Computer instructions for guiding differential diagnosis through information maximization

• US 7,877,272 B2
• Filed: 02/27/2008
• Issued: 01/25/2011
• Est. Priority Date: 03/02/2005
• Status: Active Grant

• Current Assignee: Cerner Innovation, Inc. (Oracle Corporation)
• Original Assignee: Siemens Medical Solutions USA Incorporated (Siemens AG)
• Inventors: Mueller, Marianne Larissa, Krishnan, Sriram, Rosales, Romer E., Rao, R. Bharat
• Primary Examiner(s): O'Connor; Gerald J.
• Assistant Examiner(s): Raj; Rajiv J
• Application Number:
  US12/038,169
• Publication Number:
  US 20080147445A1
• Time in Patent Office: 1,063 Days
• Field of Search: 705 2- 3
• US Class Current: 705/2
• CPC Class Codes: G16H 10/60 for patient-specific data, e.g. for electronic patient records
• CPC Class Codes: G16H 15/00 ICT specially adapted for medical reports, e.g. generation or transmission thereof
• CPC Class Codes: G16H 20/00 ICT specially adapted for therapies or health-improving plans, e.g. for handling prescriptions, for steering therapy or for monitoring patient compliance
• CPC Class Codes: G16H 50/20 for computer-aided diagnosis, e.g. based on medical expert systems
• CPC Class Codes: G16H 50/70 for mining of medical data, e.g. analysing previous cases of other patients
• CPC Class Codes: G16Z 99/00 Subject matter not provided for in other main groups of this subclass

Abstract

A system and method for diagnosis and treatment decisions based on information maximization is disclosed. Utilizing patient information as well as clinical records from other patients can reduce the uncertainty in both diagnosis and treatment options. The information maximization may consider additional data such as risk, cost, and comfort in making a proper medical decision.

Claims

1. In a non-transitory computer readable storage medium having stored therein data representing instructions executable by a programmed processor for determining at least one question of interest, the storage medium comprising instructions for:

providing patient observations including patient medical data;

providing clinical observations based on recorded clinical data;

identifying mutual information between the question of interest and the recorded clinical data by comparing the patient observations with the clinical observations to determine dependence between the question of interest and the recorded clinical data as it relates to the patient observations, wherein the dependence between the question of interest and the recorded clinical data indicates the mutual information;

selecting relevant observation data by maximizing the dependence between the recorded clinical data about the question of interest given the patient medical data, wherein the relevant observation data is unknown with respect to the patient but is relevant for analyzing the at least one question of interest based on the identification of the mutual information

determining a solution for the at least one question of interest that comprises a result derived from the relevant observation data, wherein the solution comprises a determination of the unknown from the relevant observation data such that the unknown with respect to the patient is known based on the result;

obtaining additional observations based on the determined solution, wherein the additional observations comprise results of an implementation of the determined solution; and

iteratively repeating the comparing, selecting, and determining after updating the patient medical data with the additional observations, wherein each iteration utilizes updated patient medical data that includes the additional observations from a previous iteration.

2. The instructions of claim 1 wherein the at least one question of interest is a medical approach.

3. The instructions of claim 2 wherein the medical approach comprises either a diagnosis of a medical condition or a test for a medical condition.

4. The instructions of claim 1 wherein the determining a solution comprises:

identifying a plurality of options;

generating a probability for each of the options, wherein the probability comprises a level of certainty that the option is a solution to the at least one question of interest; and

selecting one of the options based on the probabilities for the options, wherein the selected option comprises the determined solution.

5. The instructions of claim 1 wherein the instruction of determining a solution for the at least one question of interest comprises maximizing information for each of at least one observable variable and the question of interest, based on the relevant observation data and the patient observations.

6. In a non-transitory computer readable storage medium having stored therein data representing instructions executable by a programmed processor for making a medical decision regarding a particular patient and a medical question of interest, the storage medium comprising instructions for:

receiving patient information comprising data related to the particular patient;

receiving clinical data from pre-existing medical records, wherein the clinical data comprises data from a plurality of other patients;

determining, from the clinical data, mutual information between the medical question of interest and the clinical data based on a comparison of the patient information with clinical data from a subset of patients with similar conditions as the particular patient, wherein the mutual information comprises an unknown variable for the particular patient, but comprises a known variable from at least one of the plurality of other patients, further wherein the unknown variable assists with identifying the medical decision;

selecting relevant observation data by maximizing a dependence between the unknown variable for the particular patient and the known variable from at least one of the plurality of other patients in view of the patient information, wherein the relevant observation data is relevant for analyzing the medical question of interest based on the determination of the mutual information;

utilizing the relevant observation data for identifying the medical decision, wherein the medical decision comprises uncovering the unknown variable for the particular patient such that the patient information is updated to include results for the medical decision and the unknown variable becomes known based on the results;

obtaining additional observations based on the identified medical decision, wherein the additional observations comprise results of an implementation of the identified medical decision; and

iteratively repeating the determination, selection and utilization with the updated patient information, wherein the identified results for the medical decision are updated based on a previous iteration'"'"'s results.

7. The instructions of claim 6 wherein the medical decision comprises a potential medical test and the medical results comprises results from the potential medical test.

8. The instructions of claim 6 wherein the medical decision comprises a potential diagnosis and the results for the medical condition comprises test results for that potential diagnosis.

9. The instructions of claim 6 wherein the medical decision is a determination of breast cancer and the results for the medical decision comprise at least one of film mammography, digital mammography, ultrasound, PET scan, magnetic resonance imaging, and any combination thereof.

10. The instructions of claim 6 wherein the utilizing comprises calculating an uncertainty of the medical question of interest after obtaining results for the medical decision.

11. The instructions of claim 10 wherein the uncertainty is based on a maximization of the information provided by at least one possible medical decision about the medical question of interest.

12. The instructions of claim 11 wherein the identifying of the medical decision comprises:

considering a plurality of medical decisions related to the medical question of interest;

identifying observable variables in each of the plurality of medical decisions;

calculating, in consideration with the patient information, the observable variables for each of the plurality of medical decisions; and

selecting the medical decision based on a maximization of the mutual information using the observable variables.

13. The instructions of claim 12 wherein the maximization comprises determining which of the medical decisions maximize dependence with the medical question of interest.

14. The instructions of claim 12 wherein the maximization comprises an identification of mutual information with the medical question of interest.

15. The instructions of claim 12 wherein the iterative comparison comprises iteratively calculating and selecting the medical decision.

16. The instructions of claim 6 wherein the utilizing comprises calculating a benefit to the particular patient, wherein the benefit includes a consideration of risk, cost, effectiveness, and patient discomfort provided by the patient.

17. A computing system for reducing uncertainty of a medical condition for a particular patient, the system comprising:

a source of patient data, wherein the patient data comprises medical data from a patient history for the particular patient;

a source of clinical data, wherein the clinical data comprises a collection of patient data from other patients;

a memory coupled with the patient data source and the clinical data source that receives and stores the patient data and the clinical data;

a processor coupled with the memory and configured for;

identifying mutual information between the medical condition and the clinical data by comparing the clinical data from a subset of patients with similar conditions to the particular patient with the patient data, wherein the comparison determines a dependence between the medical condition and the clinical data as it relates to the patient data, wherein the dependence between the medical condition and the clinical data indicates the mutual information;

determining relevant clinical data with a maximization of the dependence between the medical condition and the clinical data in view of the patient data, wherein the relevant clinical data is unknown with respect to the particular patient, but is relevant for analyzing the medical condition of the particular patient based on the identification of the mutual information;

analyzing the relevant clinical data to determine a medical decision for the medical condition, wherein the medical decision comprises a determination of the unknown from the relevant clinical data with respect to the particular patient, such that the unknown with respect to the particular patient is known based on the medical decision;

updating the patient data based on the determined medical decision; and

repeating the comparing, determining, and analyzing, wherein the repeating comprises utilizing the updated patient data; and

a display coupled with the processor that provides the medical decision.

18. The system of claim 17, wherein the medical decision comprises a potential diagnosis of a medical condition.

19. The system of claim 17, wherein the medical condition comprises a potential medical test and the medical decision comprises results from the medical test.

20. The system of claim 19, wherein the medical condition is breast cancer and the medical decision comprises at least one of film mammography, digital mammography, ultrasound, PET scan, magnetic resonance imaging, and any combination thereof.

21. The system of claim 17, wherein the patient data comprises at least one of patient information, demographics, medical history, symptoms, test results, and any combination thereof.

22. The system of claim 17, wherein the medical decision is replaced with an alternative medical decision based on an uncertainty for the medical decision wherein the uncertainty of the medical decision is determined based on the relevant clinical data.

23. The system of claim 22, wherein the alternative medical decision replaces the medical decision and an uncertainty of the alternative medical decision is determined and updating by the repeating.