Computerized system and method for estimating levels of obesity in an insured population

Abstract

A computerized system and method for estimating levels of obesity in an insured population using claims data. The model uses health risk assessment data comprising age, height, and weight information as well as information about health conditions and health behaviors for a member population. Claims data is used to train a two-stage model on the member population. The first stage comprises a support vector machine, a rule-based module, and a generalized linear model that estimates the probability of obesity. The second stage comprises a regression neural network that operates on the output of the first stage and a subset of the input feature vector. Cost and utilizations in these areas, along with overall health measures as well as demographics and social factors, are inputs to a set of pattern recognition engines that perform regression. The output is the estimated body mass index of the member.

1 . A computerized device comprising a trained inference model that produces an estimate of obesity in an insured population comprising:
 (a) inputting to said computerized device identification data for a member for which height and weight data is not available;   (b) creating in said computerized device a feature vector comprising for said member for which height and weight data is not available:
 (1) demographic data for said member; 
 (2) a plurality of claim data metrics based on an analysis of said member claim data; and 
 (3) at least one census metric; 
   (c) presenting said feature vector to said trained obesity inference model in said computerized device, said model trained according to:
 (1) a body mass index for each member of a subset of said insured population for which height and weight data is available, said body mass index calculated from said height and weight data for each member of said subset; 
 (2) feature vector data for each member of said subset of said insured population comprising for each member:
 (i) demographic data for said member; and 
 (ii) a plurality of claim data metrics based on an analysis of said member claim data; and 
 (iii) at least one census metric; 
 
   (d) outputting from said trained obesity inference model of said computerized device an estimated body mass index for said member for which height and weight data is not available.   
     
     
         2 . The computerized device of  claim 1  wherein said plurality of claim data metrics comprises at least one metric for each of a plurality of obesity factors. 
     
     
         3 . The computerized device of  claim 2  wherein said obesity factors are selected from the group consisting of:
 polycystic ovary syndrome, fat in various parts of the body, fatty liver disease, edema, dysmetabolic syndrome X, mental stress, depression and anxiety, gall bladder disease, back problems, pancreatitis and bile duct obstruction, fatty breast issues, abnormal weight gain, Cushing's syndrome, colon and associated cancers, intestinal malabsorption, male breast cancer, breast cancer, neoplasm of the master hormonal gland, osteoarthrosis, vitamin deficiency and absorption problems, hypothyroidism, varicose veins, goiter, lymphangitis, hypoglycemia and other glycogen problems, phlebitis, hernia tendency, lumbago, asthma, eating disorders, diabetes, sleep problems, hypertension, migraines, abnormal lipid panel, head pain, and an obesity diagnosis. 
 
     
     
         4 . The computerized device of  claim 2  wherein said at least one metric for each of a plurality of obesity factors comprises a number of claims for said obesity factor. 
     
     
         5 . The computerized device of  claim 1  wherein said plurality of claim data metrics comprises a total number of claims for each of a plurality of obesity factors. 
     
     
         6 . The computerized device of  claim 1  wherein said plurality of claim data metrics comprises a total number of claims for a specified period of time. 
     
     
         7 . The computerized device of  claim 1  wherein said plurality of claim data metrics comprises a co-morbidity score. 
     
     
         8 . The computerized device of  claim 1  wherein said plurality of claim data metrics comprises at least one chronic condition metric. 
     
     
         9 . The computerized device of  claim 8  wherein said chronic condition metric is selected from the group consisting of chronic kidney disease, depression, diabetes, HIV, hypertension, and cancer. 
     
     
         10 . The computerized device of  claim 1  wherein said at least one census metric is selected from the group consisting of averages in census block areas for racial distribution, ethnicity, income, marital status, education, employment, and family size. 
     
     
         11 . The computerized device of  claim 1  wherein said height and weight data for said members in said subset of said insured population is obtained from health risk assessment data provided by said members. 
     
     
         12 . A computerized device comprising a trained inference model that produces an estimate of obesity in an insured population, comprising:
 (a) memory at said computerized device storing for a member of said insured population member health data exclusive of height and weight data and for said member a feature vector comprising:
 (i) demographic data for said member; 
 (ii) a plurality of claim data metrics based on an analysis of said member claim data; and 
 (iii) at least one census metric; 
   (b) a processor in said computerized device executing instructions to:
 (1) present said feature vector in parallel to said trained obesity inference model wherein said model is trained according to:
 (A) a body mass index for each member of a subset of said insured population for which height and weight data is available, said body mass index calculated from said height and weight data for each member of said subset; 
 (B) feature vector data for each member of said subset of said insured population comprising for each member:
 (i) demographic data for said member; and 
 (ii) a plurality of claim data metrics based on an analysis of said member claim data; and 
 (iii) at least one census metric; 
 
 
 (2) output from said obesity inference model an estimated body mass index for said member of said insured population. 
   
     
     
         13 . The computerized device of  claim 12  wherein said plurality of claim data metrics comprises at least one metric for each of a plurality of obesity factors. 
     
     
         14 . The computerized device of  claim 13  wherein said obesity factors are selected from the group consisting of:
 polycystic ovary syndrome, fat in various parts of the body, fatty liver disease, edema, dysmetabolic syndrome X, mental stress, depression and anxiety, gall bladder disease, back problems, pancreatitis and bile duct obstruction, fatty breast issues, abnormal weight gain, Cushing's syndrome, colon and associated cancers, intestinal malabsorption, male breast cancer, breast cancer, neoplasm of the master hormonal gland, osteoarthrosis, vitamin deficiency and absorption problems, hypothyroidism, varicose veins, goiter, lymphangitis, hypoglycemia and other glycogen problems, phlebitis, hernia tendency, lumbago, asthma, eating disorders, diabetes, sleep problems, hypertension, migraines, abnormal lipid panel, head pain, and an obesity diagnosis. 
 
     
     
         15 . The computerized device of  claim 13  wherein said at least one metric for each of a plurality of obesity factors comprises a number of claims for said obesity factor. 
     
     
         16 . The computerized device of  claim 12  wherein said plurality of claim data metrics comprises a total number of claims for each of a plurality of obesity factors. 
     
     
         17 . The computerized device of  claim 12  wherein said plurality of claim data metrics comprises a total number of claims for a specified period of time. 
     
     
         18 . The computerized device of  claim 12  wherein said plurality of claim data metrics comprises a co-morbidity score. 
     
     
         19 . The computerized device of  claim 12  wherein said plurality of claim data metrics comprises at least one chronic condition metric. 
     
     
         20 . The computerized device of  claim 19  wherein said chronic condition metric is selected from the group consisting of chronic kidney disease, depression, diabetes, HIV, hypertension, and cancer. 
     
     
         21 . The computerized device of  claim 12  wherein said at least one census metric is selected from the group consisting of averages in census block areas for racial distribution, ethnicity, income, marital status, education, employment, and family size. 
     
     
         22 . The computerized device of  claim 12  wherein said height and weight data for said members in said subset of said insured population is obtained from health risk assessment data provided by said members. 
     
     
         23 . A computerized method for estimating obesity in an insured population comprising:
 (a) identifying in said computer a plurality of members for which height and weight data is available;   (b) calculating in said computer for each of said plurality of members a body mass index;   (c) creating in said computer a feature vector comprising for each of said members for which height and weight data is available:
 (1) demographic data for said member; 
 (2) a plurality of claim data metrics based on an analysis of said member claim data; and 
 (3) at least one census metric; 
   (d) training an obesity inference model in said computer, said model trained according to:
 (1) said body mass index for each of said plurality of members; and 
 (2) said feature vector for each of said plurality of members; 
   (e) receiving at said computer feature vector data for a member for which height and weight data is not available;   (f) providing said member feature vector data to said obesity inference model; and   (g) receiving from said obesity inference model an estimated body mass index for said member for which height and weight data is not available.   
     
     
         24 . The computerized method of  claim 23  wherein said plurality of claim data metrics comprises at least one metric for each of a plurality of obesity factors. 
     
     
         25 . The computerized method of  claim 24  wherein said obesity factors are selected from the group consisting of:
 polycystic ovary syndrome, fat in various parts of the body, fatty liver disease, edema, dysmetabolic syndrome X, mental stress, depression and anxiety, gall bladder disease, back problems, pancreatitis and bile duct obstruction, fatty breast issues, abnormal weight gain, Cushing's syndrome, colon and associated cancers, intestinal malabsorption, male breast cancer, breast cancer, neoplasm of the master hormonal gland, osteoarthrosis, vitamin deficiency and absorption problems, hypothyroidism, varicose veins, goiter, lymphangitis, hypoglycemia and other glycogen problems, phlebitis, hernia tendency, lumbago, asthma, eating disorders, diabetes, sleep problems, hypertension, migraines, abnormal lipid panel, head pain, and an obesity diagnosis. 
 
     
     
         26 . The computerized method of  claim 24  wherein said at least one metric for each of a plurality of obesity factors comprises a number of claims for said obesity factor. 
     
     
         27 . The computerized method of  claim 23  wherein said plurality of claim data metrics comprises a total number of claims for each of a plurality of obesity factors. 
     
     
         28 . The computerized method of  claim 23  wherein said plurality of claim data metrics comprises a total number of claims for a specified period of time. 
     
     
         29 . The computerized method of  claim 23  wherein said plurality of claim data metrics comprises a co-morbidity score. 
     
     
         30 . The computerized method of  claim 23  wherein said plurality of claim data metrics comprises at least one chronic condition metric. 
     
     
         31 . The computerized method of  claim 30  wherein said chronic condition metric is selected from the group consisting of chronic kidney disease, depression, diabetes, HIV, hypertension, and cancer. 
     
     
         32 . The computerized method of  claim 23  wherein said at least one census metric is selected from the group consisting of averages in census block areas for racial distribution, ethnicity, income, marital status, education, employment, and family size. 
     
     
         33 . The computerized method of  claim 23  wherein said height and weight data for said plurality of members is obtained from health risk assessment data provided by said members.