Intelligent systems in patient monitoring and therapy management. A survey of research projects

Artificial Intelligence With Robotics in Healthcare: A Narrative Review of Its Viability in India

This short review focuses on the emerging role of artificial intelligence (AI) with robotics in the healthcare sector. It may have particular utility for India, which has limited access to healthcare providers for a large growing population and limited health resources in rural India. AI works with an amalgamation of enormous amounts of data using fast and complex algorithms. This permits the software to quickly adapt the pattern of the data characteristics. It has the possibility to collide with most of the facets of the health system which may range from discovery to prediction and deterrence. The use of AI with robotics in the healthcare sector has shown a remarkable rising trend in the past few years. Functions like assistance with surgery, streamlining hospital logistics, and conducting routine checkups are some of the tasks that may be managed with great efficiency using artificial intelligence in urban and rural hospitals across the country. AI in the healthcare sector is advantageous in terms of ensuring exclusive patient care, safe working conditions where healthcare providers are at a lower risk of getting infected, and perfectly organized operational tasks. As the healthcare segment is globally recognized as one of the most dynamic and biggest industries, it tends to expedite development through modernization and original approaches. The future of this lucrative industry is looking forward to a great revolution aiming to create intelligent machines that work and respond like human beings. The future perspective of AI and robotics in the healthcare sector encompasses the care of elderly people, drug discovery, diagnosis of deadly diseases, a boost in clinical trials, remote patient monitoring, prediction of epidemic outbreaks, etc. However, the viability of using robotics in healthcare may be questionable in terms of expenditure, skilled workforce, and the conventional mindset of people. The biggest challenge is the replication of these technologies to the smaller towns and rural areas so that these facilities may reach the larger segment of the entire population of the country. This review aims to examine the adaptability and viability of these new technologies in the Indian scenario and identify the major challenges.

A Review of Artificial Intelligence and Robotics in Transformed Health Ecosystems

Health care is shifting toward become proactive according to the concept of P5 medicine-a predictive, personalized, preventive, participatory and precision discipline. This patient-centered care heavily leverages the latest technologies of artificial intelligence (AI) and robotics that support diagnosis, decision making and treatment. In this paper, we present the role of AI and robotic systems in this evolution, including example use cases. We categorize systems along multiple dimensions such as the type of system, the degree of autonomy, the care setting where the systems are applied, and the application area. These technologies have already achieved notable results in the prediction of sepsis or cardiovascular risk, the monitoring of vital parameters in intensive care units, or in the form of home care robots. Still, while much research is conducted around AI and robotics in health care, adoption in real world care settings is still limited. To remove adoption barriers, we need to address issues such as safety, security, privacy and ethical principles; detect and eliminate bias that could result in harmful or unfair clinical decisions; and build trust in and societal acceptance of AI.

The Beginning of a New Era: Artificial Intelligence in Healthcare

The healthcare sector is considered to be one of the largest and fast-growing industries in the world. Innovations and novel approaches have always remained the prime aims in order to bring massive development. Before the emergence of technology, the healthcare sector was dependent on manpower, which was time-consuming and less accurate with lack of efficiency. With the recent advancements in machine learning, the condition has been steadily revolutionizing. Artificial intelligence (AI) lies in the computer science department, which stresses on the intelligent machines’ creation, that work and react just like human beings. Currently, the applications of AI have been expanding into those fields, which was once thought to be the only domain of human expertise such as healthcare sector. In this review, we have shed light on the present usage of AI in the healthcare sector, such as its working, and the way this system is being implemented in different domains, such as drug discovery, diagnosis of diseases, clinical trials, remote patient monitoring, and nanotechnology. We have also briefly touched upon its applications in other sectors as well. The public opinions have also been analyzed and discussed along with the future prospects. We have discussed the merits, and the other side of AI, i.e. the disadvantages in the last part of the manuscript.

An Analysis of Problems with Auditory Alarms: Defining the Roles of Alarms in Process Monitoring Tasks

It has become a standard practice to use auditory alarm devices to enhance human monitoring performance in monitoring tasks. However, the effectiveness of such practice has been-challenged from time to time, which leads to the fundamental question of what roles alarms should and could assume. This paper reviews reported observations of interactions between human operators and alarm mechanisms in patient care, aviation, and process control. Based on the reviews, we propose that the roles of alarms in process monitoring tasks should be viewed more as a way of informing process status and less as a way of interpreting the significance of process status. The roles can best be understood in the skill-, rule-, and knowledge-based performance framework. Implications to alarm and auditory designs are discussed. Specifically, design of alarm devices should be guided by the principle of information provision regardless of whether an alarm may be true or false indication of “alarming” events.

A Cognitive Modeling Approach to Decision Support Tool Design for Anesthesia Provider Crisis Management

Prior research has revealed existing operating room (OR) patient monitors to provide limited support for prompt and accurate decision making by anesthesia providers during crises. Decision support tools (DSTs) developed for this purpose typically alert the anesthesia provider to existence of a problem but do not recommend a treatment plan. There is a need for a human-centered approach to the design and development of a crisis management DST. A hierarchical task analysis was conducted to identify anesthesia provider procedures in detecting, diagnosing, and treating a critical incident and a cognitive task analysis to elicit goals, decisions, and information requirements. This information was coded in a computational cognitive model using GOMS (Goals, Operators, Methods, Selection rules) Language. An OR monitor interface was prototyped to present output from the cognitive model following ecological interface design principles. A preliminary assessment of the DST was performed with anesthesiology and usability experts. The anesthesiologists indicated they would use the tool in the perioperative environment and would recommend its use by junior anesthesia providers. Future research will focus on formal validation of the DST design approach and comparison of tool output to actual anesthesia provider decisions in real or simulated crises.

Predictive data mining on monitoring data from the intensive care unit

The widespread implementation of computerized medical files in intensive care units (ICUs) over recent years has made available large databases of clinical data for the purpose of developing clinical prediction models. The typical intensive care unit has several information sources from which data is electronically collected as time series of varying time resolutions. We present an overview of research questions studied in the ICU setting that have been addressed through the automatic analysis of these large databases. We focus on automatic learning methods, specifically data mining approaches for predictive modeling based on these time series of clinical data. On the one hand we examine short and medium term predictions, which have as ultimate goal the development of early warning or decision support systems. On the other hand we examine long term outcome prediction models and evaluate their performance with respect to established scoring systems based on static admission and demographic data.

Specificity improvement for network distributed physiologic alarms based on a simple deterministic reactive intelligent agent in the critical care environment

Automated physiologic alarms are available in most commercial physiologic monitors. However, due to the variability of data coming from the physiologic sensors describing the state of patients, false positive alarms frequently occur. Each alarm requires review and documentation, which consumes clinicians' time, may reduce patient safety through 'alert fatigue' and makes automated physician paging infeasible. To address these issues a computerized architecture based on simple reactive intelligent agent technology has been developed and implemented in a live critical care unit to facilitate the investigation of deterministic algorithms for the improvement of the sensitivity and specificity of physiologic alarms. The initial proposed algorithm uses a combination of median filters and production rules to make decisions about what alarms to generate. The alarms are used to classify the state of patients and alerts can be easily viewed and distributed using standard network, SQL database and Internet technologies. To evaluate the proposed algorithm, a 28 day study was conducted in the University of Michigan Medical Center's 14 bed Cardiothoracic Intensive Care Unit. Alarms generated by patient monitors, the intelligent agent and alerts documented on patient flow sheets were compared. Significant improvements in the specificity of the physiologic alarms based on systolic and mean blood pressure was found on average to be 99% and 88% respectively. Even through significant improvements were noted based on this algorithm much work still needs to be done to ensure the sensitivity of alarms and methods to handle spurious sensor data due to patient or sensor movement and other influences.

On line extraction of temporal episodes from ICU high-frequency data: a visual support for signal interpretation

This paper presents a method to extract on line temporal episodes from high-frequency physiological parameters monitored in ICU, as a visual support for signal interpretation. Temporal episodes are expressions such as: "systolic blood pressure is steady at 120 mmHg from time t(0) until time t(1); it increases from 120 to 160 mmHg from time t(1) to time t(2) ...". Three words are used to describe the data evolution: {steady, increasing, decreasing}. The method deals with noisy data and missing values. It uses a segmentation algorithm that was developed previously and a classification of the segments into temporal patterns. The results obtained on simulated data are quite satisfactory. They show that the method is able to detect rapid variations as well as slow trends. Episodes extracted from real S(p)o(2) data recorded over a period of 44 h from 10 different adult patients are analysed. The visual representation of the temporal episodes is a powerful tool to help the physicians analyse in a glance the evolution in time of the variables monitored. It can help carer personnel to make quicker decisions in alarm situations.

On-Line Segmentation Algorithm for Continuously Monitored Data in Intensive Care Units

An on-line segmentation algorithm is presented in this paper. It is developed to preprocess data describing the patient's state, sampled at high frequencies in intensive care units, with a further purpose of alarm filtering. The algorithm splits the signal monitored into line segments--continuous or discontinuous--of various lengths and determines on-line when a new segment must be calculated. The delay of detection of a new line segment depends on the importance of the change: the more important the change, the quicker the detection. The linear segments are a correct approximation of the structure of the signal. They emphasise steady-states, level changes and trends occurring on the data. The information returned by the algorithm, which is the time at which the segment begins, its ordinate and its slope, is sufficient to completely reconstruct the filtered signal. This makes the algorithm an interesting tool to provide a processed time history record of the monitored variable. It can also be used to extract on-line information on the signal, such as its trend, in the short or long term.

Towards knowledge-based systems in clinical practice: development of an integrated clinical information and knowledge management support system

Given that clinicians presented with identical clinical information will act in different ways, there is a need to introduce into routine clinical practice methods and tools to support the scientific homogeneity and accountability of healthcare decisions and actions. The benefits expected from such action include an overall reduction in cost, improved quality of care, patient and public opinion satisfaction. Computer-based medical data processing has yielded methods and tools for managing the task away from the hospital management level and closer to the desired disease and patient management level. To this end, advanced applications of information and disease process modelling technologies have already demonstrated an ability to significantly augment clinical decision making as a by-product. The wide-spread acceptance of evidence-based medicine as the basis of cost-conscious and concurrently quality-wise accountable clinical practice suffices as evidence supporting this claim. Electronic libraries are one-step towards an online status of this key health-care delivery quality control environment. Nonetheless, to date, the underlying information and knowledge management technologies have failed to be integrated into any form of pragmatic or marketable online and real-time clinical decision making tool. One of the main obstacles that needs to be overcome is the development of systems that treat both information and knowledge as clinical objects with same modelling requirements. This paper describes the development of such a system in the form of an intelligent clinical information management system: a system which at the most fundamental level of clinical decision support facilitates both the organised acquisition of clinical information and knowledge and provides a test-bed for the development and evaluation of knowledge-based decision support functions.

Temporal abstraction and inductive logic programming for arrhythmia recognition from electrocardiograms

This paper proposes a novel approach to cardiac arrhythmia recognition from electrocardiograms (ECGs). ECGs record the electrical activity of the heart and are used to diagnose many heart disorders. The numerical ECG is first temporally abstracted into series of time-stamped events. Temporal abstraction makes use of artificial neural networks to extract interesting waves and their features from the input signals. A temporal reasoner called a chronicle recogniser processes such series in order to discover temporal patterns called chronicles which can be related to cardiac arrhythmias. Generally, it is difficult to elicit an accurate set of chronicles from a doctor. Thus, we propose to learn automatically from symbolic ECG examples the chronicles discriminating the arrhythmias belonging to some specific subset. Since temporal relationships are of major importance, inductive logic programming (ILP) is the tool of choice as it enables first-order relational learning. The approach has been evaluated on real ECGs taken from the MIT-BIH database. The performance of the different modules as well as the efficiency of the whole system is presented. The results are rather good and demonstrate that integrating numerical techniques for low level perception and symbolic techniques for high level classification is very valuable.

Model-based interpretation of cardiac beats by evolutionary algorithms: signal and model interaction

This paper presents a new approach for cardiac beat interpretation, based on a direct integration between a model and observed ECG signals. Physiological knowledge is represented by means of a semi-quantitative model of the cardiac electrical activity. The interpretation of cardiac beats is formalized as an optimization problem, by minimizing an error function defined between the model's output and the observations. Evolutionary algorithms (EAs) are used as the search technique in order to obtain the set of model parameters reproducing at best the observed phenomena. Examples of model adaptation to three different kinds of cardiac beats are presented. Preliminary results show the potentiality of this approach to reproduce and explain complex pathological disorders and to better localize their origin.

Effective data validation of high-frequency data: time-point-, time-interval-, and trend-based methods

Real-time systems for monitoring and therapy planning, which receive their data from on-line monitoring equipment and computer-based patient records, require reliable data. Data validation has to utilize and combine a set of fast methods to detect, eliminate, and repair faulty data, which may lead to life-threatening conclusions. The strength of data validation results from the combination of numerical and knowledge-based methods applied to both continuously-assessed high-frequency data and discontinuously-assessed data. Dealing with high-frequency data, examining single measurements is not sufficient. It is essential to take into account the behavior of parameters over time. We present time-point-, time-interval-, and trend-based methods for validation and repair. These are complemented by time-independent methods for determining an overall reliability of measurements. The data validation benefits from the temporal data-abstraction process, which provides automatically derived qualitative values and patterns. The temporal abstraction is oriented on a context-sensitive and expectation-guided principle. Additional knowledge derived from domain experts forms an essential part for all of these methods. The methods are applied in the field of artificial ventilation of newborn infants. Examples from the real-time monitoring and therapy-planning system VIE-VENT illustrate the usefulness and effectiveness of the methods.

Design and validation of an intelligent patient monitoring and alarm system based on a fuzzy logic process model

The process of patient care performed by an anaesthesiologist during high invasive surgery requires fundamental knowledge of the physiologic processes and a long standing experience in patient management to cope with the inter-individual variability of the patients. Biomedical engineering research improves the patient monitoring task by providing technical devices to measure a large number of a patient's vital parameters. These measurements improve the safety of the patient during the surgical procedure, because pathological states can be recognised earlier, but may also lead to an increased cognitive load of the physician. In order to reduce cognitive strain and to support intra-operative monitoring for the anaesthesiologist an intelligent patient monitoring and alarm system has been proposed and implemented which evaluates a patient's haemodynamic state on the basis of a current vital parameter constellation with a knowledge-based approach. In this paper general design aspects and evaluation of the intelligent patient monitoring and alarm system in the operating theatre are described. The validation of the inference engine of the intelligent patient monitoring and alarm system was performed in two steps. Firstly, the knowledge base was validated with real patient data which was acquired online in the operating theatre. Secondly, a research prototype of the whole system was implemented in the operating theatre. In the first step, the anaesthetists were asked to enter a state variable evaluation before a drug application or any other intervention on the patient into a recording system. These state variable evaluations were compared to those generated by the intelligent alarm system on the same vital parameter constellations. Altogether 641 state variable evaluations were entered by six different physicians. In total, the sensitivity of alarm recognition is 99.3%, the specificity is 66% and the predictability is 45%. The second step was performed using a research prototype of the system in anaesthesiological routine. The evaluation of 684 events yielded a sensitivity, specificity and predictability of the alarm recognition of more than 99%.

The interpretation of time-varying data with DIAMON-1

Applying the methods of Artificial Intelligence to clinical monitoring requires some kind of signal-to-symbol conversion as a prior step. Subsequent processing of the derived symbolic information must also be sensitive to history and development, as the failure to address temporal relationships between findings invariably leads to inferior results. DIAMON-1, a framework for the design of diagnostic monitors, provides two methods for the interpretation of time-varying data: one for the detection of trends based on classes of courses, and one for the tracking of disease histories modelled through deterministic automata. Both methods make use of fuzzy set theory taking account of the elasticity of medical categories and allowing discrete disease models to mirror the patient's continuous progression through the stages of illness.