System for delivering scenario specific, problem solving, decision support from non-intelligent computer systems

• US 7,493,299 B2
• Filed: 08/01/2002
• Issued: 02/17/2009
• Est. Priority Date: 02/06/2000
• Status: Expired due to Fees

• Current Assignee: Martin P. Entwistle
• Original Assignee: Palo Alto Medical Foundation For Health Care Research and Education (Sutter Health)
• Inventors: Entwistle, Martin Patrick
• Primary Examiner(s): STARKS, WILBERT L
• Application Number:
  US10/211,782
• Publication Number:
  US 20030037018A1
• Time in Patent Office: 2,392 Days
• Field of Search: 706/50, 706/15, 706/16, 706/23, 706/25, 706/26, 706/17, 706/45, 707/103.R, 435/6, 380/282, 600/300
• US Class Current: 706/50
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 5/022 Knowledge engineering; Knowledge acquisition
• CPC Class Codes: G06Q 10/10 Office automation; Time management

Abstract

A computational decision making system is provided for delivering scenario specific information. Each decision making scenario has a number of variables associated with it and variable values. The system matches a set of input variables to the grouped data objects to find a match or the closest matching grouped data object or objects. Decision information embodying the outcome of a decision making process is associated with each grouped data object and the said decision information associated with the best matching grouped data object or objects is outputted.

Claims

1. A computer system comprising:

a matching function module to compute a match or closest match between a set of inputted variable information with a plurality of predefined sets of variables, each set of variables associates with a predefined clinical scenario for a patient with a specific set of medical conditions, each variable being defined by a parameter to be measured or described and by a value of the parameter, wherein each predefined set of variables is encapsulated and retained by the system as a grouped data object and wherein specific clinical management decision information is selected for output dependent on which grouped data object or objects include variables that match or most closely match the inputted set of variable information and wherein the match or closest match is computed by means of multivariable pattern matching or filtering;

a data function module to create and retain further data objects defining additional predefined information in response to the at least one of the selected specific clinical management decision information and inputted variables; and

an output module to output the specific clinical management decision information to a user.

2. The computer system of claim 1 wherein the clinical management decision information output to a user includes a recommendation.

3. A method of performing a patient specific clinical management decision making process using a computational system, the method comprising:

identifying a plurality of discrete data sets of specific clinical management decision information;

identifying a plurality of decision variables and their values which are relevant as affecting the outcome of a clinical management decision making process in relation to each of the discreet data sets, each variable being defined by a parameter to be measured or described and by a value of the parameter;

collecting the decision variables into one or more computer readable, logically grouped decision data objects;

creating a clinical medicine decision knowledge base, said knowledge base containing the decision data objects and a set of corresponding discrete data sets of specific clinical management decision information embodying the outcome of the decision making process;

receiving, from an input device, input data representative of the values of a plurality of input variables associated with a predefined clinical scenario for a patient with a specific set of medical conditions, wherein the input variables correspond to the decision variables;

structuring the input data to form an input data object that has a structure complementary to the structure of the decision data objects;

using a matching function to compute a best matching decision data object to the input data object using multivariable pattern matching or filtering to compare the decision variables and input variables associated with the decision data object and input data object and identifying the corresponding discrete data set of specific clinical management decision information to the best matching decision data object or objects; and

outputting the discrete data set to a user.

4. The method of performing a decision making process of claim 3, wherein said decision data object is structured and handled according to object-oriented or object-relational methodologies.

5. The method of performing a decision making process of claim 3, wherein if a plurality of equally best matching decision data objects are identified, the method further includes the step of requesting the input of at least one additional input variable and computing which decision data object or objects best match the input data object comprising the at least one additional input variable.

6. The method of performing a decision making process of claim 3, further including passing the input data object to a clinical medicine decision knowledge base, wherein the knowledge base is structured so that when the knowledge base is queried using an input data object, a further data object or objects are returned containing relevant clinical management decision information.

7. The method of performing a decision making process of claim 3, wherein each discrete data set corresponds to one of a particular medical diagnosis or condition, and a combination of medical diagnoses or conditions.

8. The method of performing a decision making process of claim 3 further including providing a recommendation with the output to a user.

9. A computer system comprising:

a clinical medicine decision knowledge base module comprising a plurality of decision data objects, wherein each of said decision data objects includes a plurality of decision variables, each variable being defined by a parameter to be measured or described and by a value of the parameter and wherein said knowledge base includes a set of specific clinical management decision information embodying the outcome of the decision making process corresponding to each decision data object;

a data input device to input data indicative of the value of a plurality of input variables of the same type as the plurality of decision variables, the input variables being associated with a predefined clinical scenario for a patient with a specific set of medical conditions;

a processor; and

a computer readable medium comprising instructions to cause the processor to structure the input variables as a grouped input data object, compute a matching or closest matching decision data object or objects to the input data object using multivariable pattern matching or filtering to perform a comparison of input variables with decision variables and output specific clinical management decision information corresponding to the matching or closest matching decision data object or objects.

10. The computer system of claim 9, wherein the set of decision information is structured as a plurality of data objects.

11. The computer system of claim 9, wherein the decision information is stored in a hierarchy, whereby selected data portions of the decision information are linked to further detailed data, and the further detailed data is selectively outputted or selectively extracted for output.

12. The computer system of claim 9, wherein the decision information includes a media selected from the group consisting of text, tables, graphics, sounds, and combinations thereof.

13. The computer system of claim 9, wherein the clinical medicine decision knowledge base module contains distinct decision data objects having similar variable values or overlapping variable ranges, wherein each set of specific clinical management decision information corresponding to each decision data object having similar variable values or overlapping variable ranges do not conflict with each other.

14. The computer system of claim 9, wherein if a plurality of equally matching decision data objects are computed, an output module of the system outputs a request for at least one additional input variable and computes which decision data object or objects best matches the input data object comprising the at least one additional input variable.

15. The computer system of claim 9, wherein each decision data object belongs to a class, wherein each class is determined by the type or value of variables required to identify a match with the decision data objects and wherein the knowledge base is structured hierarchically according to class.

16. The computer system of claim 9, wherein the clinical medicine decision knowledge base module is implemented using any one or a combination of object, object-relational and relational bases having an appropriate structure.

17. The computer system of claim 9, wherein the decision making system is used for one of a medical diagnosis or condition, and a combination of medical diagnoses or conditions.

18. The computer system of claim 9, wherein the data input device extracts the input variable values from an input form to create the input data object or objects.

19. The computer system of claim 9, wherein the output is transmitted or interrogated by one or more third party applications.

20. The computer system of claim 9, wherein the system includes an editorial tool module adapted to allow the input, management, update and customization of the knowledge base.

21. The computer system of claim 20, wherein the editorial tool organizes information in the knowledge base module so that the decision data object fits the object structure which is defined for each input form.

22. The computer system of claim 9 wherein the clinical management decision information output to a user includes a recommendation.

23. A computer implemented patient specific clinical management decision method on computer readable medium comprising instructions to cause a computer to:

identify a plurality of discrete data sets of specific clinical management decision information;

identify a plurality of decision variables which are identified as relevant in affecting the outcome of a clinical management decision making process in relation to each discrete data set, each variable being defined by a parameter to be measured or described and by a value of the parameter;

collect the decision variables into one or more computer readable logically grouped decision data objects, said one or more decision data objects structured and handled according to object oriented or object relational methodologies;

create a clinical medicine decision knowledge base, said knowledge base containing the decision data objects and a set of corresponding specific clinical management decision information embodying the outcome of the decision making process;

provide a data input means suitable for receiving data indicative of the value of a plurality of input variables associated with a predefined clinical scenario for a patient with a specific set of medical conditions; and

provide a computer readable instruction set suitable for causing a computer to structure the input variables as a grouped input data object, compute a matching or closest matching decision data object or objects to the input data object by multivariable pattern matching or filtering using a matching function and output the corresponding specific clinical management decision information to the matching or closest matching decision data object or objects to a user.

24. The computer implemented method of claim 23, wherein the method includes structuring and handling the decision data objects in a computational environment according to object oriented or object relational methodologies.

25. The computer implemented method of claim 23 further including providing a recommendation with the output to a user.