Conventional telemedicine has limitations due to the existing time constraints in ... article, the authors present a solution for 'virtual telemedicine' to cope with the ...
International Journal of Information Technology and Web Engineering, 5(1), 43-55, January-March 2010
43
Virtual Telemedicine using Natural Language Processing Imran Sarwar Bajwa, The Islamia University of Bahawalpur, Pakistan
Abstract Conventional telemedicine has limitations due to the existing time constraints in the response of a medical specialist. One major reason is that telemedicine based medical facilities are subject to the availability of a medical expert and telecommunication facilities. On the other hand, communication using telecommunication is only possible on fixed and appointed time. Typically, the field of telemedicine exists in both medical and telecommunication areas to provide medical facilities over a long distance, especially in remote areas. In this article, the authors present a solution for ‘virtual telemedicine’ to cope with the problem of the long time constraints in conventional telemedicine. Virtual Telemedicine is the use of telemedicine with the methods of artificial intelligence. Keywords: Telemedicine, Telecommunication for health, Information retrieval, Text Processing, Expert system
1. INTRODUCTION Telecommunication is the most used technology all over the world in current age and still establishing a long way. This technology has made things to do in an easy and fast manner. Now enhancements in technology have made our thoughts to drag fields of life into advance technology. From last few years, alphabet „e‟ is being used with almost everything i.e. e-mail, elearning, e-commerce, e-banking and e-services. The proposal of „e-health‟ is still new and asks for more development. Medical is the field that is emerging continuously to make health facilities more affective and facilitating.
Telemedicine [1] is the need of current age to provide health facilities in the remote areas where medical experts, doctors and physicians are not available. Telemedicine uses telecommunication technology to provide medical treatment and services. Telemedicine connects patients with doctors where distance is a critical factor and exchanges the information of diagnosis, treatment and other health care activities. Telemedicine becomes more significant if the patient is far away from the medical experts and faces transportation challenges. On the other hand it is helpful way of getting medical treatment at home. Health care facilities can be improved for a specific community: children,
DOI: 10.4018/jitwe.2010010103 Copyright © 2010, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.
44 International Journal of Information Technology and Web Engineering, 5(1), 43-55, January-March 2010
old people, plague disease, etc. Telemedicine can become moiré effective in emergency cases and areas of natural disasters. Still, this is cost effective and efficient way of providing high level and skilled medical facilities to the people living in remote areas [2], who can easily access the physicians and medical specialist.
1.1 Conventional Telemedicine Telemedicine typically works in two ways [3]: store and forward method and real time method.
Store and forward method gathers patient‟s medical information locally and then patient query is emailed to a physician. Afterwards, physician prescribes a treatment and then emails the response of the medical query in 24 to 48 hours. On the other hand, in real time telemedicine, video conferencing and live data transmission methods are involved for communication between patient and medical expert. UCD Health system [4] is one of the examples of video conference based health systems.
Figure-1.1: UCD Health system – Patient side [4]
Figure-1.2: UCD Health system – Physician side [4]
Copyright © 2010, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.
International Journal of Information Technology and Web Engineering, 5(1), 43-55, January-March 2010
In store and forward method is quite approving solution but it requires lot of time to get diagnostic results in return. Time constraint can be up to 24 to 48 hours. In real time telemedicine, there are so many constraints that make its effective usability difficult. While in countries like Pakistan [1] where video conferencing is a pricey client, real mode is not appropriate solution. Secondly, high bandwidth is required for data transmission. On the other hand, the availability of the medical expert is also required, when the patients need. Virtual telemedicine is the process to provide the telemedicine features online using a virtual physician in place of the real doctor. Other famous telemedicine types are home telemedicine and individual telemedicine [3].
1.2. Virtual Telemedicine As we have discussed in the previous section, that store and forward method is reasonably practicable but the time constraint of store this method is not realistic. As some times due to the serious condition of the patient, he/she may not wait for up to 48 hours [11]. Some intelligent mechanism is required to improve the usability and affectivity of the conventional telemedicine process. An intelligent system is required that may provide immediate response. In conventional telemedicine, an additional component is proposed in this research: virtual physician. Virtual physician is a web-based application that answers without delay the medical queries. To make this facility more comprehensive, an additional functionality of consultation is also involved. In this facility, if the knowledgebase of the virtual physician cannot answer a medical query an automatic email is sent to a medical expert and the response of the query is updated in the knowledgebase for future queries. In this article, the section 2 presents the review of related work done be the various researchers in the field of telemedicine and its applications in different areas of healthcare. Section 3 highlights the architecture of the designed medical expert system and the NLP based
algorithm that process the textual information. Section 4 describes the implementation details and the section 5 presents a case study to elaborate the use of the designed system and the results of the performed experiments with the analysis are also provided in later half of the same section.
2. LITERATURE REVIEW Field of telemedicine is being proved the technology of the electronic age. Although the telemedicine was first time used in 1959 but major development work was initiated in this field for the last 8 to 10 years. Telemedicine has been used for the e-health solution of diseases: diabetes, cardiac, trauma, and general physician related diseases. P. Douglas [11] was one of the earlier researchers who realized the importance and need of the telemedicine based medical facilities. His study elaborated the use of store and forward method of medical information transformation. His work also emphasizes the need of efficient use of the resources to make the telemedicine based health care system more effective and useful. Albert and Jason conducted two preliminary studies [5] in year 2007 to examine the performance of remote display protocol (RDP) used in telemedicine systems. In first study, RDP was deployed in a wide-area network [6] and in second one, the performance of RDP was analyzed over Wi-Fi [7]. They also presented a thin client based home telemedicine architecture that was providing remote training for patients on broadband. Dena S. [9] discussed uses and benefits of telemedicine typically for rural areas in America. She presented that considerable technical, organizational, and financial obstacles have kept the rural communities deprived of benefits of the technology. This paper focuses on these issues and suggests a feasible solution for establishing successful rural telemedicine programs. DIABTel [10] Telemedicine Service is another telemedicine based system that provides daily care to diabetic patients. Major concern of the research was to provide
Copyright © 2010, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.
45
46 International Journal of Information Technology and Web Engineering, 5(1), 43-55, January-March 2010
telemonitoring of patient's blood glucose data and also support remote care from doctors to diabetic patients. Tayab D. [2] proposed a cost effective and multipurpose model of the telemedicine system. The proposed system had two major parts: a telemedicine unit for the patient side and another base unit for medical expert side. Major issue of discussion was the use of high-speed network forms for interconnectivity of the complete system. Dena Puskin, Barbara and Stuart presented a framework [8] of a telehealth system that was able to identify and understand the interaction between telemedicine services. Exploration of health information technology [8] (HIT) applications on local, regional and national levels was the major emphasis of this research. UC Davis used a telepharmacy program in UCD Health System [4] that was based on a video conferencing. The author cites many challenges to telemedicine in the recent times i.e. system expertise, imprecise administration, contractual organization, etc. Tele-echocardiology [12] is another field of major research in telemedicine. This field of research deals with the real time diagnosis if heart diseases without the support of in-house pediatric cardiologists. The major emphasis of the research was to evaluate the impact of the telemedicine in providing the health care facilities to the cardiac patient in community hospitals where cardiac specialists are not available frequently. In the recent times where wireless technologies are grasping their roots in other fields of life, at the same time telemedicine is also getting benefits of it. An advanced wireless sensor network (WSN) [] for health monitoring is introduced by G. Virone in DCS, University of Virginia. The research presents a proposal „smart healthcare‟ with the benefits of low cost and ad-hoc deployment of model sensors of for
an improved quality of health care. A. Diver [17] has recently introduced his work to emphasize the significance of image analysis as an additional support for assure the modern telemedicine needs. A pilot study based on twenty patients of trauma has been presented to highlight the limited plastic surgery experience of a doctor in the serious cases. Some outcomes of the work are introduction of user-friendly technology, clinically appropriate telemedicine applications, well trained and professional telemedicine users, etc.
3. USED METHODOLOGY Virtual telemedicine is replacing the physician in telemedicine with a virtual physician. Telemedicine is designed for remote and rural areas [12] whereas virtual telemedicine can be used in both rural and urban areas. In conventional telemedicine, there are simply two nodes: patient and doctor. Patient communicates with the doctor through some telecommunication medium; telephone, e-mail, internet, video-conferencing, etc. A simple representation of a conventional telemedicine system has been shown in figure 3.1. Major issues that are concerned with the development of a conventional telemedicine system can be divided into four categories [13]. First of all there is need of infrastructure that is based on hardware, software and connectivity mechanism of multiple nodes (patient and doctor). On the other hand basic medical equipment is required at the patient end where a literate person can transmit patient‟s information to the medical expert. Still there are important issues like accurate information exchange, security, transmission bandwidth, protocols, data sets etc.
Copyright © 2010, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.
International Journal of Information Technology and Web Engineering, 5(1), 43-55, January-March 2010
47
Figure - 3.1: A simple telemedicine system
Figure-3.2: A virtual telemedicine system The time constraint of conventional telemedicine system is typically longer. An idea of virtusal telemedicine has been presented to cover up this time constraint and make telemedicine more effective and efficient. Virtual telemedicine is the extension of conventional. A new component „medical expert system‟ has been deployed in the conventional
telemedicine system. This medical expert system is a natural language processing based expert system. In this research this expert system has been named „Virtual Medical Expert System‟.
3.1. Designed System Architecture
Copyright © 2010, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.
48 International Journal of Information Technology and Web Engineering, 5(1), 43-55, January-March 2010
This virtual medical expert based system is shown in figure 3.2. This system has robust ability of reading the patient‟s symptoms and immediately diagnosing the disease and also prescribing the appropriate medication for the patient. A natural language processing (NLP) based medical expert system is the base of the proposed health care system. The designed rule based expert system has following major components [19]. a- Graphical User Interface b- Medical Expert Knowledge base c- Medical Inference engine d- Medical Explanation Module
A graphical user interface is a facility for the user to interact with the Expert system. A wizard of forms is used to get textual input from the user and then after processing the textual information the output is shown to the user in the form of reports. b- Medical Expert Knowledge base MEKB is an intelligent knowledge base that uses Markov Logics (ML) to save domain knowledge. Markov Logic is simple extension to first-order logic. In Markov Logic, each formula has an additional weight fixed with it [5], in variation of first order logic.
a. Graphical User Interface Input Text (Patient‟s Symptom Report)
Morphological
Tokenization
Analysis
Semantic Analysis
Pragmatic Analysis
Medical Inference Engine
Diet Details
Lexical Analysis
MEKB
Medication
Dose Details
POS Tagging
Exercise Details
Side Effects
Figure 3.3 - Working of the designed system In ML, a formula's associated weight reflects the log probability and it also satisfies the formula. strength of a constraint. The higher weight of a . Use of Markov Logic enables intelligent storage formula represents the greater the difference in and retrieval of information using logical Copyright © 2010, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.
International Journal of Information Technology and Web Engineering, 5(1), 43-55, January-March 2010
connectives and quantifiers. The benefit of using Markov Logics is that the queries which even do match up to 80% will also be answered as this is not the case in typical knowledgebase that used production rules. This approach will increase the response rate of the knowledge base and makes it more effective and efficient MEKB handles two types of rules Factual knowledge and Heuristic Knowledge [19].
Factual Knowledge is the descriptive information. It is basic knowledge related to the domain i.e. Bacterial causes flue” Or “Dust allergy causes cough”.
Heuristic Knowledge is typically observed or pragmatic knowledge. This type of knowledge is extracted from the factual knowledge i.e. “if patient has temperature then it can be chest infection”.
c. Medical Expert System This is another very important part of the designed medical expert system. It is the brain of the medical expert system. The major duty if this part is to make logical deductions based upon the extracted knowledge from the medical expert knowledge base (MEKB). This inference engine not only makes decision but also extracts new information on the behalf of provided information from MEKB. This new information can also become par of the medical inference engine, if required. d. Medical Explanation Module This is another very important module of the designed system. This module provides the facility of explaining and reasoning of the system to the user. User can make different queries regarding the system domain and system.
3.2. Algorithm for Query Processing For diagnosis and treatment of the patient, two techniques are used in the proposed system. First and major technique to develop virtual telemedicine is “Rule Based Approach” in
which is the most efficient way to represent human activity in the form of rules. Used algorithm has two major parts. First part has been designed to read the patient‟s symptoms of diseases and analyze according to the given knowledge base and diagnose the accurate disease. Second part of the designed algorithm prescribes the suitable medicine to the patient. Following steps are followed by the algorithm to diagnose a particular disease: Step –I Health care person collects the patient‟s disease information along with the symptoms of disease and records in the simple English form. Step –II The patient‟s case information in the textual form is given to the designed virtual telemedicine system. Step – III Natural language processing is performed to read the given text and extract the related information. Used NLP steps to analyze text are [15]: o Tokenization (Separating tokens): The input sentence is tokenized into complete words o Morphology (to identify and analyze morphemes). The input of the previous step is further processed to identify the complete words and then identify their parts of speech (POS) category. o Lexical analysis (to identify grammatical types of the tokens): The POS tagged words are further processed to identify their particular role in the sentence and grammatical rules also assist this type of analysis. o Semantic analysis (To understand the meanings of the sentences): Different constituents of a sentence are analyzed here to extract the both implicit and explicit meanings of the input text. o Pragmatic analysis (to find out meanings in a particular context): This is an additional step that is used if the meanings of the input text are not clear it is analyzed into its particular context to make the things more clear and concise.
Copyright © 2010, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.
49
50 International Journal of Information Technology and Web Engineering, 5(1), 43-55, January-March 2010
Step –IV Pattern matching of the extracted information is performed through the medical inference engine with the information in medical expert knowledge base to find out patient‟s disease. Step –V If a match is found then treatment of the disease is recommended with the appropriate medication and health instructions. If match is not found then the query is forwarded to the medical expert. In a case, if the virtual medical expert does not find any particular solution of the patient‟s query from its knowledge base, rather related to the disease diagnosing or prescribing medicine, an automatic email is sent to the medical expert. Medical expert examines the query with available facts and makes some decisions and replies the local medical assistant. The designed system also updates the Medical Expert Knowledge Base (MEKB) so that if the same query comes in future, it may be resolved locally.
4. IMPLEMENTATION DETAILS Rural areas of Pakistan [1] are relatively backward in terms of technology. A number of challenges are to face in setting up a system for virtual telemedicine. Some of the major challenges are following:
A complete infrastructure is required to actually set up the proposed virtual telemedicine framework. A satellite based wireless internet work or WiFi system supporting speed of 1.0 Gbps or above is required. 3G cellular technology is also getting very popular these days [16] in the field of telehealth. This technology can help out in fast video sharing, video male and video conferencing. On the other side, a telemedicine center at the remote area needs basic eequipments [12] i.e.
Virtual telemedicine software Camera (s), lights, projector Digital X-Ray System UPS system Computer hardware, system and application software and accessories
5. EXPERIMENTS AND RESULTS A number of experiments were performed to test the designed health care system. A medical assistant was involved to use the system. A multiple step procedure is involved to use the designed medical health care system. The steps are following: 1. Patient Registration 2. Patient Record File Generation 3. Processing User Details 4. Generating Patient Report
Budget and financial constraints are more significant [14]. First of all expensive medical equipments are required at the telemedicine centers. High bandwidth for communication is also an expensive solution.
Brief description of all these phases with the help of a case study has been provided in the later part of the section.
At the site, adequate human resources are required [15] i.e. technicians to implement the proposed virtual Telemedicine system, a medical assistant having medical training to perform basic tests of the patients and some health workers having basic literacy of computer and capable of using computers.
A patient is needed to register with his personal details i.e. name, age, sex, address, family history, previous cases, etc for using the proposed virtual telemedicine system. Figure 4.1 shows the form that is used to register the patient first.
5.1. Patient Registration
Copyright © 2010, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.
International Journal of Information Technology and Web Engineering, 5(1), 43-55, January-March 2010
51
Medical assistant can also use digital stethoscope and electrocardiograph file with ECG recorder or images with the examination camera [13]. A text file containing the patient‟s case details is prepared.
5.3. Processing User Details The input text file containing the patient‟s history and symptoms is given to the designed system foe processing. In first step, input is read and tokenized e.g. the output of a sentence “The patient has high fever.” is [The] [patient] [has] [high] [fever] [.] Figure 5.1 – Form to register a patient
5.2. Patient Record File Generation After registration, the medical expert performs basic tests of a patient to get the reading of temperature, blood pressure, blood group, sugar level and ESG (if required). Then he records the common symptoms of the patient in the system. Besides these tests, the data i.e. color of tongue, color of eyes, heart beat, face color, etc is also captured and is updated in the system. The data form is shown in the figure 4.2
After tokenizing the text, morphological analysis is performed of given text to define the structuring and transformation of the words. POS Tagging is also performed to identify different parts of speech e.g. [The]
[patient]
Det.
Noun
[has]
[high]
Verb adjective
[fever] [.] Noun
After POS tagging, the text is lexically and syntactically analyzed and a parse tree is generated for semantic analysis. Figure 5.3 shows the generated parse tree of the above example.
S NP
VP NP
Det.
Noun
Verb
The
patient
H.V.
Adej
Noun
has
high
fever
Figure 1.0- Parse tree generated for the example Figure 5.2 – Form to update patient‟s status
There are two rationales for performing the syntactic analysis; to validate the phrases and sentence according to grammatical rules defined
Copyright © 2010, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.
52 International Journal of Information Technology and Web Engineering, 5(1), 43-55, January-March 2010
by the English language and finding out the semantical constituents of natural language. Moreover, the semantical analysis helps in identifying the main parts of a sentence i.e. object, subject, actions, attributes, etc. [The] [patient] Subject
[has] Verb
[high]
[fever]
[.]
Object
In this step, associations are identified by doing semantic analysis. It is determined in this specified that which actions have been performed by which object and a set of attributes belong to which object e.g. in the above example it is extracted that a person is having a high fever.
5.4. Generating Patient Report Afterwards, the extracted information of the last phase is matched with the knowledge in MEKB. Inference engine extracts the desired information and processes the patient‟s symptoms to infer the disease. If disease is found then the respective medication of the disease and additional information i.e. diet and exercise details are also provided. The designed system will not only provide the treatment strategy of the disease but also recommends tests if necessary for the confirmation of the disease. If the tests were recommended by the system to the patient, patient/administrator will have to provide the results of the tests to the system so that system may recommend the right treatment of the disease. If the system is not able to answer the patient then an automatic e-mail will be forwarded to the medical expert. The medical expert will carefully examine the case by consulting all the test reports and data sent by the local medical assistant and diagnosis the disease and also prescribes the appropriate medication. When the medical assistant receives the response, the medical expert‟s opinion is also updated in the knowledge-base of the system.
base cannot reply then the patient‟s data will be emailed to expert. The correctness of the decision made by the software and the medical expert is based on the accuracy of the data captured by the medical assistant. The quality and accurateness of the images and video of the patient is also quite important. To validate the precision and affectivity of the designed system symptom reports of three groups of ten patients were defined. For each group three reports i.e. easy, average an difficult were generated for each group. The symptom reports were carefully prepared and processed for each patient using the designed health care system. For correct and wrong diagnosis of a symptom report various points were given. Table 5.1 shows the details of the results. Group Group Group Total 1 2 3
%
Easy
10/10
9/10
9/10
2.8
93.33
Average
9/10
8/10
9/10
2.6
86.66
Difficult
7/10
8/10
8/10
2.3
76.66
Average Accuracy: 85.5% Table 5.1 Virtual Telemedicine Based Healthcare System Following are some benefits over using the proposed framework virtual telemedicine.
Improved and immediate access the specialty care Upgraded emergency medical services Reduction in un-necessary duplication of services Less dependency on the medical expert Easier diagnostic consultation Expanded disease cure education More patient health queries Remote medical consultation Reduction in health care cost Automated patient record keeping
A prescription will be generated after the patient‟s data is submitted. If the knowledge Copyright © 2010, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.
International Journal of Information Technology and Web Engineering, 5(1), 43-55, January-March 2010
6. CONCLUSION & FUTURE WORK Virtual Telemedicine is the new concept which actually works faster than that of the traditional telemedicine systems. An expert system has been deployed in place of a medical expert that has ability to immediate respond. This immediate response can help to treat patients in time and more effectively. 90% queries can be entertained locally. The accuracy achieved with the designed system is 85.5%. The Virtual expert system becomes more robust and intelligent with the passage of time as the knowledge-base grows and the level accuracy will also improve. For the under developed and developing countries like Pakistan, Bangladesh, Sri Lanka etc the usability of the Virtual telemedicine will be more useful and beneficial. The experiments were performed on a simulator and it is acceptable that these results may vary when the system will be run real time. In future enhancements the algorithms is needed to be improved to increase the accuracy level of the system. Medical explanation module is also needed to enhance its usability.
References [1] M. Z. Khalid, A. Akbar, A. Kumar , A. Tariq, M. Farooq, [2008] “Using Telemedicine as an Enabler for Antenatal Care in Pakistan”, Proc. 2nd International Conference: E-Medical System, Oct 2008, Tunisia, pp 1-8 [2] Tayab Din Memon, BS Chowdhry, AK Baloch, “Design and Implementation of a Telecardiologic System”, MUET, Research Journal, Volume 23, No. 4, Oct 2004. [3] William R., David H., Susan M., Tracy L., Kathryn Pyle, Mark Helfand, [2006] “Telemedicine for the Medicare Population: Update”, AHRQ publication No. 06-E007 [4] Thomas S. Nesbitt, [2007] “Meeting the Health Care Needs of California’s Children: The Role of Telemedicine”, Digital Opportunity for Youth Issue Brief, Number 3: September 2007
[5] Albert M. Lai, Jason Nieh, Justin B., [2007] “REPETE2: A Next Generation Home Telemedicine Architecture”, AMIA 2007 Symposium Proceedings, pp 1020-1022 [6] Lai AM and Nieh J., [2006] “On the Performance of Wide-Area Thin-Client Computing”, ACM Transaction on Computer Systems. May 2006, p. 215-209 [7]. Lai AM and Nieh J., [2005] “Web Content Delivery Using Thin-Client Computing”, In: Chanson ST, Xu TJ, editors. Web Content Delivery. Springer; 2005. pp. 325-346 [8] Dena Puskin, Barbara and Stuart, [2006] “Telemedicine, Telehealth, and Health Information Technology”, An ATA Issue Paper, The American Telemedicine Association, May 2006 [9] Dena S. Puskin, [1995] “Opportunities and challenges to telemedicine in rural America”, Journal of Medical Systems, Volume 19, Number 1 / February, 1995, pp 59-67 [10] E. J. Gomez, F. Del Pozo, M. Hernando, “Telemedicine for diabetes care: The DIABTel approach towards diabetes telecare”, Informatics for Health and Social Care, Volume 21, Issue 4 October 1996 , pages 283 – 295 [11] D A Perednia and A. Allen, “Telemedicine Technology and Clinical Applications”, The Journal of the American Medical Association, 273(6), 8 Feb 1995, pp. 483–88. [12] C A Sable, et al., “Impact of Telemedicine on the Practice of Pediatric Cardiology in Community Hospitals”, Pediatrics, January 2002, Vol. 109, No. 1 pp. e3. [13] Rashid E, Ishtiaq O, Gilani S, et al. “Comparison of store and forward method of teledermatology with face-to-face consultation” J Ayub Med Coll Abbottabad 2003 Apr-Jun;15 (2):34-6. [14] J. Jiehui, Z. Jing, [2007] „Remote patient monitoring system for China.‟ IEEE Potentials, Vol. 26, Issue 3, pp 26-29, IEEE, May/June 2007. [15] Bajwa I., Choudhary I. [2006] “A Rule Based Paradigm for Speech Language Context
Copyright © 2010, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.
53
54 International Journal of Information Technology and Web Engineering, 5(1), 43-55, January-March 2010
Understanding”, J Donghua University (English Edition) Jun 2006, 23(6), pp.39-42
Applications Subproject-4, Special Report”, Telemedicine Journal and e-Health, Vol 8, No 2, 2002.
[16] Hersh W, Helfand M, Wallace J, [2002] “A systematic review of the efficacy of telemedicine for making diagnostic and management decisions”, J Telemed Telecare 8(4):197-209. [17] Andrew J Diver, Harry Lewis, Derek J Gordon, (2009), “Telemedicine and Trauma Referrals – a Plastic Surgery Pilot Project”, Ulster Med J 2009 78(2), pp. 113-114 [18] Demartines N, Otto U, Mutter D, Labler L, van Weymarn A Vix M, et al. “An evaluation of telemedicine in surgery: tele-diagnosis compared with direct diagnosis” Arch Surg 2000 135(7), pp. 849-53. [19] Shahbaz F., Maqbool F., Razzaq S., Irfan K., Zia T., [2008] “The Role of Medical Expert Systems in Pakistan”, World Academy of Science, Engineering and Technology 2008 Vol. 37 pp. 296298 [20] Whitten PS, Mair FS, Haycox A, May CR, Williams TL, Hellmich S. et al. “Systematic review of cost effectiveness studies of telemedicine interventions” BMJ 2002; 324(7351) pp. 14341437. [21] Yellowlees P. “Successful development of telemedicine systems – seven core principles” J Telemed Telecare 1997; 3(4), pp. 215-222. [22] M P Cutchin, “Virtual Medical Geographies: Conceptualizing Telemedicine and Regionalization”, Progress in Human Geography, 26, 1, 2002, pp. 19–39 [23] G Zahlman and S Laxminarayan, “Special Issue on Telemedical Systems, Guest Editorial”, IEEE Transactions on Information Technology in Biomedicine, Vol 3, Number 2, June 1999. [24]L Kun (1998), “Biomedical Information Technology: Opportunities for the Future”, Proceedings ITAB „98, Washington, DC, (Ed S Laxminarayan and E M Tzanakou), IEEE Press. [25] A Lacroix, et al., “International Concerted Action on Collaboration in Telemedicine: Recommendations of the G-8 Global Healthcare Copyright © 2010, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.