Urologic Oncology: Seminars and Original Investigations 42 (2024) 48−56

Review Article
Exploring the opportunities and challenges of implementing artificial
intelligence in healthcare: A systematic literature review
Betelhem Zewdu Wubineha,d,*, Fitsum Gizachew Deribab,d, Michael Melese Woldeyohannisc
a
Faculty of Information and Communication Technology, Wroclaw University of Science and Technology, Wroclaw, Poland
b
School of Computing, University of Eastern Finland, Joensuu, Finland
c
School of Information Science, Addis Ababa University, Addis Ababa, Ethiopia
d
School of Computing and Informatics, Wachemo University, Hosaena, Ethiopia
Received 11 November 2023; accepted 25 November 2023

Abstract
Recent progress in the realm of artificial intelligence has shown effectiveness in various industries, particularly within the healthcare sec-
tor. However, there are limited insights on existing studies regarding ethical, social, privacy, and technological aspects of AI in the health
sector, which is the gap our study aims to fill. This study aimed to synthesize empirical studies on the challenges and opportunities of using
AI by conducting a systematic review. We reviewed 33 articles published between 2015 and 2022 in the PubMed, IEEE Xplore, and Science
Direct databases. The results show that artificial intelligence has the promise of improving health care and faces obstacles when imple-
mented. Most of the reviewed studies indicated that the use of AI provides several opportunities, including teamwork and decision-making,
technological advancement, diagnosis and patient monitoring, drug development, and virtual health assistance. However, the findings show
that the use of AI in the health sector hinders multifaceted challenges, including ethical and privacy-related issues, lack of awareness, unreli-
ability of technology, and professional liability. The findings highlight that artificial intelligence has the potential to transform healthcare
and that addressing these challenges is crucial to fully utilize its potential. Ó 2023 The Author(s). Published by Elsevier Inc. This is an open
access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)

Keywords: Artificial intelligence; Healthcare; Machine learning

1. Introduction Medical errors, which are preventable adverse events result-

ing from medical care, pose a significant public health concern
Healthcare is a cyclical process in which healthcare profes- and are one of the leading causes of death [3]. Health practi-
sionals gather, analyze, and integrate data from various sources, tioner can make various form of errors, such as inappropriate
then use clinical knowledge and empirical research to confirm blood transfusions, misdiagnosis, under and overtreatment, sur-
these findings and arrive at conclusions that are implemented gical injuries, self-inflicted harm, and fatalities [4]. According
in practice [1]. Currently, healthcare professionals are facing a to the patient safety experts at Johns Hopkins University, medi-
variety of occupational pressures from different angles [2], cal errors are associated with more than 250,000 deaths in the
including heavy workloads, administrative tasks, inadequate United States every year, based on data on medical death rates
social support from co-workers, uncertainty about patient care, in an 8-year period [5]. Moreover, the demand for healthcare
and a tendency towards emotional reactions, making it chal- far outweighs the supply of physicians, resulting in a shortage
lenging to adapt to changes. To enhance quality and safety, of physicians to meet the growing need [6].
improve productivity, and promote staff satisfaction, health Furthermore, the healthcare industry is experiencing
systems need to review their policies and strategies. rapid technological transformation, leading to the genera-
tion of vast amounts of data in various health databases, but
*Corresponding author. little effort is being made to harness this information to pre-
E-mail address: [email protected] (B.Z. Wubineh). vent diseases [7]. To optimize patient care and utilization of

https://doi.org/10.1016/j.urolonc.2023.11.019
1078-1439/Ó 2023 The Author(s). Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/)
 B.Z. Wubineh et al. / Urologic Oncology: Seminars and Original Investigations 42 (2024) 48−56 49

medical resources, it is imperative to move from manual fundamental steps outlined by Kitchenham [17]. The
healthcare systems to automated ones. The modern technolo- authors adhered to these guidelines to identify the purpose
gies such as AI for data analysis is needed to address this chal- of the review, draft the protocol, define the inclusion and
lenge. Do this, AI, particularly, machine learning and its exclusion criteria, and perform the search process to extract
applications in healthcare have garnered significant attention, data and evaluate the quality of the articles before synthe-
identifying and forecasting high-risk populations for adverse sizing the review.
health outcomes and creating public health interventions
being used in the public health sector [8]. The applications of 2.1. Define the search strategy
AI in healthcare are numerous and diverse, including improv-
ing patient experience and results, providing value, improving The scope of the database search, the search strategy,
quality, safety, supporting evidence-based decision making, and the search string were used to identify the first set of
and optimizing health system performance [9]. By using relevant primary studies. The scope of the search considers
machine learning to explore computer algorithms that both the publication year and the location. The scope of our
enhance task performance based on past experiences [10], AI search was restricted to studies published between 2015
can minimize diagnostic and treatment errors that occur in and December 2022. Targeted research publications were
human clinical practice. For example, it is possible to predict found using the following well-known databases: IEEE
unfavorable hemodynamic events and other crises in Xplore, ScienceDirect, and PubMed. We searched the liter-
patients during surgery, which is a reality in many criti- ature using both automatic and manual methods. The auto-
cal care units [11]. This prediction was used to build matic search was conducted by searching in the selected
intraoperative clinical decision support, allowing the DB utilizing a pre-established search string. The search
practice of active systems in healthcare. In addition, string was constructed from a research question using key-
machine learning is used to find drug-medication inter- words. Then a pilot search was used to enhance the search
actions during pre- and postmarketing surveillance due string. The search string is as follows: (“Artificial
to the accessibility of multidimensional drug characteris- intelligence” OR “machine learning”) AND (“challenge*”
tics [12]. OR “opportunity*”) AND (“healthcare”). The snowball
Various researchers have explored the principles, diffi- method was used to track studies that are overlooked. Fig. 1
culties, and practical applications of machine learning in describes the study review protocol.
long-term clinical care, highlighting the challenges and
acceptance of AI in digital technology [13,14]. Addition- 2.2. Study selection criteria
ally, some researchers have analyzed the use of telemedi-
cine and AI to support the self-isolation of COVID-19 The author discussed the criteria for choosing studies in
patients, mapping the challenges of AI-based telemedicine detail as an approach to extract relevant studies while
during COVID-19 outbreaks [15,16]. Most of the studies retaining interesting, related studies. The articles identified
focused on specifically exploring the principles, difficulties, were filtered using a list of criteria (excluding criteria) to
or practical applications of AI. However, despite these obtain documents that were not pertinent to the study and
efforts, there is limited information on a comprehensive are presented in Table 1. Automatic and manual searches
study on ethical, social, healthcare provider, privacy, and yielded a total of 1,249 articles, 20 of which were duplicates
technological challenges and opportunities related to that were resolved before the exclusion criteria were
patient monitoring, technological advances in healthcare, applied for 1,177 articles. Then, 52 papers were found after
medical imaging, diagnosis, treatment, and drug develop- applying the criterion, which was done by reading the
ment. To address this gap, this study answers the following paper’s title and abstract. Ten of the papers are not full-
research questions: length and 42 have been retained for further review. Finally,
9 papers were of low quality and 33 were included in our
RQ1. What are the perceived challenges while implement- study.
ing AI technology in the healthcare sector?
RQ2. What opportunities does AI technologies do have in 2.3. Quality assessment criteria
the healthcare sector?
Following the application of the exclusion criteria, the
The purpose of our study is to explore the challenges and 42 remaining articles were further evaluated using a clear
opportunities of ML and AI in healthcare by synthesizing quality checklist shown in Table 2. Kitchman [17] criteria
empirical studies. were the basis for the quality assessment instruments
applied in the study. Based on factors that affect the results,
2. Methodology this criterion was divided into 4 main categories: reporting,
relevance, rigor, and credibility. First, based on the purpose,
The authors performed this review according to the pre- clarity, and coherence of the studies, the quality of the
ferred reporting items for systematic reviews and the eight reporting was evaluated. The degree to which studies add
50 B.Z. Wubineh et al. / Urologic Oncology: Seminars and Original Investigations 42 (2024) 48−56

Fig. 1. Review protocol adopted from [17].

value to research and practice was then used to examine the The final factor used to determine credibility is how sig-
question of rigor. The relevance of the studies was also nificant and logical the findings and conclusions of the stud-
evaluated based on how completely and thoroughly all the ies are. Regarding how well the paper responds to the issues
questions the publication claimed to address were posed, the answers to the quality checklist questions were
addressed.

Table 2
Table 1 Quality assessment criteria.
Study exclusion criteria.
# Question
# Criterion
Q1 Are the aims of the study clearly stated?
EC1 Duplicate paper from multiple sources Q2 Are the scope and content clearly defined?
EC2 Paper without full text available Q3 Are the research method and design well presented?
EC3 Paper not written in English Q4 Are the study variables likely to be valid and reliable?
EC4 The article is not published in a journal or conference. Q5 Are all the study questions answered?
EC5 Secondary studies such as case study, Review, Survey Q6 Are the primary studies discussing and stating the finding clearly?
EC6 The article is published before 2015 Q7 Is the research document process adequately documented?
EC7 The article does not describe artificial intelligence/machine Q8 Are the conclusions related to the purpose of the study?
learning and its challenge or opportunity for healthcare care. Q9 Are the limitations of the research clearly stated?
 B.Z. Wubineh et al. / Urologic Oncology: Seminars and Original Investigations 42 (2024) 48−56 51

given on a numerical scale numbered 0 for "no" 0.5 for 1 paper was published in 2016 from a source other than
"somewhat" and 1 for "yes," and the evaluation studies are those included in this study, followed by 2 articles in 2017
presented in the Appendix. and 1 article in 2018 from other sources. In 2019, 4 articles
were published (2 from other sources, 1 from PubMed and
2.4. Data extraction 1 from ScienceDirect). The year 2020 saw the highest num-
ber of publications, with 17 papers (11 from other sources,
To conduct our investigation, we carefully reviewed 5 from PubMed, and 1 from IEEE). In 2021, 6 papers were
includes 33 primary studies to gather vital information published (4 from other sources and 2 from PubMed), while
based on the codification of extraction form. To accurately in 2022, 2 papers were published (from ScienceDirect).
extract data from each study, we created a data element that
included essential information such as the study authors 3.2. Challenges of AI in healthcare
and year of publication, aim of the study. During a pilot
data extraction, we identified and extracted all the fields To explore the challenges encountered in healthcare, we
required to answer our research questions. These fields pose RQ: What are the perceived challenges while imple-
encompassed a broad range of topics related to the health- menting AI technology in the healthcare sector? We identi-
care industry, including artificial intelligence and machine fied 4 main issues. These challenges include privacy
learning techniques, challenges encountered while using concerns, ethical considerations and data issues, social and
AI, and opportunities. patient awareness, and healthcare and technology organiza-
tions as depicted in Table 3.
2.5. Data synthesis Ethical and privacy-related challenge: The ethical and
privacy challenges are a significant issue in the healthcare
Our aim in data synthesis is to effectively organize and industry, which requires strict compliance with laws, regu-
present the results of the primary studies to address our lations, and rules [18]. Integration of AI in medicine
research questions. Since this study is qualitative in nature, presents a wide range of challenges in terms of privacy and
we performed a synthesis of the data collected based on the ethics, including issues with patient data protection, ethical
purpose of the research and the results of the selected pri- boundaries of innovation, and the actual impact of technol-
mary studies. We analyze both individual studies and the ogy on both doctors and patients [19]. To realize the full
collection of studies to identify patterns or similarities. To potential of AI in healthcare, strategic decisions must be
accomplish this, we classified studies with similar meanings made that balance competing interests and values, which
under a single concept. The quality scores for each study include issues of privacy, accountability, intellectual prop-
were meticulously evaluated. erty rights, and transparency [20]. Understanding the
broader perspective and changes regarding privacy con-
3. Result and discussion cerns in healthcare is crucial, as trust is essential. Ethical
3.1. Statistics of study challenges arise when using AI in healthcare [21].
Lack of awareness on AI technologies: Building trust
The year-wise distribution of the primary studies with patients is crucial in healthcare, and this is especially
selected from various journals is shown in Fig. 2. true when it comes to AI-based technologies. Doctors and
Fig. 2 illustrates the distribution of articles on ML and policy makers acknowledge that the public may be reluctant
AI in healthcare from 2016 to 2022. The graph shows that to rely on AI-generated recommendations, given the long-
standing importance of face-to-face interaction with physi-
cians [22]. To complicate matters, there is often a lack of
understanding among the general population about the
potential benefits and limitations of AI in healthcare, lead-
ing to unrealistic expectations that can make it difficult for
physicians to adopt these technologies [23].

Table 3
Challenges of AI in healthcare.

Identified theme of challenges No of studies Studies

Ethical and privacy related challenges 4 (26.7%) [18−21]

Lack of awareness on AI technologies 2 (13.3%) [22,23]
Unreliability and trustworthiness of AI 5 (33.3%) [22,24−27]
technologies
Healthcare providers and professional 4 (26.7%) [1,25,28,29]
liability
Fig. 2. Distribution of the selected studies for each year.
52 B.Z. Wubineh et al. / Urologic Oncology: Seminars and Original Investigations 42 (2024) 48−56

Fig. 3. Description of AI challenges in healthcare.

Unreliability and trustworthiness of AI technologies: understanding of how decisions are made, leading to mis-
Technological challenges pose obstacles in the transition trust in technology. Additionally, the scarcity of high-qual-
from conventional computer architectures to AI architec- ity data for AI model training and evaluation can lead to
tures, as the emergence of graphics processors, field-pro- inaccurate predictions, diagnoses, and biases in algorithms.
grammable gate arrays, and special-purpose AI chips The description of challenges in AI is shown in Fig. 3.
require advanced computing and storage devices to offer AI Healthcare providers and professional liability: Clini-
applications [24]. However, deploying these infrastructures cal judgment has always been the domain of trained and
and storages can be complex and expensive for healthcare certified healthcare professionals. However, the increased
organizations [25]. Furthermore, some healthcare professio- use of AI decision support systems to assist with clinical
nals, including physicians, nurses, and assistant nurses, lack tasks may impact the professional responsibility of health-
adequate focus on digital technology, which can limit the care providers towards their patients [28]. One concern is
adoption of AI in healthcare [26]. Another critical issue is that healthcare professionals may become too dependent on
the lack of transparency and interpretability of AI algo- AI technology and be less willing to question errors or dou-
rithms, particularly when processing unstructured data [22]. ble-check results [1]. While AI systems are important to be
These technology-related concerns are prevalent among user-friendly, their integration into routine clinical practice
hospital administrators and physicians and pose a challenge can present difficulties depending on their intended use
to AI systems. Furthermore, deploying an AI model requires [25]. Physicians can face challenges when learning to inte-
interoperability standards to enable seamless integration with grate and use technology, as they may have varying levels
data from various electronic health records [27]. In health- of technology literacy and limited hands-on experience
care, these challenges must be addressed to ensure the using AI applications in practice [29].
smooth integration of AI technology and promote its wide-
spread use. Besides this, in AI system, not only use algo- 3.3. Opportunities to use AI in the healthcare
rithms but also after using there is a model deployment
phase. There may be a lack of interoperability standards, in RQ2. What opportunities do AI technologies have in the
order to when a model functions, it must be able to connect healthcare sector? The study identifies 5 fundamental
to the data in various electric health records [27]. opportunities of using AI in health care depicted in Table 4.
In general, the opaqueness and complexity of AI algo- Teamwork and decision-making: Current healthcare
rithms pose significant challenges. This can hinder environments require healthcare providers to collaborate
 B.Z. Wubineh et al. / Urologic Oncology: Seminars and Original Investigations 42 (2024) 48−56 53

Table 4 setting using a virtual reality simulator, as mentioned in

Opportunities of AI in healthcare. [34], which allows them to practice treatment beforehand
Identified opportunities No of studies Studies
and consider all eventualities to create precise and secure
surgical procedures. Additionally, technology is changing
Teamwork and decision making 2 (10%) [30,31] the way citizens perceive healthcare, as mentioned in [35].
Technological advancements 5 (25%) [32−36]
AI is also being used to control workflows, image analysis,
Diagnosis and patient monitoring 4 (20%) [37−40]
Drug development 3 (15%) [41−43] robotic surgery, virtual assistants, and clinical decision sup-
Virtual health assistants 6 (30%) [44−49] port, as discussed in [36]. The description of opportunities
in AI is described in Fig. 4.
Diagnosis and patient monitoring: AI is used for per-
and work in teams, which requires strong communication, sonalized medicine and predictive analytics to identify dis-
shared decision-making, coordinated actions, and progress eases early and monitor their progression, resulting in
evaluation, as highlighted in [30]. AI chatbots can be used better patient outcomes and reduced healthcare costs
to handle problems related to coordinating and scheduling [37,38]. In addition, AI can aid in the development of new
medical appointments, providing reminders, and notifying treatments, advancing healthcare research and innovation.
providers of a patient’s condition based on symptoms, as AI can significantly improve the efficiency and effective-
discussed in [31]. ness of patient care, with several specific opportunities for
Technological advancements: AI-enabled health moni- implementation. For example, predictive analytics can be
toring technology can help the elderly by ensuring timely used to identify patients who are at risk of developing com-
delivery of care, allowing healthcare providers to expand plications and predict the likelihood of deterioration in the
patient care efforts outside office hours and promote self- patient [39]. By intervening early, healthcare providers can
management, as discussed in [32]. For example, sensor improve patient outcomes and prevent critical conditions
technology can simplify self-monitoring for patients with from developing [40].
heart failure using user-friendly hardware, as highlighted in Drug development: Artificial intelligence has the poten-
[33]. Various health-related technologies also have the tial to greatly enhance the efficiency and effectiveness of
potential to handle laboratory practices for medical profes- the drug development process. One specific opportunity is
sionals and students. For example, novice surgeons can virtual screening, which can analyze large amounts of data
practice their techniques in a repeatable and risk-free on drug interactions and identify new drug targets [41].

Fig. 4. The description of AI opportunities in healthcare.

54 B.Z. Wubineh et al. / Urologic Oncology: Seminars and Original Investigations 42 (2024) 48−56

This can greatly accelerate drug discovery, allowing However, accessing and utilizing patient data is a challenge
researchers to identify potential drug candidates more due to privacy concerns, and the implementation of AI in
quickly and at a lower cost. In addition, AI can analyze data healthcare is also hampered by a lack of awareness of the
on drug candidates to identify the most promising com- patient, technological limitations, and professional liability
pounds for further development [42] and explore the repur- issues. Ethical issues, such as data privacy and bias, are
posing of existing drugs for new potential uses [43]. among the most significant challenges we found when using
Virtual health assistants (VHA): The use of virtual AI in healthcare. Ensuring the protection of patient data and
reality (VHA) with AI has the potential to revolutionize avoiding the perpetuation of social biases is crucial for the
healthcare delivery. Studies, such as the one by Salway et responsible and equitable implementation of AI in health-
al. [44], have identified that the availability of health care care. In addition, there is a lack of awareness and under-
24 hours a day can greatly increase accessibility to health- standing of AI among patients and the public, which can
care. AI can also help identify patient symptoms and pro- lead to resistance and mistrust. On the positive side, AI
vide them with appropriate care recommendations, such as presents several opportunities in healthcare, including
self-care advice or a referral to a healthcare professional improved teamwork and decision-making, advances in
[45]. In addition, AI can help patients manage their medica- technology, medical imaging and diagnosis, patient moni-
tions by providing reminders and alerts about when to take toring, and virtual health assistance. AI supports healthcare
their medications and possible side effects [46,47]. Finally, professionals by analyzing large amounts of medical data
remote patient monitoring using artificial intelligence can and assisting in informed decision making, leading to better
improve healthcare access, especially in rural or under- patient outcomes and resource efficiency. AI has the poten-
served areas [48,49]. tial to revolutionize healthcare by enabling more accurate
diagnoses, personalized treatment plans, and efficient utili-
4. Limitation and Future Direction zation of resources. However, it is crucial to have realistic
expectations and address ethical considerations, such as
Our study has several limitations. First, this study lim- transparency, fairness, and accountability, in AI develop-
ited to three databases and five keywords, which may miss- ment, deployment, and use. Future research should also
ing relevant studies. Further studies are required by focus on developing standards for the evaluation of AI algo-
exploring additional databases and incorporating more rithms and investigating the ethical and legal implications
search keywords. For example, incorporating search terms of AI in healthcare.
on challenges of AI may be retrieved by searching with
terms such as obstacles, barriers, difficulties, and similar Declaration of Competing Interest
ones.
Several issues need further exploration, arising from our The authors declare that they have no known competing
study for future work. For instance, what strategies can be financial interests or personal relationships that could have
employed to manage and alleviate concerns about profes- appeared to influence the work reported in this paper.
sional liability when AI is involved in healthcare decision
support and treatment planning? How might the lack of Funding
awareness among healthcare professionals and the general
public regarding AI impact the successful integration of This work did not receive external funding.
this technology into healthcare systems? In addition, it is
necessary to evaluate the benefits and potential drawbacks Supplementary materials
of incorporating virtual health assistance powered by AI in
patient care, considering aspects such as accessibility, Supplementary material associated with this article can
patient engagement, and trust. be found in the online version at https://doi.org/10.1016/j.
urolonc.2023.11.019.
5. Conclusions
References
The use of AI in healthcare care has become increasingly
significant in recent years, offering numerous advantages.
[1] Hazarika I. Artificial intelligence: opportunities and implications for
This study highlighted various challenges and opportunities the health workforce. Int Health 2020;12(4):241–5. https://doi.org/
in the use of AI in the healthcare industry. We found that 10.1093/INTHEALTH/IHAA007.
most studies indicate the potential of AI to assist healthcare [2] Grace MK, VanHeuvelen JS. Occupational variation in burnout
professionals in their duties, automate repetitive tasks, and among medical staff: evidence for the stress of higher status. Soc Sci
improve medication precision, as well as enhance decision Med 2019;232(May):199–208. https://doi.org/10.1016/j.socsci-
med.2019.05.007.
making and teamwork. In addition, the study revealed that [3] Rodziewicz TL, Hipskind JE. Medical error prevention. StatPearls
AI has proven beneficial in medical imaging, diagnosis and [Internet]. Treasure Island (FL): StatPearls Publishing; 2020:1–37.
treatment, virtual health assistance, and drug development. http://www.ncbi.nlm.nih.gov/pubmed/29763131.
 B.Z. Wubineh et al. / Urologic Oncology: Seminars and Original Investigations 42 (2024) 48−56 55

[4] Carver N, Gupta V, Hipskind JE. Medical error. StatPearls [Internet]. [24] Iliashenko O, Bikkulova Z, Dubgorn A. Opportunities and challenges
Treasure Island (FL): StatPearls Publishing; 2022. of artificial intelligence in healthcare. E3S Web Conf 2019:02028.
[5] Makary MA, Daniel M. Medical error—the third leading cause of https://doi.org/10.1051/e3sconf/201911002028.
death in the US. BMJ 2016;353:i2139. [25] Singh RP, Hom GL, Abramoff MD, Campbell JP, Chiang MF. Cur-
[6] Lorkowski J, Jugowicz A. Shortage of physicians: a critical review. rent challenges and barriers to real-world artificial intelligence adop-
Medical Research and Innovation 2021:57–62. tion for the healthcare system, provider, and the patient. Transl Vis
[7] Olaronke I, Oluwaseun O. Big data in healthcare: prospects, chal- Sci Technol 2020;9(2):1–6. https://doi.org/10.1167/tvst.9.2.45.
lenges and resolutions. In: FTC 2016 - Proc. Futur. Technol. Conf.; [26] Hee Lee D, Yoon SN. Application of artificial intelligence-based
2017. p. 1152–7. https://doi.org/10.1109/FTC.2016.7821747. technologies in the healthcare industry: opportunities and challenges.
[8] Alanazi A. Using machine learning for healthcare challenges and Int J Environ Res Public Health 2021;18(1):1–18. https://doi.org/
opportunities. Informatics Med Unlocked 2022;30(February):1–5. 10.3390/ijerph18010271.
https://doi.org/10.1016/j.imu.2022.100924. [27] Hofer IS, Burns M, Kendale S, Wanderer JP. Realistically integrating
[9] Chen M, Decary M. Artificial intelligence in healthcare: an essential machine learning into clinical practice: a road map of opportunities,
guide for health leaders. Healthc Manag Forum 2020;33(1):10–8. challenges, and a potential future. Anesth Analg 2020;130(5):1115–8.
https://doi.org/10.1177/0840470419873123. https://doi.org/10.1213/ANE.0000000000004575.
[10] Kassa SM, Wubineh BZ. Use of machine learning to predict califor- [28] Carter SM, Rogers W, Win KT, Frazer H, Richards B, Houssami N.
nia bearing ratio of soils. Advances in Civil Engineering 2023;25. The ethical, legal and social implications of using artificial intelli-
[11] Mahajan A, Esper SA, Cole DJ, Fleisher LA. Anesthesiologists’ role gence systems in breast cancer care. Breast 2020;49:25–32. https://
in value-based perioperative care and healthcare transformation. doi.org/10.1016/j.breast.2019.10.001.
Anesthesiology 2021;134(4):526–40. https://doi.org/10.1097/ [29] Liyanage H, Liaw ST, Jonnagaddala J, et al. Artificial intelligence in
ALN.0000000000003717. primary health care: perceptions, issues, and challenges. Yearb Med
[12] Liu N, Chen CB, Kumara S. Semi-supervised learning algorithm for Inform 2019;28(1):041–6. https://doi.org/10.1055/s-0039-1677901.
identifying high-priority drug-drug interactions through adverse [30] Morley L, Cashell A. Collaboration in health care. J Med Imaging
event reports. IEEE J Biomed Health Informatics 2020;24(1):57–68. Radiat Sci 2017;48(2):207–16.
https://doi.org/10.1109/JBHI.2019.2932740. [31] Jadczyk T, Wojakowski W, Tendera M, Henry TD, Egnaczyk G,
[13] Ashton JJ, Young A, Johnson MJ, Beattie RM. Using machine learn- Shreenivas S. Artificial intelligence can improve patient management
ing to impact on long-term clinical care: principles, challenges, and at the time of a pandemic: The role of voice technology. J Med Inter-
practicalities. Pediatr Res 2023;93(2):324–33. https://doi.org/ net Res 2021;23(5):e22959.
10.1038/s41390-022-02194-6. [32] Ho A. Are we ready for artificial intelligence health monitoring in
[14] Giansanti D, Di Basilio F. The artificial intelligence in digital radiology: elder care? BMC Geriatr 2020;20:1–7.
part 1: the challenges, acceptance and consensus. Healthcare (Basel) [33] Bachtiger P, Plymen CM, Pabari PA, et al. Artificial intelligence, data
2022;10(3):1–13. https://doi.org/10.3390/healthcare10030509. sensors and interconnectivity: future opportunities for heart failure.
[15] Deriba FG, Salau AO, Tefera BT, Demilie WB. Artificial intelligence Card Fail Rev 2020;12(6):e11.
assisted decision making in predicting COVID-19 patient’s path. J [34] Licci M, Thieringer FM, Guzman R, Soleman J. Development and
Pharm Negat Results 2023:1250–5. validation of a synthetic 3D-printed simulator for training in neuroen-
[16] Huang JA, Hartanti IR, Colin MN, Pitaloka DAE. Telemedicine and doscopic ventricular lesion removal. Neurosurg Focus 2020;48(3):
artificial intelligence to support self-isolation of COVID-19 patients: E18.
recent updates and challenges. Digit Health 2022;15:8. https://doi. [35] Bohr A, Memarzadeh K. The rise of artificial intelligence in health-
org/10.1177/20552076221100634. care applications. Artif Intell Healthc 2020:25–60. https://doi.org/
[17] Budgen D, Kitchenham BA, Charters SM, Turner M, Brereton P, 10.15420/cfr.2019.14.
Linkman SG. Presenting software engineering results using struc- [36] Lu S, Burton SL. Man vs robots? Future challenges and opportunities
tured abstracts: a randomised experiment. Empir Softw Eng 2008;13 within artificial intelligence (AI) health care education model. Proc
(4):435–68. https://doi.org/10.1007/s10664-008-9075-7. RAIS Conference I 2017;6(November):49–58.
[18] Vayena E, Blasimme A, Cohen IG. Machine learning in medicine: [37] Matheny ME, Whicher D, Israni ST. Artificial intelligence in health
addressing ethical challenges. PLoS Med 2018;15(11):4–7. https:// care: a report from the National Academy of Medicine. JAMA
doi.org/10.1371/journal.pmed.1002689. 2020;323(6):509–10.
[19] Bartoletti I. Ethical and privacy challenges. In Artificial Intelligence [38] Jeddi Z, Bohr A. Remote patient monitoring using artificial intelli-
in Medicine: 17th Conference on Artificial Intelligence in Medicine, gence. In Artificial Intelligence in Healthcare. Academic Press;
AIME 2019, Poznan, Poland, June 26-29, Proceedings 17 (pp. 7-10). 2020:203–34. https://doi.org/10.1016/B978-0-12-818438-7.00009-5.
Springer International Publishing; 2019. https://doi.org/10.1007/978- [39] Assaf D, et al. Utilization of machine-learning models to accurately
3-030-21642-9_2. predict the risk for critical COVID-19. Intern Emerg Med
[20] Cohen IG, Evgeniou T, Gerke S, Minssen T. The European artificial 2020;15:1435–43.
intelligence strategy: Implications and challenges for digital health. [40] Gajarawala SN, Pelkowski JN. Telehealth benefits and barriers. J
Lancet Digit Health 2020;2(7):e376–9. https://doi.org/10.1016/ Nurse Pract 2021;17(2):218–21. https://doi.org/10.1016/j.nur-
S2589-7500(20)30112-6. pra.2020.09.013.
[21] Gerke S, Minssen T, Cohen G. Ethical and legal challenges of artifi- [41] Gupta R, Srivastava D, Sahu M, Tiwari S, Ambasta RK, Kumar P.
cial intelligence-driven healthcare. Artif Intell Healthc 2020:295– Artificial intelligence to deep learning: machine intelligence
336. https://doi.org/10.1016/B978-0-12-818438-7.00012-5. approach for drug discovery. Mol Diversity 2021;25:1315–60.
[22] Sun TQ, Medaglia R. Mapping the challenges of artificial intelli- https://doi.org/10.1007/s11030-021-10217-3.
gence in the public sector: evidence from public healthcare. Gov [42] Bakowski MA, Beutler N, Wolff KC, et al. Drug repurposing screens
Inf Q 2019;36(2):368–83. https://doi.org/10.1016/j.giq.2018.09 identify chemical entities for the development of COVID-19 inter-
.008. ventions. Nat Commun 2021;12(1):3309.
[23] Vollmer S, Mateen BA, Bohner G, et al. Machine learning and artificial [43] Mohapatra S, Nath P, Chatterjee M, et al. Repurposing therapeutics
intelligence research for patient benefit: 20 critical questions on transpar- for COVID-19: rapid prediction of commercially available drugs
ency, replicability, ethics, and effectiveness. BMJ 2020;368:1–12. https:// through machine learning and docking. PLoS One 2020;15(11
doi.org/10.1136/bmj.l6927. November):1–13. https://doi.org/10.1371/journal.pone.0241543.
56 B.Z. Wubineh et al. / Urologic Oncology: Seminars and Original Investigations 42 (2024) 48−56

[44] Salway S, Mumtaz Z, Bhatti A, Barnes A, Dawson J, Jhangri GS. [47] Gudala M, Ross MET, Mogalla S, Lyons M, Ramaswamy P, Roberts
Scaling up the ‘24/7 BHU’ strategy to provide round-the-clock mater- K. Benefits of, barriers to, and needs for an artificial intelligence−pow-
nity care in Punjab, Pakistan: a theory-driven, coproduced implemen- ered medication information voice chatbot for older adults: interview
tation study. Health Res Policy Syst 2022;20(1):139. study with geriatrics experts. JMIR Aging 2022;5(2):e32169.
[45] Hill MG, Sim M, Mills B. The quality of diagnosis and triage advice [48] Marcin JP, Shaikh U, Steinhorn RH. Addressing health disparities in
provided by free online symptom checkers and apps in Australia. rural communities using telehealth. Pediatr Res 2016;79(1):169–76.
Med J Aust 2020;212(11):514–9. [49] Davalagi SB, Palicheralu BS, Murthy SSN, Hurlihal S. Acceptance of
[46] Zhao M, Hoti K, Wang H, Raghu A, Katabi D. Assessment of medi- artificial intelligence (AI)-based screening for breast health in urban
cation self-administration using artificial intelligence. Nat Med slums of central Karnataka, India−SWOC analysis. J Fam Med Prim
2021;27(4):727–35. https://doi.org/10.1038/s41591-021-01273-1. Care 2022;11(10):6023–8.