Role of Clinical Decision Support Systems (CDSS) in Reducing the Operating Cost and Medical Errors Paper

Role of Clinical Decision Support Systems (CDSS) in Reducing the Operating Cost and Medical Errors in Health Organization

Abstract

An overview of current technological innovations and applications in Artificial Intelligence in Medicine (AIM) and computer-based clinical diagnosis support systems (CDSS) is presented based on relevant literature. Great strides in AIM and CDSS were apparent in developed countries especially with the increased capacity of computers for storage and processing. Wireless technology has also contributed greatly to the progress in both AIM and CDSS.

The Great Ormond Street Hospital (GOSH) was used as a historical case study for the judicious application of innovation and technology in improving patient care and medical learning. A questionnaire distributed to GOSH medical and administrative staff on the efficacy of AIM and CDSS in the hospital returned an overwhelmingly positive response. Role of Clinical Decision Support Systems (CDSS) in Reducing the Operating Cost and Medical Errors Paper

The health situation of Bangladesh and the level to which AIM and CDSS is utilized in the health centres was investigated. A visit to various hospitals revealed that most of them use standard diagnostic equipment such as MRI, digital X-rays and CT scans but the extent of computer use is for administrative functions.

https://healtheappointments.com/operating-room-essays/

Studies on the application of nanotechnology for diagnostic and therapeutic applications are being explored are  identified as the direction of combined AIM and CDSS technology innovation. One of the challenges is the development of a biodegradable medium for the technology.

It is concluded that while AIM and CDSS is progressing greatly in breadth of applications in the developed countries, the same cannot be said for developing countries such as Bangladesh. The dearth of the most basic of health services will probably push back any real application of existing CDSS and AIM technology in Bangladesh until the barriers to effective government and non-government organisation collaboration are removed.

I. Introduction

Healthcare is a field to which computer technology has been a boon. There have been great strides in technology in all aspects of medicine encompassing treatment, communications, research and medication.

With the advent of faster, more compact, and cost-effective computers with ever-increasing storage capacities, the development of huge databases of clinical histories of patients for both clinical and research use has become extremely accessible.  This has made it easier for researchers to clearly identify the true nature of specific diseases, the efficacy of drugs and their adverse effects as well as therapeutic and preventive procedures. Clinicians now have access to significant data that could help in finding similarities between previous and current cases that could make patient care better.

However, databases are by nature merely bits of data in one area, and without organization and ease of access, it would be very difficult to make sense or use of it. It is necessary to characterize each piece of information in relation to other relevant data and make generalizations. (Szolovits 1982) Moreover, the greatest problem in healthcare the world over is escalating costs of professional expertise, treatment, medication and healthcare delivery.

It is the purpose of this paper to provide an overview of the recent developments in artificial intelligence and computer-based clinical diagnostic support systems in developed countries, the impact on cost and medical errors, an overview of the case of Great Ormond Street Hospital as an illustration of the use of AIM and CDSS in a hospital setting, and the extent of use of these technologies in Bangladesh as confirmed by an ocular visit.

II. Literature Review

A. Artificial Intelligence in Medicine (AIM)

Winston defines artificial intelligence as: “…the study of ideas which enable computers to do the things that make people seem intelligent … The central goals of Artificial Intelligence are to make computers more useful and to understand the principles which make intelligence possible.” in his textbook “Artificial Intelligence.” Artificial Intelligence in Medicine (AIM) does not require that machines are able to “think” per se but to use the knowledge base of human experts to accurately and methodically standardize clinical practice. Role of Clinical Decision Support Systems (CDSS) in Reducing the Operating Cost and Medical Errors Paper.

According to Clancey and Shortcliffe, AIM is “…primarily concerned with the construction of AI programs that perform diagnosis and make therapy recommendations. Unlike medical applications based on other programming methods, such as purely statistical and probabilistic methods, medical AI programs are based on symbolic models of disease entities and their relationship to patient factors and clinical manifestations.” (Coiera 2003)

There has been criticism that AIM is set to replace medical and health professionals in the delivery of patient care. AIM, however, is limited to perform very specific functions, much like an expert who specializes in a certain field and dispenses with much of general knowledge. The medical expert relies on distilled information from the general practitioner and laboratory tests and can presume that the data pointed in the direction of the specific disorder. There is normally no need to consider other possibilities but merely to confirm or reject the initial findings based on their specialized knowledge. An expert does not even need to come in direct contact with the patient. (Szolovits 1982)

The need for human interaction is also an aspect of AIM that makes no real attempt to take over, despite some innovations such as the Remote Presence Robots and Domo, the housekeeping robot. AIM is a highly sophisticated, complex tool that appears to take over where humans fall short, but it remains merely a tool nonetheless.

One application of computers in medicine is the use of the simplest decision-making tool, flowcharts. Flowcharting makes a recording of the sequence of patient-clinician dialogue, procedures and laboratory analyses, diagnoses and therapy and outcome that subsequently results. It is the basis for many triage or emergency protocols, patient interviewing programs and therapeutic guidelines in the acid-base area. The main problem with flowcharting is its tendency to grow too large for ease-of-use when the number of variables grows. It also neglects to make any associations with different pieces of data and fails to follow any logical characterization of the database. (Szolovits 1982)

In combination, three techniques could be modified to develop AI techniques that could be used as a clinical tool. One technique is the use of expertise and common sense. It is more difficult than it sounds to encode the way humans think or access information. Many clinical cases are a product of events and experiences outside recorded clinical data. It takes a clinician’s “common sense” or the association of ideas based on past experiences and of the patient to arrive at an accurate diagnosis of the problem. Such is the well-known figure of the family doctor, who knows his patients so well that diagnosis is not much of a problem in most cases.

Some programmers have tried to incorporate such human experience and train of thought into a program, such as the Present Illness Program (PIP) which makes use of a specific train of inferential reasoning. However, it would be unrealistic to expect a program to provide for each and every circumstance with any accuracy. (Szolovits 1982) Many of the first AIM failed to develop beyond the research phase, mainly because of lack of support from clinicians. (Coiera 2003)

AI has problem-solving capabilities that rely on data and reasoning programs. One example of a reasoning program is the General Problem Solver (GPS) which established deductive pathways to break down the problem into manageable chunks. Role of Clinical Decision Support Systems (CDSS) in Reducing the Operating Cost and Medical Errors Paper. Other examples of reasoning programs are PLANNER, CONNIVER and Truth Maintenance System (TMS). (Szolovits 1982)

But AI systems are far from static inasmuch as processing information is concerned. If programmed correctly, AI systems can simulate a human capacity for learning, and since it can handle and keep track of large amounts of information, that capacity could lead to significant clinical discoveries and new knowledge through detection of certain patterns in the data. (Coiera 2003)

The first of the AIM programs include: CASNET system of Rutgers University, a diagnostic and therapeutic program for eye diseases including glaucoma; EXPERT, a system used for thyroid disorders and in rheumatology; MYCIN system of Stanford University used for the diagnosis and treatment of bacterial infections of the blood and other infectious disease; INTERNIST system of the University of Pittsburgh used for diagnosis in general internal medicine; diagnosis and treatment for acid/base and electrolyte imbalances called ABEL; and a program of MIT for physician use of digitalis.

AIM should not be confused with computer-based business applications in medicine such as the keeping of administrative and financial records. Such activities, while important from a business point of view, has very little impact on health management or the problems of decision-making in the clinical setting. (Szolovits 1982)

However, AIM has become more focused on a broader spectrum of medicine that providing computer support for routine clinical situations and are more appropriately referred to as clinical decision support systems (CDSS). AI-based systems can accomplish various functions. One popular scenario is where AI resides in an independent being such as robots and performs human tasks. Shades of Battlestar Galactica notwithstanding, such scenarios are unnecessarily unctuous. (Coiera 2003)

B. Clinical Decision Support Systems (CDSS)

Computer-based Clinical Decision Support System (CDSS), sometimes referred to as expert systems,  is defined as program or software which is designed to assist in clinical decision making for specific patients by providing an objective checklist of prognostic information that may be relevant to the case and provide suggested preventive, therapeutic and maintenance strategies. (Prescription in Ischaemic Stroke Management (PRISM) Study Group 2002) CDSS are intended primarily as an assistive tool for tasks that requires the manipulation and analysis of raw data such as patient records, to which a contraindication could be detected for a planned treatment that would otherwise be overlooked. It may also detect variations in previous and current patient data such as blood levels that could have significance to treatment. (Coiera 2003)

The Centre for Health Evidence’s Dr. Robert Hayward describes CDSS functionally as “…systems [that] link health observations with health knowledge to influence health choices by clinicians for improved health care”.  It includes a medical database, mechanisms derived from evidence-based medicine and linked by medical logic modules in terms that trigger certain processes depending on the input such as patient variables. (Clinical decision support system 2007)  Role of Clinical Decision Support Systems (CDSS) in Reducing the Operating Cost and Medical Errors Paper.

Initially, CDSS used the Arden syntax in conjunction with HL7 standards and medical logic modules, which contains the one medical decision depending on the trigger of the data input into the system.  Recently, CDSSs are being designed using artificial neural networks based on Bayesian statistical theory able to improve outcomes for individual patients. The network is capable of presenting various possible diagnoses complete with probability estimates and propose best clinical practice methodologies. (Farukhi 2006)

The most common of CDSS technology in everyday use is knowledge-based or expert systems which contain clinical knowledge on a very specific task which is processed down to reasonable conclusion from a set of rules. There are several applications of this knowledge base. These include:

1.      Alerts and reminders – these are systems attached to a machine such as an ECG that monitors a patient’s condition and warns of variations in the condition. It can also process laboratory results and send feedback either via e-mail or on-screen. Clinicians can be notified to execute important tasks on or before it becomes necessary.
2.      Diagnostic assistance – expert systems can assist in making a diagnosis when the case is complex or the clinician has little or no experience in the situation.
3.      Therapy analysis and planning – CDSS can scan for inconsistencies and errors in a prescribed treatment plan as well as in physician order entry. It eliminates errors due to overlooked data. The system can also be used to actually form a treatment plan, but requires a more extensive knowledge base and structure of accepted treatment protocols and guidelines.
4.      Prescription of medications – CDSS or specifically prescribing decision support systems (PDSS) are usually used to process pharmacokinetics, dosage appropriateness and contraindications such as allergies of a patient. PDSS have pre-set routines and may be set up to transmit prescriptions to a pharmacy electronically.
5.      Information retrieval – CDSS may also be used to justify a certain diagnosis or treatment plan by providing ease of data retrieval and information filter from a database or the Internet.
6.      Image recognition and interpretation – expert systems can be used to interpret clinical images as complex as an angiogram, which is an important tool in mass reading, catching abnormal images that may escape human detection. (Coiera 2003)
To organize data in order to structure medical decision making problems, decision theory is used. It assumes that existing states can be evaluated and provide a normative, rational theory of decision making. (Szolovits 1982) As part of machine learning systems, CDSS can be programmed to analyze one or more drugs that is weakly efficacious and suggest ways to chemically improve it, leading to the discovery of new drugs or drug combinations. Aside from developing better drugs which may lead to shorter treatment periods, the system reduces the research costs of pharmacological development, indirectly reducing healthcare costs. (Coiera 2003)

Expert systems are also most valuable in clinical laboratories, where the system interprets test results as the pathologist peruses the hard copy. Examples of such systems are:

1        The PUFF system, which interprets pulmonary function tests, was produced in 1977 by the Pacific Presbyterian Medical Centre in San Francisco which is still in use today.

2        Pathology Expert Interpretative Reporting System (PEIRS) which reports on thyroid function tests, arterial blood gases, urine and plasma catecholamines, human chorionic gonadotrophin, glucose tolerance tests, cortisol, gastrin, cholinesterase phenotypes and parathyroid hormone related peptide. It has a diagnostic accuracy of 95% with a capacity of 100 reports daily (Coiera 2003)

The establishment of standard clinical protocols is dependent on clinical trials, which in turn generate evidence-based guidelines that provide some leeway for individual differences. In order for clinical trials to be run effectively, there must be a consistency in clinical care to identify elements that are unnecessary or even harmful. There are instances that standard clinical practice appears to run counter to patient-specific variations but this perception is largely based on subjective clinical judgment. In theory, most established protocols allow for small interventions that have little or no effect on outcome. Using an explicit set of specific instructions should render clinician judgment extraneous provided the patient data is accurate, at the same time establishing the standard for such patient-specific therapy. (Morris 2000) However, in practice, patient treatment is often dependent on clinician decisions which may not comply sufficiently with protocols for the most effective treatment possible.

The evaluation of CDSS with regards to their role in improving healthcare has been mixed but over all, it has been found to improve patient safety through fewer medication and treatment errors, quality of care by making more efficient use of resources such as clinician time and clinical pathways and guidelines and efficiency in health case delivery through reductions in test duplication, faster order processing and cost-effective drug prescribing. (Coiera 2003)

Walton, Dovey, Harvey and Freemantle executed a review of eleven studies on the effects of CDSS which used mathematical models of the pharmacokinetics of the specific drug within the body, the complexities of which depended on the degree of involvement in the body. The effect off CDSS on process and outcome of care was varied. Seven studies exhibited a significant effect on drug dosage, in which two showed a proportional increase of patients whose drug concentrations in the body fell within therapeutic range. There was also better physiological control in six of the studies, i.e. increased effectiveness of maintaining postoperative blood pressure with a computer assisted pump. Role of Clinical Decision Support Systems (CDSS) in Reducing the Operating Cost and Medical Errors Paper.Undesirable side effects of treatment such as achievement of toxic drug concentrations was reduced in six of the studies while in the two studies that included cost-benefit analysis, there was lower mean direct cost of computer-assisted treatment ($7,102 versus $ 13,758), largely attributable to shorter hospital stays.  In measuring medical care outcome, five of the six studies that included this measure showed significant improvement for asthma, infection treatment and postoperative pain management. Overall, patients under computer-support achieved therapeutic drug levels in less time (effect size 0.69 and -0.44, respectively) although there was no significant effect on total amount of drug used (-0.43) or on the achieved and targeted physiological parameter levels (-1.22). (1999)

Between 1966 and 1996, 18 studies regarding the effectiveness of computer-aided drug dosage programs in calculating the most appropriate doses in a hospital setting was reviewed. A meta-analysis of the data indicate that there is an overall benefit to using computers because it is more accurate in adjusting to individual patients than manual determination by clinicians. Clinicians also exhibit more confidence in using higher doses when the case requires it because there is a smaller risk of inadvertent overdosing. Patients were subjected to shorter treatment periods and had fewer side effects. (Walton, Dovey, Harvey and Freemantle 1999)

It is strongly suggested by the survey conducted by Johnston, Langton, Haynes and Mathieu that CDSS can improve a clinician’s performance. The survey further established that preventive care interventions such as reminders from computerized medical records for influenza or tetanus vaccination significantly improved the success statistics of a clinician. (1994)

Aside from the benefits to research, the high compliance rates to explicit computerized protocols may be attributed to the pursuit of excellence and efficiency in patient care, pressure to comply with technological advancements, prestige of association with research centres and/or economic benefits from patients enrolling in the program. However, many clinicians are reluctant to subscribe to changes in clinical practice because they perceive it to be a negative reflection on human decision-making that the machines are “taking over.” Some have unrealistic expectations of the computerized protocol, that it should be perfect and provide for all possible, no matter how improbable, scenarios. There is also an overweening pride found in many physicians who believe they should be the final authority in medical care, and since practitioners are often excluded from the protocol development process, this may be understandable. (Morris 2000) However, it must be remembered that the protocols themselves are base on clinical trials supervised by clinicians, and that they merely standardize the practice to minimize errors in judgment. It is a tool that clinicians use to make treatment more efficient but only upon the crucial first step of the clinical diagnosis.

Barriers to CDSS use are mostly functional. Data collection is an arduous and time-consuming process, and it is only in limited cases that the costs of documentation can be justified. Some CDSS are also poorly designed, making it inconvenient to use. Regional health information organizations (RHIO) are a good source of health outcome information because it is a coalition of healthcare stakeholders who share information. CareSpark is an example of an RHIO who made significant progress in improving patient care with the use of CareEngine, an ActiveHealth Management CDSS. It brought together clinical data with best practices and constantly updates itself. (Lamont 2007) There are some circumstances, however, where standardization may be inappropriate, such as when the clinical problem is infrequent. There is also a tendency to “cookbook” treatment, stifling innovation because it is easier and may be connected to reimbursement guidelines. (Morris 2000)

There was also an initial resistance to CDSS because it involves changes in the traditional process of care and the perception that it may not be worthwhile to use. However, this reluctance has largely been put to rest, and many systems are currently in clinical use, especially PDSS. Role of Clinical Decision Support Systems (CDSS) in Reducing the Operating Cost and Medical Errors Paper.

The task of determining the proper therapeutic drug dosages is not a simple task. It requires a significant knowledge of pharmacokinetics and a good working acquaintance with math. Drug interaction is subject to a host of variables that make it more than likely that errors of judgment will occur because the complexity of the task is made more difficult with the knowledge that there is only a thin line between a therapeutic and toxic dose. The tendency of most clinicians is to work with a safe margin of error which unfortunately frequently results in the under dosage of their patients. (Walton, Dovey, Harvey and Freemantle 1999)

Computers are especially useful for processes with predetermined sets of outcomes; meaning, a set of variables can be computed to result in a predictable set of values. For example, a man who is 54 years old, height 6’2” and weighing 70 kilos with diabetes mellitus would need x amount of a drug. The variables above are the most general but the number of variables will not matter in the least so long as the computer is programmed to process it. Moreover, errors are minimal so long as the parameters are within a valid range. (Walton, Dovey, Harvey and Freemantle 1999)

Humans, on the other hand, are apt to do two things: forget or become confused after the seventh level of instruction; and become careless or inattentive doing repetitive tasks, increasing the chances of error. While there is no doubt that the human clinician is invaluable because interpersonal relations is necessary for diagnosis, the value of an objective, calculating tool to determine dosage and other clinical processes should not be discounted. (Walton, Dovey, Harvey and Freemantle 1999)

Clinical decision-making is often rendered ineffective when the variables that need to be considered exceed the capacity of the human mind to process accurately. The clinical environment is unfortunately saturated with large numbers of variables and leads to clinical error, inconsistent clinical practice and non-compliance with standard guidelines. Morris suggests computer-based decision support system that combines evidence-based guidelines and individual patient variations in the delivery of recommended treatment. There are two point-of-care computerized protocols that employ the explicit method that take into account individual differences in the delivery of patient care developed at LDS Hospital in Salt Lake City, Utah for mechanical ventilation of patients with acute respiratory distress syndrome (ARDS) and bedside clinical decisions for intravenous fluid and hemodynamic support.

ORDER A PLAGIARISM-FREE PAPER NOW

The study involved 250 patients with ARDS for more than 100,000 hours. In all, the computers generated 38,546 to which clinicians only overruled in less than 1% and barotraumas was less in the group received computer-supported clinical care. Physician compliance was higher for the explicit method based systems at 94% as compared to CDSS protocols for antibiotic and diabetes guidelines at 60% (2000)

Walton, Dovey, Harvey and Freemantle observed that clinicians who relied on manual determination of drug dosage and drug administration were overly cautious, using lower initial and maintenance doses than those who had computer support. (1999) The effect of this caution is to achieve lower than therapeutic levels in patients and longer treatment periods. It is suggested that CDSS where computers were used to directly administer drugs under clinician supervision were the most effective, especially in the administration of anaesthetic agents.

The Affinity Health System by Wellinx of  Menasha, Wis. is a CDSS that integrates the decision support software with an electronic prescription module. In a study published in the Annals of Family Medicine (September/October 2004), the prescribing behaviour of 19 physicians who used the system showed that the average cost for new prescriptions decreased by $4.16, significant for antidepressants, asthma drugs, hypertension and headache medication. Role of Clinical Decision Support Systems (CDSS) in Reducing the Operating Cost and Medical Errors Paper. The system allows clinicians to select a diagnosis that has prewritten prescriptions that are supported by evidence-based therapeutic reviews, cost comparison, new studies and safety warnings.  (Study: e-prescribing, decision support systems can lower drug costs 2004)

C. Role of CDSS in reducing the operating cost of a health organization

CDSS is touted as a way to provide high-quality, cost-effective health care and the most common tools used are information sharing and physician education, Hawthorne effect or outcomes accountability, utilization review/case management/disease management, ancillary resource management, supply management/formulary and inventory control/operational efficiency, guidelines, protocols, and policies and procedures, standard orders, critical elements, critical pathways and interactive computer assists, benchmarking/identification of best-performance profiles; and continuous quality improvement. (Rosenstein 1999)

Expenditures in health care in the United States are estimated to reach 15.9% of the 2010 national GDP. In the creation of a clinical information system that would maximize the implementation of technological innovations for medical use, it needs to consider clinical task support, clinical management control, competition support, and clinical decision support. (Farukhi 2006)

A reduction in diagnostic costs was observed when a CDSS reminder system was used because it eliminated errors such as redundant tests patient (Johnston, Langton, Haynes and Mathieu 1994) The prevention of adverse drug reactions and inappropriate medication dosing is the most widely-documented application of CDSS. It also helps to establish the standards for quality of care. (Trowbridge and Weingarten 2001)

CDSS is especially useful in patient specific data recall, the increased efficiency of care resulting in subsequent reducing reduction in costs for diagnostic as well as hospital time. It modifies clinician behaviour and the process and quality of care. (Trowbridge and Weingarten 2001)

AIM is also used in disseminating medical expertise in an inexpensive medium to areas where such expertise would otherwise be unavailable, such as developing countries. Such collaboration of medical experts would otherwise be impracticable. (Szolovits 1982) One good example is the use and modification of existing technologies by Great Ormond Street Hospital in the UK, and tertiary healthcare institution that is also the foremost teaching hospitals in Europe for the treatment and care of children.

Health providers usually focus on cost savings as an incentive to engage in decision-support activities, usually assessed through their impact on length of stay (LOS), which translates to the cost savings from shortened LOS. As an example a typical hospital has an average variable cost per day of $300 and with 5,000 admissions annually, reducing the LOS from 3.5 to 3 days means a reduction in total inpatient days by 2,500. (Rosenstein 1999)

D. AIM and CDSS in the reduction of medical errors

There are estimates from the National Academy of Sciences’ Institute of Medicine that medical mistakes account for 98,000 deaths a year, more than that attributed to AIDS or car accidents. (Farukhi 2006)

A good example where CDSS can be life-saving is when diseases are not endemic to an area makes its appearance and medical staff cannot easily recognize it. The Asian bird flu which came out of Hong Kong in 1997 has spread across Europe and Asia, and if it makes its way to the US, the VisualDx of Logical Images in Rochester, N.Y. which includes an acute pulmonary infections program that provides access to almost 13,000 photos which can help doctors distinguish between an infection the bird flu virus H5N1 as well as 40 other respiratory diseases. It is in use in 350 hospitals in the US and other countries. It is also available online from West Pennsylvania Hospital’s website. The rapid diagnosis of such an event could mean the difference between one case and an epidemic. Role of Clinical Decision Support Systems (CDSS) in Reducing the Operating Cost and Medical Errors Paper.

Screen shot of VisualDx

Source: http://asp.usatoday.com/_common/_scripts/big_picture.aspx?width=490&height=352&storyURL=/tech/news/techinnovations/2006-08-07-visualdx-bird-flu_x.htm&imageURL=http://images.usatoday.com/tech/_photos/2006/08/08/software-large.jpg

Currently, the features of VisualDx has been expanded, to include data on bioterrorism, chemical and radiation injuries, oral diseases and diseases in newborns, eye diseases, geriatric illnesses and environmental and occupational disease. VisualDx uses visual recognition coded for easy retrieval. Licensing varies from US $10,000 to $45,000 annually. (Manning 2006)

The development of a standard electronic medical record (EMR) would greatly reduce medical errors and enable extensive information sharing among medical and health institutions as well as patients. The Systemized Nomenclature of Medicine and Arden Syntax (SNOMED) were developed to enable the compilation and sharing of clinical information across different platforms and across sites. (Farukhi 2006)

One CDSS is the Essentris OnWatch system which monitors patients for respiratory compromise, hemodynamic instability, fluid status imbalance, actionable lab results, and potential infection and adverse drug events in real time and may warn clinicians automatically. It enables medical staff to identify at-risk patients, access crucial information, and produce retrospective analytic reports to evaluate the effectiveness of the process. (Clinical decision support tool set 2004)

Meridian Health of Neptune, N.J. embedded evidence-based practice guidelines into its Siemens’ INVISION(R) computerized physician order entry (CPOE) system that improved quality of care delivered to their patients.  INVISION is used by many medical staff and is particularly interactive because it has the capacity to check physician’s orders against national best practice protocols.  (Meridian Health receives national recognition for innovative use of siemens computerized physician order entry 2005)
Medical software provider Artificial Medical Intelligence (www.artificialmed.com) has launched Emscribe Dx, a document scanning automated coding solution that identifies medical phrases and tags them for appropriate diagnostic and procedure codes. It is able to generate the Electronic Health Record (EHR) through data extraction from existing documents, eliminating the need to encode the data into the existing IT infrastructure and allows it to synchronize data from other systems and present as well as process records in one window. (Artificial Medical Intelligence announces latest edition of EMscribe Dx for computer assisted coding solution of icd9 diagnostic and procedure codes 2005)

Isabel is a CDSS that was developed because Isabel Healthcare’s co-founder Jason Maude’s daughter Isabel nearly died because of a misdiagnosis of infections secondary to chicken pox. Led by Dr. Joseph Britto of St. Mary’s Hospital in London, Isabel was designed to manage the biomedical knowledge that would assist in leading to a correct initial diagnosis and increase patient safety. Its database includes the description of more than 10,000 diseases and 4,000 drugs supported by regularly updated medical textbooks, journals and other articles. Many US hospitals currently use Isabel because it quickly provides likely diagnoses. (Lamont 2007)

NIST prototype hip replacement “phantom”

(Source: http://www.rxpgnews.com/Hip/

NIST_measuring_device_aims_to_up_

hip_operation_success_25719.shtml)
Institute of Standards and Technology (NIST) is working with the American Academy of Orthopaedic Surgeons (AAOS) to improve Computer Assisted Orthopaedic Surgery (CAOS) tracking instruments used by surgeons to plan hip replacement surgery. The “phantom,” a facsimile of the artificial socket, ball and femur substitutes that surgeons use to replace the joint and bone in hip operations and monitored by measuring instruments, is used to test the accuracy of CAOS devices prior to actual surgery, reducing the need to redo the operation because of poor positioning of the implants due to imprecise measurements.Role of Clinical Decision Support Systems (CDSS) in Reducing the Operating Cost and Medical Errors Paper.  It has yet to be approved by the Federal Drug Administration but once it is, clinical trials would well follow and researchers anticipate the extension of its application in surgical procedures on the knee and shoulder as well. (NIST measuring device aims to up hip operation success 2007)

E. Latest applications of CDSS and AIM in developed countries and its benefits

CDSS range from the simple to the most complex. They are used primarily to facilitate diagnosis, such as Dxplain which provides the clinician with a detailed differential diagnosis based on the patient’s information. Others may include antibiotic management and drug dosage calculators that help reduce medical errors and increase the quality of patient care and safety. A CDSS is distinguished from practice guidelines and critical pathways because it is patient-specific. It is also most effective when in tandem with computerized order entry and electronic medical records. (Trowbridge and Weingarten 2001)

1. Clinical Diagnosis systems

LDS Hospital is an example of an institution that builds on its success. Over twenty years, its clinical-epidemiology unit has worked closely with the clinical and medical staff as it designs, implements and review computer-based CDSS to diagnose infections, administer care systems, identify and prevent adverse reactions to drugs, and improve the use of anti-infective agents. The HELP system, which started operation in 1980, integrates the decision support function and the hospital information system (HIS) including admission, discharges and order entry. Such integration helps in adapting more closely to clinical working processes. (Coiera 2003) After each review, observed flaws and strengths were taken into account to improve on the design or function of the CDSS to further benefit patients. It is widely accepted today and many hospitals use CDSS and order-entry programs as a matter of course. (Garibaldi 1998)

SYSTEM
DESCRIPTION
ACUTE CARE SYSTEMS

(Dugas et al. 2002),
Decision support in hepatic surgery
POEMS (Sawar et al., 1992)
Post-operative care decision support
VIE-PNN (Miksch et al., 1993)
Parenteral nutrition planning for neonatal ICU
NéoGanesh (Dojat et al., 1996)
ICU ventilator management
SETH (Darmoni, 1993)
Clinical toxicology advisor
LABORATORY SYSTEMS

GERMWATCHER (Kahn et al.,1993)
Analysis of nosocomial infections
HEPAXPERT I, II (Adlassnig et al., 1991)
Interprets tests for hepatitis A and B
Acid-base expert system (Pince, et al., 1990)
Interpretation of acid-base disorders
MICROBIOLOGY/PHARMACY (Morrell et al., 1993)
Monitors renal active antibiotic dosing
PEIRS    (Edwards et al., 1993)
Chemical pathology expert system
PUFF (Snow et al., 1988)
Interprets pulmonary function tests
Pro.M.D.- CSF Diagnostics (Trendelenburg, 1994)
Interpretation of CSF findings
EDUCATIONAL SYSTEMS

DXPLAIN (Barnett et al., 1987)
Internal medicine expert system
ILLIAD (Warner et al., 1988)
Internal medicine expert system
HELP (Kuperman et al., 1991)
Knowledge-based hospital information system
QUALITY ASSURANCE AND ADMINISTRATION

COLORADO MEDICAID UTILIZATION REVIEW SYSTEM
Quality review of drug prescribing practices
MANAGED SECOND SURGICAL OPINION SYSTEM
Aetna Life and Casualty assessor system
MEDICAL IMAGING

PERFEX (Ezquerra et al., 1992)
Interprets cardiac SPECT data
(Lindahl et al. 1999).
classification of scintigrams
Table 1: A wide variety of expert systems have been placed into routine clinical use. These systems are typical examples. (Source: http://www.coiera.com/aimd.htm) Role of Clinical Decision Support Systems (CDSS) in Reducing the Operating Cost and Medical Errors Paper.

DXplain is a CDSS from the Laboratory of Computer Science at the Massachusetts General Hospital first developed in 1986, and the stand-alone version became available in 1996. The web-based version replaced all other methods of distribution since the Internet became widely used. It combines the function of medical textbook and referencing system. It can describe more than 2000 diseases complete with aetiology, signs, symptoms, pathology and prognosis. It has an interactive format for data input and, it can then generate a list of diagnoses and the justification for each. (Dxplain)

Professor Mahdi Mahfouf in the University of Sheffield’s Department of Automatic Control and Systems Engineering is developing a system that will simulate a doctor’s decision making process in determining treatment for patients in intensive care. It processes all possible interactions between drugs and a specific patient within seconds. It may be overridden by the doctor, and the system absorbs the doctor’s input to be used for future cases that present in a similar way. (Artificial intelligence to help intensive care doctors 2007)

The Map of Medicine is software that provides “decision-trees” for the use of general practitioners for disease management based on hospital practice guidelines and established clinical pathways. The brainchild of UCL Medicine Professor Owen Epstein, the tool enhances the efficiency of the general practice, minimizing unnecessary referrals. Developed through Medic-to-Medic, which is now owned by Informa, the software tool complies with the requirements of the UK’s National Health Service program “Connecting for Health.” (Map of Medicine adopted across NHS 2007) The algorithms are color-coded to indicate primary and secondary care and comprise of more than 340 clinical pathways of the major specialties. (From UCLB Business Online (www.uclb.com) »Breakthrough clinical pathway software, developed by UCL and the Royal Free Hospital, is now commercialised by Informa 2007)

2. Nanotechnology

Lemelson-MIT Prize winner Ray Kurzweil, inventor of the flatbed scanner and the first text-to-speech reaading machine states that humans can achieve near immortality with computers and biotechnology. He predicts that nanoechnology will reach its optimum application in the 2020s when nanobots, microscopic computer-based robots  capable of travelling in the bloodstream, will extend life by destroying pathogens, repairing DNA and reversing the aging process. (Spring 2004)

A recent innovation by the University of Texas M. D. Anderson Cancer Centre researchers is the use of biologically compatible viral and gold particles to fabricate a “nanoshuttle” that can identify and latch on to disease sites such as damaged arteries and from there send signals that can be interpreted by imaging devices. It can also be adapted to carry drugs, genes or stem cells. (Using biologically compatible materials to fabricate a nanoshuttle 2007)

A similar study by the College of Veterinary Medicine at Ohio State University found that silica nanoparticles when injected in animals could be detected by ultrasound and actually improved the image. Role of Clinical Decision Support Systems (CDSS) in Reducing the Operating Cost and Medical Errors Paper.The researchers are working on developing biodegradable nanopoarticles that could be used on humans and could be programmed to go to specific sites of the body and carry medicines. It is slated to be an important step in the early detection of cancerous growths. (Nanotechnology can identify disease at early cellular level 2007)

The University of North Carolina (UNC) in Chapel Hill researchers have produced customized biocompatible nanoparticles that are designed to attack specific types of cancer on the molecular level by delivering anticancer drugs directly to the abnormal cells, eliminating the side effects of chemotherapy. The particles also enhance imagery of cancer cells to improve diagnosis and may be used for gene therapy. The researchers have overcome the problem of fabricating stable nanoparticles capable of being directed towards a specific site by using the Particle Replication in Non-wetting Templates (PRINT) technique. (“Custom” nanoparticles could improve cancer diagnosis and treatment 2007)

A molecular computer built by Israeli researcher Itamar WIllner of the Hebrew University of Jerusalem in Israel uses enzymes to perform calculations and is designed to be eventually implanted in the human body to monitor drug release. Because it is much smaller than silicon computers, it is already widely used in calculations and will eventually be used in bio-sensing equipment for intelligent drug delivery.  (Enzyme computer could live in human body 2007)

3. Robotic Assistants

The HeartLander, a robotic device developed by Cameron Riviere, is designed to deliver medicine or attach medical devices on a beating heart.  It is less invasive than open heart surgery and does not require surgeons to interrupt the heartbeat. About 20 mm long, the robot resembles a caterpillar and moves like one as well across the surface of the heart at seven inches per minute. (Robot created to treat ailing hearts 2007)

The University of Calgary has developed with neurosurgeon Dr. Garnette Sutherland and associates the robotic operative machine called NeuroArm. Working in conjunction with a magnetic resonance imaging machine, NeuroArm is controlled from a computer and the enhanced detail and control enables surgeons to have unprecedented precision, accuracy, dexterity and stamina during microsurgery such as neurosurgery. It is currently being tested. (First image-guided surgical robot created 2007)

Dr. Garnette Sutherland with neuroArm. (Credit: Ken Bendiktsen, University of Calgary)

NeuroArm was developed in collaboration with MacDonald, Dettwiler and Associates Ltd. (MDA), who also created Canadarm and Canadarm2. It is collaboration between experts in the fields of medicine, engineering, physics, and education. It also involved the private sector, government agencies and funding organizations in what has been dubbed a flagship program. NeuroArm enables surgeons to work accurately in very small spaces, a significant ability in microsurgery. (World’s first image-guided surgical robot to enhance accuracy and safety of brain surgery 2007)
Retinal implants, electrodes attached to the defective retina of blind patients and connected to a mini camera was designed to give sight, but the clinical trials were not encouraging. Bonn University scientists, however, are working to correct this with a software system called Retina Encoder enables the prosthetic to generate signals that can be visually interpreted by the brain. There are some problems with the translator, however, which developers are hoping will improve with clinical trials. Moreover, vision would be limited to the perception of shapes of large objects. (Retinal implants may be significantly enhanced with new software 2007)

Shades of I, Robot, Massachusetts Institute of Technology have developed a robotic housekeeper called Domo designed to help elderly or people of limited mobility within a house setting. It is capable of grasping objects and placing them on shelves, providing assistance in simple tasks such as putting away the dishes. Domo can locate human faces, reacts to motion and responds to pressure on its hands, arms and neck. Originally funded by NASA, it is now supported by Toyota which would like to develop the project for improve robots for the home. Applications for assembly lines, agriculture and space travel are also being investigated. (Scientists build robotic housekeeper 2007; Futuristic Robot Adapts To People, New Places 2007) Role of Clinical Decision Support Systems (CDSS) in Reducing the Operating Cost and Medical Errors Paper.

The development of a manmade heart to replace a defective has been long-sought after in the medical profession. Unlike the Scarecrow in ‘The Wizard of Oz’ who wants to feel, the successful repair or replacement of the heart is a matter of life and death, and because of the scarcity of organic replacements, an artificial heart would be the next best thing. The first artificial heart design was patented in 1963 by Paul Winchell and used by Robert Jarvik for the Jarvik-7 Artificial Heart. The first recipient, Barney Clark, survived just 112 days after implantation in 1982. While the failure rate was high, the Jarvik-7 was used as a stop-gap device as patients waited for donor hearts (Artificial heart 2007)

A more successful prosthetic heart was the AbioCor Implantable Replacement Heart, received by Robert Tools on July 2, 2001 in surgery performed at the Jewish Hospital in Louisville, Kentucky. He survived 17 months.

In 2004, the US Food and Drug Administration approved the use of the implantable Total Artificial Heart (TAH-t) from Syncardia Cardio West for patients suffering from end stage biventricular failure with no organ donor in sight. The success rate for the transplant itself is 79% and survival rate is between one to five years.  It is available in several institutes including  Montreal Heart Institute (Quebec, Canada), Groupe Hospitalier La Pitié-Salpêtrière (Paris, France), Deutsches Herzzentrum Berlin / German Heart Institute Berlin (Berlin, Germany). In September 6, 2006, AbioMed of Danvers, Mass. Succeeded in having the first fully- implantable heart prosthetic AbioCor device with the FDA under Humanitarian Device Use.

Biopsies can now be performed with more precision due to the development of PneuStep, a remotely controlled motor and robot that does not use metal or electricity and works in conjunction with an MRI which can more accurately locate and collect tissue samples. This is especially useful when performing a biopsy from somewhere like the prostate. PneuStep enables smooth and precise movements and is controlled with MRI guidance. (Robotics: engineers announce plastic, air- and light-driven device more precise than human hand 2007)

London hospital St. Mary’s NHS Trust and Imperial College is the pilot for the Remote Presence (RP6) Robots, named Sister Mary and Dr. Robbie that allow doctors to visually examine patients remotely using wireless technology. Deployed in the General Surgery Ward and A;E Department, the RP6 are also being used for training purposes at the Academic and Clinical Skills Unit. Doctors may also view X-rays and laboratory results and read patient records. The RP6 allows doctors to have a presence in several places at once when necessary.Role of Clinical Decision Support Systems (CDSS) in Reducing the Operating Cost and Medical Errors Paper.  The pilot, the only one in the UK, is to evaluate patient response to robots and will spearhead an innovative application of telemedicine technology. (Robo-doc to start making ward rounds at St Mary’s Hospital 2005)

UCL Medicine’s Professor Owen Epstein has built a “Virtual Consulting Room” that allows doctors to access specialist knowledge through the use of 370 patient journeys, clinical flowcharts and 1200 care pathways relevant to Accident & Emergency, general internal medicine, obstetrics & gynaecology, oncology & palliative care, paediatrics, radiology and surgery. Out of 58 London-based general practitioners who tried the system, 88% gave positive feedback and 82% found it easy to use. (Carim 2006)
Three-dimensional imaging has been enhanced by taking advantage of advanced computer architecture and memory bandwidth. The Mayo Clinic and IBM scientists have sought to make the image processing faster and better aligned so that images produced can more easily interpreted by radiologist, especially in diseases such as cancer. (Processing of 3-D medical images improved 2007)

F. Case study: GOSH

A case study for the utilization, application and integration of CDSSs may be illustrated by a perusal of London’s Great Ormond Street Hospital (GOSH) systems. A simple questionnaire was given out to 100 of the healthcare staff, including administrative staff  and medical technicians and established that the following technologies were overwhelmingly assistive in the carrying out of health services to the patients of GOSH and the primary source for research and development for future innovations. To the question “Do you believe that computer-assisted technology has been of help in carrying out patient care?” 96% assented, with the remaining 4% failing to turn in a response. To the question “Which of the specific technologies do you personally believe is particularly efficacious?” the response has been to medical practitioners video recording and editing of procedures such as the heart transplants as a means of improving patient treatment through virtual first-hand experience, and the electronic medical records for the administrative staff. Improved and interactive imaging systems are also cited as “critical” to accurate diagnosis and consultative activities. (67%)

The history of GOSH is a tale of innovations and risks taken in the pursuit of providing the best possible care for its patients. Working in partnership with the University College London’s Institute of Child Health, GOSH is the foremost provider of specialist children’s health care in the UK and the most comprehensive research and teaching centre in the UK. It houses the only Biomedical Academic Centre for paediatrics dealing with cardiac and brain problems as well as the largest centre for children with cancer in Europe. (About Great Ormond Street Hospital (GOSH))

GOSH treats only children referred by specialists from other institutions, and the patients are likely to have a cacophony of health problems, many of them rare or congenital or both.  It was first established in 1852 when a survey revealed that a disproportionate number of children were dying due to lack of prioritization of sick children in overcrowded hospitals. Gynecologist Dr. Charles West saw the need for a fully-equipped hospital for the care of sick children and with the support of philanthropists and health reformers, he established The Hospital for Sick Children at 49 Great Ormond Street a mansion that had been the home of Dr. Richard Mead, Queen Anne’s physician. Funded entirely from subscriptions, donations and fundraisers, the hospital had only initially 10 beds and the clinical staff, comprised of Dr. West, Dr. William Jenner and a surgeon, and donated their services without pay. In 1858, a fundraiser in which Charles Dickens spoke out for support of the hospital, the money raised was enough to expand the hospital by the purchase of the neighboring house, expanding the capacity to 75 beds. Before the creation of the NHS, individual beds had a sponsor, wealthy benefactors who paid the expenses in behalf of themselves or as endowments. The 1875 building built as a replacement for the original hospital was again demolished in 1990 to make way for The Variety Club, which houses the Intensive Care unit. A new building opened in 1893 which housed, among other things, an X-Ray Department in the basement.  (Baldwin 2007) Role of Clinical Decision Support Systems (CDSS) in Reducing the Operating Cost and Medical Errors Paper.

Financing had always been a problem for GOSH but the inception of the NHS brought this to an end. In 1994, GOSH became an independent NHS Trust and spearheaded many innovations in paediatric research. Post-war London had a smaller population to service, and GOSH evolved into a tertiary children’s hospital, accepting increasingly non-routine treatment cases. Purpose-built additions to the Barrie Wing that would house Pharmacy, Radiography, Cardiac and Dental departments were built in 1965.

Dr. West determined to pursue the improvement of healthcare for children by encouraging clinical research in paediatrics and the training of paediatric nurses. Formal nurse training was introduced on-site in 1878 based on the handbooks of Dr. West, eventually establishing the Charles West School of Nursing, which would later be transferred to South Bank University. (Baldwin 2007)

Clinical breakthroughs persisted from 1945, including the successful treatment of tuberculosis with Streptomycin, new procedures in plastic surgery, cardiac surgery and neurosurgery for children and new drug treatment for child cancers, cystic fibrosis, eating disorders and complex genetic disorders. The hospital produced specialists in various new fields such as child rheumatism, metabolic disorders, orthopaedic treatments, respiratory conditions, asthma and genetic abnormalities. In 1908, a new out-patient building was built with funding provided by William Waldorf Astor (Baldwin 2007)

Currently, GOSH treats more than 100,000 patients annually and provides specialist care for children from all over the world. The GOSH continues to be “The Child First and Always” as it develops its specialist strengths into the 21st century with the help of computer and biotechnology innovations. (Baldwin 2007)

The first hospital to carry out transplants in children in 1988, then experimental, it has since then become the largest in Europe with 30 transplant cases annually. GOSH celebrated the dual milestones of the 100th lung transplant and 200th heart transplant in June 2005. Twenty-eight years after the first successful GOSH bone marrow transplant (BMT) in the UK by Professor Roland Levinsky in 1979, the GOSH BMT unit also celebrated its 1,000th child BMT in 2007. GOSH pioneered the transplant approaches that specifically benefited children, including the mini BMT for immunodeficient patients who are too ill for conventional BMT and gene therapy.  (Gosh celebrates its 1000th bone marrow transplant 2007) With its reputation as one of the best teaching hospitals in the UK, GOSH has provided the medical teaching population with a tool enabling them to practically see first-hand the procedures of open-heart surgery performed on babies. GOSH Head of the Cardiothoracic Unit Professor Elliot and colleagues make use of DV technology and video editing software to capture and present an actual operation with close up views of the heart with the use of a head camera. The captured video is then edited with the use of digital video editing software Final Cut Studio. The software allows images to be enhanced for easier viewing and is stored on a shared Xserve storage drive for access within the department. The next step for Professor Elliot is to catalogue these movies into an encyclopaedia that can be used as a reference and teaching tools over the Internet.  (Kent 2007)

GOSH designs and develops in-house customized medical instruments through its specialist mechanical and engineering section in which electrical safety testing is routinely carried out. The Rigel system enables efficient testing protocols that strives to comply with the targetted 5% downtime for all hospital equipment. (Advanced Electromedical Test Technology 2006)

Being the foremost in medical technology innovation is not always about specific clinical success. The ability to provide for a holistic approach to patient care includes a well-oiled administrative system. Because GOSH has numerous clinical IT systems that run on different modalities and systems, integration is a genuine concern. One aspect to the challenge to integration for clinicians and various users was the accessing of different user accounts for the different systems they needed to log on to on a regular basis.

GOSH chose an identity and access management system from Sentillon where users needed only to sign in once to access all applications.Role of Clinical Decision Support Systems (CDSS) in Reducing the Operating Cost and Medical Errors Paper.  Moreover, the system only needed users to select a patient once to access all the data associated with that patient. This one-off accessibility eliminates time wasted with multiple log-ins and hunting data for specific patients from different systems. The problem, however, is the potential loss of confidentiality. However, efforts are being made to develop smart technology to eliminate this danger. (More 2007)

The applications referred to above are numerous, and the following are some of the more innovative and crucial systems used in GOSH.

Picture archiving and communication systems (PACS) is a computer-based medical imaging software which stores, retrieves, distributes and presents in a variety of modalities including ultrasonography, MRI, positron emission tomography, endoscopy and mammography.

PACS is used primarily to eliminate the need for film-based archives and enables medical practitioners to share images online for tele-diagnosis, distance education and tele-radiology. It reduces the costs associated with traditional imaging and is widely distributed by major medical imaging and IT companies. One format which PACS has difficulty interpreting is the Digital Imaging and Communications in Medicine (DICOM) image format. The PACS network uses a central server as an archive and which is connected locally or over a wide area such as the Internet over a Virtual Private Network or Secure Sockets Layer.

Client workstations are used to manipulate images and hospitals use a film digitizer for backwards compatibility. It has the capability to interface with multiple systems. This interface reduces the risk of incorrectly identifying patient records, improves workflow patterns and storage retrieval. (Picture archiving and communication system 2007)

Great Ormond Street Hospital in London contracted TOMCAT Clinical Systems from Belfast to provide a paediatric Cardiac Information System to manage the cardiac treatment of patients. The TOMCAT system is in wide use in hospitals in the UK and Ireland because of its ease of use and flexibility. It is a modular system that covers the cardiothoracic and cardio-respiratory area and uses a diagrammatic data capture interface that integrates with hospital information systems. (Great Ormond Street selects Tomcat for cardiac system 2004)

CareVue is a computer-based clinical information system (CIS) registered by Philips which integrates patient care with data collection, organization, assessment and analysis of clinical data with the capability to interface with the health care staff. It allows caregivers to access and update patient charts instantly and eliminates redundancy as well as illegibility in charting. CareVue can trend clinical events and can interface with other medical devices including the cardiac monitor and ventilator.  It may be configured to replace transfer summaries, nursing progress notes and respiratory therapy progress notes. (What is CareVue?)

The JAC Pharmacy Stock Control and Drug Accounting System is a comprehensive package that collects data on drugs, patients, physician patters, suppliers and finance in order to manage stock control, order and delivery processing, invoice and credit handling, dispensing, ward stock, reporting, accounting, manufacturing and trading. All functions are automated with manual override options. (JAC’s Hospital Pharmacy Stock Control System)

The Federal Drug Administration-approved da Vinci Surgical System developed by Intuitive Surgery Inc in the UK enables doctors to fine tune complex heart surgery using smaller and more precise movements. The procedures are rendered less invasive and can be used for different types of surgery including general laparoscopic surgery, thoracoscopic surgery and laparoscopic prostate surgery (Robotic arms set to take over heart ops 2002)

GOSH makes use of iBase Logic Systems Manager that enables clinicians to view medical images over the Internet. GOSH clinicians participated in its development together with Addenbrooke’s Hospital in Cambridge, Birmingham Children’s Hospital and UCL. It is constantly being upgraded based on feedback from institutions that use it in anticipation of the changing needs in clinical image systems management. The image management tool links to the patient administration system of the institution using it and is NHS-compliant. (iBase medical image software under new management 2006) Role of Clinical Decision Support Systems (CDSS) in Reducing the Operating Cost and Medical Errors Paper.

Aside from integration, another administrative concern with clinical implications is the care and maintenance of the equipment used. With the wide selection of medical technology in use in GOSH, the Biomedical Engineering department ensures patient safety through a maintenance of the machines with Advanced electromedical test technology developed by Rigel Medical based on a “braincell” concept.”  It uses smart RFID tags placed on the 18,000 electromedical devices in use at the hospital which documents the identification and periodic testing of the medical equipment in compliance with IEC technical standards. (Advanced Electromedical Test Technology 2006)
GOSH has been in the forefront of medical innovations from different perspectivesincluding specialist development, medical technology research and design, teaching and telemedicine as well as medical administrative systems. From its beginnings as a 10-bed hospital for sick children in 1852 to its current status as the best tertiary children’s hospital in the world, GOSH is the perfect example of how a noble cause, a proper concern for patients and a strong support system can translate to far-reaching innovations in medicine and technology.
G. Level of application of CDSS and AIM in Bangladesh

1. General health status in Bangladesh

Bangladesh is a coastal nation found in South East Asia bordered by India, the Bay of Bengal and Myanmar. Achieving independence from Pakistan in 1971, a parliamentary government replaced military rule but the resulting government structures proved to be inflexible, where party disagreement was incompatible with a progressive attitude about key social issues such as high mortality and morbidity rates of mothers and children. (Pearson 1999) It is a fertile country frequently flooded because of its low-lying areas and many rivers. Development of public infrastructures is difficult because of the constant flooding and the people rely on the waterways to get around.

Key demographic indicators in Bangladesh

(Source: http://www.bangladeshgateway.org/healthpolicy.php)

·         Total Population – over 130 million

·         Population density – 881/sq km

·         People below poverty line – 60%

·         77% live in rural areas

·         At current growth rate doubling (population) time is 25-30 years

·         Per capita GDP – Tk.18, 896

Some Health infrastructure information:

·         UHFWC – 3375

·         31-50 bed UHC – 397

·         Various types of district level hospitals – 80

·         Government medical college hospitals – 13

·         Postgraduate hospitals – 6

·         Specialized hospitals – 25

·         Doctor to population ratio – 1:4719

·         Nurse to population ratio – 1:8226

·         Total hospital beds – 40,773 (over 29000 in GOB)

Some health and FP indicators:

·         CDR – 5.2 /1000 population

·         Annual Growth rate – 1.48 percent

·         MMR – 3.92 /1000 live births

·         IMR – 62 /1000 live births

·         Under 5 MR – 83 /1000 live births

·         TFR – 2.9

·         CPR – 53.8%

·         Life expectancy at birth – 68 (m) and 69 (f)

·         Fully immunized children – 52%

·         TB (smear positive new) detection rate – 31.2%
Bangladesh is densely populated, the majority of which are Bengali and Hindu, with a GNP per capita of approximately US $230. Mostly illiterate, women are considered inferior, and many are unaware of their rights. Cultural beliefs, such as gender roles and birth giving, account for a large portion of how health care is accessed and received. To provide basic health care to less than 40% of the population, the government spent a 1997 average of $ 10.50 per person, translating to a total of 3.9% of the GDP, of which 64% was on drugs. Theoretically free, providers often required user charges. Management of foreign subsidy was often inefficient and failed to achieve its goals. The influx of people to urban areas further taxed the available staff and funding resources beyond sustainable levels. (Pearson 1999)

The Bangladesh government, in its pursuit of its declared health policy, identified the following principles: creation of citizen awareness of the equitable right and  need to obtain health, nutrition and reproductive health services regardless of creed, religion or gender; involvement of the citizenry in planning, management, funding and monitoring of health service procedures; facilitation of collaborative health-related efforts between the government and NGOs; decentralization of the health services; legal support of the service providers; and establishment of a self-sufficient health sector in the delivery of primary health care and essential services. Role of Clinical Decision Support Systems (CDSS) in Reducing the Operating Cost and Medical Errors Paper. (Health Policy of Bangladesh) It is a highly hierarchical society and many professionals come from privileged, urban backgrounds that have little contact with the rural and urban poor. Public service is not a native concept of social duty.

In Bangladesh, the government is comprised of six local governments, which is divided in 64 districts, further broken down into 460 Thanas before proceeding to unions and villages. At the Thana level, primary care is delivered via Thana health centers (THCs) which have theaters, radiology units, laboratories, dental units and delivery units, each with 31 beds. Of the more than 400 created, many are currently in poorly maintained condition and have substandard healthcare staff. In consequence, THCs are under utilized and fail to deliver the essential package of services (ESP) it was designed to provide. These include reproductive and child healthcare, communicable disease management and limited curative care. As a result, hospitals are in heavy use, operating on over-capacity. (Pearson 1999)

Approximately half of the female population in Bangladesh suffers from chronic energy deficiency and 43% of pregnant women are iodine deficient. The mortality rate of pregnant women, new mothers and newborns (75%) is high largely as a consequence of inadequate emergency obstetric services and trained health care staff. Many new mothers die of excessive postpartum bleeding, postpartum sepsis and eclampsia as well as complications arising from unsafe abortion (25%), obstructed labor, and violence and injuries. Other complications that claim the lives of nine million women a year are attributable to pregnancy and childbirth, such as fistulae, uterine prolapse, or painful intercourse (Bangladesh 2005)

2. Interventions

The Department for International Development Health Systems (DFID) recognized the need to improve the health sector in Bangladesh, and partnered with government agencies to address the need for: organization and management by decentralization and human resources management; closer ties between public and private sectors; support of NGOs; stronger medical and nursing education; and a more developed health economics capability. Reorganization of the management hierarchy of health services and family planning services were undertaken to eliminate the duplication of service provision. Gender inequality is also an issue, where female family planning workers have no job security while male health workers are often given tenure. The health problems in Bangladesh are due mainly to poor hygiene and sanitation, accounting for 80% of disease. The mortality rate of mothers and infants are high and unassisted and dangerous abortions are common practice despite efforts to promote family planning practices. Life expectancy of females at birth is lower than for males. (Pearson 1999)

Government control of health services has long been a subject of concern for health reformers. However, recent collaboration among public, private and NGOs have played a big part in the improvement of the health sector, especially in the development and training of healthcare personnel and staff, especially in the matter of maternal, newborn and child health. Family welfare visitors (FWV), who provide most of the maternity health care at the union level, for example, are given an 18-month training course in one of the 12 FWV Training Institutes. Senior staff nurses undergo a 12-month midwifery course in order to provide midwifery as well as nursing care. Medical assistants who provide general health services undergo four years of training in the institutes provided by the Director General of Health Services. The government has even provided training facilities for traditional healing methods such as homeopathy, Ayurvedic, Unani and traditional birth attendants or Dais, because 90% of births are done in the home. (Bangladesh 2005) Role of Clinical Decision Support Systems (CDSS) in Reducing the Operating Cost and Medical Errors Paper.

A Child Survival Programme (CSP) between Concern, Saidpur and Parbatipur municipalities and 24 ward health committees (WHC) to improve maternal and child health care in a sustainable municipal health service. The programme included immunization, vitamin A supplements, perinatal, neonatal and antenatal care, and management of childhood diseases.  However, shortages of competent health workers as well as political and social barriers proved to be serious impediments in CSP. Another collaboration is the CORE Group, a group of more than 35 non-governmental organizations (NGOs) involved in USAID Child Survival and Health Grants and ORC Macro. The Child Survival Technical Support Unit assesses the sustainability of health grants in Bangladesh and concluded that in order to succeed; grants must be used in collaboration with WHCs and a strong health department management. (Datta, Kouletio and Rahman 2005)

The challenge to government to improve health statistics is made worse by the increase in urban population to 29 million in 2001, doubling the figure for the 1999 population recorded at 12 million. However, some improvement has been seen in the plight of children in the 10 years between 1994 and 2004 due to the prevention and control of measles, poliomyelitis, diphtheria and diarrhoeal disease. However, the situation is still grim enough even for a developing country. (Bangladesh 2004)  About 58% of married women use some form of modern contraception while 11% use traditional methods of family planning, and the total fertility rate has decreased by more than half from 1975 (6.3%) to 2004 (3%).  Children’s prospects have also improved somewhat for the neonatal, post-neonatal, infant, child and under-five mortality rates. (Bangladesh 2005)

3. Use of Technology to Improve Healthcare

One of the world poorest countries, Bangladesh will be offered online medical diagnostic assistance by a company called Bangladesh Telemedicine Services (BTS). BTS will link 200 specialists from the Bangladesh capital city Dhaka and the system allows ECGs and other clinical images to be sent for expert interpretation.

Initially, 50 telemedicine centres will be set up by BTS in the rural areas in a 120-mile radius around Dhaka where medical services are below minimum standards at a cost of US $8,000 each. It is free for local doctors to use but patients will be required to pay, and 40% will be given back to BTS.  Run as a commercial enterprise, backers of the project hope to recoup expenses with cost-savings from early detection of diseases.

According to BTS Managing Director Dr Sikder Zakir, the Bangladeshi government spends US$1 Billion on healthcare yet the majority of the citizens do not receive the care they need. One of the reasons for this is the lack of expertise of many health workers to identify diseases at an early stage. (Hermida 2002)

An India-based hospital group Apollo Hospitals Group franchised the first private multi-specialty hospital in Bangladesh in Dhaka, hoping to tap into the US $ 330 million spent by citizens for overseas medical treatment. While private healthcare is already available, the capacity is not enough to meet demand and many have inadequate facilities to treat disease such as cardiovascular diseases. Of the 450 beds, 45 are reserved for charity cases. (Rowe 2005)

Ocular visits to Apollo Hospitals Dhaka, Square Hospital Dhaka, Lab Aid Cardiac and Specialized Hospital and United Hospital Ltd. Dhaka revealed that in general, hospitals and health care centres in Bangladesh are equipped with standard diagnostic equipment. These include machines for magnetic resonance imaging (MRI), computed axial tomography (CAT) scans, digital X-rays and ultrasound. However, the level of services and capabilities are more of the standard of a community hospital, even in the newest of the major hospitals in the capital, Square Hospital which opened in 2006. Moreover, the best facilities are concentrated in Dhaka and in private hospitals.

III. Research Methodology

The foregoing overview was a result of research of peer-reviewed articles from journals such as Journal of Managed Care, Annals of Family Medicine and The New England Journal of Medicine focusing on the various applications and innovations in artificial intelligence in Medicine (AIM) and computer-based clinical diagnosis support systems (CDSS). News articles on updates for AIM and CDSS in various institutions as well as announcements and reviews of various technological innovations, such as nanotechnology was also utilized in this overview.

The information for the Great Ormond Street Hospital International Children’s Hospital (GOSH-ICH) case study was culled from the website of GOSH-ICH as well as from news articles and independent industrial publications which describe the AIM and CDSS technology currently being used in GOSH-ICH. Role of Clinical Decision Support Systems (CDSS) in Reducing the Operating Cost and Medical Errors Paper. A simple questionnaire was given out to the clinical and administrative staff of GOSH to provide a general level of satisfaction regarding the efficacy of the new technologies in use at the tertiary hospital.

The Bangladesh overview was taken from publications from the World Health Organization and USAID, Government of Bangladesh-sponsored releases and independent news articles. To determine the extent to which CDSS and AIM technology was present, ocular visits to the largest private hospitals in Dhaka was undertaken because these were the most likely to have such equipment.

IV. Discussion

As illustrated in an earlier part of the paper, there have been massive and exciting strides in medical technology in the field of AIM and CDSS. However, such developments are mostly limited to developed countries such as the US and the UK, countries which can afford to refine and improve existing technology.

It is but human to speculate how much benefit the developments costs of any single innovation mentioned would have given towards improving the healthcare in Bangladesh and similar countries with the provision of the most basic of health care services. An innovation such as a robotic housekeeper may be intellectually reasonable, but hardly a matter of urgency, such as the situation of mothers and infants in Bangladesh who are dying because of the lack of basic human services such as health care. However, what is the burden of developing countries in this matter?

According to Bangladesh Telemedicine Services Managing Director Dr Sikder Zakir, the Bangladeshi government spends US $1 Billion in healthcare, yet the statistics do not make this apparent. The possibility that this is due to corruption is very strong and makes it very unfortunate for the general population.

The initiative of using telemedicine to bring medical knowledge and expertise in isolated and rural areas Bangladesh is a step in the right direction, and is completely feasible and sustainable for expansion to a degree that makes it significant on the national level. Moreover, the fact that it is a commercial enterprise makes the incentive for it to work so much stronger for those managing it, and reduces the possibility of more corruption, as would be the case if it was a charitable endeavour.

However, based on the results of past efforts to provide assistance to improve the healthcare in Bangladesh clearly indicates that government involvement is a major factor in the success of any project. While the current government appears to be cooperating, there is no guarantee that the next elected set of officials will be just as accommodating.

The Apollo Hospital is also a clear indication that the economic situation in Bangladesh will eventually make it feasible to have a wider use of CDSS and other modern and beneficial technology in the future. However, the barriers to better health care still exist, such as the brain drain from government hospitals to private practices is legitimate, the overpopulation of urban areas, gender inequalities and accessibility of rural populations to healthcare.

There is no one single solution to the situation in Bangladesh in which basic health care is still limited to mere survival. Any initiative should take into account four directions: the government, the population, traditional social roles and the technology.

V. Challenges and future speculation of CDSS and AIM

The field of nanotechnology seems of particular interest for the computer-based medical research community because it combines the three most crucial aspects of medical technology: diagnosis, isolation of the problem and efficient treatment. Role of Clinical Decision Support Systems (CDSS) in Reducing the Operating Cost and Medical Errors Paper. The use of microscopic computers inside the body may seem like an invasion of the worst kind for technophobes, but it is no different from implanting a steel bar to support a broken bone while healing or the oral or intravenous intake of chemicals designed to combat disease. Moreover, nanotechnology nicely combines the functions of AIM and CDSS in one package, and the extension of the applications appears limitless. The challenge for this aspect is to come up with a biologically suitable medium for the technology to make it safer for implantation and eventual absorption or elimination from the human body.

Another intervention is the development of human-like robotic assistants for both the hospital and home settings. While diagnostic and treatment capabilities are essentially limited, the availability of a robot nursing aid for the wheelchair bound would provide great relief for those who prefer the atmosphere of home rather than a care institution. Another application for a sufficiently developed robot with wireless connection capabilities would be to provide difficult to reach areas with medical information and consultative assistance without requiring an expert to be physically present. While telemedicine centres provide this service, the maintenance of such facilities will be eliminated and mobile robots can be sent to any area anytime when needed.

VI. Conclusions and Recommendations

Computer-based clinical decision support systems (CDSS), or expert systems, have been a natural evolution from artificial intelligence in medicine (AIM) research with the quality and availability of computer technology and increased storage capacity. The expansion of the development from stand-alone, task-oriented AI processes, CDSS has come to encompass all aspects of medicine, from business management to microsurgery.

A review of studies that evaluated the actual benefits of CDSS and AIM to clinical management in terms of patient outcomes and cost control has been somewhat mixed due mainly to some resistance to changing clinical routine in patient care as well as the tendency for developers to document only the initial phases of implementation. However, overall feedback has been positive.

An overview of current CDSS and AIM developments in developed countries exhibits acceleration in software as well as hardware developments for healthcare use in all settings, including the home although many are in the clinical study phase. For those in current use in institutions, integration and maintenance are issues because of the sheer number of systems that an institution may be which have different modalities or platforms. GOSH illustrates how both may be achieved.

Bangladesh is also considered as a representative of CDSS use in developing countries, but the literature makes no mention of any CDSS or AIM technology in use in the general population except for a telemedicine project in the suburbs of Bangladesh capital Dhaka. Such technology is only possibly available in private hospitals such as the newly franchised Apollo Hospital in Dhaka but the capacity of such private institutions is insufficient to cater to demand. In consequence, those who can afford go abroad to seek proper medical treatment, and those who cannot afford get no medical treatment most of the time.

It is suggested that help from developing countries would be most welcome in terms of technology and expertise and would greatly alleviate the plight of the average Bangladeshi. However, it must be conceded that government control of the healthcare sector must be taken into serious consideration based on the results of previous efforts to provide assistance.Role of Clinical Decision Support Systems (CDSS) in Reducing the Operating Cost and Medical Errors Paper.  There is a parallelism between the situation in Bangladesh and the evolution of GOSH, which came into being because children were being marginalized. In the case of Bangladesh, it is children and women, specifically mothers. Gosh had been initially financed by donations and sponsors and gradually attained independence as a trust. The same could be applied to institutions in Bangladesh. It must be noted though that the healthcare problems in Bangladesh are not confined to one area, such as was the case for GOSH, and government intervention was not much of an issue.

Telemedicine will do much to solve the problem of access as well as training for medical workers, but it only deals with two facets of the problem. The biggest problems, government involvement and education of the population require a different approach. What is needed is the establishment of a healthcare body that is independent of changes in the political structure, preferably an NGO or a consortium of NGOs such as CORE. The current situation in Bangladesh is improving because the government is being cooperative. It is up to health experts and concerned organizations to take advantage and push for a more stable organization that can sustain healthcare development in the country. Once basic needs are met, and then perhaps CDSS and AIM may come in to raise the standards of healthcare for mothers and children in Bangladesh.

References

‘About Great Ormond Street Hospital (GOSH)’ 2007, GOSH-ICH.  Retrieved May 8, 2007 from  http://www.ich.ucl.ac.uk/about_gosh/

‘Advanced electromedical test technology’ 2006, Rigelmedical. Retrieved May 8, 2007 from http://www.rigelmedical.com/press/advanced_electromedical_test_technology.asp

‘Artificial heart’ 2007, Wikipedia, The Free Encyclopedia. Retrieved May 9, 2007, from http://en.wikipedia.org/w/index.php?title=Artificial_heart;oldid=129936686

‘Artificial Medical Intelligence announces latest edition of EMscribe DX for computer assisted coding solution of ICD9 diagnostic and procedure codes; provides electronic health record functionality in today’s environment’ 2005, Business Wire. Retrieved May 5, 2007 from http://findarticles.com/p/articles/mi_m0EIN/is_2005_Oct_17/ai_n15692822

‘Bangladesh’ 2005, Improving Maternal, Newborn and Child Health in the South-East Asia Region.  Retrieved May 3, 2007 from http://unpan1.un.org/intradoc/groups/public/documents/APCITY/UNPAN022523.pdf

‘Clinical decision support system’ 2007. Wikipedia. Retrieved May 5, 2007 from http://en.wikipedia.org/wiki/Clinical_decision_support_system

‘Clinical decision support tool set’ 2004, Health Management Technology. Retrieved May 3, 2007 from http://findarticles.com/p/articles/mi_m0DUD/is_11_25/ai_n6355973

‘’Custom’ nanoparticles could improve cancer diagnosis and treatment’ 2007, Rxpgnews. Retrieved May 4, 2007 from http://www.rxpgnews.com/nanotechnology/Custom_nanoparticles_could_improve_cancer_diagnosi_3781_3781.shtm

‘Dxplain’ n.d., Massachusetts General Hospital Laboratory of Computer Science. Retrieved May 5, 2007 from http://lcs.mgh.harvard.edu/projects/dxplain.html

‘Enzyme computer could live in human body’ n.d., Rxpgnews. Retrieved May 4, 2007 from http://www.rxpgnews.com/biotechnology/Enzyme_computer_could_live_in_human_body_3522_3522.shtml

‘First image-guided surgical robot created’ 2007, Science Daily. Retrieved May 4, 2007 from http://www.sciencedaily.com/upi/index.php?feed=Science&article=UPI-1-20070418-12515300-bc-canada-neuroarm.xml. Role of Clinical Decision Support Systems (CDSS) in Reducing the Operating Cost and Medical Errors Paper.

‘Futuristic robot adapts to people, new places’ 2007, Science Daily. Retrieved May 4, 2007 from http://www.sciencedaily.com/releases/2007/04/070411124418.htm

‘Great Ormond Street Hospital celebrates 100th lung and 200th heart transplant,’ 2005 News-Medical.Net. Retrieved from May 8, 2007 from http://www.news-medical.net/?id=11222

‘GOSH celebrates its 1000th Bone Marrow Transplant’ 2007, GOSH-ICH. Retrieved May 12, 2007 from http://www.ich.ucl.ac.uk/pressoffice/pressrelease_00505

‘Health policy of Bangladesh’ n.d., Bangladesh Development Gateway. Retrieved May 3, 2007 from http://www.bangladeshgateway.org/healthpolicy.php

‘iBase medical image software under new management’ 2006, E-Health Insider. Retrieved of May 8, 2007 from http://www.e-health-insider.com/news/item.cfm?ID=1696

‘Meridian Health receives national recognition for innovative use of Siemens computerized physician order entry’ 2005, Business Wire. Retrieved May 5, 2007 from  http://findarticles.com/p/articles/mi_m0EIN/is_2005_Oct_13/ai_n15688018

ORDER A PLAGIARISM-FREE PAPER NOW

‘Nanotechnology can identify disease at early cellular level’ 2007, Rxpgnews. Retrieved May 4, 2007 from http://www.rxpgnews.com/nanotechnology/Nanotechnology_can_identify_disease_at_early_cellu_4162_4162.shtml

‘NIST measuring device aims to up hip operation success’ 2007, Rxpgnews. Retrieved May 4, 2007 from http://www.rxpgnews.com/Hip/NIST_measuring_device_aims_to_up_hip_operation_success_25719.shtml

‘Processing of 3-D medical images improved’ 2007, United Press International. Retrieved May 4, 2007 from http://science.monstersandcritics.com/news/article_1289348.php/Processing_of_3-D_medical_images_improved

‘Retinal implants may be significantly enhanced with new software’ 2007, Science Daily. Retrieved May 4, 2007 from http://www.sciencedaily.com/releases/2007/04/070412113210.htm

‘Robo-doc to start making ward rounds at St Mary’s Hospital’ 2007, Science Daily. Retrieved May 4, 2007 from http://www.sciencedaily.com/releases/2005/05/050519083715.htm

‘Robot created to treat ailing hearts’ 2007 Science Daily. Retrieved May 4, 2007 from http://www.sciencedaily.com/upi/index.php?feed=Science;article=UPI-1-20070419-15024100-bc-us-heartbot.xml

‘Robotics: engineers announce plastic, air- and light-driven device more precise than human hand’ 2007, Science Daily. Retrieved May 4, 2007 from http://www.sciencedaily.com/releases/2007/04/070406145338.htm

‘Scientists build robotic housekeeper’ 2007, United Press International. Retrieved May 4, 2007 from http://www.upi.com/NewsTrack/Science/2007/04/10/scientists_build_robotic_housekeeper/

‘Study: E-prescribing, decision support systems can lower drug costs’ 2004, I-Healthbeat.org. Retrieved May 4, 2007 from http://www.ihealthbeat.org/index.cfm?Action=dspItem;itemID=106042

‘Using biologically compatible materials to fabricate a nanoshuttle’ 2007, Rxpgnews. Retrieved May 4, 2007 from http://www.rxpgnews.com/drugdelivery/Using_biologically_compatible_materials_to_fabrica_3254_3254.shtml

‘World’s first image-guided surgical robot to enhance accuracy and safety of brain surgery’ 2007, Science Daily. Retrieved May 4, 2007 from http://www.sciencedaily.com/releases/2007/04/070417114732.htm

Baldwin, N 2007, ‘The whole story’ ,GOSH-ICH. Retrieved May 8, 2007 from http://www.ich.ucl.ac.uk/about_gosh/history/story/index.html

Carim, L 2006, ‘Virtual consultant can improve patient care’, UCL Communications. Retrieved May 4, 2007 from  http://www.ucl.ac.uk/news/news-articles/0612/06122201

Coiera, E 2003, Guide to Health Informatics 2nd Edition, Oxford University Press, USA.

Datta, D, Kouletio, M & Rahman, T 2005, ‘Developing urban health systems in Bangladesh’, Participatory Learning and Action. vol. 51, pp 90-99.

Farukhi, F 2006, ‘Clinical decision support systems’, Public Health Management & Policy, 7th Edition. Retrieved May 5, 2007 from http://www.case.edu/med/epidbio/mphp439/Clinical_Decision.htm. Role of Clinical Decision Support Systems (CDSS) in Reducing the Operating Cost and Medical Errors Paper.

Garibaldi, R 1998, ‘Computers and the quality of care – a clinician’s perspective’, The New England Journal of Medicine, vol. 338, no. 4, pp 259-260.

Hermida, A 2002, ‘Telemedicine lifeline for villages’, BBC News. Retrieved May 5, 2007 from  http://news.bbc.co.uk/2/hi/technology/2288619.stm

Jenders, R 2006 ‘Improving patient-centered workflow with clinical decision support systems’, Cedars-Sinai Medical Centre. Retrieved May 3, 2007 from http://cslxinfmtcs.csmc.edu/ucla/iqpc-SAN-2006-10.pdf

Johnston, M, Langton, J, Haynes, R & Mathieu, A 1994, ‘Effects of computer-based clinical decision support systems on clinician performance and patient outcome: a critical appraisal of research’, Annals of Internal Medicine, vol. 120, no. 2, pp 135-142.

Kent, S. 2007 ‘The heart of the matter’, Apple. Retrieved May 8, 2007 from http://www.apple.com/uk/business/profiles/gosh/

Liu, J, etal 2006, ‘Nanoparticles as image enhancing agents for ultrasonography’, Physics in Medicine and Biology, vol. 51, pp 2179-2189

Manning, A 2006, ‘Diagnosis at the click of a mouse,’ USA Today. Retrieved May 5, 2007 from http://www.usatoday.com/tech/news/techinnovations/2006-08-07-visualdx-bird-flu_x.htm?csp=34

McMullin, S, Lonergan, T & Rynearson, C 2005, ‘Twelve-month drug cost savings related to use of an electronic prescribing system with integrated decision support in primary care’, Journal of Managed Care, vol. 11, no.4, pp. 322-332.

McMullin, S, Lonergan, T, Rynearson, C, Doerr, T, Veregge, P & Scanlan 2004, ‘Impact of an evidence-based computerized decision support system on primary care prescription costs’, Annals of Family Medicine,  vol. 2, n. 5.

Morris, A 2000, ‘Developing and implementing computerized protocols for standardization of clinical decisions’, Annals of Internal Medicine, vol. 132 , no.5, pp 373.383.

Pearson, M 1999, ‘Bangladesh: health briefing paper’, Department for International Development Health Systems. Retrieved May 4, 2007 from http://www.dfidhealthrc.org/publications/Country_health/Bangladesh.pdf

Prescription in Ischaemic Stroke Management (PRISM) Study Group 2003, ‘Cluster-randomized, controlled trial of computer-based decision support for selecting long-term anti-thrombotic therapy after acute ischaemic stroke’, Q J Med, vol. 96, pp. 143-153.

Rosenstein, A 1999, ‘Inpatient clinical decision-support systems determining the ROI’, Healthcare Financial Management. Retrieved May 4, 2007 from http://findarticles.com/p/articles/mi_m3257/is_2_53/ai_53953607

Rowe, H 2005, ‘New Bangladeshi hospital targets millions spent on overseas healthcare’, Aegis. Retrieved May 5, 2007 from http://www.aegis.com/NEWS/AFP/2005/AF050140.html

Szolovits, P 1982, ‘Artificial intelligence and medicine’, Chapter 1 in Szolovits, P. (Ed.) Artificial Intelligence in Medicine. Westview Press, Boulder, Colorado.  1982.

Trowbridge, R & Weingarten, S 2001. ‘Making health care safer: a critical analysis of patient safety practices.’ Evidence Report/Technology Assessment: Number 43. AHRQ Publication No. 01-E058, July 2001. Agency for Healthcare Research and Quality, Rockville, MD.
Vassallo, D, etal 2001, ‘An evaluation of the first year’s experience with a low-cost telemedicine link in Bangladesh’, Journal of Telemedicine and Telecare, vol. 7, no. 3, pp. 125-138.

Walton, R, Dovey, S, Harvey, E & Freemantle, N 1999, ‘Computer support for determining drug dose: systematic review and meta-analysis’. BMJ vol.  318, pp 984-990. Role of Clinical Decision Support Systems (CDSS) in Reducing the Operating Cost and Medical Errors Paper.