Method and system for predicting adherence to a treatment

Abstract

Data characterizing an individual is received. Thereafter, one or more variables are extracted from the data so that, using a predictive model populated with the extracted variables, a likelihood of the individual adhering to a treatment regimen can be determined. The predictive model is trained on historical treatment regimen adherence data empirically derived from a plurality of subjects. Subsequently, data characterizing the determined likelihood of adherence can be promoted.

What is claimed is: 
     
         1 . A treatment administration system comprising a computer-implemented model having a learning component and a predictive component, the system configure for:
 determining treatment adherence patterns, in response to receiving first data and second data;   extracting a combined set of variables from the first data, the first data collected from treatment adherence patterns associated with one or more treatments administered to the plurality of treatment subjects;   determining a likelihood of adherence to a first treatment regimen for a first treatment subject, based on the combined set of variables populating the computer-implemented model,
 the likelihood of adherence being determined in response to computing one or more values for the combined set of variables by using data associated with at least one treatment subject at a plurality of stages of the first treatment regimen, 
 wherein the first treatment regimen is administered to the first treatment subject according to a first score generated for the first treatment subject based on the determined likelihood of adherence of the first treatment subject. 
   
     
     
         2 . The system of  claim 1 , wherein a second treatment regimen is administered to a second treatment subject according to a second score generated for the second treatment subject based on the determined likelihood of adherence for the second treatment subject. 
     
     
         3 . The system of  claim 1 , wherein the model is trained on historical treatment regimen adherence data empirically derived from the plurality of treatment subjects. 
     
     
         4 . The system of  claim 2 , wherein the first score and the second score are respectively associated with corresponding disjoint score ranges. 
     
     
         5 . The system of  claim 4 , wherein one or more messages are delivered to the first treatment subject and the second treatment subject based on the first score and the second score, respectively, depending on a corresponding score range for the first score and the second score. 
     
     
         6 . The system of  claim 5 , wherein at least a first message is delivered to the first treatment subject via a first delivery channel determined for the first treatment subject and at least a second message is delivered to the second treatment subject via a second delivery channel determined for the second treatment subject. 
     
     
         7 . The system of  claim 5 , wherein the one or more messages provide sequential guidance to the first treatment subject and the second treatment subject to respectively increase a likelihood of adherence to the first treatment regimen and the second treatment regimen. 
     
     
         8 . The system of  claim 7 , wherein a first message is delivered to the first treatment subject according to a first predetermined time schedule associated with the first treatment regimen and a second message is delivered to the second treatment subject according to a second predetermined time schedule associated with the second treatment regimen. 
     
     
         9 . The system of  claim 1 , wherein:
 the second data characterizes the first treatment subject, at least a portion of the second data received subsequent to start of administration of the first treatment regimen to the first treatment subject,   at least one variable associated with the second data affects the likelihood of the first treatment subject adhering to the first treatment regimen, and   the first score being updated based on the at least one variable.   
     
     
         10 . The system of  claim 9 , wherein an updated treatment regimen is administered to the first treatment subject based on the updated first score and the combined set of variable is determined by combining at least two variables when the first treatment regimen has more interdependence with the combined set of variables than with the at least two variables. 
     
     
         11 . A computer program product comprising a non-transitory machine-readable medium storing instructions that, when executed by at least one programmable processor, cause the at least one programmable processor to perform operations comprising:
 determining treatment adherence patterns, in response to receiving treatment data and demographic data associated with a plurality of treatment subjects;   extracting a combined set of variables from the treatment data, the treatment data collected from treatment adherence patterns associated with one or more treatments administered to the plurality of treatment subjects;   determining a likelihood of adherence to a first treatment regimen for a first treatment subject, using the computer-implemented model populated with the combined set of variables,
 the likelihood of adherence being determined in response to computing one or more values for the combined set of variables by using data associated with at least one treatment subject at a plurality of stages of the first treatment regimen; 
   the first treatment regimen being administered to the first treatment subject according to a first score generated for the first treatment subject based on the determined likelihood of adherence of the first treatment subject.   
     
     
         12 . The computer program product of  claim 11 , wherein a second treatment regimen is administered to a second treatment subject according to a second score generated for the second treatment subject based on the determined likelihood of adherence for the second treatment subject. 
     
     
         13 . The computer program product of  claim 11 , wherein the model is trained on historical treatment regimen adherence data empirically derived from the plurality of treatment subjects. 
     
     
         14 . The computer program product of  claim 12 , wherein the first score and the second score are respectively associated with corresponding disjoint score ranges. 
     
     
         15 . The computer program product of  claim 14 , wherein one or more messages are delivered to the first treatment subject and the second treatment subject based on the first score and the second score, respectively, depending on a corresponding score range for the first score and the second score. 
     
     
         16 . The computer program product of  claim 15 , wherein at least a first message is delivered to the first treatment subject via a first delivery channel determined for the first treatment subject and at least a second message is delivered to the second treatment subject via a second delivery channel determined for the second treatment subject. 
     
     
         17 . The computer program product of  claim 15 , wherein the one or more messages provide sequential guidance to the first treatment subject and the second treatment subject to respectively increase a likelihood of adherence to the first treatment regimen and the second treatment regimen, and
 a first message is delivered to the first treatment subject according to a first predetermined time schedule associated with the first treatment regimen and a second message is delivered to the second treatment subject according to a second predetermined time schedule associated with the second treatment regimen.   
     
     
         18 . A treatment administration method for a computer-implemented model having a learning component and a predictive component, the method executed on one or more computer processors, the method comprising:
 determining treatment adherence patterns, in response to receiving treatment data and demographic data associated with a plurality of treatment subjects;   extracting a combined set of variables from the treatment data, the treatment data collected from treatment adherence patterns associated with one or more treatments administered to the plurality of treatment subjects;   determining a likelihood of adherence to a first treatment regimen for a first treatment subject, using the computer-implemented model populated with the combined set of variables,
 the likelihood of adherence being determined in response to computing one or more values for the combined set of variables by using data associated with at least one treatment subject at a plurality of stages of the first treatment regimen; 
   the first treatment regimen being administered to the first treatment subject according to a first score generated for the first treatment subject based on the determined likelihood of adherence of the first treatment subject.   
     
     
         19 . The method of  claim 18 , wherein:
 the second data characterizes the first treatment subject, at least a portion of the second data received subsequent to start of administration of the first treatment regimen to the first treatment subject,   at least one variable associated with the second data affects the likelihood of the first treatment subject adhering to the first treatment regimen, and   the first score being updated based on the at least one variable, and   an updated treatment regimen is administered to the first treatment subject based on the updated first score and the combined set of variable is determined by combining at least two variables when the first treatment regimen has more interdependence with the combined set of variables than with the at least two variables.   
     
     
         20 . The method of  claim 18 , wherein a second treatment regimen is administered to a second treatment subject according to a second score generated for the second treatment subject based on the determined likelihood of adherence for the second treatment subject, and the model is trained on historical treatment regimen adherence data empirically derived from the plurality of treatment subjects, the first score and the second score respectively associated with corresponding disjoint score ranges.