Search Images Maps Play YouTube Gmail Drive Calendar More »
Sign in
Screen reader users: click this link for accessible mode. Accessible mode has the same essential features but works better with your reader.

Patents

  1. Advanced Patent Search
Publication numberWO2009151617 A2
Publication typeApplication
Application numberPCT/US2009/003518
Publication date17 Dec 2009
Filing date10 Jun 2009
Priority date10 Jun 2008
Also published asEP2304608A2, WO2009151617A3
Publication numberPCT/2009/3518, PCT/US/2009/003518, PCT/US/2009/03518, PCT/US/9/003518, PCT/US/9/03518, PCT/US2009/003518, PCT/US2009/03518, PCT/US2009003518, PCT/US200903518, PCT/US9/003518, PCT/US9/03518, PCT/US9003518, PCT/US903518, WO 2009/151617 A2, WO 2009151617 A2, WO 2009151617A2, WO-A2-2009151617, WO2009/151617A2, WO2009151617 A2, WO2009151617A2
InventorsDennis J. Underwood, Kevin E. Gilpin, Rishi Sikka, Nicolas R. Encina
ApplicantPraxeon, Inc.
Export CitationBiBTeX, EndNote, RefMan
External Links: Patentscope, Espacenet
Accessing health-related information using personal attribute profiles
WO 2009151617 A2
Abstract
An information retrieval system is disclosed for digitally stored health information that can include a knowledge model operative to classify health information and a health profile input interface operative to receive a health profile from the user. The system can also include health profile storage responsive to the health profile input interface and operative to store the health profile, and access logic responsive to both the user health profile storage and the knowledge model and operative to access health information for the user based on both the health profile and the knowledge model.
Claims  (OCR text may contain errors)
1. An information retrieval system for digitally stored health information, comprising: a knowledge model operative to classify health information, a health profile input interface operative to receive a health profile from the user, health profile storage responsive to the health profile input interface and operative to store the health profile, and access logic responsive to both the user health profile storage and the knowledge model and operative to access health information for the user based on both the health profile and the knowledge model.
2. The system of claim 1 further including a query input interface operative to receive a search query from a user, and wherein the access logic is further responsive to the query to access health information search results.
3. The system of claim 1 wherein the health profile input interface and the health profile storage are operative to store a medical profile that includes - demographic information, conditions, and treatments.
4. The system of claim 3 wherein the health profile includes substantially all of the information from a record formatted according to the Continuity of Care Record standard.
5. The system of claim 1 wherein the health profile input interface is operative to upload the health profile in a machine-readable format.
6. The system of claim 1 wherein the health profile input interface and the health profile storage are operative to receive a health profile that includes self- reported interest areas.
7. The system of claim 1 wherein the health profile input interface and the health profile storage are operative to receive a health profile that includes substantially only self-reported interest areas.
8. The system of claim 1 wherein the health profile input interface and the health profile storage are operative to receive a health profile for a demographic group.
9. The system of claim 1 wherein the health profile input interface and the health profile storage are operative to receive a health profile for a family group.
10. The system of claim 1 wherein the health profile input interface and the health profile storage are operative to receive a health profile for an health- oriented advocacy group.
11. The system of claim 1 wherein the health profile input interface and the health profile storage are operative to receive a health profile for an health- oriented self-help group.
12. The system of claim 1 further including profile adjustment logic operative to adjust the profile.
13. The system of claim 12 wherein the profile adjustment logic is adaptively responsive to activities by users of the system to adjust the profile.
14. The system of claim 12 wherein the profile adjustment logic is operative to adjust user profiles based on global topics model for the system.
15. The system of claim 14 wherein the profile adjustment logic is responsive to ongoing changes in global topic models for the system to adjust user profiles.
16. The system of claim 12 wherein the health profile input interface includes input logic operative to enable a user to access and modify information resulting from adjustments to the profile.
17. The system of claim 1 wherein the search logic is operative to prioritize search results based on both the profile and the search query.
18. The system of claim 1 wherein the search logic is operative to select and exclude results based on both the profile and the search query.
19. The system of claim 1 further including a list-based search result presentation interface responsive to the search logic and operative to present the health information search results to the user.
20. The system of claim 1 further including an alert generation interface responsive to the search logic and operative to alert the user to the availability of health information.
21. The system of claim 1 further including provider matching logic responsive to the search logic and operative to match users with providers based on both the health profile and the query.
22. The system of claim 1 wherein the knowledge model includes information about at least medical conditions and treatments,
23. The system of claim 1 further including information access monitoring logic operative to monitor information access by users of the system and to derive evaluation information for information items based on how they are accessed and wherein the access logic is further responsive to the information access monitoring logic to access health information.
24. The system of claim 23 wherein the information access monitoring logic is responsive to an explicit recommendation control for system users.
25. The system of claim 1 further including recommendation display logic operative to display recommendation information to the user for health information accessible to the access logic.
26. The system of claim 1 wherein results from the monitoring logic are associated with individual user profiles.
27. The system of claim 1 wherein the knowledge model is operative to classify information about clinical trials and wherein the access logic is operative to access information about clinical trials for the user based on both the health profile and the knowledge model.
28. The system of claim 1 wherein the knowledge model is operative to classify information about health-related regulatory information and wherein the access logic is operative to access information about health-related regulatory information for the user based on both the health profile and the knowledge model.
29. The system of claim 28 wherein the knowledge model is operative to classify information issued by a government entity that regulates medical treatments and wherein the access logic is operative to access information issued by the government entity for the user based on both the health profile and the knowledge model.
30. The system of claim 1 wherein the knowledge model is operative to classify information on a regional level and wherein the access logic is operative to access health information for the user based on both region identification information in the health profile and the knowledge model.
31. The system of claim 1 wherein the knowledge model is operative to classify information on an organizational level and wherein the access logic is operative to access health information for the user based on both organizational identification information in the health profile and the knowledge model.
32. The system of claim 1 wherein the access logic is operative to access outcome information for the user based on both information in the health profile and the knowledge model.
33. The system of claim 1 wherein the access logic is operative to access treatment protocol information for the user based on both information in the health profile and the knowledge model.
34. The system of claim 1 wherein the access logic is operative to access health news for the user based on both information in the health profile and the knowledge model.
35. The system of claim 1 wherein the access logic is operative to access public health alerts for the user based on both information in the health profile and the knowledge model.
36. The system of claim 1 wherein the access logic is operative to access environmental alerts for the user based on both information in the health profile and the knowledge model.
37. The system of claim 1 wherein the knowledge model is based on document fingerprinting.
38. An information retrieval system for digitally stored health information, comprising: a health profile input interface operative to receive a health profile from the user, health profile storage responsive to the health profile input interface and operative to store the health profile, wherein the health profile input interface and the health profile storage are operative to store a medical profile for the user that includes the user's demographic information, conditions diagnosed for the user, and treatments prescribed for the user, and access logic responsive to the user health profile storage and operative to access health information for the user based on the demographic information, the conditions, and the treatments in the health profile.
39. The system of claim 38 wherein the health profile includes substantially all of the information from a record formatted according to the Continuity of Care Record standard.
40. The system of claim 38 wherein the health profile input interface is operative to upload the health profile in a machine-readable format.
41. An information retrieval system for digitally stored health information, comprising: means for classifying health information, means for receiving a health profile from the user, means for storing the health profile responsive to the means for receiving a health profile, and means for accessing health information for the user based on both the health profile and the knowledge model and being responsive to both the means for storing the health profile and the means for classifying health information.
Description  (OCR text may contain errors)

ACCESSING HEALTH-RELATED INFORMATION USING PERSONAL ATTRIBUTE PROFILES

Cross-Reference to Related Applications

This application is related to U.S. Provisional Application Nos. 60/719,100 and 60/719,101 , both filed on September 21 , 2005, to U.S. Patent Application no. 11/525,245, filed on September 21 , 2006, to U.S. Patent Application no. 61/131 ,559, filed on June 10, 2008, and to published PCT application no. WO07035912, which are all herein incorporated by reference.

Field of the Invention

This application is directed to systems for accessing information, including systems that employ profiles and queries to retrieve health-related information.

Background of the Invention

A number of specialized medical search engines currently exist. These use specialized search logic to allow users to search for information related to their health.

Summary of the Invention

In one general aspect, the invention features an information retrieval system for digitally stored health information that includes a knowledge model operative to classify health information and a health profile input interface operative to receive a health profile from the user. The system also includes a health profile storage responsive to the health profile input interface and operative to store the health profile, and access logic responsive to both the user health profile storage and the knowledge model and operative to access health information for the user based on both the health profile and the knowledge model. In preferred embodiments the system can further include a query input interface operative to receive a search query from a user, with the access logic being further responsive to the query to access health information search results. The health profile input interface and the health profile storage can be operative to store a medical profile that includes demographic information, conditions, and treatments. The health profile can include substantially all of the information from a record formatted according to the Continuity of Care Record standard. The health profile input interface can be operative to upload the health profile in a machine-readable format. The health profile input interface and the health profile storage can be operative to receive a health profile that includes self-reported interest areas. The health profile input interface and the health profile storage can be operative to receive a health profile that includes substantially only self- reported interest areas. The health profile input interface and the health profile storage can be operative to receive a health profile for a demographic group. The health profile input interface and the health profile storage can be operative to receive a health profile for a family group. The health profile input interface and the health profile storage can be operative to receive a health profile for a health-oriented advocacy group. The health profile input interface and the health profile storage can be operative to receive a health profile for a health-oriented self-help group. The system can further include profile adjustment logic operative to adjust the profile. The profile adjustment logic can be adaptively responsive to activities by users of the system to adjust the profile. The profile adjustment logic can be operative to adjust user profiles based on global topics model for the system. The profile adjustment logic can be responsive to ongoing changes in global topic models for the system to adjust user profiles. The health profile input interface can include input logic operative to enable a user to access and modify information resulting from adjustments to the profile. The search logic can be operative to prioritize search results based on both the profile and the search query. The search logic can be operative to select and exclude results based on both the profile and the search query. The system can further include a list-based search result presentation interface responsive to the search logic and operative to present the health information search results to the user. The system can further include an alert generation interface responsive to the search logic and operative to alert the user to the availability of health information. The system can further include provider matching logic responsive to the search logic and operative to match users with providers based on both the health profile and the query. The knowledge model can include information about at least medical conditions and treatments. The system can further include information access monitoring logic operative to monitor information access by users of the system and to derive evaluation information for information items based on how they are accessed with the access logic being further responsive to the information access monitoring logic to access health information. The information access monitoring logic can be responsive to an explicit recommendation control for system users. The system can further include recommendation display logic operative to display recommendation information to the user for health information accessible to the access logic. The results from the monitoring logic can be associated with individual user profiles. The knowledge model can be operative to classify information about clinical trials and wherein the access logic can be operative to access information about clinical trials for the user based on both the health profile and the knowledge model. The knowledge model can be operative to classify information about health-related regulatory information and wherein the access logic can be operative to access information about health-related regulatory information for the user based on both the health profile and the knowledge model. The knowledge model can be operative to classify information issued by a government entity that regulates medical treatments and wherein the access logic can be operative to access information issued by the government entity for the user based on both the health profile and the knowledge model. The knowledge model can be operative to classify information on a regional level and wherein the access logic can be operative to access health information for the user based on both region identification information in the health profile and the knowledge model. The knowledge model can be operative to classify information on an organizational level and'wherein the access logic can be operative to access health information for the user based on both organizational identification information in the health profile and the knowledge model. The access logic can be operative to access outcome information for the user based on both information in the health profile and the knowledge model. The access logic can be operative to access treatment protocol information for the user based on both information in the health profile and the knowledge model. The access logic can be operative to access health news for the user based on both information in the health profile and the knowledge model. The access logic can be operative to access public health alerts for the user based on both information in the health profile and the knowledge model. The access logic can be operative to access environmental alerts for the user based on both information in the health profile and the knowledge model. The knowledge model can be based on document fingerprinting.

In another general aspect, the invention features an information retrieval system for digitally stored health information that includes a health profile input interface operative to receive a health profile from the user, and health profile storage responsive to the health profile input interface and operative to store the health profile. The health profile input interface and the health profile storage are operative to store a medical profile for the user that includes the user's demographic information, conditions diagnosed for the user, and treatments prescribed for the user. The system further includes access logic responsive to the user health profile storage and operative to access health information for the user based on the demographic information, the conditions, and the treatments in the health profile.

In preferred embodiments the health profile can include substantially all of the information from a record formatted according to the Continuity of Care Record standard. The health profile input interface can be operative to upload the health profile in a machine-readable format. In a further general aspect, the invention features an information retrieval system for digitally stored health information that includes means for classifying health information, means for receiving a health profile from the user, means for storing the health profile responsive to the means for receiving a health profile, and means for accessing health information for the user based on both the health profile and the knowledge model and being responsive to both the means for storing the health profile and the means for classifying health information.

In another general aspect, the invention features an attribute system that provides a profile operative when accessing digitally stored information or content through search or browsing such that the digital information is prioritized and organized according to the profile, along with community activity and recommendation and along with an optional user input. In this context, content is digital data and information which may be in the form of, but not limited to, text documents or articles, including documents created from any combination of text, tables, images, video, audio, equations and relationships and drawings, annotated images, annotated video, annotated audio or any combinations of these, in a variety of formats including all of the common digital formats available. Annotation may be full or partial transcripts of audio, images or video and meta- tagged classifications or combinations of these. Content may also refer to documents, images, video or audio that are dynamically created from one or more sources in response to given directions and rules. The system features an information interface operative to access digitally stored information and a topic model interface operative to access one or more digitally stored topic models. In one embodiment, the system can create a digital computer representation of the attribute profile and separately process the digital information through fingerprinting logic that is responsive to the topic models and content. The fingerprinting logic is described in more detail in published PCT application no. WO07035912 Prioritizing and organizing the search or browsing information according to the attribute profile is done by comparing and matching the content of the information to the attribute profile through a system that is able to assess through scoring, the relevance of the information to the user's or group's attribute profile and the search query. Match logic is operative to identify one or more of the information sources that are relevant to the search query, based on the query, the attribute profile and content fingerprints.

In addition to using fingerprint logic operations and fingerprint matching to evaluate content element relevance, additionally relevance scoring can be modified according to other factors derived from activity and recommendation measures and group preferences and weighting factors. User activity and user recommendation of content is the process whereby other users add, explicitly or implicitly by their actions, measured weighting factors to content. Weighting factors can be determined through a scoring mechanism that measure community activity through actions such as opening, emailing, commenting or recommending articles. User activity and user recommendation weighting factors can result in an incremental increase in perceived value or decrease in the perceived value of an element of content. Overall user activity and user recommendation for any particular element of content is determined by integrating over all users weighted by their assigned or perceived contribution or influence factor. Overall the position of an element of content in a list of elements of content can be algorithmically determined, singly or in combination, from weighting factors based on user or group attribute profiles, weighting factors based on relevance to other elements of content, weighting factors based on user activity and recommendation, weighting factors based on perceived contribution quality, weighting factors based on the quality and reliability of the content source weighting factors based on measures of the current value of the content such as but not limited to publication date and weighting factors based on region, language, length and complexity other demographic factors.

The attributes of the profile may be specific as in diseases such as hypertension, body-mass index and type Il diabetes and drugs such as Cozaar™ and insulin or may be general in the form of diabetes, exercise, yoga, overweight, etc. Attributes may be personal and provide profiles that are specific to an individual or may be general and applied to groups of individuals, usually interacting and organized according to, but not limited to, rules or guidelines, relationships, expectations, obligations, preferences, membership or any combination of these. Groups organize and usually interact according to common needs, common interests and values. They may and be persistent over periods of time as in a community or may dynamically form and disappear as needs, interests and values change or as goals are met. Groups usually have dynamic membership with individuals joining and leaving as their personal needs, interests and values change or as the group goals, interests and values change.

The group attribute profile(s) is the result of an open adaptive system in which contributions to the group profile continuously change as individuals join and leave the group. Group attribute profiles are the result of combining all of, or a portion of, member attribute profiles, determined by the group guidelines, in such a way as to provide the group with a characteristic attribute profile that fully represents common needs, common interests and values and that usually differentiates the group from other groups and individuals. Individuals may join groups in an independent manner or collectively, dynamically and continuously changing the group attribute profile(s) through adaptive processes. Group attribute profiles can also be changed in a discontinuous, specific way through editing, deleting, adding and otherwise changing and adjusting group attributes or changing group membership.

Individual attribute profiles can be changed in a dynamic, continuous way through individual behavior or in a discontinuous way through editing, deleting, adding and otherwise changing and adjusting individual attributes. Continuous changes to an individual's attribute profile include, but is not limited to, changing group membership, direct interaction with others, through an individual's actions regarding reading, emailing, commenting on and recommending content. As an individual's attribute profile changes so does the group attribute profile of the groups to which the individual belongs. As a result the combined collection of attribute profiles for all individuals in this system and all groups in this system compose an adaptive open system in which changes are reflected throughout in a complex but determined way. Profiles can be divided into separate parts in a variety of ways. A user profile could be divided into a user-entered part and a system-modifiable part, for example.

Delivery of on-line information can be through communication protocols that can broadly be categorized as "pull technologies" and "push technologies". Pull protocols rely on the client, in this case the individual, to request data and information. Examples are internet search engines, internet destination sites etc. On the other hand push protocols originate with the application, publisher, content source, system that deliver data or information to the client, in this case the individual, through a variety of channels such a, but not limited to, email, RSS, etc. In one embodiment registered users can browse current news and alerts, issue queries, pre-formed from information already provided or constructed de novo by the user, comment, discuss, annotate and recommend articles, email and save documents, articles, image, audio, video etc. The invention can provide for delivery of content and information based on the user attribute profile and preferences through push technology such as email and RSS.

In one embodiment of the invention there is a user interface that provides a way for users to register and interact with the method and system described herein thereby operating on a computer system that enables searches to be performed over content, enables content to be browsed and organized and enables refinements to the underlying attribute profiles and methods and systems. The details of one embodiment of this user interface can be found in Figures 8-12. Systems according to the invention may be advantageous in that they allow searches to be more precisely tailored to the needs individuals. This is in contrast to typical prior art services, applications and systems such as internet services and applications in which users are indistinct from each other from the point of view of the computer application; e.g., searching for health information on the world-wide web accesses data and information from all available sources and presents the results usually as a list. The results in such systems represent the same ordered view of information for every searcher asking the same question regardless of the searchers individual context; i.e., the characteristics, attributes, needs and prior behavior of each searcher are not taken into account. It is difficult for users to define the context of their use of internet technology and applications. The invention can therefore provide a way for users or groups to create maintain and use a computer generated representation of one or more classes of their attributes that function either singly or in various combinations as a proxy for aspects and characteristics of their personal or group identity.

Users may have a number of different attribute profiles that reflect their different roles when using systems and methods described in this application. Users may act as individuals and use their personal attribute profile for the particular domain of interest or they may wish to use a group attribute profile or a combination of personal and group attribute profile depending on specific situations, requirements and needs. Group attribute profiles are proxies for aspects and characteristics of group identity/classification or preference. Each set of attributes accessed can be from a variety of domain-aware classes such as medical history, professional background, interests and hobbies, demographics, geographical location, shopping preferences, etc. The form of the computer representation of these attributes can be generated automatically from detected user and group behavior in distinct situations, from interaction with other systems or services, or can be generated, appended or amended through known methods of editing, incorporating and appending through direct interaction or machine inference. In one aspect, the invention can employ domain-aware classifications and definitions to extract relevant terms, ideas, meanings, and messages, which can be expressed as a computer fingerprint, such that subsequent behavior by the user or group may be precisely determined and adapted to a user's or group's attributes. Details of the fingerprint can be found in published PCT application no. WO07035912 In one embodiment, there may be a collection of group attribute profiles for any one domain or class representing different aspects of the group such as but not limited to demographics, region, etc. In this embodiment, individuals can use any or any combination of these group attribute profiles depending on the situation, need or preference. In certain instances, users may not have access to all group attribute profiles depending on group policies, rules and accepted understanding. Individuals can belong to none, one or more groups.

Brief Description of the Drawings.

Fig. 1 , is a block diagram of an illustrative embodiment of an information search system incorporating features according to different aspects of the invention that can include a Topic Identifier, a Result Organizer Process and a User Attribute Profile;

Fig. 2 is a is a block diagram of an illustrative embodiment of a Search Engine and Result Organizer Process incorporating features according different aspects of the invention that can include an Attribute Organizer Process and a Topic Organizer Process;

Fig. 3 is a flowchart illustrating back-end operation of the embodiment of Fig. 1 ;

Fig. 4 is a flowchart illustrating front-end operation of the embodiment of Fig. 1 ; Fig. 5 is a block diagram of an illustrative embodiment of an user attribute profiling system incorporating features according different aspects of the invention;

Fig. 6 is a block diagram of an illustrative embodiment of a user matching and comparison system based on the user attribute profiling system in Fig. 5, incorporating features according to different aspects of the invention;

Fig. 7 is a block diagram of an illustrative embodiment of a service search system incorporating features according to different aspects of the invention that can include a Topic Identifier, a Result Organizer Process and a User Attribute Profile.

Fig. 8 is a screen shot showing a user interface that allows a user to register and log into and use MyDailyApple(TM) an example of a web-based service that brings together relevant news, blogs and medical research based on the user's or group's health interests;

Fig. 9a is a screen shot showing the registration process for new users to MyDailyApple(TM);

Fig. 9b is a screen shot showing the manner in which Topics can be entered and suggested spelling and completion offered in an automated way from the underlying ontological model;

Fig. 10 is a screen shot showing the registered user's Account as used by the system and methods in the action of determining relevance for news and alerts and also in searching and organization; Fig.11 is a screen shot showing the registered user's Account Details and Profiles and illustrates how more than one Profile is listed and how Profiles can be uploaded from 3rd party applications having a health profile;

Fig.12a is a screen shot showing the registered user's MyDailyApple(TM) News Home Page a news and alerts page after registration;

Fig.12b is a screen shot showing the fuller listing of a particular news and alerts article (see Figure 12a);

Fig. 12c is a screen shot showing user Recommendations for a particular news article.

Fig.12d is a screen shot showing the Learn More output page from a particular news or alerts article.

Detailed Description of an Illustrative Embodiment

Referring to Fig. 1 , an illustrative embodiment of a search system incorporating features according to different aspects of the invention can include a Topic Identifier 12 with a Document Interface and a Query Interface. The Document Interface can be used to access documents 14 ... 14n, and the Query Interface can be used to access queries 16 ... 16n. The Topic Identifier also has an Output Interface for Document Fingerprints 18 ... 18n and Query Fingerprints 20 ... 2On, as well as a Topic Model Interface for accessing stored Topic Models 22 ... 22m. The Topic Identifier preferably operates in the same manner on documents and queries to yield fingerprints of the same format.

The system also includes an Augmentation and Analysis Module 24. This module preferably receives Query Fingerprints from the Topic Identifier, although it can also operate on Document Fingerprints. The Analysis and Augmentation Module can also include an interface for Curation and Augmentation Input 32.

The system includes an Indexer 26 with an input that can receive Document Fingerprints, which may or may not have been the subject of Analysis and Augmentation Module. The Indexer can produce an Index 28, such as an inverted search index optimized to be searched by a Search Engine 30. The Search Engine can be an off-the-shelf search engine that receives Query Fingerprints, which may or may not have been subject to augmentation, and finds similar Document Fingerprints.

Referring to Fig. 2, an illustrative embodiment of the Search Engine 30 and Result Organizer Process 34 incorporating different aspects of the invention can include the system whereby the Search Engine can provide its output to a Result Organizer Process that is controlled by a Attribute Organizer Processor 52 and a Topic-based Organizer Processor 54 and outputs to a Result Interface (not shown). The User Attribute Profile 36 has an interface to the Attribute Organizer Process 52 and an interface to the Attribute Filter Processor that provides a supervisor role to the Search Engine through the Attribute Filter Processor 38. In this embodiment, results from the Search Engine and the Attribute Organizer Processor can be displayed on a same results page. The Topic-based Organizer Processor has an interface that accesses the Topic Models and also provides a supervisor role to the Search Engine through the Topic Filter Process 50. In this embodiment, results from the Search Engine and the Topic-based Organizer Processor can be displayed on a same results page. In some applications of the search system embodied in the illustrative embodiment, the user can also provide feedback about search results that can be used by the system or its curators. In one embodiment, the system is implemented using software running on one or more Windows®-based computers and the user interface uses a standard web browser. Other structures and platforms can of course be used to implement systems according to the invention, including various combinations of dedicated hardware and special-purpose software running on general-purpose hardware. In addition, the various elements and steps described can be reorganized, divided, and combined in different ways without departing from the scope and spirit of the invention. It should also be noted that not all of the items presented in the discussion of the illustrative embodiment will be required for all implementations of the invention.

Referring to Fig. 3, illustrative back-end operations for the search system will now be outlined generally. The identifier first accesses documents to be indexed (step 100). These can be digitally stored documents of a wide variety of types, such as web pages, text files, structured and unstructured data repositories, emails, xml, presentations, news feeds, and publications, transcripts of or annotations of images, video and audio, in English or foreign languages. They can be accessed in a variety of ways such as from local storage, or from public, private, or vendor sites available through any suitable channel. The identifier then recognizes topics in the documents based on one or more of the stored topic models (step 102). The set of recognized topics for each document are then expressed as a document fingerprint (step 104). The document fingerprints generated by the identifier can then be indexed by the indexer (step 106). Periodically, the topic models may need to be updated (step 108). This can happen, for example, when two genes that were believed to be different are proven to be the same. These updates then usually need to be propagated into the index. The system preferably does this by maintaining a history, which enables it to only update and re- index fingerprints for documents that are affected by the changes to the topic model (see steps 110-114).

Referring to Fig. 4, illustrative front-end operations for the search system will now be outlined generally. The topic identifier first accesses a query to be searched in the index (step 120). This query can take the form of a naturally posed free- form textual passage, and can be indexed in the same way as are the documents. Traditional Boolean query capabilities are preferably also provided. The topic identifier can recognize topics (step 122) and generate the query fingerprint (step 124) in much the same way as it does the document fingerprints. The query can then be analyzed and augmented (step 126), such as to handle topic restriction, and topic expansion. The search engine then uses the augmented query fingerprint to search the index for documents that match it (step 128). The Topic Models and the Attribute Profile can independently act as supervisor processes, 130 and 132, in the Search Engine. Results of the search can be displayed in a simple list format.

The Topic-based Result Organizer Processor (step 136b, not shown) and/or the Attribute Organizer Process (step 136c, not shown) can also apply various types of operations to the results. This process can first apply a relevancy test to define a limited subset of the documents. In this embodiment, the relevancy test can be a statistical test.

The system can output raw results and/or processed results, such as to a web browser based interface (step 136). Users can then interact with the results in a variety of ways that will be discussed in more detail below. They can also provide feedback to the system on their satisfaction with the results. This feedback can be received and used by the system or its curators to refine the search or to improve further searches (step 138).

Where the system uses more than one model, the fingerprint index contains information resulting from different approaches to recognizing information in a document. It may even be possible to have two different types of models for a same subject area. Where a chemistry topic model and a person topic model are used together, for example, the index can allow for matches that are based both on an advanced understanding of chemistry as well as an advanced understanding of how people are related within organizations. This can reduce the chance that a highly relevant document is missed by approaching the documents in different ways.

Referring to Fig. 5, an illustrative embodiment of an. attribute system incorporating features according to different aspects of the invention can include a Topic Identifier with an Attribute Interface 254. The Attribute Interface can be used to access Attributes 252a ... 252n. The Topic Identifier also has an Output Interface for Attribute Fingerprints 256a ... 256n as well as a Topic Model Interface for accessing stored Topic Models 360a ... 360m. The Topic Identifier preferably operates in a similar manner on attributes as it does on documents and queries (Figure 1 ) to yield fingerprints of the same format.

The system also includes an Augmentation and Analysis Module 258. This module preferably receives Attribute Fingerprints from the Topic Identifier. The Analysis and Augmentation Module can also include an interface for Curation and Augmentation Input 300.

The system includes an Indexer 320 with an input that can receive User Attribute Profiles, which may or may not have been the subject of Analysis and Augmentation Module. The Indexer can produce an Index 340, such as an inverted search index optimized to be searched by a Search Engine.

Referring to Figure 6, an illustrative embodiment of a user matching and comparison system based on the User Attribute Profiles created by the User Attribute Profiling System in Fig. 5, incorporating features according to different aspects of the invention can include a Attribute Profile Comparison System 400 with a User Attribute Profile Interface. The User Attribute Profile Interface can be used to access the Attribute Profiles for users 1 ... n.

The Attribute Profile Comparison System 400 incorporating different aspects of the invention can include the system whereby the Attribute Profile Comparison System can provide its output to a Introduction Management Service 540 and or an Indexer 420. The Attribute Profile Comparison System can be composed of a Similarity Processor 480 that uses standard techniques and methods to calculate the relatedness and similarity of User Attribute Profiles by examining the profiles and comparing the degree to which related attributes have similar content. The fingerprinting techniques described previously provide an effective and efficient way to rapidly compare the similarity of many profiles, and so compare the similarity or relatedness of the associated users. Based on User Supervisor Control 520 and the output from the Similarity Processor, a Match Based Processor 500 can issue an Introduction Management Service 540 that negotiates, under User Supervisor Control, interactions between users. The Introduction Management Service can be a 3rd party service. In addition to the negotiation of specific introductions between users, the Analysis Processor 460 can utilize the relatedness measures from the Similarity Processor for all users 1 ... n to analyze by clustering or other methods of analysis the relationship between all users across all attribute classes within every User Attribute Profile. The Analysis Processor has an interface with the Indexer 420. The Indexer can produce an Index 440, such as an inverted search index optimized to be searched by a Search Engine.

Referring to Figure 7, an illustrative embodiment of a service matching and comparison system based on the Service Attribute Profiles created by the User Attribute Profiling System according to Fig. 5 in which this specific embodiment generates the Attribute Profiles from Service Attributes of Services 1 ... n, incorporating features according to different aspects of the invention which can include a Service Attribute Profile Comparison System 800 with a Service Attribute Profile Interface. The Service Attribute Profile Interface can be used to access the Attribute Profiles for Services 1 ... n.

The Service Attribute Profile Comparison System incorporating different aspects of the invention can include the system whereby the Service Attribute Profile Comparison System can provide its output to an Indexer 820. The Indexer can produce an Index 840, such as an inverted search index optimized to be searched by a Search Engine.

Referring to Figure 8, an illustrative embodiment of a user interface that allows a user to register and log into and use MyDailyApple(TM) an example of a web- based service that brings together relevant news, blogs and medical research based on the user's or group's health interests. The linked 801 prompts users to register. Users that have previously registered can sign in directly using their login information 802. Top Stories 803 are the top news stories according to activity measures (recommendations by other system users or groups) feeding into the relevance score for each news and alert item. This is important because link-based techniques do not work for news and user recommendations fill this void in the relevance technique.

Referring to Figure 9a, a screen shot showing the registration process for new users to MyDailyApple(TM). Account Details 901 prompts the new user for usual account information including a Screen Name which can remain as the default "anonymous". Health Profile 902 prompts the new user for a Profile Name, default if "Myself, and a set of Topics which refer to the diseases, conditions, drugs, symptoms and signs relating to disease and conditions for the Profile "Myself. Other Profile Names and associated Topics can be entered in subsequent Health Profile additions and editing. Sign Up completes the registration for the new user.

Referring to Figure 9b, a screen shot showing the manner in which Topics can be entered and suggested spelling and completion offered in an automated way from the underlying ontological model. This is to assist users in entering a complete and valid Health Profile. In some examples when a Health Profile Topic is unrecognized for any reason, the unrecognized Topic is highlighted and prompted for user action. Figure 10, shows the registered user's Account as used by the system and methods in the action of determining relevance for news and alerts and also in searching and organization. The Account is specific for each Profile Name 1001 , in this case "Myself. Included and available for editing is Demographic Information. In addition the Medical Information 1002 entered as part of the registration process or through addition of a new Profile Name or through editing of a Profile Name, is listed and available for editing. Any changes on this page can be updated by issuing an "Update" order.

Figure 11 , shows the registered user's Account Details 1101 and Profiles 1102 and illustrates how more than one Profile is listed and how Profiles can be uploaded from 3rd party applications having health profiles (e.g., Google Health), Personal Health profiles or Records (PHR), Electronic Medical Records (EMR), Continuity of Care Records (CCR) etc. Importing Heath Profiles provides a way for new users to include their Health Profiles that are maintained with other applications.

Figure 12a, shows the registered user's MyDailyApple(TM) News Home Page. "My Personal News" 1201 lists the most relevant news and alerts to the user's Health Profile "Myself 1202 (a link to another Health Profile, "Jimmy" is also shown). In addition to "My Personal News", "Most Popular News", 1207 lists health news stories that are the most popular among all users of MyDailyApple™, ranked in order of all other user activity as measured by, but not limited to, users opening the article, emailing the article, recommending the article etc.. Each article has a "Score" 1213 which measures the community interest in the article and the relevance to your Health Profile. Each article in this embodiment is listed by title, date and source. Each title is enabled such that a user can open a fuller listing by a mouse click 1205 which changes the listing to provide a snippet of text information taken from the article and other functionality (Figure 12b). MyDailyApple(TM) News Home Page provides access to a Search application 1206 and Account information and editing and How It Works v informational pages 1203 enabling the registered user access to additional functionality. "What's in the News" 1204 is a list of the current medical and health news organized by, but not limited to, Drug and Disease topics. Each topic provides a link to all of the current news and alerts discussing that topic and organized by, but not limited to, date. The font of the "What's in the News" represents the number of underlying news and alerts articles such that larger fonts and holding represents more news and alerts article than smaller fonts that are un-bolded.

Figure 12b shows a fuller listing of a particular news and alerts article 1205 (see Figure 12a). A more detailed description of the article, a "snippet" 1212, is taken from the article and shown to the user. In addition in this embodiment, added functionality allows the user to go to the original or "Full Text" 1208, enables to the user to provide recommendations 1209 on the article under their screen name which is then available to the rest of the MyDailyApple(TM) audience, enables the user to "Learn More" 1210 by submitting an automatic search using the full information of the article including title, source and snippet (see Figure 12c) and enables the user to "Email" 1211 by email the article in its full listing form to selected email recipients.

Figure 12c, shows a news article with user recommendations 1214.

Figure 12d, shows a Learn More 1210 output page from a particular news or alerts article. Health and medical information related to the query article is organized in this embodiment, according to whether the information is Reference material, News, Guidelines and Recommendations, Research and Clinical Trials. Organization is created dynamically dependent on the preferences, needs and expectations of the user and is part of the registered user or group's Health Profile. Various aspects of the invention that can be used in connection with systems such as the one shown in Figs. 1-12 will now be discussed in more detail.

Examples

Example 1 : Creation of a User Health Attribute Profile

This is an example of the creation of a health attribute profile. A patient provides a summary of their medical history through an on-line process of answering medical questions, providing a narrative history, uploading medical analyses, medical records etc. or by any means whereby relevant information pertaining to their medical condition(s) or treatment(s) and therapies or, in the case of acting as a proxy, the relevant information pertaining to their medical condition(s) or treatment(s) and therapies of those they represent. The content of their Personal Health Record, herein called a "Health Profile", provides consumers of health information with a convenient way to tailor health information around their medical and health needs. It is important, when searching for health information, to get the best possible information from the best sources and to have the information organized and tailored to your specific health profile and needs. Understanding the medical history of a consumer, including their current and past conditions and treatment, diagnostic information, symptoms and signs, their demographic profile and family history and their health preferences, provides a unique opportunity to focus on-line health exploration in areas that are most relevant to their needs and of most interest to them. Services and applications such as searches, health news and alerts, advertising and on-line social environments can all be tuned and tailored to a consumer's needs using their Health Profile.

A Health Profile is a user attribute profile generated from the patient's medical history that is used to give them access to the best medical information available that's specifically relevant to their needs enabling them to make sound health decisions. This profile of their medical history is used to power a suite of on-line tools and services.

There are three distinct conceptual steps in the creation of a Health Profile:

First, it begins with the patient's medical history. Signs and symptoms can be entered in by the patient or their representative in a combination of discrete fields and narrative history. The history can be supplemented, with patient permission, from an Electronic Medical Records (EMR) application. Stream of medical, pharmacy and laboratory claims data, once again with patient permission, may also be used to augment the Health Profile. Uploaded documents including medical records, discharge summaries, laboratory reports, radiology reports and EKG results can also contribute.

Second, the patient's medical history is then used to create a health attribute profile using the systems described in this application. In one such case user health attribute profile is described as a fingerprint representing all of the concepts and relationships within their medical history built from a multi-ontology model. Details of the fingerprint and its representation are contained in published PCT application no. WO07035912. This model of medical terminology has been disambiguated, curated and refined and is used by health care professionals for advanced medical searching at Curbside.MD (www.curbside.md), and by patients at MyDailyApple(TM) (www.MyDailyApple^MJ.com).

Last, the user health attribute profile or Health Profile is used to power health applications and tools. The Health Profile can be matched against semantic fingerprints generated from the medical literature/evidence based medicine to power medical news and alerts. Semantic fingerprints from trusted medical sources including, but not limited to, Mayo Clinic, WebMD, NIH and other sources of trusted medical, treatment and guidance information, news from trusted news services and alert services and blogs, forums, discussions and transcripts of lectures, demonstrations, procedures, interviews could also power an alert and news delivery system. The Health Profile could operate in the background during keyword searching to reprioritize results. The Health Profiles also can match personalized health information and advertising to individual patients. In addition, the users Health Profile can also be used to promote social networks by identifying individuals with similar health attribute profiles and health care providers with expertise among a cohort of individuals with similar health attribute profiles.

The entire process of health profile creation, matching and sharing is secure, patient controlled and compliant with rules, guidelines, usage, requirements, regulations etc. such as, but not limited to, HIPAA regulations.

Example 2: Exploring Internet Information with User Health Attribute Profile

This is an example of how a Health Profile can be used in internet exploration of health information. Consider the following narrative medical history as an example of a personalized Health Profile:

"/ am a 46 year old African American man. I have a history of diabetes, hypertension and depression. I am currently taking Avandia, insulin, Atenolol and Paxil for these conditions. I am not allergic to any medications. My father passed away from a heart attack when he was 50 years old. My mother is in good health with no major medical problems. I have an older brother had a heart attack when he was 40 years old. "

Even though this is a short description, it contains a wealth of medical information that can be used to help access the best medical information from reliable, credible sources.

Consider the following vignettes:

1. Search for "heart attack":

a. Without the Health Attribute Profile the results of such a search does not reflect items of interest based on the user's African American race, gender and co-morbidities. Lack of synonym and hierarchical expansion limits breadth of material that can be returned in search.

b. With a Health Attribute Profile, his profile as an African American man with a history of diabetes leads him to an increased risk of an atypical presentation of an acute coronary syndrome. His health attribute profile will re-order the search results to rank higher articles highlighting his increased risk of an atypical presentation of this disease. Synonym expansion and parentage (as described in published PCT application no. WO07035912) of the term heart attack to acute myocardial infarction will bring forth additional highly relevant material. Hierarchical understanding of heart attack/AMI as a child of acute coronary syndrome will bring forth additional highly relevant material. 2. World Wide Web search for "Robitussin":

a. Without the Health Attribute Profile the possibility of a drug-drug interaction may be completely missed since there is no reference of patient's current medication usage in the absence of the user health attribute profile.

b. With a Health Attribute Profile there is the potential for the development of serotonin syndrome as a result of an adverse interaction between dextromethorphan (active ingredient in Robitussin) and Paxil. This interaction may be particularly pronounced in some African Americans with a particular cytochrome phenotype. His Health Attribute Profile, reflecting his current medication use and race, will reorganize the search results to highlight this potentially fatal interaction. Understanding of drug concepts will highlight the ingredients of multi-drug OTC and prescription products. This will bring forth additional highly relevant material in the search results.

3. News and Alerts:

a. Without the Health Attribute Profile news and alerts are presented without relevance to patient's medical circumstances, demographics and co-morbidities.

b. With a Health Attribute Profile, African Americans with acute myocardial infarction (AMI) receive poorer quality care and have an increased risk of mortality. News regarding heart attacks relevant to the patient's demographic and co- morbidities will be presented. Synonym expansion of the term heart attack to acute myocardial infarction will bring forth additional highly relevant material. Hierarchical understanding of heart attack/AMI as a child of acute coronary syndrome will bring forth additional highly relevant material.

4. Health information and advertisements:

a. Without the Health Attribute Profile ads are based strictly on the keyword without respect to the patient's market segmentation. Ads are not altered and may actually indicate dangerous products for this patient.

b. With a Health Attribute Profile certain antihypertensive agents (i.e. diuretics) have been shown to have increased efficacy in African American populations. Advertisements can be tailored based on market segmentation of the patient's race, gender, age, co-morbidities and current medications. In searches for "Robitussin", advertisements for alternative anti-tussive agents without dextromethorphan can be presented.

Example 3: Using a User Health Attribute Profile to find Health Services and Collaborating with Health Professionals

Patients derive support and empowerment through social networks of individuals with similar medical problems and concerns. Unfortunately, disclosing sensitive medical information and finding those individuals with similar medical issues remains a challenge. This process of information disclosure and the identification of reliable, authentic social networks is enabled through the Health Profile.

Users of the Health Profile can choose which medical information they wish to disclose to outside parties. For instance, an individual with HIV may choose to exclude disclosure of their HIV status. This information then could be excluded from a Health Profile used for social networking. This Health Profile would reflect all their pertinent health attributes, except any areas of self-chosen privacy. This confidential health attribute profile then could be confidently and securely shared to identify other members with similar health issues and concerns. Individuals would be matched against the richness of their entire medical history, not just on the basis of a single disease or medication. Once a similar network has been found, the users would always have the option of deciding to engage with this community or decline. Interaction is not forced but is always voluntary and in accord with an individual's privacy concerns.

A similar mechanism also could be used to match patients with appropriate care providers. Currently, patients rely largely on word-of-mouth to find their health care providers. Rudimentary on-line tools exist to find physicians, but they lack substantial depth or information on expertise. The Health Profile can remedy this information failure.

This process is initiated by associating health providers and institutions with the Health Profiles of all of the patients under their care. This allows us to generate a composite health attribute profile that can be affiliated with any given health care provider. This composite profile, reflecting the medical history of an entire population, then can be matched against any individual Health Profile. This allows a patient to find providers handling patients of similar demographic profile and medical complexity. The present application describes a system that adapts, refines, organizes and enhances access to information based on a combination of personal and group attributes, preferences and prior behavior for individuals and groups. In one embodiment, the invention relates to a system and method for providing and using a computer generated representation of one or more classes of attributes that function either singly or in various combinations as a proxy for aspects and characteristics of a personal or group identity/classification or preference. Each set of attributes accessed can be from a variety of domain-aware classes such as medical history, professional background, interests and hobbies, demographics, geographic location, shopping preferences, etc. The attributes may be specific as in diseases such as hypertension, body-mass index and types Il diabetes and drugs such as Cozaar™ and insulin or may be general in the form of diabetes, exercise, yoga, overweight, etc. The form of the computer representation of these attributes can be generated automatically from detected user or group behavior in distinct situations using machine inference, from interaction with other systems or services or can be generated, appended or amended through known methods of editing, incorporating and appending through direct interaction. In one aspect, the invention can employ domain- aware classifications and definitions to extract relevant terms, ideas, meanings, and messages, which can be expressed as a computer representation or fingerprint, such that subsequent behavior by the user or group may be precisely determined and adapted to a user's or group's attributes. Details of the fingerprint can be found in published PCT application no. WO07035912 One application of this invention is modifying the response of internet based technologies and applications to the attributes of the user or group.

The present invention has now been described in connection with a number of specific embodiments thereof. However, numerous modifications which are contemplated as falling within the scope of the present invention should now be apparent to those skilled in the art. Therefore, it is intended that the scope of the present invention be limited only by the scope of the claims appended hereto. In addition, the order of presentation of the claims should not be construed to limit the scope of any particular term in the claims.

What is claimed is:

Patent Citations
Cited PatentFiling datePublication dateApplicantTitle
US20030097280 *24 Oct 200222 May 2003Fitzpatrick Constance Y.Health information gathering system
US20030208381 *29 Mar 20016 Nov 2003Walter Ervin DennisPatient health record access system
US20070276702 *25 May 200729 Nov 2007Sandeep DaniSystem and Method for Collaboration and Communication in Health Management
Classifications
International ClassificationG06Q10/00, G06Q50/00, G06F17/30
Cooperative ClassificationG06F19/324, G06Q50/22, G06Q10/06, G06F19/322, G06F19/3443, G06F19/326
European ClassificationG06Q10/06, G06Q50/22, G06F19/32C, G06F19/34J, G06F19/32E
Legal Events
DateCodeEventDescription
17 Feb 2010121Ep: the epo has been informed by wipo that ep was designated in this application
Ref document number: 09762913
Country of ref document: EP
Kind code of ref document: A2
11 Dec 2010NENPNon-entry into the national phase in:
Ref country code: DE