit radiologists.

Figure 4: Benefits of AI to Radiologists (Source: https://www.barco.com/es/news/2018-01-26-

what-can-artificial-intelligence-do-for-radiologists)

Apart from the few AI applications already in existence, there are other ways AI can be applied

in healthcare based on the properties and functionalities that define it. According to Colling et al.

(2019, p. 143), AI has been the greatest technical achievement to hit the medical industry in the

20th and 21st centuries. One of the future applications defined is AI in digital pathology, aiding in

image processing and analysis methods. Digital Image Analysis, commonly referred to as DIA,

8|Page
is a very sensitive and significant aspect of digital pathology, with the ability to make a

groundbreaking impact on the ways and approaches employed when analysing the various

images. The idea behind image analysis is to use any computer aided processing techniques and

extract information from images and other user driven features. This process often involves

cutting edge technologies being applied and mathematical analysis and designs being produced

from the end product (Colling et al., 2019, p. 144). The machine learning algorithms can make

this process easier by automatically analysing images and giving the output that best fits the data

offered. These algorithms can be trained on various data outputs and the patterns expected to

make certain predictions or how they are to be diagnosed based on the imaging. This would

greatly fasten the image analysis process, making it easier and faster to receive treatment.

Figure 5. Digital pathology AI development approach(Source:

https://onlinelibrary.wiley.com/doi/epdf/10.1002/path.5310)

9|Page
According to Reddy, Fox, and Purohit (2019, p. 22), the potential role of AI in the delivery of

healthcare and medical research is increasingly becoming evident. In clinical decisions, AI has

possible applications and can guide doctors in making clinical decisions that are life-saving. Not

only are these systems capable of guiding doctors, but they are also capable of helping patients

attain personalised treatment conclusions by drawing from the large-scale data based on previous

cases (Reddy et al., 2019, p. 23). Some of these decisions are seen to be harder to make in

conventional clinical decision-making. AI can infer the patient's health status and recommend

alternative treatment paths even where the measurements are not available. The neural networks

are being trained on the medical diagnosis and appear to be capable of predicting medical

conditions that may guide clinicians (Reddy et al., 2019, p. 23). In this case, in the future, AI can

effectively match and surpass pathologists' performance and radiologists to better healthcare

delivery in the process shown below.

10 | P a g e
Figure 6: Application of AI in Clinical Decision Support (Source:

https://www.researchgate.net/figure/The-artificial-intelligence-AI-based-integrated-clinical-

decision-support-system-of_fig6_313237078)

(Noorbakhsh-Sabet, Zand, Zhang, Abedi, & Translational Informatics Institute, 2019, p. 3)

discuss precision medicine as another possible application of Ai in the healthcare industry. The

need for valuable and accurate treatment methods has been the topmost priority for the medical

industry for the last few years. The ability to accurately detect what a patient is experiencing and

offer the right treatment at the right time can save very many lives. There have been several

cases of patients being misdiagnosed or treated for the wrong conditions due to human error in

recent years. With AI, the patient data stored in the system is reviewed and analysed, and the best

treatment for a condition is developed for the patient. For instance, chronic diseases can be

detected and managed using AI. One example is machine learning algorithms to detect

retinopathy, where the two validation datasets are trained using deep learning and can look for

and detect various signs that lead up to this conclusion (Noorbakhsh-Sabet et al., 2019, p. 4).

11 | P a g e
Figure 7. Identification of Retinopathy, using AI (deep learning) (Source:

https://eandv.biomedcentral.com/articles/10.1186/s40662-020-00206-2)

There are many biomedical signals recorded by wearable devices based on the deduction by Yu,

Beam, and Kohane (2018, p. 726). Some of the signals captured include heart rate, tremor, voice,

and limb movement. These signals can be useful in the detection of diseases and also inferring

health conditions. Through AI, it is possible to achieve the early detection of diseases using heart

rate and the skin temperature recorded by wearables. Wearables include the

photoplethysmography sensor, which makes it possible to monitor pulmonary diseases,

cardiovascular diseases, sleep apnea, and anaemia (Yu et al., 2018, p. 726). With the inclusion of

AI, it is possible to detect specific conditions early from these vitals and help save lives. In the

future, wearables will become common and effective when used with AI to detect health

conditions early in advance.

12 | P a g e
2.3 Challenges in Application of AI in healthcare

Despite the many benefits associated with Ai and the positive impact it is likely to have in

healthcare, there are challenges associated with its application. Some of these challenges include:

 The system is likely to be biased against some of the subgroups and certain populations

based on what the Algorithms are trained on and the applied form of sampling. While AI

refers to human intelligence being replicated by machines, they are only able to showcase

what they have been trained on and are therefore likely to end up being biased in one way

or another based on the sampling data used in the training experience (Reddy et al., 2019,

p. 25).

 The decision-making process in AI is based on how machines process information. It is,

therefore, necessary for people to review the information that is put out and always be

very alert on the decision-making process employed by the machine (Reddy et al., 2019,

p. 25). The complexities involved in the machine learning algorithms and deep learning

may create opaque aspects in the decision-making process, a problem referred to as the

black-box problem.

 The application of AI in the medical field also creates a loophole on who is to be held

responsible and accountable for the use and applications of AI. While the algorithms and

systems using AI have been designed by people, it is hard to define who is to be held

accountable if the systems fail or directly or indirectly cause a fatality to the patient (Lee

& Yoon, 2021, p. 11).

 According to (Yu et al., 2018, p. 728), it is almost impossible to develop a certification of

the AI technologies before the larger-scale deployment of such technologies. This means

that FDA has to be willing to carry out pre-market approvals that they have reviewed and

13 | P a g e
 consider safe while waiting for the clinical impact of the devices. This is a challenging

grey area to be in terms of medical submissions.

3.0 Proposed Methodology and Methods

3.1 Research Methodology

The most appropriate and ideal methodology for this project is the qualitative approach. The

research is aimed at the applications of Artificial Intelligence in the healthcare sector and the

different ways that systems can be designed to increase efficiency in the sector. According to the

research questions, the nature of the research does not demand the use of numerical data to

answer them. The qualitative methodologies have been identified as valuable and credible for

empirical research (Miller, Chan, & Farmer, 2018, p. 1). The qualitative approach can generate

ideas for the research and aid in gathering insights regarding the application of AI in healthcare

(Center for Research Quality, 2015). The approach has been proven to be flexible. It can

effectively focus on obtaining rich insights from how AI systems can be effectively designed and

applied to deliver quality healthcare. As per the description of the personal lens by (Clark &

Vealé, 2018, p. 483 CT), the qualitative approach is the most appropriate approach since it will

help filter and interpret the information on the usage of AI in the healthcare sector.

3.2 Research Method

Many qualitative research methods can be utilised for this particular project. The most efficient

and applicable method is secondary research. This method uses data that is existing with the

applicable information and insights that are required for the project. For the project, secondary

data is ideal because it will help save time compared to other research methods. The university’s

online library was a good source of peer-reviewed articles and journals. Additionally, Google

Scholar, websites, and other authentic science journals repositories were used to obtain the

14 | P a g e
secondary data needed for the research. The secondary information used for the research can

easily be validated since their references will be recognised. As the researcher for this project, I

will be considered the instrument responsible for the interpretation, observation, filtering of the

data through my lens. The secondary method will be used to collect the data needed to answer

the research questions.

3.3 Data Triangulation

Triangulation refers to the process of using multiple data sources to have an understanding of the

concept that is being studied (Korstjens & Moser, 2018, p. 122). Since the research approach

used for the study is a qualitative one, triangulation is needed to verify the validity of the

secondary data sources used. The process of data triangulation will be utilised in the research on

the use of AI in the healthcare sector to gather various peer-reviewed articles to provide the

needed information. Triangulation will be achieved through the collection and storage of the

articles on EndNote software. The software has a plugin that is integrated into Microsoft Office.

3.4 Ethical Issues

 Privacy and Confidentiality are of great concern in the study. All of the information

gathered from the secondary materials and used in the study has to be protected through

paraphrasing (Guthrie, 2012, p. 15).

 Intellectual property refers to the creations of one's mind, which includes inventions,

designs and articles. The materials used for the study are the authors' intellectual

property, which should be addressed through referencing using APA 7 th edition to

acknowledge all the sources.

 The use of the secondary data should not result in any form of damage or distress to the

people who are affiliated with the sources, as discussed by Oliver (2010, p. 44)

15 | P a g e
The ethical issues identified affected the choices of the research tools utilised in the study since

there was a need to use software to ensure sources are well acknowledged.

4.0 Conclusion

The research is mainly focused on understanding the applications of AI in healthcare and how it

can be used to design effective systems. To effectively understand the research and its objectives,

the following questions are formulated: How has AI been employed and used in the healthcare

industry to diagnose, treat, or prevent different diseases? In what other ways can AI technology

be exploited to provide better care in the health industry? To effectively answer these questions,

the study uses a qualitative approach emphasising the secondary research method. Using this

approach, it is possible to achieve conclusive and elaborative research that sheds light on the

application of AI in the healthcare sector and how it improves the sector. In addition to the

research methodology and methods, the ethical issues that arise have been solved by effectively

referencing the sources used and paraphrasing the information obtained from the sources.

5.0 Proposed Timeline

16 | P a g e
17 | P a g e
6.0 References
Badnjević, A., Avdihodžić, H., & Gurbeta Pokvić, L. (2021). Artificial Intelligence In Medical
Devices: Past, Present And Future. Psychiatria Danubina, 33(Suppl 3), S336-S341.
Retrieved from https://lesa.on.worldcat.org/v2/oclc/9041909570. Retrieved from
https://lesa.on.worldcat.org/v2/oclc/9041909570
Center for Research Quality, C. (2015). Overview of Qualitative Research Methods. YouTube-
Video. Retrieved from https://youtu.be/IsAUNs-IoSQ. Retrieved from
https://youtu.be/IsAUNs-IoSQ
Clark, K. R., & Vealé, B. L. (2018). Strategies to enhance data collection and analysis in
qualitative research. Radiologic technology, 89(5), 482CT-485CT.
Colling, R., Pitman, H., Oien, K., Rajpoot, N., Macklin, P., Snead, D., . . . Verrill, C. (2019).
Artificial intelligence in digital pathology: a roadmap to routine use in clinical practice.
Journal of Pathology, 249(2), 143-150. doi:10.1002/path.5310. Retrieved from
https://lesa.on.worldcat.org/v2/oclc/8261056178
Guthrie, G. (2012). Basic research methods: An entry to social science research: SAGE
Publications India.
Jiang, F., Jiang, Y., Hui, Z., Dong, Y., Li, H., Ma, S., . . . Wang, Y. (2017). Artificial intelligence
in healthcare: past, present and future. Stroke and Vascular Neurology, 2(4), 230-243.
doi:10.1136/svn-2017-000101. Retrieved from
https://lesa.on.worldcat.org/v2/oclc/9127111173
Korstjens, I., & Moser, A. (2018). Series: Practical guidance to qualitative research. Part 4:
Trustworthiness and publishing. European Journal of General Practice, 24(1), 120-124.
Lee, D., & Yoon, S. N. (2021). Application of Artificial Intelligence-Based Technologies in the
Healthcare Industry: Opportunities and Challenges. International journal of
environmental research and public health, 18(1). doi:10.3390/ijerph18010271. Retrieved
from https://lesa.on.worldcat.org/v2/oclc/8873669425
Miller, R. M., Chan, C. D., & Farmer, L. B. (2018). Interpretative Phenomenological Analysis: A
Contemporary Qualitative Approach. Counselor Education and Supervision, 57(4), 240-
254. doi:10.1002/ceas.12114. Retrieved from
https://lesa.on.worldcat.org/v2/oclc/7948146006

18 | P a g e
Noorbakhsh-Sabet, N., Zand, R., Zhang, Y., Abedi, V. B., & Translational Informatics Institute,
G. H. S. D. P. (2019). Artificial Intelligence Transforms the Future of Health Care. The
American Journal of Medicine, 132(7), 795-801. doi:10.1016/j.amjmed.2019.01.017.
Retrieved from https://lesa.on.worldcat.org/v2/oclc/7990656263
Oliver, P. (2010). The student's guide to research ethics: McGraw-Hill Education (UK).
Pesapane, F., Codari, M., & Sardanelli, F. (2018). Artificial intelligence in medical imaging:
threat or opportunity? Radiologists again at the forefront of innovation in medicine.
European Radiology Experimental, 2(1), 1-10. doi:10.1186/s41747-018-0061-6.
Retrieved from https://lesa.on.worldcat.org/v2/oclc/7892828773
Reddy, S., Fox, J., & Purohit, M. P. (2019). Artificial intelligence-enabled healthcare delivery.
Journal of the Royal Society of Medicine, 112(1), 22-28.
doi:10.1177/0141076818815510. Retrieved from
https://lesa.on.worldcat.org/v2/oclc/7934822214
Trishan, P., & Szolovits, P. (2018). Artificial intelligence, machine learning and health systems.
Journal of Global Health, 8(2). doi:10.7189/jogh.08.020303. Retrieved from
https://lesa.on.worldcat.org/v2/oclc/7931777652
Tsui, E.-Y. A. (2020). Application of Artificial Intelligence (AI) in Surgery. Retrieved from
https://www.imperial.ac.uk/news/200673/application-artificial-intelligence-ai-surgery/
Yu, K.-H., Beam, A. L., & Kohane, I. S. (2018). Artificial intelligence in healthcare. Nature
Biomedical Engineering, 2(10), 719-731. doi:10.1038/s41551-018-0305-z. Retrieved
from https://lesa.on.worldcat.org/v2/oclc/7869187809

19 | P a g e