Applications of Artificial Intelligence in E-Healthcare Systems
2: Shiv Nadar University, Delhi-National Capital Region (NCR), India
3: Samarkand State University, Uzbekistan
Increased use of artificial intelligence (AI) is being deployed in many hospitals and healthcare settings to help improve health care service delivery. Machine learning (ML) and deep learning (DL) tools can help guide physicians with tasks such as diagnosis and detection of diseases and assisting with medical decision making.
This edited book outlines novel applications of AI in e-healthcare. It includes various real-time/offline applications and case studies in the field of e-Healthcare, such as image recognition tools for assisting with tuberculosis diagnosis from x-ray data, ML tools for cancer disease prediction, and visualisation techniques for predicting the outbreak and spread of Covid-19.
Heterogenous recurrent convolution neural networks for risk prediction in electronic healthcare record datasets are also reviewed.
Suitable for an audience of computer scientists and healthcare engineers, the main objective of this book is to demonstrate effective use of AI in healthcare by describing and promoting innovative case studies and finding the scope for improvement across healthcare services.
Inspec keywords: health care; natural language processing; electronic health records; learning (artificial intelligence); data analysis; diseases
Other keywords: medical computing; diseases; patient diagnosis; natural language processing; data analysis; artificial intelligence; learning (artificial intelligence); health care; medical information systems; electronic health records
Subjects: Biology and medical computing; Natural language interfaces; Medical administration; Machine learning (artificial intelligence); Data handling techniques
- Book DOI: 10.1049/PBHE040E
- Chapter DOI: 10.1049/PBHE040E
- ISBN: 9781839534492
- e-ISBN: 9781839534508
- Page count: 306
- Format: PDF
-
Front Matter
- + Show details - Hide details
-
p.
(1)
-
1 Introduction to AI in E-healthcare
- + Show details - Hide details
-
p.
1
–18
(18)
Artificial intelligence includes the improvement of computer frameworks that are suitable for performing assignments that regularly require human knowledge, for example, object detection, solving complex problems and so on. Nowadays, humans encounter various novel diseases, with symptoms that may not be identifiable at an early stage, causing high loss of life. AI can improve healthcare and so extend the human life span. Specifically, with rapid enhancements in computer handling, these AI-based frameworks are currently improving the accuracy and productivity in identifying and treating diseases across different specializations. AI is evolving in the e-healthcare industry in a number of areas, including: collecting medical records, AI-based diagnostics, automatic follow-up, etc. Data science (AI, machine learning, deep learning) has been extensively developed for the e-healthcare sector to provide improved quality in decision-making, finding daily infection patterns, identifying early risks and prediction, and so saving lives and promoting good health. Electronic health records enable the provision of better data-driven decisions, building good machine learning models for prediction, greater operational speed and accuracy, and reducing the amount of valuable time spent by patients and healthcare workers. In this chapter we discuss the complete road map to construct powerful artificial intelligence (AI) in e-healthcare systems using tools and intelligent algorithms. AI e-healthcare applications can be used to manipulate, transmit, and keep patient information, using medical records to enable fast access anywhere in the world. This works on the Internet which is able to monitor payments and maintain records through e-healthcare areas.
-
2 The scope and future outlook of artificial intelligence in healthcare systems
- + Show details - Hide details
-
p.
19
–36
(18)
The goal described in this chapter is to imitate human cognitive functions in the context of artificial intelligence. This translates into healthcare, enhanced by the availability of health data and the rapid advancement of analytical technology. This chapter discusses and addresses the current scenario of AI in overall healthcare applications. Different forms of healthcare data may be used. Machine learning for complex information, namely, classical vector support and neural network systems and current dedicated knowledge as well as linguistic processing for unstructured data are common AI techniques. Cancer, neurology, and cardiology are the key fields of health issues that could be analyzed and predicted using AI methods. AI transforms the healthcare area, providing guidelines for the diagnosis and treatment, communication with, and collaboration of patients. Exploring AI and big data analytical applications provides users with insights and enables users to prepare and use resources, particularly to meet unique health challenges.
-
3 Class dependency-based learning using Bi-LSTM coupled with the transfer learning of VGG16 for the diagnosis of tuberculosis from chest X-rays
- + Show details - Hide details
-
p.
37
–54
(18)
Tuberculosis is an infectious disease that leads to the death of millions of people across the world. The mortality rate from this disease is high in patients suffering from immunocompromised disorders. early diagnosis can save lives and avoid further complications. However, the diagnosis of TB is a very complex task. The standard diagnostic tests still rely on traditional procedures developed in the 20th century. These procedures are slow and expensive. Therefore, this chapter presents an automatic approach for the diagnosis of TB from posteroanterior chest X-rays. This is a two-step approach, in which in the first step the lung regions are segmented from the chest X-rays using the graph cut method, and then in the second step the transfer learning of VGG16 combined with bidirectional LSTM is used for extracting high-level discriminative features from the segmented lung regions and then classification is performed using a fully connected layer. The proposed model is evaluated using data from two publicly available databases, namely the Montgomery County set and the Schezien set. The proposed model achieved accuracy and sensitivity of 97.76%, 97.01% and 96.42%, 94.11% on the Schezien and Montgomery County datasets, respectively. This model enhanced the diagnostic accuracy of TB by 0.7% and 11.68% on the Schezien and Montgomery County datasets, respectively.
-
4 Drug discovery clinical trial exploratory process and bioactivity analysis optimizer using deep convolutional neural network for E-prosperity
- + Show details - Hide details
-
p.
55
–75
(21)
Drug discovery is a complex and a time-consuming process as its success depends on various factors. An optimum and precise combination of drug compounds is key to treating individual health conditions. The current coronavirus pandemic witnessed an increase in adoption of the E-Prosperity system. E-Prosperity helps doctors perform various kinds of research and drug discovery operations by continuous monitoring of quarantined patients as well as frontline workers to keep them safe. According to WHO, around 82 million COVID-19 cases were reported worldwide at end of 2020. This created high demand for E-Prosperity system for disease diagnosis and monitoring. Another challenge is the discovery of drug; various researches are being carried out to find the effective drug for coronavirus infection. But most of the researchers face difficulties such as lack of physical apparatus or system to handle large number of patients remotely, short time to perform four-phases of drug clinical trials, and to find the optimal drug composition to cure the disease without causing any side effects to the patients. The major challenge we face here is to handle huge amounts of data generated during the clinical trial process, considering the correct parameters required for efficient discovery. Machine learning provides some advanced solutions, such as reinforcement learning, generative learning, and more. Deep learning is an emerging technology that helps overcome the drawbacks like dimensionality reduction and feature extraction. In the proposed model, we used a novel model called Optimized Deep Convolution Neural Network Outer-Loop, which is based on data to improve the clinical trial results and to suggest effective decisions in the field of drug discovery using the information collected over the electronic devices.
-
5 An automated NLP methodology to predict ICU mortality CLINICAL dataset using multiclass grouping with LSTM RNN approach
- + Show details - Hide details
-
p.
77
–96
(20)
A large amount of data is generated in the fields of medicine, as healthcare is a major component of a tech-oriented lifestyle. Even now, the pandemic situation in 2020 led to a huge burden as well as medical emergencies. A need for a detailed understanding of various healthcare electronic clinical data (HECD) using Natural Language Processing (NLP) techniques has been recognized. High protection for clinical data that are being analyzed, the patient processed information generated from diagnosis using various methods, such as in ICU (intensive care unit), mortality levels, and resource allocation are performed instantly. Experts from the medical field use a patient's information from the laboratory results, such as X-ray, MRI, CTC scan, etc., to diagnose the disease using various inputs. The known vital parameters from clinical results, such as blood pressure (BP) and heart rate, always fluctuate and are inconsistent as well as the mismatch of information based on age, name, etc., are to be focused on to resolve the major challenges. Here we used the ICU clinical data as a sample to consider the features based on multiclass word phrase classification from unstructured data using a long short-term approach (LSTM). We categorize the data according to classes and group them into valid information wherein records can be minimal and help reduce the text. Also, valuable information can be protected and sensitive data can be kept confidential by phenomenal support, which is automated using NLP approaches. There are typical steps for treatment based on the critical level of the patients, in such stage time-series is mostly handled as development variables for mortality extraction from ICU data. A supervised structured data from clinical processing involve a bi-directional LSTM approach for processing the samples and brings efficient results from the records to make a reliable model from multiclass prediction.
-
6 Applying machine learning techniques to build a hybrid machine learning model for cancer prediction
- + Show details - Hide details
-
p.
97
–114
(18)
With the arrival of the latest advancements in the area of medicine, massive amounts of information related to cancer are being collected and available on the market to the medical analysis community. However, forecasting disease is one of the foremost, fascinating, and difficult responsibilities of physicians. Machine learning methods are becoming a preferred tool for medical researchers. In our proposed method, a hybrid machine learning model (HMLM) is built using ensemble learning techniques, using stacking of models such as multilayer perceptron (MLP), decision tree (DT), support vector machine (SVM), and logistic regression (LR). Initially, preprocessing is done and then the output is passed to the feature selection method, where each feature is ranked in accordance with the dependent attribute. Once the features are selected, HMLM is used for classification. The output is then validated using a confusion matrix and also by calculating the score of the model. The proposed model performs well with the reduced number of features and gives higher accuracy than existing models.
-
7 AI in healthcare: challenges and opportunities
- + Show details - Hide details
-
p.
115
–140
(26)
Artificial intelligence (AI) is the part of computer science which focuses on developing automated advanced technologies that can execute specific tasks that would normally require human knowledge. AI is a particular concept of science with many unique methods and advanced machine learning techniques which has effected fundamental changes in nearly almost all fields of the technology sector. AI can be classified into two parts, the first is narrow AI, which is always based on executing tasks in the correct way, and although certain systems appear smart, there are some constraints and limitations. The second is artificial general intelligence (AGI) which describes a device that has intellectual ability and which can use that ability to make improvements, in much the same was as a human. Healthcare 4.0 is about collecting massive quantities of data and bringing them to use in apps, aiming for better health business decisions as well as huge cost reductions and control of those costs. Virtual assistants can respond to questions, track patients, and have fast responses because they are available 24/7. Today, many personal assistant apps allow for more contact among patients and healthcare providers during doctors' visits, reducing the risk of readmission to the clinic or repeated medical visits. The world has been struggling greatly with the COVID-19 pandemic, although each step of advanced technical knowledge and creative problem solving is bringing us closer to ending this global situation. Both machine learning and artificial intelligence are helping to solve the problems caused by COVID-19 and to rid society of this pandemic. Machine learning allows advanced technologies to imitate humans, and consumes vast amounts of data in order to recognize technical patterns and perceptions rapidly. Institutions were fast to implement machine learning abilities in the battle against COVID-19 in many ways, including scaling consumer interactions, identifying how COVID-19 grows, and speeding up testing and diagnosis.
Many of these trends are converging, including data collection, increased AI usage, and the implementation of web apps. As a result of the emphasis on implementation and transformation, health services are becoming much more reliable. Such programs will be capable of transforming multiple stages of medication, as well as smart devices and the change to personalized medicine (PM). We address herein a number of AI challenges and opportunities in healthcare, which could help experts and stakeholders in taking the best possible decisions.
-
8 Impression of artificial intelligence in e-healthcare medical applications
- + Show details - Hide details
-
p.
141
–166
(26)
The aim of Artificial Intelligence (AI) is to depict human intelligence. It's causing a pattern change in healthcare, due to the ease of electronic medical records and the rapid growth in analytical tools. Not long ago, AI techniques sent shockwaves through the healthcare industry, encouraging a lively debate as to whether AI doctors could eventually treat patients suffering in the future. Human doctors are difficult to be replaced by machines anytime soon, but AI can help physicians in good clinical decision-making and even substitute human judgement in some areas of healthcare. The notable recent implementations of AI in healthcare have been achieved by the widespread availability of health information and the rapid increase in availability of big data analytic methods. Medically important information hidden away in vast amounts of data can be revealed using effective AI methods, which can aid in medical decision. Until AI systems could be used in healthcare, they must be "educated" using data produced from patient evaluation, such as transmitting, diagnostics, and medication assignment, so that they might learn related groups of subjects, correlations among subject features, and expected results. Demography, patient records, electronic records from surgical equipment, physical exams, and established testing laboratory and digital images are all examples of medical studies. AI has been utilized in a range of technological areas, including IoT, computer vision, automated vehicles, and natural language processing, according to the rapid advancement of AI operating systems technologies. Most excitingly, biomedical researchers are now consciously attempting to use AI to enhance diagnosis and patient outcomes, thus increasing the overall utility of the healthcare sector. The increase in interest is evident, particularly since the past 5 years, and strong development can be expected in the near future. The impact AI can have on the medical sector in the future is significant.
-
9 Heterogeneous recurrent convolution neural network for risk prediction in the EHR dataset
- + Show details - Hide details
-
p.
167
–187
(21)
The increasing adoption of electronic health record (EHR) systems has brought tremendous opportunities in medicine while enabling more personalized prognostic models. Predicting the risk of chronic diseases using EHRs has attracted considerable attention in recent years. Diabetes is a chronic disease, and using EHRs to predict the onset of diabetes could improve the quality and efficiency of medical care. However, most work to date has investigated the binary classification problem for predicting the onset of diabetes, but little attention has been given to assessing the risk of developing comorbidities that are major causes of morbidity and mortality. Leveraging massive EHRs brings tremendous promise to advancing clinical and precision medicine informatics research. However, it is very challenging to directly work with multifaceted patient information encoded in EHR data. Deriving effective representations of patient EHRs is a crucial step to bridge raw EHR information and the endpoint analytical tasks, such as risk prediction or chronic disease subtyping. To overcome these problems two new approaches, Heterogeneous Recurrent Convolution Neural Network (HRCC) and Multi Level Spatial Coherence Optimization Approach (MLSCO), are developed in this work. To understand the predictive performance of this approach several performance metrics are used. Compared with traditional machine learning models, the deep learning-based approach achieves superior performance on risk prediction tasks. The experimental results show that this method provides a superior predictive effect than other traditional machine learning models.
-
10 A narrative review and impacts on trust for data in the healthcare industry using artificial intelligence
- + Show details - Hide details
-
p.
189
–204
(16)
Artificial intelligence (AI) and robotics are quickly entering the healthcare industry, assuming several important responsibilities, including the capability to diagnose and treat patients. Although the majority of research in its technological development has been focused on interpersonal contact, this has led to a slew of technological challenges. As a result, the scope of technologies' effect on human-human interactions and relationships in the healthcare industry is limited rather than treating healthcare information technology. The hypothesis of AI in health care is that trust is critically important for healthcare partnerships, and because of this, healthcare AI could have dramatic consequences on these bonds of trust. In helping the technology of healthcare in AI, it is believed that data from the traditional task and medical devices have been classified into various levels for maintaining healthcare in the AI domain. Despite this, the application of AI will necessitate successful planning and strategy implementation to help and comprehensively apply these technologies.
-
11 Analysis of COVID-19 outbreak using data visualization techniques: a review
- + Show details - Hide details
-
p.
205
–243
(39)
Coronaviruses are a family of infections viruses that can cause sickness due to infection. They can shift from basic cold and hack to a more serious malady. The SARS-CoV and MERS-CoV syndromes are the most serious instances that the world has ever encountered. This pandemic is spreading worldwide, and it is critical for us to analyze and comprehend its spread using various approaches. Our work is primarily concerned with determining the global spread pattern of the virus. This groundbreaking work presents a study of the COVID-19 outbreak using various visualization techniques and data analysis techniques. This study also shows the comparison of cases between China, from where this pandemic actually arise, and the rest of the world. Additionally, this work analyzes the impact and dissemination of COVID-19 using several prediction and time-series forecasting techniques, such as SARIMA and ARIMA models.
-
12 Artificial intelligence-based electronic health records for healthcare
- + Show details - Hide details
-
p.
245
–266
(22)
Artificial intelligence (AI) is the ability of machines to perform tasks like a person and assist in solving existing real-life problems. These machines have their own ability to solve problems by constructive use of machine learning, deep learning, and neural network algorithms. AI created by persons adds many intelligent features through which they can resolve existing real-life problems. An electronic health record (EHR) is a collection of different types of documents generated by any health-associated body tests or wearable smart devices stored electronically in a digital format. Healthcare is a very vast structure and there are millions of health matters. So, it is very difficult to collect all health documents and keep them in a structured format for future use. It is a major concern because many deaths occur in a year since they do not have access to health documents. In this chapter, constructive use of AI in EHR structure is discussed.
EHR is the digital form of all documents in a single place. It is highly secure and is only pre-owned by authorized users. It assists to treat forbearing with multiple health matters and manage all their documentation for future use. In such cases electronic health documentation works like a key to saving several lives. AI is a very advanced technology that assists in collecting all advices about the forbearing and how keep them secure. According to Gartner's report, AI grew between 2018 and 2019 from 4% to 14%, and in the 2020 Report it increased to 40%. AI plays an important and huge role in e-healthcare and also assists to keep the information secure in EHR. This chapter describes the role of AI in e-healthcare and EHR with the architecture of healthcare documentation. We study the requests, use of AI in EHR, healthcare, and privacy of all health documents.
-
13 Automatic structuring on Chinese ultrasound report of Covid-19 diseases via natural language processing
- + Show details - Hide details
-
p.
267
–282
(16)
Artificial intelligence (AI) technology saves human lives and reduces the risks taken by doctors. In the proposed work identification of lung sound and blood cells is automated. AI enables the extraction of de-noised lung sounds to understand the patient's health and maintain social distance between doctors and patients. The COVID-19 virus affects a large number of humans day-to-day. So doctors and caretakers are not able to maintain the social distance with the patients. In this paper, we provide intelligent interaction from patients' facial movements and sound recognition (SR) using human-robot interaction (HRI). COVID-19 attacks hemoglobin. In general, blood cells, in particular red blood cells, are counted manually using a hemocytometer along with some laboratory equipment and chemical compounds. This device detects a patient's intrusive physiological signals like whether heart rate (HR) is continuous and measures HR and SR in red blood cells using multinomial tactic regression (MTR) algorithm for better performance. Automatic lung sound extraction is done using a convolutional structural network (CSN). Natural language processing can be easily used to detect and improve the performance of the lung and for the reduction of lung sound noise.
-
Back Matter
- + Show details - Hide details
-
p.
(1)