tools all still require advanced medical skills, in order to rate the symptoms and choose additional tests to deduce the probabilities of different diagnoses. DSS only supports the decision making process. The human user is required to evaluate all the factors before making a decision. On the other hand, an expert system works on the basis of the knowledge acquired from an expert and applies a set of probability-based rules to arrive at a decision regarding a particular problem. Therefore, they are suitable for usage by healthcare providers having basic training in healthcare.
Related Work & Competitors
Mobile devices and PDAs are already being used in the healthcare field and are becoming critical to patient care because they give the clinicians mobility so that they are not glued to a particular location in the hospital. Because mobile devices and PDAs allow clinicians to see everything in real time, they have the most current data during rounds. Pharmacists are also using them in real time to document interventions, check for adverse drug reactions, and provide up-to-date information on patients. Epocrates Inc and bioMerieux Inc are providing clinicians with the software applications for their PDAs. Currently, 500,000 clinicians use the Epocrates suite of products, which includes >30,000 community and hospital pharmacists. Their two most useful mobile products are Epocrates Rx and Epocrates Essential. Epocrates Rx features drug monographs, health plan formularies, a drug-interaction checker, and calculators. Epocrates Essential features the same components as Epocrates Rx with the addition of an infectious-disease treatment guide, alternative (herbal) medicines, an intravenous (IV) compatibility checker, disease monographs and symptom assessment, and diagnostic and laboratory tests. Thus far, the company has received positive feedback on its suite of products.
The objective of bioMerieux's Stellara is to reduce medical errors and adverse drug reactions. Stellara is powered by Expert System Platform Technology from TheraDoc Inc. The Stellara software systems- Etage 1, 2, 3, and 4-integrate individual electronic patient records with coded clinical data and focused medical informatics that represent an onboard clinical decision-support database. Whereas both companies do not have plans to add new models to their products now, they do plan to add new features and functionalities to their product lines. Another company Lexi-Comp provides a suite of expert systems for PDAs (mainly for pharmacists, dentists, nurses and physicians) but their solutions are costly (approx. $350). Unlike the user-interface of MobDoc, their user-interface is not graphical and is in English. MobDoc is different from the other products because
• It will be tailored especially for the maternal and child health programs.
• The questions for gathering the patient data are presented in graphical manner and in local languages and
hence the software is very user-friendly.
• It will have the facility to generate reports for monitoring the progress and effectiveness of the health
care programs.
• The software design is based upon platform-independent model and hence can be easily extended for
different architectures and operating systems.
• It will be developed and marketed at a very low cost.
Larry Weed, the creator of Problem-Oriented Medical Record (POMR) said that like we need X-rays as extensions of human eye, it is necessary to use electronic extensions of human memory and analytic capacity at the time of action. In one study of physicians in outpatient clinics the subjects of the survey recalled only 50% of patient information, 5 minutes after the appointment. Sixty percent of the physicians surveyed did not know the names of their patients' drugs, and 20% did not recall the purpose of the patients' medications. It has been estimated that 180,000 patients are dying each year as a result of medical error. Another source suggests that 1.3 million injuries may occur in the U.S. annually during hospitalization. Although many hospital injuries are unpredictable and unavoidable, 20% to 70% may be preventable. These findings illustrate the limitations of the human memory and underscore the urgent need for alternatives. Expert systems and decision support systems are therefore soon going to become a necessity in the medical field and will significantly improve the quality of health care. In the rural areas, where the health care infrastructure is usually missing, these systems can provide economical health service to masses and can reduce the dependence on the availability of a human expert (that is,

a doctor). In the cases which are not emergencies, the expert system can have the following benefits over a doctor:
• Has the combined knowledge of many doctors and researchers
• Has no time limit
• Does not suffer from forgetfulness or imprecise recall
• Operates at same high standards every day
• Has no interest in selling any particular treatment
• Can show you the full reasoning behind its findings
The MobDoc will have all of the above mentioned advantages of an expert system along with the features
that the user-interface will be extremely graphical and easy to use.
Summary of the Invention
The technology behind this invention is named as Unified Model for Expert Systems, Medical Diagnosis edition (UMES-MD). Through this technology, provision for normalization and translation of domain knowledge across different language, cultures and formats will be provided. The solution will provide maximum independence from technical platform and connectivity and hence will increase the scope of availability of medical expertise. The knowledge acquisition through ongoing usage of the tool, mining of historical data, and data stored in data warehouses will keep the system young, evolving and intelligent. Because the knowledge engineering and data curation process will be a continuous one, involving multiple subject-matter experts, the medical expertise will also evolve by taking advantage of powerful inference capabilities provided by the tool. The knowledgebase will be approved and tested by a body of medical experts to ensure high-quality and timely responses. The system will be developed in compliance with the already existing technical standards in the healthcare domain for the purpose of easy extension and integration with other existing modules and systems (e.g., OpenEMed, OpenEHR, OpenMEDIS).
The UMES-MD model, as shown in Figure 2, is a good solution platform as it addresses the CAKE Phenomena mentioned earlier. Further, the model extension for medical diagnostics provides features and functions custom to the medical diagnostic domain, thereby providing greater architectural coverage and making it an implementation-ready architecture. The architecturally significant elements of UMES-MD can be divided into four layers:
• Expert System Knowledgebase Layer: This is the central repository of knowledge, containing structured
as well as unstructured information. It has the following salient features:
o Data is stored in a data warehouse format, having facts and dimensions. For medical diagnostics, Diagnosis is the fact, with dimensions such as Symptoms, Condition, Treatment, Patient, Pharmaceutical, Practice, Location, Time, etc.
o Rules store associations between various factors in the format [(COMBINATION X) IMPLIES (COMBINATION Y)]. The rules may be provided externally, or derived from the knowledgebase repository. Each rule can have attributes, hierarchy information, a time lag factor and a rating factor and a status.
o Unstructured content such as images, photos, audio and video can be associated with the data and rules of the knowledgebase.
• Expert System Logic Layer: The logic layer of the expert system has three main subsystems:
o Inference Engine: Applies rule sets and chaining algorithms to the knowledgebase to come up with inferences or conclusions.
o Explanation System: Provides an explanation of the reasoning and inferences in a human-readable format.
o Knowledgebase Editor: Used to add, modify and approve/reject rules. Also provides supplementary query and reporting capability.

Unified Interface Layer: The purpose of this layer is to provide access to the expert system in a simplified manner and has maximum independence from technology, language, location, etc. This is achieved using the following subsystems:
o Services Interface
o Batch & Data Interface
o Normalization & Translation Engine
o Monitoring Service
Medical Diagnostic Extension Layer: This is a set of applications built on top of the Unified Interface Layer, providing advanced features and functions for medical diagnosis by leveraging the underlying expert system. The applications may exist on a variety of technical operating platforms, and may have local databases for storing their private data. Applications may be customized, enhanced or extended. However, a core framework for the following applications is provided, to enable effective use of the UMES-MD model. The core framework consists of a reference architecture for data structures, program logic, system interfaces and user interfaces.'
o Medical Expert Diagnosis
o Medical Expert Telemedicine
o Medical Expert Pharmacology
o Medical Expert Network
o Medical Expert Reference
o Medical Expert Trend Analysis
(Figure Removed)
Following is a short summary of the medical edition of the invention. The solution is based upon the assumption that some means of mobile communication or satellite communication is already available in the remote areas. A rule-based expert system capable of doing differential diagnosis can be loaded on the mobile devices so that the basic first aid and primary healthcare services can be provided to the patients even if a doctor is unavailable. The community healthcare workers, who have at least a basic training of reading and writing, can be trained to use the software application and assist the patients in their locality. To reduce the training time of these healthcare workers, it is important that the application interface supports local languages and dialects. The use of pictures to capture the patient condition can reduce the communication overheads involved in the diagnostic process. This is especially useful in case of emergencies.
On the basis of the captured patient data, the application (MobDoc) suggests a diagnosis and the course of treatment for the patient. In case of an emergency, the application suggests the healthcare worker to take the patient to the nearest hospital. While the patient is the way to the hospital, the patient data would have already reached the hospital so that high quality and timely attention is made available to the patient. The mobile devices can be connected to light wireless servers which run on rechargeable batteries for storing patient data. The wireless server can in turn be connected to a central server from where data can be collected for trend analysis and monitoring the progress of healthcare programs.
The data from the mobile application is stored in a database. Apart from providing the immediate diagnosis and treatment, using the application, the patient health can also be monitored on a regular basis. In order to retrieve the patient records, one needs to enter the patient name, age and location through a form. On submitting the form, the patient data is returned to the end-user. The stored patient history can also be used to check for drug-drug intervention and allergies when the doctors are prescribing new medicines or changing the medical regimen. Main features of the application are:
• to provide diagnosis and treatment
• data storage and sharing
• record keeping and monitoring of patient health
Technical Details
The entire system for managing the patient information flow is known as MobDoc. The MobDoc System has the following main components:
• MOBGUI: This is the main GUI of the expert system using which the community healthcare workers
gather the data related to the patient and his/her medical condition. The data submitted from this form is
currently being stored as XML file but will be stored in the local database in future. The data is also
transmitted to the wireless server.
• MOBShell: This is the knowledgebase of the expert system. Knowledge engineering is being done to
develop the same. The knowledgebase size is small at the moment and can be directly loaded on the
MOBs. However, when the knowledgebase size starts growing, scalability can become an issue. Therefore,
in the next stage of the development of the solution the knowledgebase will be loaded on the wireless
servers.
• MOBDB: This is the Microsoft SQL database used to store the patient information. In order to keep the
solution light-weight a local copy of the database can be installed on the wireless servers in the vans. The
information is sent from the MOB to the MOBDB stored on the wireless server. The synchronization of
the local copy with the central database occurs when there is network connectivity.
• MOBMonitor: This is a component that is always running in the background and is used to send/receive
information to/from the central server and the main database independently of the state of the network.
If the network is not available for any reason it just keeps waiting until it is available.
• DBCIient: The DBCIient is a GUI used to search the MOBDB for information about the patient from
previous runs.
• MOBNetWebService: This is a web service used to share the patient records amongst the collaborating
hospitals or care givers.
• MOBUpdateWebService: This service is used to update the knowledgebase with the latest diagnostic
rules and treatment information.
The screenshots shown below give an overview of the application. The data flow diagram of the system is shown in Figure 3 and the architecture is shown in Figure 4. The Dispatcher in the diagram is like a medical call-center that will be contacted for routing the calls from the community healthcare workers in the event of an emergency.
(Figure Removed) Figure 4. Layered Architecture of the GUI (Adapted from- Lhotka R, Expert C# 2005 Business Object 2nd Springer-Verlag, New York 2006.)
The multimedia-based medical expert system, MobDoc, which can assist the community healthcare providers to help the expecting mothers and infants, gives us a competitive edge. The patient conditions can be captured visually and in local languages leading to drastic reduction in the training and complexity of usage of the system by the midwives and nurses. The capability to take voice notes and pictures of the patient will also be provided. The application can be adapted to the local languages and dialects on the basis of the GPS coordinates of the mobile device on which the application is running.
Apart from being mostly graphical, the proposed solution will be having text in local languages (e.g., Guajarati, Hindi) so that the system is easy to understand and use. Information like how much dosage of a particular medicine is required for a patient or the inventory of supplies with the healthcare providers can be checked in real time using the expert system and the information stored on the wireless servers. The reports for monitoring the effectiveness and progress of the health care programs can be easily generated from the central server. This solution will be developed using domain-specific modeling so that it can be customized in local languages and can run on heterogeneous architectures.
The midwifery training module which is video-based is another salient feature of the application. The videos can be customized with voice-overs in local languages and dialects for the training and convenience of the healthcare providers and patients. This will also be helpful in explaining the medical condition to the patients. In case the healthcare worker is not sure about a procedure or a process, they can always refer to the videos or training modules in the application.
References
1. http://www.britannica.com/EBchecked/topic/400270/MYCIN
2. "Decision support system for the diagnosis of schizophrenia disorders",
http://www.ncbi.nlm.nih.Rov/pubmed/16400472
3. "Decision support for diagnosis of lyme disease", http://www.ncbi.nlm.nih.gov/pubmed/16160260
4. "Evaluation of a Computer Assisted Decision Support System (DSS) for Diagnosis and Treatment of Ventilator
Associated Pneumonia (VAP) in Intensive Care Unit (ICU)",
http://Rateway.nlm. nih.gov/MeetingAbstracts/ma?f=102248792. html
5. "DiagnosisPro differential diagnosis reminder tool", http://en.diaRnosispro.com/

Statement of Claim
I, Sukrit Sondhi, claim,
1. The invention of a portable electronic device that can accept information of medical conditions and provide diagnostic guidance, in a combination of text, graphical, audio and video formats.
2. The device of the above claim, having ability to extract and discover medical knowledge from historical medical records worldwide, and seeking approval from relevant authorities or experts before use.
3. The device of the above claims, having ability to incorporate expert or popular opinions into the diagnostic guidance, in real time.
4. The device of the above claims, having standardized mechanisms to translate information between multiple languages and formats for input, output or internal use.
5. The device of the above claims, having ability to use a wide variety of existing hardware, software and connectivity platforms as its operating environment.
6. The limitation of existing devices to integrate best-of-breed options for Compatibility, Availability, Knowledge and Expertise (henceforth, CAKE) in the field of Medical Diagnostics into a single, cohesive system.
7. The unique design of the device of claim, to be the first to overcome the said limitations, as described in the accompanying complete specification and illustrated in the diagrams included.