Elsevier

Computers in Biology and Medicine

Volume 44, 1 January 2014, Pages 27-36
Computers in Biology and Medicine

Designing an efficient electroencephalography system using database with embedded images management approach

https://doi.org/10.1016/j.compbiomed.2013.10.022Get rights and content

Abstract

Many diseases associated with mental deterioration among aged patients can be effectively treated using neurological treatments. Research shows that electroencephalography (EEG) can be used as an independent prognostic indicator of morbidity and mortality. Unfortunately, EEG data are typically inaccessible to modern software. It is therefore important to design a comprehensive approach to integrate EEG results into institutional medical systems. A customized EEG system utilizing a database management approach was designed to bridge the gap between the commercial EEG software and hospital data management platforms. Practical and useful medical findings are discoursed from statistical analysis of large amounts of EEG data.

Introduction

According to the Population Policy White Paper released by the Republic of China Department of the Ministry of the Interior, more than 14% of Taiwan's population will be 65 or older by the year 2018. Additionally, due to declining birth rates, the Department of Health (DOH) has issued a warning stating that the Taiwanese population has progressed from “aging” to “aged”, and will eventually become a “super-aged” society by 2015. This issue is not only limited to Taiwan but also to other Asian countries such as Japan and China, making senior healthcare an increasingly important concern across the region. Most neurological issues associated with aging can be detected and properly treated through neurological analysis, making electroencephalography (EEG) testing very common in hospitals. However, most hospitals have different types of medical equipment and systems procured from various companies, which cannot be integrated by each other. Proper integration of these various systems is needed to raise efficiency, availability rates and improve overall healthcare utilization. Even though approaches such as the Picture Archiving and Communication System (PACS) and the Health Information System (HIS) are now very common in hospitals, some important and common diagnoses, such as EEG and EKG/ECG, are not available through these medical systems.

Conventional EEG analysis measures and records the electrical activity of a patient's brain. Special sensors called electrodes are attached to the patient's head at multiple sites and connected by wires to a computer. Brain cells called neurons communicate by producing tiny electrical impulses. The EEG hardware records this electrical activity, stores the raw data in a computer hard-drive, and displays the information on a monitor or records it on a paper as a series of wavy lines. The data measured by the scalp EEG, also referred to as an EEG graph (electroencephalogram), are used for clinical and research purposes. Since the patterns on the electroencephalogram are indicative of known medical conditions (especially brain-related diseases), a physician can examine the recorded patterns of electrical activity and detect abnormalities. The human brain is an important and complicated organ that controls the central nervous system. It is responsible for many important functions such as seeing, hearing, tasting, smelling, and sense of balance. The human brain contains more than 100 billion neurons, each linked to as many as 10,000 other neurons. Thus, the brain and the rest of the nervous system comprise a vital network for the human body. Taking EEG measurements is now a common and essential medical procedure in most hospitals for detecting brain-related diseases, such as seizure disorders, epilepsy, newborn monitoring, brain injury, brain tumor, hemorrhage, cerebral infarct, encephalitis, attention deficit, syncope, senile dementia, Alzheimer's disease, hallucinations, and brain tumors. As such, it needs to be properly integrated with other medical systems to enhance healthcare quality and efficiency.

Many nations are dealing with the impact of a recent combination of declining birthrates and increased life expectancy. This not only impacts economic growth, but also increases governmental healthcare spending, giving rise to many healthcare-related issues as the populations in these societies age. The research outlined in this paper originates from a mid-sized town in Taiwan with a population of over 86,000 residents. Because of its geographical advantages and comfortable weather, the community has actively recruited senior emigration. The local hospital has a great deal of experience in handling senior-related healthcare issues ranging from senile dementia Alzheimer's disease, dizziness, vertigo, headaches, and metabolic encephalopathy. Most of these issues can be detected and properly managed with neurological diagnosis and treatment. Medical research has shown that EEG can be used as an independent indicator of mortality and can be used to assess healthcare utilization [1]. In his research, Stecker concluded that “…the costs associated with caring for a patient with an abnormal EEG were roughly 3 times that of a patient with a normal EEG. The risk of death was 3.7 times higher in patients with an abnormal EEG than in patients with a normal EEG [1]”. This finding brought to light the beneficial use of EEG in senior healthcare. Even though EEG is not designed only for the senior healthcare, the lowest birthrate among the countries worldwide has made Taiwan face an accelerated aging population and declining fertility. According to the Department of the Ministry of the Interior, the crude birthrate was 25.64% in 1971 and currently is 8.7% in 2012, which means that by 2016, the elderly population will first outnumber the juvenile population. Currently, every 7.2 young adults are responsible for the care of one senior citizen, but this number will drop to 2 to 1 by 2027. This aging situation discovered in our clinical experience is even worse. The database in our system shows the average age of the neurological patients is 54 (see Section 4 below), and many of them came to the hospital in wheelchairs. These senior patients generally suffered from different neurological diseases and could not express themselves well. It is believed that this situation will get worse as the birthrate continues to decline. This important fact cannot be neglected. It is understandable that the EEG is very versatile and is not limited to elderly. However, because of the austere growing aging population in Taiwan, the government authorities sense the crisis and put more efforts to link the EEG to senior healthcare. Thus, making EEG testing efficient and effective is particularly important, motivating the development of an integrated EEG system for efficient diagnoses of healthcare needs. Since the most fundamental and crucial means of neurological therapy is the EEG machine, attention has focused on efficient management and extensions of EEG data.

Current EEG analysis presents several difficulties that need to be overcome in order to improve effectiveness. The first issue is storage and retrieval. Digital EEG results are typically stored on the hard drive of a dedicated personal computer and are stored and formatted more like the file-system manager. The literature has pointed out this has great limitations for efficient data management. In addition, regular EEG examinations result in data files from 40 MB to 100 MB (for sleep disorders or unconscious patients). Such large files are generally difficult for physicians to examine directly on a standard monitor. Transferring such large data files over the Internet is also time-consuming. Physicians have to examine the EEG data by visualizing the entire footage to identify abnormalities. Depending on the size of the raw data, it could take considerable time to complete the entire results. Many approaches have been explored to overcome this problem. For example, several notable engineering algorithms, such as signal processing and neural networks, have been utilized to analyze EEG signals in an attempt to create automatic abnormality detection and classification [2], [3], [4], [5], [6], [8], [9]. However, accuracy and validity have arisen as serious concerns since these approaches have trouble distinguishing “false alarms” from true abnormalities, and unexpected “data noise” can affect results, e.g., if a patient blinks or moves his eyeballs or head, electrical signals known as ocular artifacts are produced [7], [9], [10]. Furthermore, EEG wave patterns vary due to the complexity of the brain itself. The results produced from these algorithms necessitate careful verification from neurologists, making the use of these approaches impractical.

The second problem with most current EEG systems is a lack of integration with most existing hospital computer databases, e.g., PACS and HIS systems. Formatting differences between EEG results and hospital information management systems prohibit further processing by existing HIS/PACS systems. To the best of our knowledge, no PACS systems are capable of graphically displaying EEG results. Most EEG systems exist as isolated systems and are unable to communicate with other medical databases. This leads to a third problem, namely, the lack of any interface between EEG and other systems, which results in inconvenience and inefficiency in EEG diagnosis reports. This is why most EEG diagnosis reports take the form of text descriptions without graphics to show any abnormalities. These limitations have made the patience referral or follow-up checking difficult since the physicians cannot visualize the EEG data in any graphic form from the reports.

The fourth problem of current EEG testing is that inefficient operation limits the number of examinations a given hospital can provide. For current EEG procedures, physicians must wait until a patient who is undergoing an EEG test completes the testing before a diagnosis report can be prepared. Since the PC storing the examination results stays in the same room as the EEG machines, physicians try to avoid any additional interruptions while the machine is in use. The physician or technician must wait until EEG examination is complete before they can begin diagnosis using the results. This limitation has inhibited the promotion of efficient healthcare solutions.

Although all of the problems described above emphasize the importance and potential impact of using EEG databases, and although research has been carried out trying to overcome the current problems [11], [12], [13], [14], several major hurdles remain. The development of a comprehensive EEG system to bridge the gaps in the current usage of the hardware and the results provided from EEG machines is necessary.

Section snippets

Method and implementation

A 32-channel Nicolet Alliance Works system was used for the initial 24 h of video/EEG monitoring, metabolic encephalopathy, senile dementia and other common EEG tests. The examination results were recorded in digital format and stored on the hard drive of the personal computer included with the EEG unit. There are three major problems with the use of this EEG system, especially in countries where English is not the predominant language: first, although the hard drive is replaceable, the

System design life cycle

The systems design life cycle is a conceptual process used in project management that describes the stages involved in an information systems development, from an initial feasibility study through maintenance of the completed application. It follows the same pattern as most new inventions: finding a problem with motivation, planning and evaluation, analyzing and designing the solution, implementing the solution and maintenance. Many successful project developments have followed the system

Results

The EEG database system described in this study has undergone extensive testing and modification since its introduction in late November 2005, and its final version has been available online since July 2006. Even though this in-house EEG system has been released for years, it is still serving two regional hospitals as it was designed following the procedures of the system design life cycle. Furthermore, in order to have accurate statistical analysis, our in-house system requires a longer time

Discussion

As mentioned earlier, this in-house EEG system allows physicians to inspect patients' EEG abnormality patterns (saved as images in the PACS) by different querying options (such as the diagnosis). This function helps neurologists and researchers to oversee and observe the EEG image patterns under the same category (or disease). When the database accumulates large amounts of data, this will help to disclose accurate and vital medical judgments. For examples, the database recorded the most common

Conclusion

A growing number of governments have focused attention on healthcare issues to ensure public safety and enhance the lives of their citizens. Currently the healthcare industry is one of the largest industries in many developing countries [20]. Integrating different medical systems to provide efficient diagnoses with proper medications for healthcare is therefore a crucial issue. Research also shows that new and different medical systems are introduced to hospitals when new technologies are

Conflict of interest statement

None declared. All authors declare that there are no conflicts of interest in relation to this manuscript.

References (26)

Cited by (0)

1

Tel.: +886 492910960x4673; fax: +886 492915205

View full text