End of life predictive model

• US 8,244,654 B1
• Filed: 12/19/2008
• Issued: 08/14/2012
• Est. Priority Date: 10/22/2007
• Status: Active Grant

• Current Assignee: Healthways SC, LLC (Warner Bros. Discovery, Inc.)
• Original Assignee: Healthways Incorporated
• Inventors: Hobgood, Adam, Montijo, Michael, Hamar, G. Brent, Dobbs, Angela, Hamlet, Karen, McGinnis, Matthew
• Primary Examiner(s): Rivas, Omar Fernandez
• Application Number:
  US12/340,491
• Time in Patent Office: 1,334 Days
• Field of Search: None
• US Class Current: 706/21
• CPC Class Codes: G06N 3/02 Neural networks
• CPC Class Codes: G06Q 40/08 Insurance
• CPC Class Codes: G16H 50/20 for computer-aided diagnosis, e.g. based on medical expert systems
• CPC Class Codes: G16H 50/30 for calculating health indices; for individual health risk assessment
• CPC Class Codes: G16H 50/50 for simulation or modelling of medical disorders
• CPC Class Codes: G16H 50/70 for mining of medical data, e.g. analysing previous cases of other patients

Abstract

A statistical processing system to predict end of life status in current patients of a given population. The processing system includes a server configured with programming instructions implementing a plurality of statistical models to predict a health status or outcome for the highest risk of death based on proprietary factors and the patient being administered no further therapeutic treatment.

Claims

1. A system for identifying patients needing healthcare, comprising:

a first module, executed by a processor, for preparation of patient data for input to statistical modeling submodules;

a second module, executed by the processor, for receiving and evaluating outcome data of the statistical modeling submodules and establishing one or more predictive values meeting a predetermined criteria and applying the one or more predictive values to a first set of neural net models; and

a third module, executed by the processor, configured to receive and apply one or more results from the first set of neural net models to a second set of neural net models and generate a results set indicative of patients needing healthcare,

wherein the patient data comprises data from a pre-determined population set.

2. The system of claim 1, wherein the patient data is government health data, insurance healthcare data, or commercial healthcare data.

3. The system of claim 1, wherein the predetermined criteria is the risk of death in a predetermined time frame.

4. The system of claim 1, wherein the results set identifies patients eligible for or needing end of life care.

5. The system of claim 1, wherein the system identifies patients eligible for or needing hospice care or other end of life care.

6. The system of claim 1, wherein the system identifies patients who are predicted to die within a predetermined time frame.

7. A system for establishing factors for a healthcare predictive model, comprising:

a processor; and

a storage in data communication with the processor, such storage configured to store;

a set of distinct variables used for statistical analysis, wherein the variables are representative of medical conditions of patients;

a first module configured to identify a plurality of individual factors predictive of a certain outcome, based on a set of distinct variables;

a second module configured to identify combination sets of the individual factors and test the combination sets for performance against one or more population sets;

a third module configured to determine one or more combination sets meeting a criteria, wherein the third module comprises one or more neural networks; and

a fourth module configured to apply a selected neural network against medical condition data of a pre-determined population of patients.

8. The system of claim 7, wherein the medical condition data comprises government health data, insurance data, or commercial healthcare data.

9. The system of claim 7, wherein the criteria to determine the selection of combination sets comprises identifying combination sets with a high correlation between patients who were predicted to have a high risk of death within a predetermined period of time and for which those patients did die.

10. The system of claim 7, wherein the system identifies patients eligible for or needing hospice care or other end of life care.

11. The system of claim 7, wherein the system identifies patients who are predicted to die within a predetermined time frame.

12. The system of claim 7, wherein the system is additionally comprised of a testing module.

13. The system of claim 7, wherein the system is additionally configured to predict patient death within a pre-determined timeframe.

14. The system of claim 7, wherein one of the factors comprises either a count of distinct durable medical equipment claims or a total amount paid for home health claims.

15. The system of claim 7, wherein one of the factors comprises either a natural log of the payment for diagnosis related groups, or a natural log of skilled nursing facility charges.

16. The system of claim 7, wherein one of the factors comprises a count of skilled nursing facility claims.

17. The system of claim 7, wherein one of the factors comprise a factor selected from the group consisting of a count of skilled nursing facility claims, a count of distinct hospital visits with congestive heart failure diagnosis, a count of distinct claims with pain-related diagnoses, a member age, and a total amount paid for physician claims.

18. The system of claim 7, wherein the factors comprises a count of skilled nursing facility claims and at least one factor selected from the group consisting of a count of distinct hospital visits with congestive heart failure diagnosis, a count of distinct claims with pain-related diagnoses, a member age, and a total amount paid for physician claims.

19. The system of claim 7, wherein one of the factors comprises a factor selected from the group consisting of a count of distinct durable medical equipment claims, a total amount paid for home health claims, a natural log of the payment for diagnosis related groups and a natural log of skilled nursing facility charges.

20. A method of identifying patients in need or eligible for hospice care or other end of life care, the method comprising:

accessing, with a processor, a set of data stored in a storage, such data comprising patient healthcare for a plurality of patients over a predetermined time period;

organizing the set of data with the processor;

applying a predictive model, with the processor, to the set of data so as to identify any patients at risk of death in a next predetermined time period; and

contacting patients having an estimated end of life status within a predetermined threshold to offer the patient end of life care.

21. The method of claim 20, wherein the set of data is insurance claim data.

22. The method of claim 20, wherein the set of data is government health data.

23. The method of claim 20, further comprising determining, based on the predictive model, that an appropriate action for at least one patient is to be admitted to a hospice.

24. The method of claim 20, further comprising determining, based on the predictive model, that an appropriate action for at least one patient is to be administered interventions related to end of life care.

25. The method of claim 20, wherein the predetermined time period is a year, and wherein the predictive model estimates end of life risk by inputting patient data into a model configured to utilize factors that show the highest correlation of predicting death in a patient and wherein the patient did die within the proscribed time parameters.

26. The method of claim 20, wherein the predetermined time period is in a range of one day to about one year.

27. A computer implemented system for identifying patients needing end-of-life care, the system comprising:

a computing environment;

a storage in data communication with the computing environment and configured to store patient data and a set of distinct factors;

a software program operating on the computing environment and configured to;

identify combination sets of individual factors predictive of a certain outcome and that have a high correlation in the patient data with the target outcome;

execute a plurality of neural networks using the combination sets of individual factors;

select one of the plurality of the neural networks based on the results of executing the neural networks;

apply the selected neural network to the patient data;

output a list of patients at risk for end of life within a predetermined period of time; and

recommend end of life care for a group of patients identified as being at the end of life.

28. The system of claim 27, wherein one of the factors comprises either a count of distinct durable medical equipment claims or a total amount paid for home health claims.

29. The system of claim 27, wherein one of the factors comprises either a natural log of the payment for diagnosis related groups, or a natural log of skilled nursing facility charges.

30. The system of claim 27, wherein one of the factors comprises a count of skilled nursing facility claims.

31. The system of claim 27, wherein one of the factors comprise a factor selected from the group consisting of a count of skilled nursing facility claims, a count of distinct hospital visits with congestive heart failure diagnosis, a count of distinct claims with pain-related diagnoses, a member age, and a total amount paid for physician claims.

32. The system of claim 27, wherein the factors comprises a count of skilled nursing facility claims and at least one factor selected from the group consisting of a count of distinct hospital visits with congestive heart failure diagnosis, a count of distinct claims with pain-related diagnoses, a member age, and a total amount paid for physician claims.

33. The system of claim 27, wherein one of the factors comprises a factor selected from the group consisting of a count of distinct durable medical equipment claims, a total amount paid for home health claims, a natural log of the payment for diagnosis related groups and a natural log of skilled nursing facility charges.

34. A server storing software that, when executed, performs the acts of:

accessing and organizing a set of data comprising patient healthcare data for a plurality of patients over a predetermined time period;

and contacting patients having an estimated end of life status within a predetermined threshold to offer the patient end of life care.

35. A method of incrementally updating a set of factors for input to a predictive modeling program wherein the set of factors are indicative of patient healthcare records, the method comprising:

choosing, with a processor, a first set of factors for input into the predictive modeling program;

receiving, with the processor, output from the predictive modeling program, such output identifying patients at risk for death in a predefined period;

determining outcome data, with the processor, after the predefined time period has completed, which patients did die;

comparing the first set of factors with the outcome data;

determining, with the processor, which of the first set of factors achieved a greatest accuracy;

adjusting the first set of factor based on the outcome data periodically; and

recommending end of life care for a group of patients identified as being at the end of life.

36. The method of claim 35, wherein the adjustment of the factors is done annually.

37. The method of claim 35, wherein the factors comprise number of hospital stays and length of hospital stays.

38. The method of claim 35, wherein the predetermined period is used for predicting end of life within one year.

39. The method of claim 35, wherein one of the factors comprises a count of skilled nursing facility claims.

40. The method of claim 35, wherein the factors comprise one or more factors selected from the group consisting of a count of skilled nursing facility claims, a count of distinct hospital visits with congestive heart failure diagnosis, a count of distinct claims with select pain-related diagnoses, a member age as of end of year used in model, and a total amount paid for physician claims.

41. The method of claim 35, wherein the factors comprise a count of skilled nursing facility claims and at least one factor selected from the group consisting of a count of distinct hospital visits with congestive heart failure diagnosis, a count of distinct claims with select pain-related diagnoses, a member age as of end of year used in model, and a total amount paid for physician claims.

42. The method of claim 35, wherein one of the factors comprises a factor selected from the group consisting of a count of distinct durable medical equipment claims, a total amount paid for home health claims, a natural log of the payment for diagnosis related groups and a natural log of skilled nursing facility charges.

43. A system for predicting end of life for a patient within a predetermined time frame using a set of predictive models, the system comprising:

means for storing and executing a set of predictive models on a computing device;

means for identifying factors which have a correlation to end of life as input to the models;

means for updating the factors based on previous successful predictions of the models;

means for outputting a report identifying particular groups of patients meeting a certain criteria based on at least the identified factors; and

means for recommending end of life care for a group of patients identified as being at the end of life.

44. The system of claim 43, wherein the report outputting means comprises a spreadsheet, list or other report identifying specific patients at risk for death within a specified time frame.

45. The system of claim 43, wherein the predictive models comprise configured neural net models.

46. The system of claim 43, wherein one of the factors comprises either a count of distinct durable medical equipment claims or a total amount paid for home health claims.

47. The system of claim 43, wherein one of the factors comprises either a natural log of the payment for diagnosis related groups, or a natural log of skilled nursing facility charges.

48. The system of claim 43, wherein one of the factors comprises a count of skilled nursing facility claims.

49. The system of claim 43, wherein one of the factors comprise a factor selected from the group consisting of a count of skilled nursing facility claims, a count of distinct hospital visits with congestive heart failure diagnosis, a count of distinct claims with pain-related diagnoses, a member age, and a total amount paid for physician claims.

50. The system of claim 43 wherein the factors comprise a count of skilled nursing facility claims and at least one factor selected from the group consisting of a count of distinct hospital visits with congestive heart failure diagnosis, a count of distinct claims with pain-related diagnoses, a member age, and a total amount paid for physician claims.

51. The system of claim 43, wherein one of the factors comprises a factor selected from the group consisting of a count of distinct durable medical equipment claims, a total amount paid for home health claims, a natural log of the payment for diagnosis related groups and a natural log of skilled nursing facility charges.

52. A system for predicting end of life for a patient within a predetermined time frame from the output of a set of configured predictive models, the system comprising:

a storage component;

a software program configured to;

identify factors as input into the models;

update the factors based on previous successful predictions of the models;

output a report identifying a particular group of patients meeting a certain criteria based on at least the identified factors; and

recommend end of life care for a group of patients identified as being at the end of life.

53. The system of claim 52, wherein one of the factors comprises a count of skilled nursing facility claims.

54. The system of claim 52, wherein one of the factors comprise a factor selected from the group consisting of a count of distinct hospital visits with congestive heart failure diagnosis, a count of distinct claims with select pain-related diagnoses, a member age as of end of year used in model, and a total amount paid for physician claims.

55. The system of claim 52, wherein the factors comprises a count of skilled nursing facility claims and at least one factor selected from the group consisting of a count of distinct hospital visits with congestive heart failure diagnosis, a count of distinct claims with select pain-related diagnoses, a member age as of end of year used in model, and a total amount paid for physician claims.

56. The system of claim 52, wherein one of the factors comprises a factor selected from the group consisting of a count of distinct durable medical equipment claims, a total amount paid for home health claims, a natural log of the payment for diagnosis related groups and a natural log of skilled nursing facility charges.