Patent Issued for Maintaining stability of health services entities treating influenza (USPTO 11468996): Cerner Innovation Inc.

2022 OCT 27 (NewsRx) -- By a News Reporter-Staff News Editor at Vaccine Daily -- From Alexandria, Virginia, NewsRx journalists report that a patent by the inventors McNair, Douglas S. (Seattle, WA, US), filed on May 4, 2020, was published online on October 11, 2022.

The patent’s assignee for patent number 11468996 is Cerner Innovation Inc. (North Kansas City, Missouri, United States).

News editors obtained the following quote from the background information supplied by the inventors: “Influenza is a contagious respiratory illness with a long history of causing human morbidity and mortality. Despite extensive surveillance of seasonal influenza, its economic costs remain difficult to quantify. Although statistical methods have been proposed for estimating the excess hospitalization rate and mortality rate of influenza, few economic studies have attempted to measure the health insurance losses arising from acute-care hospitalizations resulting from influenza.

“Major influenza pandemics tend to occur three to four times each century and have a number of characteristics that differ from intermittent influenza epidemics. By definition a pandemic affects a large number of countries worldwide. A pandemic virus, which infrequently encounters the world human population, results in a large number of hospitalized cases and excess mortality. The novelty and virulence of the pandemic virus also makes prevention and control measures difficult as existing vaccines are not effective and production of new vaccine may take six months or more. Antiviral drugs are in general the only virus-specific intervention during the initial response. Neuraminidase inhibitor medications such as oseltamivir and zanamivir have the advantage of conferring almost immediate protection and their use does not interfere with response to inactivated influenza vaccine.

“Although it is well recognized that countries must prepare for the next influenza pandemic, the uncertainty regarding the characteristics of the virus, the populations who will be most seriously affected, and the most cost-effective policies make preparation difficult. The potential losses for health services providers, for health insurers, for reinsurers, and for banks in epidemic and pandemic situations may be catastrophic. Financial ruin for even a modest fraction of these entities would have serious and lasting economic consequences for broader society in the affected countries. In particular, unlike other industries, health services have peculiar concerns and limitations and obligations, such as described further herein, that pose unique challenges for similar for maintaining financial stability.”

As a supplement to the background information on this patent, NewsRx correspondents also obtained the inventors’ summary information for this patent: “This Summary is provided to introduce a selection of concepts in a simplified form that are further described below in the Detailed Description. This Summary is not intended to identify key features or essential features of the claimed subject matter, nor is it intended to be used as an aid in determining the scope of the claimed subject matter.

“Embodiments of the invention are directed towards systems and methods for determining and mitigating the aggregate loss risk associated with hospitalization for epidemic or pandemic influenza for health insurers, reinsurers, provider organizations, or public policy-makers. An accurate prediction of this risk may be provided, which may be used to determine parameters for reinsurance underwriting or for issuance and trading of catastrophe bonds (“cat bonds”) or other insurance-linked securities (ILS) and derivatives to lay off substantial amounts of such risk to capital markets investors. In particular, one embodiment uses a novel log-expit transformation of the raw data and non-parametric gradient-boosting machine-learning modeling in order to determine a high-claim right-tail risk. Some embodiments further comprise securitizing epidemic or pandemic influenza acute care health services catastrophe risk.

“Accordingly, in one aspect a method is provided for distributing instruments representing securitized epidemic or pandemic catastrophe risk, implemented on a computer system at a reinsurer. The method includes receiving, at the reinsurer, a first allotment of first risk instruments of a risk class representing one or more epidemic or pandemic catastrophe risks. The risk class being issuable from the computer system at the reinsurer on a recurring basis, each of the first risk instruments having a first issue date and providing a return on an investment, the amount of the return being reduced upon the occurrence of a realization event for the corresponding represented epidemic or pandemic catastrophe risk. The method also includes distributing from the reinsurer, the first risk instruments of the first allotment to one or more investors, wherein the realization event for a given risk class is defined as an occurrence of an event meeting a predetermined impact threshold, the occurrence of the event meeting the predetermined impact threshold is determined according to an index of physical parameters issued by a neutral party, and the physical parameters are related to but separate from catastrophic loss.

“In another aspect, a method is provided for securitizing epidemic or pandemic acute-care health services catastrophe risk. The method comprises determining a forecast model for predicting aggregate loss statistical distributions based on historical insurance claims and electronic health record information for a plurality of hospital admissions over a period of time. The method also comprises determining the aggregate loss with confidence-band or Value at Risk (VaR) bounds on the losses thus determined, and establishing one or more risk classes on the system of the reinsurer, each risk class representing one or more epidemic or pandemic catastrophe risks, each risk class being recurringly issuable from the system of the reinsurer or from a financial exchange as risk instruments providing a return on an investment, the amount of the return for a risk instrument being reduced upon the occurrence of a realization event for the corresponding represented epidemic or pandemic catastrophe risk. The method further includes issuing from the reinsurer, a first collection of risk instruments of a first risk class of the one or more risk classes, wherein the realization event for a given risk class is defined as an occurrence of an event meeting a predetermined impact threshold, the occurrence of the event meeting predetermined impact threshold is determined according to an index of epidemic infection-related parameters issued by a neutral party such as the U.S. Centers for Disease Control (CDC) or the World Health Organization (WHO), and the epidemic infection-related parameters are related to but separate from catastrophic loss.

“In another aspect, a method is provided for securitizing epidemic or pandemic acute-care health services catastrophe risk. The method includes determining a time series of viral hospital admissions data and claims resulting from these in-patient care episodes and store said time series on machine-readable media; performing exploratory fitting to EVD, IG, and other skew-kurtotic distributions, and evaluate accuracy of fits in right-tail (QQ plots, confidence bands). The method also includes setting coefficients for affine transform to scale and remove offset of claims, applying log-expit transform to the raw claims data, and partition data into training and test datasets. The method further includes setting or determining: variables’ fitting constraints (monotonicity), Tweedie index parameter, learning rate for machine-learning boosting algorithm, maximum interaction depth for gradient boosting, subsampling fraction for bagging generation of boosting tree models, a number M of boosting trees to be generated and evaluated, and a number N cross-validation iterations. The method further includes performing M iterations of Tweedie boosting, determine convergence of gradient boosting model, and determining the best iteration in converged model solution.”

The claims supplied by the inventors are:

“1. A method for securitizing epidemic or pandemic acute-care health services catastrophe risk comprising: transforming a first structured data set into a first transformed data set using a transformation executed by one or more processors, the first structured data set including data representative of health record information associated with hospital admissions for a first predetermined period of time; training a model including non-parametric machine-learning processes to predict aggregate risk using the first transformed data set; generating a classification instrument using the trained model to predict aggregate risk of a second transformed data set; and communicating the classification instrument to a networked server accessible by multiple users.

“2. The method of claim 1 further comprising: transforming a second structured data set into the second transformed data set using the expit transformation executed by the one or more processors, the second structured data set including data representative of non-classified historical insurance claims and corresponding electronic health record information for a plurality of hospital admissions.

“3. The method of claim 1, wherein the method further comprises: conditioning the first transformed data set by generating a plurality of boosting trees based on the first structured data set or Tweedie boosting the first structured data set.

“4. The method of claim 1, wherein the method further comprises: transforming a second structured data set into the second transformed data set using an expit transformation.

“5. The method of claim 4, wherein the second structured data set includes data representative of health record information associated with hospital admissions for a second predetermined period of time.

“6. The method of claim 1, wherein each of the one or more risk classification instruments includes a plurality of risk classes, and wherein each risk class is associated with a potential epidemic or pandemic contagion.

“7. The method of claim 1, wherein training the model to predict aggregate risk using the first transformed data set comprises: identifying from the first transformed data set a training cohort and a validating cohort; Tweedie boosting the training cohort; supplying the tweedie boosted training cohort as input to train the model; and validating the trained model using the validating cohort.

“8. A system for distributing instruments representing securitized epidemic or pandemic catastrophe risk, the system comprising: a communication interface for receiving a structured data set including data representative of health record information associated with hospital admissions for a predetermined period of time; a data conditioner for conditioning received structured data into a transformed data set using a transformation function; a modeler for providing a model including non-parametric machine-learning processes the transformed data set as input for the model; a classifier for generating one or more risk classification instruments based on the model’s output; a publishing interface for transmitting the one or more risk classification instruments to a remote server that provides access to the one or more risk classification instruments to a plurality of users.

“9. The system of claim 8, wherein the transformation function comprises an expit transformation.

“10. The system of claim 9, wherein the transformation function further comprises generating a plurality of boosting trees based on the structured data, or Tweedie boosting the structured data.

“11. The system of claim 8, wherein each of the one or more risk classification instruments includes a plurality of risk classes, and wherein each risk class is associated with a potential epidemic or pandemic contagion.

“12. The system of claim 8, wherein the modeler trains the model by Tweedie boosting a first plurality of expit transformed pre-classified historical insurance claims and associated electronic health record information, and wherein the modeler validates the model using a second plurality of expit transformed pre-classified historical insurance claims and associated electronic health record information.

“13. A computer-readable storage media having computer-executable instructions embodied thereon that when executed by a processor, facilitate a method for securitizing epidemic or pandemic catastrophe risk, the method comprising: transforming a first structured data set into a first transformed data set using an expit transformation executed by one or more processors, the first structured data set including data representative of health record information associated with hospital admissions for a first predetermined period of time; training a model including non-parametric machine-learning processes to predict aggregate risk using the first transformed data set; generating one or more risk classification instruments using the trained model to predict aggregate risk of a second transformed data set; and publishing the one or more risk classification instruments to a networked server accessible by multiple users.

“14. The computer-readable storage media of claim 13, further comprising: transforming a second structured data set into the second transformed data set using the expit transformation, the second structured data set including data representative of non-classified historical insurance claims and corresponding electronic health record information for a plurality of hospital admissions.

“15. The computer-readable storage media of claim 13, wherein the model comprises non-parametric machine-learning processes.

“16. The computer-readable storage media of claim 13, wherein the method further comprises: conditioning the first transformed data set by generating a plurality of boosting trees based on the first structured data set or Tweedie boosting the first structured data set.

“17. The computer-readable storage media of claim 13, wherein the method further comprises: transforming a second structured data set into the second transformed data set using the expit transformation.

“18. The computer-readable storage media of claim 17, wherein the second structured data set includes data representative of health record information associated with hospital admissions for a second predetermined period of time.

“19. The computer-readable storage media of claim 13, wherein training the model to predict aggregate risk using the first transformed data set comprises: identifying from the first transformed data set a training cohort and a validating cohort; Tweedie boosting the training cohort; supplying the tweedie boosted training cohort as input to train the model; and validating the trained model using the validating cohort.”

For additional information on this patent, see: McNair, Douglas S. Maintaining stability of health services entities treating influenza. U.S. Patent Number 11468996, filed May 4, 2020, and published online on October 11, 2022. 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=11468996.PN.&OS=PN/11468996RS=PN/11468996

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