Patent Issued for Predictive model segmentation platform (USPTO 11120509): Hartford Fire Insurance Company

2021 OCT 05 (NewsRx) -- By a News Reporter-Staff News Editor at Insurance Daily News -- From Alexandria, Virginia, NewsRx journalists report that a patent by the inventors Gray, Willie F. (North Granby, CT, US), Harper, Amanda Sue (Atlanta, GA, US), Scaia, Allison M. (St. Louis, MO, US), Workin, Lisa (Eagan, MN, US), filed on August 12, 2019, was published online on September 14, 2021.

The patent’s assignee for patent number 11120509 is Hartford Fire Insurance Company (Hartford, Connecticut, United States).

News editors obtained the following quote from the background information supplied by the inventors: “An insurer may provide payments when claims are made in connection with a disability insurance policy. For example, an employee who becomes too ill to work might receive payments associated with a long term disability insurance policy purchased by his or her employer. Note that payments might continue until the employee is able to return to work. The insurer may assign a claim handler to communicate with the employee, the employer, and/or medical service providers to facilitate the employee’s return to the workplace. Moreover, different claim handlers may have different abilities and/or different workloads.

“In one approach, newly received disability insurance claims might be assigned to claim handlers in a random or round robin manner. This, however, might lead to one claim handler having a significantly more complex workload as compared to another claim handler. To avoid such a result, particular types of disability insurance claims might be more effectively assigned to particular claim handlers. For example, a relatively complicated insurance claim might be more efficiently processed by a claim handler who handles a relatively small number of insurance claims and/or is especially skilled when it comes to handling these types of insurance claims.

“Manually determining which claim handler should be assigned to each individual insurance claim, however, can be time consuming task, especially when there are a substantial number of claims to be analyzed. For example, an insurer might receive tens of thousands of new long term disability insurance claims each year (which might represent a billion dollars of potential liability). It would therefore be desirable to provide systems and methods to facilitate the assignment of disability insurance claims to claim handlers, in an automated, efficient, and accurate manner.”

As a supplement to the background information on this patent, NewsRx correspondents also obtained the inventors’ summary information for this patent: “According to some embodiments, systems, methods, apparatus, computer program code and means may facilitate the assignment of disability insurance claims to claim handlers. In some embodiments, a triage platform may receive data indicative of a disability insurance claim submitted in connection with a disability insurance policy, including at least one claim characteristic. The triage platform may determine, based on the claim characteristic, a claim segment to be associated with the disability insurance claim. A claim handler may be assigned to the disability insurance claim in accordance with the determined claim segment. Information about the disability insurance claim may then be automatically routed to the assigned claim handler.

“Some embodiments provide: means for receiving, at a triage platform, data indicative of a disability insurance claim submitted in connection with a disability insurance policy, including at least one claim characteristic; means for determining, by a computer processor of the triage platform based on the claim characteristic, a claim segment to be associated with the disability insurance claim; means for assigning a claim handler to the disability insurance claim in accordance with the determined claim segment; and means for automatically routing information about the disability insurance claim to the assigned claim handler.

“A technical effect of some embodiments of the invention is an improved and computerized method to facilitate the assignment of disability insurance claims to claim handlers. With these and other advantages and features that will become hereinafter apparent, a more complete understanding of the nature of the invention can be obtained by referring to the following detailed description and to the drawings appended hereto.

“According to some embodiments, a triage platform may receive data indicative of a disability insurance claim submitted in connection with a disability insurance policy, including at least one claim characteristic. The triage platform may determine, based on the claim characteristic, a claim segment to be associated with the disability insurance claim. A claim handler may be assigned to the disability insurance claim in accordance with the determined claim segment. Information about the disability insurance claim may then be automatically routed to the assigned claim handler.”

The claims supplied by the inventors are:

“1. A computer system comprising: (a) an indeterminate data capture module configured to extract historical indeterminate data from historical indeterminate data sources and current indeterminate data from current indeterminate data sources, wherein the indeterminate data comprises at least one of document images, claim handler notes, digitized recorded voice files, narrative speech or text, information in descriptive note fields, and signal characteristics of audible voice data files; (b) an intake server configured to receive a plurality of current data sets corresponding to a plurality of claims, each of the plurality of current data sets including one or more of a field having a value indicative of diagnosis data, a field having a value indicative of recovery profile data, a field having a value indicative of complexity data, and a field indicative of recovery likelihood data; © a segmentation platform interposed between the intake server and a plurality of remote user-accessible computers, comprising: (i) one or more data storage devices configured to store: (1) a plurality of historical data sets including a data table including at least a payment data field, a demographic data field, a medical code data field, an injury data field, and an employment data field, and (2) the plurality of current data sets, wherein the plurality of historical data sets and the plurality of current data sets stored by the one or more data storage devices include the historical indeterminate data and the current indeterminate data; (ii) a model training component computer processor configured to: receive at least some of the plurality of historical data sets from the one or more data storage devices; apply one or both of a variable association data reduction process and a granularity reduction process to the plurality of historical data sets to generate reduced historical data sets and reduce a dimensionality of the plurality of historical data sets; and train, based upon the reduced historical data sets, a predictive model to segment the plurality of current data sets, wherein the predictive model comprises one or more a neural network, a Bayesian network, an expert system, a decision tree, a collection of decision trees, and a support vector machine configured for processing involving large numbers of variables; (iii) the predictive model trained on the reduced historical data sets to segment the plurality of current data sets, wherein the predictive model being trained on the reduced historical data sets includes the predictive model being trained on the historical indeterminate data; (iv) a segment rules database configured to store segment rules to divide the plurality of claims into a plurality of segment classifications, at least one segment rule configured to: segment a data set to a higher complexity claim segment when the data set is indicative of at least two of the following conditions: the value indicative of the diagnosis data value is true, the value indicative of the recovery profile data is true, and the value indicative of the complexity data exceeds a pre-determined level; and segment a data set to a lower complexity claim segment when the data set is indicative of either: the value indicative of the recovery likelihood data is below a first pre-determined threshold value or the value indicative of the recovery likelihood data is above a second pre-determined threshold value; and (v) a triage processor configured to: (1) receive a current data set of the plurality of current data sets corresponding to a claim received by the intake server, the current data set including at least one of the field having the value indicative of the diagnosis data, the field having the value indicative of the recovery profile data, the field having the value indicative of the complexity data, and the field having the value indicative of the recovery likelihood data; (2) access the segment rules database to retrieve the at least one segment rule; (3) segment, based on application of the predictive model and the at least one segment rule to the current data set corresponding to the claim, the current data set to a claim segment, wherein the triage processor being configured to segment the current data set comprises the triage processor being configured to segment the current data set further based upon the current indeterminate data; and (4) automatically route, via the communication network, an electronic file corresponding to the current data set to the one of the plurality of remote user-accessible computers associated with a linked user identifier.

“2. The computer system of claim 1, wherein the triage processor is further configured to repeat steps (1)-(4) for additional ones of the plurality of current data sets and continuously adapt the predictive model by activating the model training component computer processor to update the predictive model with the additional ones of the plurality of current data sets and results corresponding to the additional current data sets.

“3. The computer system of claim 1, wherein the one or more data storage devices of the segmentation platform is further configured to store: user data corresponding to a plurality of users, the user data including fields indicative of values of user identifier data, user workload data, user experience data, and user expertise data; and wherein the triage processor is further configured to, before automatically routing the electronic file corresponding to the current data set to the one of the plurality of remote user-accessible computers: dynamically determine updated user current workload data in the one or more data storage devices based upon continually received current workload data for each of the plurality of users; and link the user identifier, corresponding to a user, to the current data set based upon the assigned claim segment and the user data including the dynamically determined updated current workload data for each of the plurality of users; wherein the triage processor is configured to automatically route the electronic file to the one of the plurality of remote user-accessible computers associated with the linked user identifier responsive to linking of the user identifier to the claim.

“4. The computer system of claim 3, wherein the triage processor is further configured to repeat steps (1)-(4) for additional current data sets and continuously adapt the predictive model by activating the model training component computer processor to update the predictive model with the current data set and results corresponding to the additional current data sets.

“5. The computer system of claim 1, wherein the triage processor is further configured to: determine, based on application of the predictive model to the current data set, a recovery profile for the current data set; and transmit via the communication network the recovery profile corresponding to the current data set to the one of the plurality of remote user-accessible computers associated with the linked user identifier.

“6. The computer system of claim 1, wherein the intake server is further configured to receive data corresponding to a current data set by at least one of: a submitted paper claim and a telephone call center.”

There are additional claims. Please visit full patent to read further.

For additional information on this patent, see: Gray, Willie F. Predictive model segmentation platform. U.S. Patent Number 11120509, filed August 12, 2019, and published online on September 14, 2021. Patent URL: http://patft.uspto.gov/netacgi/nph-Parser?Sect1=PTO1&Sect2=HITOFF&d=PALL&p=1&u=%2Fnetahtml%2FPTO%2Fsrchnum.htm&r=1&f=G&l=50&s1=11120509.PN.&OS=PN/11120509RS=PN/11120509

(Our reports deliver fact-based news of research and discoveries from around the world.)