Handbook of Artificial Intelligence in Healthcare Vol 2
Practicalities and Prospects
Visit the link below to download the full version of this book:
https://medipdf.com/product/handbook-of-artificial-intelligence-in-healthcare-vo
l-2-practicalities-and-prospects/
Click Download Now
Preface
This volume is a sequel of Handbook of Artificial Intelligence in Healthcare. The first
volume focuses on advances and applications of artificial intelligence (AI) method-
ologies in several specific areas, i.e. signal, image and video processing as well as
information and data analytics. In this volume, several general practicality challenges
and future prospects of AI methodologies pertaining to the healthcare and related
domains are presented in Part I and Part II, respectively. A total of 17 chapters are
included in this volume. A description of each contribution is as follows.
Decision-making and control in healthcare environments are essential activities.
AI-based tools are useful for informed decision-making by both physicians and
patients. Albu et al. present several intelligent paradigms, particularly artificial neural
networks and fuzzy logic, for modelling, prediction, diagnosis and control in health-
care applications. These intelligent tools are able to assist in decision-making and
control processes for prevention, early detection and personalized healthcare.
Triberti et al. aim to tackle the “human” challenge pertaining to AI in healthcare
practice, focusing on the potential risk in the doctor–patient relationship. Noting that
there is still limited knowledge on the usage of AI in health and medicine, they study
the guidelines for identifying people who work with AI in the healthcare context.
They argue that it is important to form an interdisciplinary team with members who
are able to value both rigorous practice and health and well-being of patients.
Belciug acknowledges the cross-fertilization of statistical analysis and AI for
devising new and impactful methods to assist in medical practice and discovery. It
is necessary to exploit statistical analysis for validating AI-based methodologies in
healthcare, in order to improve reliability and credibility of the findings. In addition,
useful plan, design and implementation of statistical analysis with respect to AI in
healthcare research are discussed.
Pedell et al. examine the benefits of introducing humanoid robots into different
active ageing and aged care settings. It is found that implementation and interaction
with robots require a well-designed plan, in order to develop trust and interest for
creating a shift in feelings of control pertaining to older adults as well as staff. In
a group setting, older adults can engage and enjoy the interaction with both the
robot and the wider group with positive effects. Successful interactions between
v
vi Preface
older adults and humanoid robots also need to be supported by motivational goal
modelling and technology probe techniques.
To combat cancer, which is a leading cause of mortality worldwide, physical
activity (PA) plays a significant role in reducing the risk of developing cancer.
Dadhania and Williams investigate the use of digital wearable tools in offering
advantages including scale, cost and data capture. Specifically, current methods of
evaluating PA in cancer patients and how wearable accelerometers are used in cancer
clinical trials are studied. The successes and challenges associated with collecting
PA data with wearable accelerometers in digital healthcare trials are discussed.
Stankova et al. develop an online application of a home-administered parent-
mediated program for children with Autism spectrum disorder for enhancement
of their communication skills. The program is organized in modules, each with
different text and visual cards, targeting impressive/expressive language, discourse
abilities and other functions. The instructional component for parents involves activ-
ities within the Moddle e-educational platform. The administration for the program
follows a strict schedule, which is also available in Moodle.
To overcome the “black-box” issue, Gerlings et al. focus their research on explain-
able AI models. Different explanation needs with respect to stakeholders in the
case of classifying COVID-19 patients are studied. The need for a constellation of
stakeholders involved in human-AI collaborative decision-making is highlighted.
The study provides insights into how AI-based systems can be adjusted to support
different needs from stakeholders, in order to facilitate better implementation in the
healthcare context.
Resta uses a neural network model, i.e. the self-organizing map (SOM), to iden-
tify the emergence of COVID-19 clusters among different regions in Italy, in an
attempt to explain different characteristics of the pandemic within the same country.
Demographic, healthcare and political data at the regional level are considered, and
the interactions among them are examined. By leveraging capabilities of the SOM
model, the relations among variables can be visualized, and an early warning system
can be developed to address further intervention in the battle against the COVID-19
pandemic.
Casacuberta and Vallverdú indicate that universal emotion leads to a conceptual
bias in the use of AI in medical scenarios. Indeed, emotional responses in medical
practices are mediated culturally. As a result, a multicultural approach is required in
the medical context, taking special consideration of emotional variations with respect
to different cultural background of patients. From the computational perspective, the
most common biases that can originate from data treatment utilizing machine learning
algorithms are discussed.
The Russian Hoc Group on Application of AI Technologies in Health Informatics
(AHG2 TC215 ISO) highlights the importance of designing and deploying AI-based
systems in accordance with established guidelines and legislation for medical appli-
cations. In this respect, the formation of unified approaches, definitions and require-
ments for AI in medicine can significantly increase efficiency of the associated
development and application. A consistent approach through global standardization
can reduce the burden of stakeholders when establishing regulatory frameworks.
Preface vii
Initiatives to define goals and directions for standardization pertaining to AI in the
healthcare areas are discussed.
Gusev et al. discuss AI research and development in Russia, where government
and expert community are working together to develop legal and technical regula-
tions. AI-based software products for diagnostic and treatment processes, including
clinical trials, are regulated comprehensively. A balance between accelerating time
to market of AI products and ensuring their safety and efficacy is required along
with appropriate consideration on the potential risks and problems. The first series
of Russian national technical standards to accelerate AI product development and
instil trust in medical practitioners are being established.
Kolpashchikov et al. address issues and challenges on the use of robotic tech-
nologies in healthcare. In addition to surgical and rehabilitation robots, non-medical
robots that are useful for healthcare organizations to reduce costs, prevent disease
transmission and mitigate the lack of workforce are reviewed. One critical issue that
prevents future development of robotic in healthcare is lack of autonomy, which is
most challenging in minimally invasive surgery where flexible robots are used in
confined spaces. Innovative solutions for producing flexible robots as well as new
robotic designs with appropriate actuators and sensors are required.
Belandi et al. conduct a review on the development of Internet of things (IoT)
and machine learning for smart healthcare systems. Utilizing smart healthcare tech-
nologies encompassing IoT and machine learning devices for monitoring home envi-
ronments is becoming popular, particularly for elderly patients with long-term non-
acute diseases who do not require hospitalization. The survey focus is placed on
two aspects, namely architectures and algorithms, of the available technologies. A
taxonomy for classification of the reviewed models and systems is provided.
Hoppe et al. highlight the lack of studies on the potential of digital business models
in the healthcare sector. Key performance indicators (KPIs), individualization, effi-
ciency and communication channels are identified as the main factors. An evaluation
with a structural equation modelling process indicates that KPIs and communication
channels have a significant influence on the potential of digital business models and
their processes in healthcare. An outlook on the benefits and challenges pertaining
to the rapid development of AI in the healthcare sector is presented.
Manresa-Yee et al. explore the transparency and interpretability issues of AI,
particularly deep neural network models. Through explainable AI, users are able to
understand the predictions and decisions from AI-based systems, increasing trust-
fulness and reliability of the systems. An overview on explanation interfaces in the
healthcare context is discussed. A survey on healthcare related to studies on expla-
nations in the form of natural text, parameter influence, visualization of data graphs
or saliency maps is presented.
Giarelis et al. introduce a graph-based text representation method for discovery of
future research collaboration in the medical field. The method combines graph-based
feature selection and text categorization for formulation of a novel representation
of multiple scientific documents. The proposed method is able to provide useful
predictions on future research collaborations, as demonstrated through the use of the
COVID-19 Open Research Data Set.
viii Preface
Shopon et al. investigate information security by combining privacy concepts
and biometric technologies. An analysis on the protection of physiological and
social behavioural biometric data through a variety of authentication applications is
given. Current and emerging research studies in the multi-modal biometric domain,
including the use of deep learning-based methods, are explained. Open questions
and future directions in this research field are discussed, offering new methods in
biometric security and privacy investigation and providing insights into the emerging
topics of big data analytics and social network research.
The editors are grateful to all authors and reviewers for their contributions. We
would also like to thank the editorial team of Springer for their support throughout the
compilation of both volumes of this handbook. We sincerely hope that the research
and practical studies covered in both volumes can help instil new ideas and plans
for researchers and practitioners to work together, as well as to further advance
research and application of AI and related methodologies for the benefits of health
and well-being of humans.
Waurn Ponds, Australia Chee-Peng Lim
Shiga, Japan Yen-Wei Chen
Adelaide, Australia Ashlesha Vaidya
Nagpur, India Charu Mahorkar
Shoreham-by-Sea, UK Lakhmi C. Jain
May 2021
Contents
Part I Practicalities of AI Methodologies in Healthcare
1 Intelligent Paradigms for Diagnosis, Prediction and Control
in Healthcare Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Adriana Albu, Radu-Emil Precup, and Teodor-Adrian Teban
1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.2 Relevant References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.3 Medical Decision-Making Based on Artificial Neural
Networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.3.1 Skin Diseases Diagnosis . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.3.2 Hepatitis C Predictions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
1.3.3 Coronary Heart Disease Prediction . . . . . . . . . . . . . . . . . . 16
1.4 Medical Image Analysis Using Artificial Neural Networks . . . . . 18
1.5 Artificial Neural Networks Versus Naïve Bayesian
Classifier . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
1.5.1 Hepatitis B Predictions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
1.5.2 Stroke Risk Prediction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
1.6 Prosthetic Hand Myoelectric-Based Modeling and Control
Using Evolving Fuzzy Models and Fuzzy Control . . . . . . . . . . . . . 27
1.6.1 Evolving Fuzzy Modeling Results . . . . . . . . . . . . . . . . . . . 28
1.6.2 Fuzzy Control Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
1.7 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
2 Artificial Intelligence in Healthcare Practice: How to Tackle
the “Human” Challenge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
Stefano Triberti, Ilaria Durosini, Davide La Torre, Valeria Sebri,
Lucrezia Savioni, and Gabriella Pravettoni
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
2.2 AI in Healthcare . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
2.3 A “third Wheel” Effect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
2.3.1 “Confusion of the Tongues” . . . . . . . . . . . . . . . . . . . . . . . . 50
2.3.2 Decision Paralysis and Risk of Delay . . . . . . . . . . . . . . . . 51
ix
x Contents
2.3.3 Role Ambiguity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
2.4 An Interface for AI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
2.5 Identifying Personnel to Work with AI . . . . . . . . . . . . . . . . . . . . . . 55
2.6 Recommendations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
2.7 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
3 A Statistical Analysis Handbook for Validating Artificial
Intelligence Techniques Applied in Healthcare . . . . . . . . . . . . . . . . . . . 61
Smaranda Belciug
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
3.2 Hypothesis Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
3.2.1 Contingency Tables or Cross-Tabulation . . . . . . . . . . . . . 66
3.2.2 Odds Ratio . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
3.2.3 Pearson’sχ 2 Test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
3.3 Normality Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
3.3.1 Kolmogorov–Smirnov Goodness of Fit (K-S) Test . . . . . 69
3.3.2 Lilliefors Test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
3.3.3 Shapiro Wilk W Test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
3.4 Statistical Benchmarking Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
3.4.1 T-test or Student’s T-test . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
3.4.2 T-test for Two Independent Groups of Observations . . . . 73
3.4.3 Equality of Variances: Levene’s Test . . . . . . . . . . . . . . . . . 75
3.4.4 Equality of Variances: Bartlett’s Test . . . . . . . . . . . . . . . . 76
3.4.5 Mann–Whitney Test or Mann–Whitney Wilcoxon
Test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
3.4.6 One Way ANOVA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
3.4.7 Tukey’s Honest Significant Difference Test . . . . . . . . . . . 79
3.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
4 Designing Meaningful, Beneficial and Positive Human Robot
Interactions with Older Adults for Increased Wellbeing
During Care Activities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
Sonja Pedell, Kathy Constantin, Diego Muñoz, and Leon Sterling
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
4.2 Social Robotics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
4.2.1 The Nao Robot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
4.2.2 The Need for Meaningful Activities and a Holistic
Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
4.3 Method: Learning from HCI Approaches for Exploring
Social HRI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
4.3.1 Situated Action . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
4.3.2 Participatory Design and Mutual Learning . . . . . . . . . . . . 90
4.3.3 Technology Probes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
4.3.4 Motivational Goal Models and Technology Probes . . . . . 91
Contents xi
4.3.5 Understanding Emotions . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
4.3.6 Iterative Visits in the Field and Data Collection . . . . . . . 92
4.4 Four Case Studies Using the Nao in the Field . . . . . . . . . . . . . . . . . 92
4.4.1 Preparing Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . 92
4.4.2 Interaction stages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
4.4.3 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
4.4.4 Case Study 1: Active Ageing Knitting Group . . . . . . . . . 93
4.4.5 Case Study 2: Dementia Respite Care as Part
of the Active Ageing Program . . . . . . . . . . . . . . . . . . . . . . 95
4.4.6 Case Study 3: Men’s Shed . . . . . . . . . . . . . . . . . . . . . . . . . 99
4.4.7 Case Study 4: Residential Care . . . . . . . . . . . . . . . . . . . . . 99
4.5 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
4.5.1 Creating a Basis Through Humor and Turning
Initial Negative Emotions into Positive . . . . . . . . . . . . . . . 101
4.5.2 Increasing Wellbeing Through Activity
and Application of Skills . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
4.5.3 Situated AI for Human Robot Interactions . . . . . . . . . . . . 103
4.5.4 Designing Social Interactions . . . . . . . . . . . . . . . . . . . . . . . 103
4.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
5 Wearable Accelerometers in Cancer Patients . . . . . . . . . . . . . . . . . . . . 109
Seema Dadhania and Matthew Williams
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
5.2 The Cancer Patient and Outcome Measures . . . . . . . . . . . . . . . . . . 111
5.2.1 Measuring Physical Activity . . . . . . . . . . . . . . . . . . . . . . . . 111
5.2.2 Measuring Physical Activity in the Cancer Patient . . . . . 112
5.3 Harnessing Wearable Technology in Oncology . . . . . . . . . . . . . . . 114
5.3.1 What Can Wearable Technology Be Used
to Measure in Oncology, and Why Are These
Parameters Relevant? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
5.4 Accelerometers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
5.4.1 Challenges with Wearable Accelerometer Data . . . . . . . . 118
5.5 Real-World Experience of Running a Digital Health Study . . . . . 122
5.5.1 Device Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
5.5.2 Successes and Challenges of Running
a Real-World Wearable Accelerometer Study . . . . . . . . . 125
5.6 Clinical Studies in Cancer Patients Using Wearable
Accelerometers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
5.7 Ethical Issues with Wearable Accelerometer Data . . . . . . . . . . . . . 135
5.7.1 Data Privacy and Security . . . . . . . . . . . . . . . . . . . . . . . . . . 135
5.7.2 Data Ownership . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
5.7.3 Insurance Premiums . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
5.8 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
xii Contents
6 Online Application of a Home-Administered Parent-Mediated
Program for Children with ASD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
Margarita Stankova, Tsveta Kamenski, Polina Mihova,
and Todor Datchev
6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150
6.2 Conceptual Framework and Aims of the Program . . . . . . . . . . . . . 151
6.2.1 Behavioral Model of Communicative Failure . . . . . . . . . 151
6.2.2 Structure of the Program . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
6.2.3 Technical Description and Parameters
of the Program . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
6.2.4 Technical Specifications of the System . . . . . . . . . . . . . . . 159
6.3 Pilot Testing of the Program—Qualitative Analysis . . . . . . . . . . . 162
6.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
7 Explainable AI, But Explainable to Whom? An Exploratory
Case Study of xAI in Healthcare . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
Julie Gerlings, Millie Søndergaard Jensen, and Arisa Shollo
7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170
7.2 Related Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
7.2.1 Adoption and Use of AI in Healthcare . . . . . . . . . . . . . . . 171
7.2.2 Drivers for xAI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172
7.2.3 Emergence of xAI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
7.2.4 AI and xAI in the Fight Against the COVID-19
Pandemic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174
7.3 Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176
7.3.1 Data Collection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176
7.3.2 Data Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178
7.4 Case Setting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
7.5 Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182
7.5.1 Development Team . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182
7.5.2 Subject Matter Expert . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
7.5.3 Decision-Makers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
7.5.4 Audience . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 190
7.6 Discussion and Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . 191
Appendix 1—Technical Aspects of LungX . . . . . . . . . . . . . . . . . . . . . . . . . 193
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195
8 Pandemic Spreading in Italy and Regional Policies:
An Approach with Self-organizing Maps . . . . . . . . . . . . . . . . . . . . . . . . 199
Marina Resta
8.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
8.2 Related Literature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200
8.3 Data and Research Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202
8.4 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205
8.5 Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207
Contents xiii
8.6 Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212
9 Biases in Assigning Emotions in Patients Due to Multicultural
Issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215
David Casacuberta and Jordi Vallverdú
9.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215
9.2 The Non-Universality of Emotions . . . . . . . . . . . . . . . . . . . . . . . . . . 216
9.3 Emotions in Medical Contexts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217
9.4 Machine Learning, Data, Emotions, and Diagnosis . . . . . . . . . . . . 219
9.4.1 What is Affective Computing? . . . . . . . . . . . . . . . . . . . . . . 219
9.4.2 Data for Automatic Emotion Detection . . . . . . . . . . . . . . . 220
9.4.3 Developing the Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . 221
9.5 Correcting Data Biases in Medical Diagnosis . . . . . . . . . . . . . . . . . 222
9.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225
Part II Prospects of AI Methodologies in Healthcare
10 Artificial Intelligence in Healthcare: Directions
of Standardization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
Georgy Lebedev
10.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
10.2 Definition of Artificial Intelligence (AI) . . . . . . . . . . . . . . . . . . . . . 232
10.3 History . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233
10.4 AI Features and Development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234
10.5 Problems and Challenges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239
10.6 AI Systems in Healthcare . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 244
10.7 Quality and Safety of AI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 248
10.8 Standardization of AI in Healthcare . . . . . . . . . . . . . . . . . . . . . . . . . 252
10.9 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256
11 Development of Artificial Intelligence in Healthcare in Russia . . . . . 259
A. Gusev, S. Morozov, G. Lebedev, A. Vladzymyrskyy,
V. Zinchenko, D. Sharova, E. Akhmad, D. Shutov,
R. Reshetnikov, K. Sergunova, S. Izraylit, E. Meshkova,
M. Natenzon, and A. Ignatev
11.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 260
11.1.1 National Strategy for AI in Healthcare
of the Russian Federation . . . . . . . . . . . . . . . . . . . . . . . . . . 261
11.1.2 The Work of Government Agencies and the Expert
Community on the Development of AI
in Healthcare . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 262
11.2 AI Regulations in Healthcare of the Russian Federation . . . . . . . . 267
11.2.1 Basic Principles of Regulations in Healthcare . . . . . . . . . 267
xiv Contents
11.2.2 Technical and Clinical Trials of Software
as a Medical Device Created with the Application
of AI Technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 269
11.2.3 State Registration of Software as a Medical Device
Created with the Application of AI Technologies . . . . . . 271
11.2.4 Post-registration Monitoring of Software
as a Medicaldevice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271
11.3 Technical Regulations of Artificial Intelligence
in the Russian Federation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 272
11.4 Practical Experience of Artificial Intelligence in Healthcare
of the Russian Federation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273
11.5 Chapter Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 276
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277
12 Robotics in Healthcare . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281
Dmitrii Kolpashchikov, Olga Gerget, and Roman Meshcheryakov
12.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281
12.2 Surgical Robots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283
12.2.1 Computer-Assisted Surgery . . . . . . . . . . . . . . . . . . . . . . . . 283
12.2.2 Mechanical Design and Control . . . . . . . . . . . . . . . . . . . . . 284
12.2.3 Application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286
12.3 Rehabilitation Robots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 287
12.3.1 Contact Therapy Robots . . . . . . . . . . . . . . . . . . . . . . . . . . . 287
12.3.2 Assistive Robotics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289
12.3.3 Non-Contact Therapy Robots and Socially
Assistive Robotics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 292
12.4 Non-Medical Robots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 293
12.5 Challenges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 294
12.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 296
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 297
13 Smart Healthcare, IoT and Machine Learning: A Complete
Survey . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 307
Valerio Bellandi, Paolo Ceravolo, Ernesto Damiani, and Stefano Siccardi
13.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 307
13.2 Architecture and Pipeline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 309
13.2.1 Research Questions and Methodology Adopted . . . . . . . 310
13.3 The General Picture of Levels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 311
13.3.1 Architectures for the Local Integration
Level—The Edge Level . . . . . . . . . . . . . . . . . . . . . . . . . . . . 312
13.3.2 Task Allocation and Resource Management—The
Fog Level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 312
13.3.3 Global Integration of Tasks and Resources—The
Cloud Level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313
13.3.4 Algorithms and Data Analytics . . . . . . . . . . . . . . . . . . . . . 314
13.3.5 Architectural Configurations . . . . . . . . . . . . . . . . . . . . . . . 315
Contents xv
13.4 Data Pipeline and Data Storage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 319
13.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 322
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 323
14 Digital Business Models in the Healthcare Industry . . . . . . . . . . . . . . . 331
Nathalie Hoppe, Felix Häfner, and Ralf Härting
14.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 332
14.2 Role of the Healthcare Sector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 332
14.3 Current Trends of Digitalization in Healthcare . . . . . . . . . . . . . . . . 333
14.4 Potential Benefits of Digital Business Models
in the Healthcare Industry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 335
14.4.1 Research Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 335
14.4.2 Industry-Dependent Determinants
of Digitalization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 336
14.4.3 Digital Technologies Along the Care Pathway . . . . . . . . . 337
14.4.4 Challenges of Digitalization in Healthcare . . . . . . . . . . . . 339
14.4.5 Study Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 340
14.4.6 Interpretation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 346
14.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 348
14.6 Outlook: The Role of AI in Healthcare . . . . . . . . . . . . . . . . . . . . . . 349
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 350
15 Advances in XAI: Explanation Interfaces in Healthcare . . . . . . . . . . . 357
Cristina Manresa-Yee, Maria Francesca Roig-Maimó,
Silvia Ramis, and Ramon Mas-Sansó
15.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 358
15.2 Related Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 359
15.3 Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 361
15.4 Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 361
15.4.1 Prediction Tasks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 362
15.4.2 Diagnosis Tasks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 363
15.4.3 Automated Tasks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 364
15.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 365
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 366
16 Medical Knowledge Graphs in the Discovery of Future
Research Collaborations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 371
Nikolaos Giarelis, Nikos Kanakaris, and Nikos Karacapilidis
16.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 371
16.2 Background Issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 372
16.2.1 Graph Measures and Indices . . . . . . . . . . . . . . . . . . . . . . . . 373
16.2.2 Graph-Based Text Representations . . . . . . . . . . . . . . . . . . 374
16.2.3 Graph-Based Feature Selection . . . . . . . . . . . . . . . . . . . . . 374
16.2.4 Graph-Based Text Categorization . . . . . . . . . . . . . . . . . . . 375
16.2.5 Graph-Based Link Prediction . . . . . . . . . . . . . . . . . . . . . . . 375
16.3 The Proposed Framework . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 376
xvi Contents
16.3.1 Graph-Based Text Representation . . . . . . . . . . . . . . . . . . . 376
16.3.2 Graph-Based Feature Selection . . . . . . . . . . . . . . . . . . . . . 377
16.3.3 Graph-Based Text Categorization . . . . . . . . . . . . . . . . . . . 378
16.3.4 Graph-Based Link Prediction . . . . . . . . . . . . . . . . . . . . . . . 378
16.4 Experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 382
16.4.1 Cord-19 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 382
16.4.2 Experimental Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 382
16.4.3 Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 384
16.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 385
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 388
17 Biometric System De-identification: Concepts, Applications,
and Open Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 393
Md. Shopon, A. S. M. Hossain Bari, Yajurv Bhatia,
Pavan Karkekoppa Narayanaswamy, Sanjida Nasreen Tumpa,
Brandon Sieu, and Marina Gavrilova
17.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 393
17.2 Literature Review and Classification of Biometric
De-identification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 394
17.3 New Types of Biometric De-identification . . . . . . . . . . . . . . . . . . . 397
17.3.1 Sensor-Based Biometric De-identification . . . . . . . . . . . . 397
17.3.2 Emotion-Based De-identification . . . . . . . . . . . . . . . . . . . . 400
17.3.3 Social Behavioral Biometrics-Based
De-identification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 403
17.3.4 Psychological Traits-Based De-identification . . . . . . . . . 405
17.3.5 Aesthetic-Based Biometric De-identification . . . . . . . . . . 408
17.4 Multi-Modal De-identification System . . . . . . . . . . . . . . . . . . . . . . 409
17.4.1 Definition and Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . 409
17.4.2 Deep Learning Architecture . . . . . . . . . . . . . . . . . . . . . . . . 410
17.4.3 Multi-Modal De-identification Methodology . . . . . . . . . . 411
17.4.4 Potential Applications of Multi-Modal Biometric
De-identification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 412
17.5 Potential Applications in Risk Assessment and Public
Health . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 413
17.6 Open Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 415
17.6.1 Open Problems of Sensor-Based Biometric
De-identification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 415
17.6.2 Open Problems of Gait and Gesture
De-identification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 416
17.6.3 Open Problems of Emotion-Based
De-identification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 416
17.6.4 Open Problems of Social Behavioral
De-identification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 416
17.6.5 Open Problems of Psychological Traits-Based
De-identification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 417
Contents xvii
17.6.6 Open Problems of Aesthetic-Based Biometric
De-identification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 417
17.6.7 Open Problems of Multi-Modal De-identification . . . . . . 417
17.7 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 418
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 418
Part I
Practicalities of AI Methodologies
in Healthcare
Chapter 1
Intelligent Paradigms for Diagnosis,
Prediction and Control in Healthcare
Applications
Adriana Albu, Radu-Emil Precup, and Teodor-Adrian Teban
Abstract Decision-making and control in healthcare applications are essential activ-
ities, which imply a large number of medical and technical aspects, and the nonlin-
earity of systems specific to these applications makes them challenging. In addition,
these activities involve humans: on the one hand, the patient, who has a medical
problem and who requires the best solution; on the other hand, the physician, who
should be able to provide, in any circumstances, a decision or a prediction regarding
the current and the future medical status of a patient. The technology, in general, and
particularly the artificial intelligence tools could help both of them, and it is assisted
by appropriate theory regarding modeling tools. Diagnosis, modeling, prediction
and control represent the mechanisms that support theoretically the healthcare appli-
cations as far as decision-making is involved. Two of the most powerful intelli-
gent paradigms that are successfully used in this field are artificial neural networks
and fuzzy logic, with their corresponding models. This chapter presents several
applications developed by the Process Control Group of the Politehnica University
Timisoara, Romania and emphasizes that these techniques, which produce intel-
ligent models, even if they are artificial, are able to make decisions and to control
various processes, being useful tools for prevention, early detection and personalized
healthcare.
Keywords Artificial neural networks · Fuzzy control · Fuzzy models · Medical
diagnosis · Medical prediction · Prosthetic hands · Recurrent neural networks
A. Albu · R.-E. Precup (B) · T.-A. Teban
Department of Automation and Applied Informatics, Politehnica University Timisoara, Bd. V.
Parvan 2, 300223 Timisoara, Romania
e-mail:
[email protected]A. Albu
e-mail:
[email protected]T.-A. Teban
e-mail:
[email protected]© The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 3
C.-P. Lim et al. (eds.), Handbook of Artificial Intelligence in Healthcare,
Intelligent Systems Reference Library 212,
https://doi.org/10.1007/978-3-030-83620-7_1
