System and method for automatic generation of a hierarchical tree network and the use of two complementary learning algorithms, optimized for each leaf of the hierarchical tree network

• US 20050240544A1
• Filed: 04/26/2005
• Published: 10/27/2005
• Est. Priority Date: 04/27/2004
• Status: Active Grant

• Current Assignee: Humana Incorporated
• Original Assignee: Humana Incorporated
• Inventors: Kil, David H., Pottschmidt, David B.
• Granted Patent: US 7,725,329 B2
• Time in Patent Office: Days
• US Class Current: 706/45
• CPC Class Codes: G06Q 10/10 Office automation; Time management
• 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 50/70 for mining of medical data, e.g. analysing previous cases of other patients
• CPC Class Codes: G16H 70/40 relating to drugs, e.g. their side effects or intended usage
• CPC Class Codes: G16Z 99/00 Subject matter not provided for in other main groups of this subclass

Abstract

A system and method that generates a hierarchical tree network and uses linear-plus-nonlinear learning algorithms to form a consensus view on a member'"'"'s future health status. Each leaf in the hierarchical tree network is homogeneous in clinical characteristics, experience period, and available data assets. Optimization is performed on each leaf so that features and learning algorithms can be tailored to local characteristics specific to each leaf.

Claims

1. A method for predicting a person'"'"'s future health status, comprising:

a. establishing a hierarchical tree network that, for a plurality of members, assigns each member to at most only one of a plurality of nodes as a function of a plurality of stratification variables;

b. providing to a computer based system, for said plurality of members and for each said member, member demographic data, available member medical claim data, and available member pharmacy claim data;

c. performing feature selection for each of said plurality of nodes to identify for each said node an optimal subset of features from a set comprising at least some of said member demographic data, said available member medical claim data, and said available member pharmacy claim data for all members assigned to that said node;

d. training a MVLR algorithm and a BRN algorithm using at least some of said member demographic data, said available member medical claim data, and said available member pharmacy claim data and storing learned parameters in a database to create a learned parameter database;

e. using the learned parameter database and, for at least one said member, using that at least one said member'"'"'s member demographic data, said available member medical claim data, and said available member pharmacy claim data, calculating a MVLR future health status score using said MVLR algorithm and calculating a BRN future health status score using said BRN algorithm, and calculating an arithmetic mean of said MVLR future health status score and said BRN future health status score to determine a final score.

2. The method for predicting a person'"'"'s future health status of claim 1, where, after determining the final score, the method further includes the step of producing a member clinical condition report, said member clinical condition report including member identification information and said final score.

3. The method for predicting a person'"'"'s future health status of claim 2, where said member clinical condition report further includes said member'"'"'s clinical condition data and at least a portion of any of said member'"'"'s medical claim information and pharmacy claim information.

4. The method for predicting a person'"'"'s future health status of claim 1, where in the step of establishing a hierarchical tree network, the plurality of stratification variables include at least some of a member'"'"'s enrollment duration, a member'"'"'s presence or lack of a catastrophic condition, a member'"'"'s presence or lack of a diabetes flag, a member having a prescription trigger or an inpatient admission trigger, and a members claim data.

5. The method for predicting a person'"'"'s future health status of claim 4, where, if said member'"'"'s enrollment duration is less than six months, said member will be assigned to one of a first set of nodes;

and, if said member'"'"'s enrollment duration is at least six months, said member will be assigned to one of a second set of nodes.

6. The method for predicting a person'"'"'s future health status of claim 5, where, if said member'"'"'s enrollment duration is less than six months and if said member has the presence of a catastrophic condition, said member will be assigned to a first subset of said first set of nodes;

where, if said member'"'"'s enrollment duration is less than six months and if said member has the lack of a catastrophic condition, said member will be assigned to one of said first set of nodes not in said first subset of said first set of nodes, said one of said first set of nodes not in said first subset of said first set of nodes comprising a second subset of said first set of nodes;

where, if said member'"'"'s enrollment duration is at least six months and if said member has the presence of a catastrophic condition, said member will be assigned to a first subset of said second set of nodes;

where, if said member'"'"'s enrollment duration is at least six months and if said member has the lack of a catastrophic condition, said member will be assigned to one of said second set of nodes not in said first subset of said second set of nodes, said one of said second set of nodes not in said first subset of said second set of nodes comprising a second subset of said second set of nodes.

7. The method for predicting a person'"'"'s future health status of claim 6, where, if said member'"'"'s enrollment duration is less than six months and if said member has the lack of a catastrophic condition and if said member has the presence of a diabetic flag and if said member has a prescription trigger, said member is assigned to a first subsubset of said second subset of said first set of nodes;

where, if said member'"'"'s enrollment duration is less than six months and if said member has the lack of a catastrophic condition and if said member has the presence of a diabetic flag and if said member has an inpatient admission trigger, said member is assigned to a second subsubset of said second subset of said first set of nodes;

where, if said member'"'"'s enrollment duration is less than six months and if said member has the lack of a catastrophic condition and if said member has the lack of a diabetic flag and if said member has a prescription trigger, said member is assigned to a third subsubset of said second subset of said first set of nodes;

where, if said member'"'"'s enrollment duration is less than six months and if said member has the lack of a catastrophic condition and if said member has the lack of a diabetic flag and if said member has an inpatient admission trigger, said member is assigned to a fourth subsubset of said second subset of said first set of nodes;

where, if said member'"'"'s enrollment duration is at least six months and if said member has the lack of a catastrophic condition and if said member has the presence of a diabetic flag and if said member has a prescription trigger, said member is assigned to a first subsubset of said second subset of said second set of nodes;

where, if said member'"'"'s enrollment duration is at least six months and if said member has the lack of a catastrophic condition and if said member has the presence of a diabetic flag and if said member has an inpatient admission trigger, said member is assigned to a second subsubset of said second subset of said second set of nodes;

where, if said member'"'"'s enrollment duration is at least six months and if said member has the lack of a catastrophic condition and if said member has the lack of a diabetic flag and if said member has a prescription trigger, said member is assigned to a third subsubset of said second subset of said second set of nodes; and

, where, if said member'"'"'s enrollment duration is at least six months and if said member has the lack of a catastrophic condition and if said member has the lack of a diabetic flag and if said member has an inpatient admission trigger, said member is assigned to a fourth subsubset of said second subset of said second set of nodes.

8. The method for predicting a person'"'"'s future health status of claim 7, where said first subset of said first set of nodes, said first subsubset of said second subset of said first set of nodes, said second subsubset of said second subset of said first set of nodes, said third subsubset of said second subset of said first set of nodes, said fourth subsubset of said second subset of said first set of nodes, said first subset of said second set of nodes, said first subsubset of said second subset of said second set of nodes, said second subsubset of said second subset of said second set of nodes, said third subsubset of said second subset of said second set of nodes, and said fourth subsubset of said second subset of said second set of nodes, each comprise a first node where said member has both pharmacy and medical claim data, a second node where said member has medical claim data only, and a third node where said member has pharmacy claim data only.

9. The method for predicting a person'"'"'s future health status of claim 4, where said member will have the presence of a catastrophic condition if said member has at least one of any cancer, an end-stage disease, a transplant, a rare disease, HIV, a combination of coronary artery disease plus chronic heart failure plus hypertension.

10. The method for predicting a person'"'"'s future health status of claim 1, where said plurality of nodes of said hierarchical tree network comprises thirty nodes,

a. a first node for any member with an enrollment duration of less than six months, the presence of a catastrophic condition, and both pharmacy and medical claim data;

b. a second node for any member with an enrollment duration of less than six months, the presence of a catastrophic condition, and pharmacy claim data only;

c. a third node for any member with an enrollment duration of less than six months, the presence of a catastrophic condition, and medical claim data only;

d. a fourth node for any member with an enrollment duration of less than six months, the lack of a catastrophic condition, the presence of a diabetes trigger, the presence of a prescription trigger, and both pharmacy and medical claim data;

e. a fifth node for any member with an enrollment duration of less than six months, the lack of a catastrophic condition, the presence of a diabetes trigger, the presence of a prescription trigger, and pharmacy claim data only;

f. a sixth node for any member with an enrollment duration of less than six months, the lack of a catastrophic condition, the presence of a diabetes trigger, the presence of a prescription trigger, and medical claim data only;

g. a seventh node for any member with an enrollment duration of less than six months, the lack of a catastrophic condition, the presence of a diabetes trigger, the presence of an inpatient admission trigger, and both pharmacy and medical claim data;

h. an eighth node for any member with an enrollment duration of less than six months, the lack of a catastrophic condition, the presence of a diabetes trigger, the presence of an inpatient admission trigger, and pharmacy claim data only;

i. a ninth node for any member with an enrollment duration of less than six months, the lack of a catastrophic condition, the presence of a diabetes trigger, the presence of an inpatient admission trigger, and medical claim data only;

j. a tenth node for any member with an enrollment duration of less than six months, the lack of a catastrophic condition, the lack of a diabetes trigger, the presence of a prescription trigger, and both pharmacy and medical claim data;

k. an eleventh node for any member with an enrollment duration of less than six months, the lack of a catastrophic condition, the lack of a diabetes trigger, the presence of a prescription trigger, and pharmacy claim data only;

l. a twelfth node for any member with an enrollment duration of less than six months, the lack of a catastrophic condition, the lack of a diabetes trigger, the presence of a prescription trigger, and medical claim data only;

m. a thirteenth node for any member with an enrollment duration of less than six months, the lack of a catastrophic condition, the lack of a diabetes trigger, the presence of an inpatient admission trigger, and both pharmacy and medical claim data;

n. a fourteenth node for any member with an enrollment duration of less than six months, the lack of a catastrophic condition, the lack of a diabetes trigger, the presence of an inpatient admission trigger, and pharmacy claim data only;

o. a fifteenth node for any member with an enrollment duration of less than six months, the lack of a catastrophic condition, the lack of a diabetes trigger, the presence of an inpatient admission trigger, and medical claim data only;

p. a sixteenth node for any member with an enrollment duration of at least six months, the presence of a catastrophic condition, and both pharmacy and medical claim data;

q. a seventeenth node for any member with an enrollment duration of at least six months, the presence of a catastrophic condition, and pharmacy claim data only;

r. an eighteenth node for any member with an enrollment duration of at least six months, the presence of a catastrophic condition, and medical claim data only;

s. a nineteenth node for any member with an enrollment duration of at least six months, the lack of a catastrophic condition, the presence of a diabetes trigger, the presence of a prescription trigger, and both pharmacy and medical claim data;

t. a twentieth node for any member with an enrollment duration of at least six months, the lack of a catastrophic condition, the presence of a diabetes trigger, the presence of a prescription trigger, and pharmacy claim data only;

u. a twenty-first node for any member with an enrollment duration of at least six months, the lack of a catastrophic condition, the presence of a diabetes trigger, the presence of a prescription trigger, and medical claim data only;

v. a twenty-second node for any member with an enrollment duration of at least six months, the lack of a catastrophic condition, the presence of a diabetes trigger, the presence of an inpatient admission trigger, and both pharmacy and medical claim data;

w. a twenty-third node for any member with an enrollment duration of at least six months, the lack of a catastrophic condition, the presence of a diabetes trigger, the presence of an inpatient admission trigger, and pharmacy claim data only;

x. a twenty-fourth node for any member with an enrollment duration of at least six months, the lack of a catastrophic condition, the presence of a diabetes trigger, the presence of an inpatient admission trigger, and medical claim data only;

y. a twenty-fifth node for any member with an enrollment duration of at least six months, the lack of a catastrophic condition, the lack of a diabetes trigger, the presence of a prescription trigger, and both pharmacy and medical claim data;

z. a twenty-sixth node for any member with an enrollment duration of at least six months, the lack of a catastrophic condition, the lack of a diabetes trigger, the presence of a prescription trigger, and pharmacy claim data only;

aa. a twenty-seventh node for any member with an enrollment duration of at least six months, the lack of a catastrophic condition, the lack of a diabetes trigger, the presence of a prescription trigger, and medical claim data only;

bb. a twenty-eighth node for any member with an enrollment duration of at least six months, the lack of a catastrophic condition, the lack of a diabetes trigger, the presence of an inpatient admission trigger, and both pharmacy and medical claim data;

cc. a twenty-ninth node for any member with an enrollment duration of at least six months, the lack of a catastrophic condition, the lack of a diabetes trigger, the presence of an inpatient admission trigger, and pharmacy claim data only; and

,

dd. a thirtieth node for any member with an enrollment duration of at least six months, the lack of a catastrophic condition, the lack of a diabetes trigger, the presence of an inpatient admission trigger, and medical claim data only.

11. The method for predicting a person'"'"'s future health status of claim 10, where, after determining the final score, the method further includes the step of producing a member clinical condition report, said member clinical condition report including member identification information and said final score.

12. The method for predicting a person'"'"'s future health status of claim 11, where said member clinical condition report further includes said member'"'"'s clinical condition data and at least a portion of any of said member'"'"'s medical claim information and pharmacy claim information.

13. The method for predicting a person'"'"'s future health status of claim 10, where said member will have the presence of a catastrophic condition if said member has at least one of any cancer, an end-stage disease, a transplant, a rare disease, HIV, a combination of coronary artery disease plus chronic heart failure plus hypertension.