Discussion:
Machine learning can significantly improve healthcare; however, the
downsides of machine AI need to be considered. Ethical dilemmas such as
the potential of human biases when creating computer algorithms need to
be addressed [34]. Health care predictions can vary by race,
genetics, and gender amongst other variations, and failure to take these
into account might over or underestimate patient risk factors. As stated
in the review by Ho et al. concerning AI analytics in healthcare, it
will become the responsibility of the clinician to ensure that AI
algorithms are developed and applied appropriately [35]. It is
imperative that healthcare continues to operate by ethically defined
guidelines to sustain trustworthiness, and that medicine continues to
prioritize the good of the patient. AI continues to be promising, as it
can decrease healthcare costs and reduce clinician workload as it
collaborates in decision making.
AI will become more intertwined in clinical practice, and there are
machine learning instruments that no longer require the revision of a
clinician to interpret their results, such is the case in IDX-DR AI
device, which detects mild diabetic retinopathy [36]. Since this
software does not require interpretation by a specialist, more primary
care physicians can use it in their practice, potentially decreasing the
workload for specialists and making the interpretation available for all
types of clinicians. This can potentially be applied to Obstetrics
regarding the interpretation of FHR and CTG interpretations.
There are instances where AI-based computer-aided design (CAD), has led
to decreased diagnostic precision in the interpretation of mammography.
AI alone was shown to be superior to a single radiologist in detecting
breast cancer. However, in practice, an individual radiologist reviews
these images, and due to bias might dismiss CAD suggestions [35].
However, when two specialists review the image, it is most likely to
undergo additional testing if the readers disagree with each other.
Thus, in this study, it was unclear if AI was cost-effective when
compared to interpretation by two radiologists. Additional disadvantages
are that AI-based CAD cannot explain the reasoning behind a decision.
Thus, in case of a misinterpretation by the software, it is difficult to
decide if wherever the manufacturer or the radiologist that interpreted
the data is at fault. Thus, creating agencies that can develop standards
that validate and ensure product quality and accuracy need to be
established. Additionally, the algorithms created need to deliver under
a broad range of settings that can adequately mirror real-world
conditions to which they are being applied clinically [37].
Real-world settings are easier to mimic using a larger amount of data
which can be obtained by accessing patient records, however, patient
confidentiality becomes a challenge when retrieving personal
information. The development of blockchain systems can potentially help
to keep patient information confidential. This would allow the
simultaneous sharing of data between centers, and incorporate it into
the AI software, and allow it to continue expanding its array of records
which would lead to improved accuracy [38].
Professionally, clinicians need to familiarize themselves with AI, to
revise it so that the machine can provide accurate information.
Furthermore, it needs to be flexible in adopting new information, and so
the machine needs to continue learning and changing accordingly.
Furthermore, the data must be representative of the population being
evaluated in a realistic clinical setting. Despite challenges to AI, it
has the overall potential of revolutionizing patient care by providing a
more accurate diagnosis, alleviating the burden of work for clinicians,
decreasing healthcare costs, and providing a baseline analysis in tests
where substantial differences in interpretation between specialists
exist. Further developments in medical AI will continue, and clinicians
must embrace them, yet be wary, and when necessary, recognize its
advantages and drawbacks to continue providing the best patient care.
Conclusions
AI helps to analyze medical data in disease prevention, diagnosis,
patient monitoring, development of new protocols and assisting
clinicians in dealing with voluminous data more accurately and
efficiently. Further studies need to be done to decrease bias when
creating algorithms and to increase adaptability in the system, enabling
the incorporation of new medical knowledge as new technology surfaces.
Practitioners must also take safety measures to ensure that the analysis
is valid and accurate, AI is not meant to replace practitioners, but
rather to serve as an adjunct in decision-making. Ethically, the use of
patient records might bridge patient confidentiality since large amounts
of data are required to enable AI systems to have access to the large
and varied population statistics which are encountered in clinical
settings, hence providing realistic and accurate predictions.
Authorship statement: Pulwasha Maria Iftikhar designed the
study. Pulwasha M. Iftikhar, Marcela Kuijpers, and Aqsa Iftikhar
performed the study, contributed to data extraction, literature review,
analyzed the data, and wrote and proofread the manuscript. Other authors
contributed to data results verification, manuscript proofreading and
amendments. All authors provided critical feedback and helped to shape
the research
Financial disclosure statement: This manuscript is original
research, has not been previously published and has not been submitted
for publication elsewhere while under consideration. Authors declare no
conflict of interest with this manuscript. The authors have no relevant
affiliations or financial involvement with any organization or entity
with a financial interest in or financial conflict with the subject
matter or materials discussed in the manuscript. This includes
employment, consultancies, honoraria, stock ownership or options, expert
testimony, grants, patents received, pending, or royalties.
Disclaimer: None
Conflict of interest: None
References
- It’s
a No Brainer: An Introduction to Neural Networks. (2019). Accessed:
January 26, 2020:https://community.alteryx.com/t5/Data-Science-Blog/It-s-a-No-Brainer-An-Introduction-to-Neural-Networks/ba-p/300479..
- Desai GS:
Artificial
Intelligence: The Future of. Obstetrics and Gynecology. J
ObstetGynaecol India. 2018, 68:326-327.
10.1007/s13224-018-1118-4
- KLAS:
Artificial intelligence success requires partnership and training.
(2019). Accessed: January 26, 2020:https://healthitanalytics.com/news/klas-artificial-intelligence-success-requires-partnership-training.
- Artificial
Intelligence in Healthcare Spending to Hit $36B. (2020). Accessed:
January 26, 2020:https://healthitanalytics.com/news/artificial-intelligence-in-healthcare-spending-to-hit-36b.
- Makary MA, Daniel M:Medical error- the
third leading cause of death in the US. BMJ. 2016, 353:2139.10.1136/bmj.i2139
- Computer
assisted fetal monitoring- Obstetrics (SURFAO-Obst). (2019).
Accessed: January 28, 2020:https://clinicaltrials.gov/ct2/show/NCT03857126.
- Guijarro-Berdiñas B, Alonso-Betanzos A:Empirical
evaluation of a hybrid intelligent monitoring system using different
measures of effectiveness. ArtifIntell Med. 2002, 24:71-96.10.1016/S0933-3657(01)00091-4
- Brocklehurst P:A study of
an intelligent system to support decision making in the management of
labour using the cardiotocograph - the INFANT study protocol. BMC
Pregnancy Childbirth. 2016, 16:10.10.1186/s12884-015-0780-0
- Dawes GS, Moulden M, Redman CWG:System
8000: Computerized antenatal FHR analysis. J Perinat Med. 2009,
19:47-51.10.1515/jpme.1991.19.1-2.47
- Kazantsev A, Ponomareva J, Kazantsev P, Digilov R, Huang P:
Development of
e-health network for in-home pregnancy surveillance based on
artificial intelligence. Proceedings of 2012 IEEE-EMBS International
Conference on Biomedical and Health Informatics. 2012, 82:84.
10.1109/BHI.2012.6211511
- Polak S, Mendyk A:Artificial
intelligence technology as a tool for initial GDM screening. Expert
Syst Appl. 2004, 26:455-460.10.1016/j.eswa.2003.10.005
- Final
Recommendation Statement: Gestational Diabetes Mellitus, Screening.
(2014). Accessed: January 26, 2020:https://www.uspreventiveservicestaskforce.org/Page/Document/RecommendationStatementFinal/gestational-diabetes-mellitu….
- Andrew LB, Isaac SK:Big Data
and Machine Learning in Health Care. JAMA. 2018, 319:1317-1318.doi:10.1001/jama.2017.18391
- Bahado‐Singh RO, Sonek J, McKenna D:Artificial
intelligence and amniotic fluid multiomics: prediction of perinatal
outcome in asymptomatic women with short cervix. Ultrasound Obstet
Gynecol. 2019, 54:110-118.10.1002/uog.20168
- Lee KS, Ahn KH:Artificial
Neural Network Analysis of Spontaneous Preterm Labor and Birth and Its
Major Determinants. J Korean Med Sci. 2019 Apr, 34:128.10.3346/jkms.2019.34.e128
- Idowu IO, Fergus P, Hussain A:
Artificial
Intelligence for Detecting Preterm Uterine Activity in Gynecology and
Obstetric Care. 2015 IEEE International Conference on Computer and
Information Technology; Ubiquitous Computing and Communications;
Dependable, Autonomic and Secure Computing; Pervasive Intelligence and
Computing. 2015, 215:220.
10.1109/CIT/IUCC/DASC/PICOM.2015.31
- Mason CW, Swaan PW, Weiner CP:Identification
of interactive gene networks: A novel approach in gene array profiling
of myometrial events during guinea pig pregnancy. Am J Obstet
Gynecol. 2006, 194:1513-1523.10.1016/j.ajog.2005.12.044
- Norwitz ER:Artificial
intelligence: Can computers help solve the puzzle of parturition?.
Am J Obstet Gynecol. 2006, 194:1510-1512.10.1016/j.ajog.2005.12.054
- MetaCore
and Genomic Analysis Tools. (2019). Accessed: January 26, 2020:https://www.nihlibrary.nih.gov/resources/tools/metacore-and-genomic-analysis-tools.
- Guh RS, Wu CJ, Weng SP:Integrating
genetic algorithm and decision tree learning for assistance in
predicting in vitro fertilization outcomes. Expert Syst Appl. 2011,
38:4437-4449.10.1016/j.eswa.2010.09.112
- Siristatidis C, Pouliakis A, Chrelias C, Kassanos D:
Artificial
Intelligence in IVF: A Need. Syst Biol Reprod Med. 2011, 57:179-185.
10.3109/19396368.2011.558607
- Manna C, Nanni L, Lumini A, Pappalardo S:Artificial
intelligence techniques for embryo and oocyte classification. Reprod
Biomed. 2013, 26:42-49.10.1016/j.rbmo.2012.09.015
- Enshaei A, Robson CN, Edmondson RJ:Artificial
Intelligence Systems as Prognostic and Predictive Tools in Ovarian
Cancer. Ann Surg Oncol. 2015, 22:3970-3975.10.1245/s10434-015-4475-6
- AI behind
potential blood test for ovarian cancer. (2017). Accessed: 26th Jan,
2020:https://www.mdlinx.com/oncology/article/1230..
- Kann BH, Thompson R, Thomas CJ, Dicker A,:Artificial
intelligence in oncology: Current application and future directions.
Cancer Network. 2019, 33:46-53.
- Using
artificial intelligence to detect cervical cancer. (2019). Accessed:
January 26, 2020:https://directorsblog.nih.gov/2019/01/17/using-artificial-intelligence-to-detect-cervical-cancer/..
- Moawad G, Tyan P, Louie M:Artificial
intelligence and augmented reality in gynecology. Curr Opin Obstet
Gynecol. 2019, 31:345.10.1097/GCO.0000000000000559
- Waran V, Narayanan V, Karuppiah R, Owen SLF:Utility of
multimaterial 3D printers in creating models with pathological
entities to enhance the training experience of neurosurgeons. J
Neurosurg. 2014, 120:489-492.10.3171/2013.11.JNS131066
- Ajao MO, Clark NV, Kelil T, Cohen SL, Einarsson JI:Case
Report: Three-Dimensional Printed Model for Deep Infiltrating
Endometriosis. J Minim Invasive Gynecol. 2017, 24:1239-1242.10.1016/j.jmig.2017.06.006
- Marescaux J, Rubino F, Arenas M, Mutter D, Soler L:
Augmented-Reality-Assisted
Laparoscopic Adrenalectomy. JAMA. 2004, 292:2211-2215.
10.1001/jama.292.18.2214-c
- Vávra P, Roman J, Zonča P:Recent
Development of Augmented Reality in Surgery: A Review. J Healthc
Eng. 2017 , 2017:4574172.10.1155/2017/4574172
- Dirie NI, Wang Q, Wang S:Two-Dimensional
Versus Three-Dimensional Laparoscopic Systems in Urology: A Systematic
Review and Meta-Analysis. J Endourol. 2018, 32:781-790.10.1089/end.2018.0411
- Song E, Yu F, Liu H:A Novel
Endoscope System for Position Detection and Depth Estimation of the
Ureter. J Med Syst. 2016, 40:266.10.1007/s10916-016-0607-1
- Char DS, Shah NH, Magnus D:Implementing
Machine Learning in Health Care — Addressing Ethical Challenges. N
Engl J Med. 2018, 378:981-983.10.1056/NEJMp1714229
- Ho CWL, Soon D, Caals K, Kapur J:Governance
of automated image analysis and artificial intelligence analytics in
healthcare. Clin Radiol. 2019, 74:329-337.10.1016/j.crad.2019.02.005
- Error! Hyperlink reference not valid. . (2019). Accessed:
January 26, 2020: Error! Hyperlink reference not valid. .
- US
FDA Approaches to Artificial Intelligence. (2018). Accessed: January
26, 2020:https://www.natlawreview.com/article/us-fda-approaches-to-artificial-intelligence..
- Maxmen A:AI
researchers embrace Bitcoin technology to share medical data.
Nature. 2018, 555:293-294.10.1038/d41586-018-02641-7