Mila Kwiatkowska

The Impact of a Telemedicine Monitoring System on Positive Airway Pressure Adherence in Patients with Obstructive Sleep Apnea: A Randomized Controlled Trial

STUDY OBJECTIVES First-line therapy for patients with moderate to severe obstructive sleep apnea (OSA) is positive airway pressure (PAP). Although PAP is a highly efficacious treatment, adherence to PAP is still a substantial clinical problem. The objective of this study was to determine whether PAP adherence can be improved with a telemedicine monitoring system. DESIGN A nonblinded, single-center, randomized controlled trial that compared standard PAP treatment versus PAP treatment and a telemedicine monitoring system SETTING University sleep disorders program in British Columbia, Canada PATIENTS Adult patients (≥ 19 yr of age) with moderate to severe OSA (apnea hypopnea index (AHI) ≥ 15 events/hr determined by polysomnography) prescribed PAP INTERVENTIONS: Patients were randomized to either standard care with an autotitrating PAP machine or an autotitrating PAP machine that transmitted physiologic information (i.e., adherence, air leak, residual AHI) daily to a website that could be reviewed. If problems were identified from information from the website, the patient was contacted by telephone as necessary. MEASUREMENTS PAP adherence after 3 mo, subjective sleep quality, and side effects RESULTS Seventy-five patients were enrolled; 39 were randomized to telemedicine and 36 to standard care. The mean age ± standard deviation (SD) was 53.5 ± 11.2 yr, mean AHI was 41.6 ± 22.1 events/hr, and 80% of patients were male. After 3 mo, mean PAP adherence was significantly greater in the telemedicine arm (191 min per day) versus the standard arm (105 min per day; mean difference = 87 min, 95% confidence interval (CI): 25-148 min, P = 0.006, unpaired t test). On days when PAP was used, mean adherence was 321 min in the telemedicine arm and 207 min in the standard arm (difference = 113 min, 95% CI: 62-164 min, P < 0.0001). Significant independent predictors of adherence included age, baseline Epworth Sleepiness Scale score, and use of telemedicine. On average, an additional 67 min of technician time was spent on patients in the telemedicine arm compared with the standard arm (P = 0.0001). CONCLUSIONS PAP adherence can be improved with the use of a web-based telemedicine system at the initiation of treatment.

Knowledge-Based Data Analysis: First Step Toward the Creation of Clinical Prediction Rules Using a New Typicality Measure

Clinical prediction rules play an important role in medical practice. They expedite diagnosis and limit unnecessary tests. However, the rule creation process is time consuming and expensive. With the current developments of efficient data mining algorithms and growing accessibility to medical data, the creation of clinical rules can be supported by automated rule induction from data. A data-driven method based on the reuse of previously collected medical records and clinical trial statistics is cost-effective; however, it requires well defined and intelligent methods for data analysis. This paper presents a new framework for knowledge representation for secondary data analysis and for generation of a new typicality measure, which integrates medical knowledge into statistical analysis. The framework is based on a semiotic approach for contextual knowledge and fuzzy logic for approximate knowledge. This semio-fuzzy framework has been applied to the analysis of predictors for the diagnosis of obstructive sleep apnea. This approach was tested on two clinical data sets. Medical knowledge was represented by a set of facts and fuzzy rules, and used to perform statistical analysis. Statistical methods provided several candidate outliers. Our new typicality measure identified those, which were medically significant, in the sense that the removal of those important outliers improved the descriptive model. This is a critical preprocessing step towards automated induction of predictive rules from data. These experimental results demonstrate that knowledge-based methods integrated with statistical approaches provide a practical framework to support the generation of clinical prediction rules.

Fuzzy logic and semiotic methods in modeling of medical concepts

The field of medicine is a quickly growing area of application for computer-based systems. However, the use of computerized methods in this knowledge-intensive and expert-based discipline brings multiple challenges. The major problem is the modeling, representing, and interpreting of diverse medical concepts. For example, some symptoms and their etiologies are described in terms of molecular biology and genetics, physiological processes are defined using models from chemistry and physics; yet mental disorders are defined in more subjective terms of feelings, behaviours, habits, and life events. Thus, the representation of medical concepts must be sufficiently expressive to model concepts which are inherently complex, context-dependent, evolving, and often imprecise. Furthermore, the representation must be formal or, at least, sufficiently rigorous in order to be processed by computers and at the same time, the representation must be human-readable in order to be validated by humans. In this paper, we describe the modeling process of medical concepts as a mapping from the real-world medical concepts into their computational models, and further into their physical implementation. First, we define the notion of a concept as a fundamental unit of knowledge and specify the fundamental principles of the computational representation of a concept. Second, we describe the characteristics of medical concepts, specifically their historical and cultural changeability, their social and cultural ambiguity, and their varied levels of precision. Third, we present a meta-modeling framework for computational representation of medical concepts. Our framework is based on fuzzy logic and semiotic methods which allow us to explicitly model two important characteristics of medical concepts: imprecision and context-dependency. We present the framework using an example of a mental disorder, specifically, the concept of clinical depression. To exemplify the changeable and evolutionary character of medical concepts, we discuss the development of the diagnostic criteria for depression. Finally, we use the example of the assessment of depression to describe the computational representation for polythetic and multi-dimensional concepts and for categorical and non-categorical concepts. We demonstrate how the proposed modeling framework utilizes (1) a fuzzy-logic approach to represent the non-categorical (continuous) nature of the symptoms and (2) a semiotic approach to represent the polythetic (contextual interpretation) and dimensional nature of the symptoms.

Application of Knowledge-Engineering Methods in Medical Knowledge Management

This paper deals with Knowledge Engineering (KE), Clinical Decision Support Systems (CDSS), and Expert Systems (ES) as essential methods and tools supporting the Knowledge Management (KM) process in medicine. Specifically, we focus on the main component of the CDSS, knowledge base (KB). We demonstrate a hybrid approach to the creation, modification, verification, and validation of KB, which combines a fuzzy rule system with data mining. We describe the design and implementation of KB for two CDSS systems. The first system, which supports the evaluation of clinical depression, uses a combination of three methods: (1) creation of fuzzy rules based on expert clinicians’ knowledge and standard guidelines, (2) construction of Artificial Neural Networks (ANN) based on patients’ data, and (3) implementation of a CAKE (Computer Aided Knowledge Engineering) tool. The second system, which supports the diagnosis of obstructive sleep apnea, uses a combination of two methods: (1) creation of fuzzy rules derived from the medical literature and the expert clinicians’ knowledge and (2) induction of decision trees from large clinical data sets. Based on these two clinical studies, we demonstrate that KE methods should be regarded as valuable methods and tools which can be successfully used in medical KM for the creation, validation, and maintenance of KB.

A Multi-factor Model for the Assessment of Depression Associated with Obstructive Sleep Apnea: A Fuzzy Logic Approach

Many patients with obstructive sleep apnea (OSA) also exhibit depressive symptoms such as fatigue, anhedonia, weight changes, and depressed or sad mood. Some of these patients are misdiagnosed with clinical depression and treated with antidepressants, which may actually impede OSA treatment. Thus, the assessment of depression is of crucial importance in sleep clinics, and is often used both before and after treatment. As there are no objective ways to measure depression, the most common form of assessment is using subjective, usually self-reporting, questionnaires. These questionnaires were created for assessing and diagnosing clinical depression and not for multiple assessments of depressive symptoms as a secondary medical condition. They are also subject to reporting inaccuracies. In this paper, we introduce STEM-D, a fuzzy logic model for assessing depression in OSA patients that incorporates the multifactorial nature of depression. We studied nine existing questionnaires and created four categories of questions. We modeled the categories using fuzzy variables, with the output variable being the severity of a patients depression. STEM-D will be used multiple times throughout treatment to monitor a patients change in depressive symptoms as a result of OSA treatment. This model will be applied in a clinical setting as part of a larger project, CPAP-T*MONITOR.

Can telemedicine improve CPAP adherence

Obstructive sleep apnoea (OSA) is a common, important disease associated with daytime sleepiness and a reduced quality of life. Moreover, untreated OSA is associated with increased healthcare utilisation and a significantly increased risk of motor vehicle crashes, hypertension, coronary artery disease and stroke.1 Treatment of patients with continuous positive airway pressure (CPAP) improves quality of life and daytime sleepiness. CPAP treatment may also substantially reduce rates of cardiac events and motor vehicle crashes.2 3 One of the major barriers to CPAP treatment is adherence. Many patients reject, discontinue or use it irregularly or suboptimally. The rate of CPAP discontinuation in the first 3 years of prescription ranges from 12% to 25%.4 When patients do not use CPAP sufficiently, clinical outcomes are compromised,5 demonstrating that optimising adherence is an important aspect of patient management. Adherence to treatment is influenced by many factors: biomedical (severity of the disorder, CPAP side effects, therapeutic response), psychological (patients self-efficacy, anxiety, claustrophobia, the patients perception of the seriousness of the disorder, belief that the treatment will be effective, patients understanding of the CPAP treatment process), social (family support), technological (simplicity of the procedure, selection of CPAP device, appropriate air pressure) and economic (cost and insurance coverage). Studies have demonstrated that adherence can be significantly improved by a comprehensive support programne and timely …

A Fuzzy Model for the Evaluation of Efficacy of Continuous Positive Airway Pressure (CPAP) Treatment

Obstructive sleep apnea hypopnea (OSAH) is a serious, chronic respiratory disorder afflicting approximately 2-4% of the general population. The standard treatment of OSA is the continuous positive airway pressure (CPAP). CPAP treatment is highly effective; however, it is not curative and its efficacy depends highly on patients life-long compliance. Modeling of the effectiveness of this therapy involves several interrelated and, often, subjective factors. This paper describes a model, called CPAP-VAL, for the evaluation of CPAP treatment based on three main factors: improvements in symptoms (nocturnal blood oxygen desaturation, excessive daytime sleepiness, hypertension, and depressive moods), CPAP treatment compliance (average hours of use, percentage of days used, and percentage of CPAP use per total hours of sleep), and patients characteristics (age, gender, and OSAH severity). The proposed model uses the fuzzy logic approach to combine subjective and objective measurements and to represent complex interrelationships between various factors. The CPAP-VAL model was designed as an evaluation component for a telehealth system, CPAP-T*MONITOR, which will support the treatment process for OSAH patients living in the rural areas.

Computational Representation of Medical Concepts: A Semiotic and Fuzzy Logic Approach

Medicine and biology are among the fastest growing application areas of computer-based systems. Nonetheless, the creation of a computerized support for the health systems presents manifold challenges. One of the major problems is the modeling and interpretation of heterogeneous concepts used in medicine. The medical concepts such as, for example, specific symptoms and their etiologies, are described using terms from diverse domains - some concepts are described in terms of molecular biology and genetics, some concepts use models from chemistry and physics; yet some, for example, mental disorders, are defined in terms of particular feelings, behaviours, habits, and life events. Moreover, the computational representation of medical concepts must be (1) formally or rigorously specified to be processed by a computer, (2) human-readable to be validated by humans, and (3) sufficiently expressive to model concepts which are inherently complex, multi-dimensional, goal-oriented, and, at the same time, evolving and often imprecise. In this chapter, we present a meta-modeling framework for computational representation of medical concepts. Our framework is based on semiotics and fuzzy logic to explicitly model two important characteristics of medical concepts: changeability and imprecision. Furthermore, the framework uses a multi-layered specification linking together three domains: medical, computational, and implementational. We describe the framework using an example of mental disorders, specifically, the concept of clinical depression. To exemplify the changeable character of medical concepts, we discuss the evolution of the diagnostic criteria for depression. We discuss the computational representation for polythetic and categorical concepts and for multi-dimensional and noncategorical concepts. We demonstrate how the proposed modeling framework utilizes (1) a fuzzy-logic approach to represent the non-categorical (continuous) nature of the symptoms and (2) a semiotic approach to represent the contextual interpretation and dimensional nature of the symptoms.

A semiotic approach to data in medical decision making

Computer-based support for medical decision making has been a subject of many research projects since the earliest days of computers. Although the early expert systems promised to automate medical diagnosis, very few systems were actually utilized in clinical settings. In the last twenty years, the intent to use computers to replace or simulate medical experts has changed to a less ambitious goal of supporting and assisting the medical decision-making process. Recently, the growing availability of electronically stored patient records has provided a new opportunity for the decision support systems to utilize the data mining tools. In all these types of decision support systems, data play a central role. This paper examines three fundamental issues surrounding data: modeling of data as representation of medical concepts, interpretation of data in multiple contexts, and utilization of data in the decision-making process. The paper introduces a semiotic approach to the analysis of the role of data in medical decision making. This approach assumes that the processes of data modeling, collection, storage, processing, and interpretation are components of a larger communication process — the medical decision-making process. Furthermore, the semiotic approach describes the medical decision-making process in a broader context of representation, interpretation, and meaning making in a social context.

A fuzzy logic approach to modeling physical activity levels of obstructive sleep apnea patients

This paper introduces a fuzzy logic approach to classifying the physical activity levels of obstructive sleep apnea (OSA) patients. Obesity is one of the risk factors of OSA, and the strong link between physical activity levels and obesity suggests that a method of assessing the physical activity levels of OSA patients is critical. We studied various objective and subjective instruments, and decided to base our fuzzy logic model on the International Physical Activity Questionnaire (IPAQ). The IPAQ model uses the duration, frequency, and intensity of various activities as inputs to classify the physical activity level of an individual. Our model will be used in a clinical setting at the Respiratory Clinic at Thompson Rivers University (TRU) to assess the physical activity level of new patients as well as patients undergoing continuous positive airway pressure (CPAP) treatment.