“Digital System For Automated Measuring Of Relative Risk Measurands And Scores Of Living Individuals And Method Thereof” in Patent Application Approval Process (USPTO 20240145096): Swiss Reinsurance Company Ltd.

2024 MAY 21 (NewsRx) -- By a News Reporter-Staff News Editor at Insurance Daily News -- A patent application by the inventors ANGELAKOPOULOU, Aspasia (Zurich, CH); DUCKER, Michael (Zurich, CH); MARTIN, Alan (Zurich, CH); RIX, Douglas (Zurich, CH); SCHOONBEE, John (Zurich, CH), filed on September 27, 2023, was made available online on May 2, 2024, according to news reporting originating from Washington, D.C., by NewsRx correspondents.

This patent application is assigned to Swiss Reinsurance Company Ltd. (Zurich, Switzerland).

The following quote was obtained by the news editors from the background information supplied by the inventors: “Automation in risk assessment and measurement is a crucial element in the life risk-transfer technology to allow automated classification of applicants of risk-transfers or to allow automated optimization of the risk cover of applicants or to automated detection of gabs in the risk cover of an applicant or portfolios of applicants. Risk-transfer systems or risk-transfer providers perform underwriting process to make decisions on applications and to price policies accordingly. With the increase in the amount of data and advances in data processing techniques, the underwriting process can be automated for faster processing of applications. In the last decade, the big data technologies revolutionize the way risk-transfer systems are enabled to collect, process, analyze, and manage data more efficiently. Thus, automation technology proliferates in various sectors of risk-transfer technology such as risk assessment, customer analytics, product development, automated marketing analytics, claims analysis, underwriting assessment and processing, fraud detection, and multi-layer risk-transfer automation. Telematics is a typical example where big data analytics is being vastly implemented and is transforming the way auto risk-transfer pricing the premiums of individual drivers. Individual life and health insurance organizations still rely on the conventional, not-automated actuarial formulas to predict mortality rates and premiums of life policies. Life risk-transfer technology has however started carrying out automated predictive data-processing to improve the operational efficacy of the automation, but there is still a lack of efficient and optimized automation operating on a high technical accuracy. Many prior art solution have been concentrated on data mining techniques to detect frauds among risk-transfer systems, which is a crucial issue due to the problem facing great losses.

“In the prior art systems, measurable risk parameter values associated with a specific, typically modifiable life-style pattern are known. They comprise e.g. heavy alcohol drinking, smoking, excess body weight and lack of physical exercise as modifiable risk factors of lifestyle, which may all contribute to the incidence of chronic diseases and premature death. There may also be synergistic and additive interactions between such factors in individuals with clustering of unfavorable lifestyle factors. Therefore, interventions aimed at reducing the number of risk factors has been recognized as an important target in both personalized medicine and public health policies. Recent studies have estimated that adopting a healthy lifestyle even at the age of 50 could add more than a decade to life suggesting significant therapeutic potential for lifestyle interventions.

“Measuring parameters such as increased gamma-glutamyltransferase (GGT), and alanine aminotransferase (ALT) enzyme activities in apparently healthy individuals may be attributed to unhealthy lifestyle factors, such as alcohol consumption or excess body weight. The increases in these liver enzymes may also associate with extra-hepatic disease risks, including metabolic syndrome, and cardio- or cerebrovascular events. While the biochemical pathways underlying such observations often remains unclear, it can be assumed that inflammatory processes, oxidative stress and generation of abnormal lipid profiles are key pathogenic factors in the sequence of events leading to hepatotoxicity or other adverse health effects, such as incident stroke, in individuals presenting with various clusters of risk factors.

“So far, it is still complex to examine the individual and joint impacts of the various unfavorable life style factors on biochemical indices of health. In particular, it is technically challenging to measure the combined effects of various lifestyle-related factors on biomarkers of liver status (ALT, GGT), inflammation (C-reactive protein) and lipid metabolism (cholesterol, HDL-cholesterol, LDL-cholesterol, triglycerides) in a large population, which includes detailed records on alcohol consumption, smoking, physical activity and health status. It can be assumed that precise measurements of the biomarker behavior in response to various types of unhealthy behaviors may improve the possibilities for automated interventions and alarm signaling aimed at adopting more favorable lifestyles and adapt a more favorable risk score measure.

“There is a need to develop measurements on health and lifestyle which can be, inter alia, validated for use in international population-based health measurements. The measured parameter value on each parameter measured, such as alcohol consumption, smoking, physical activity and coffee consumption should preferably be chosen to be assigned to mutually exclusive and collectively exhaustive categories. Data on alcohol consumption can e.g. be measured from the past 12 months prior to blood sampling and included information on the types of beverages consumed as well as the amounts and frequencies of consumption. The ethanol content in different beverages can e.g. be quantitated in grams of ethanol based on defined portion sizes or threshold values, for example, as follows: regular beer 12 grams ( 1/3 L), strong beer 15.5 grams ( 1/3 L), long drink 15.5 grams ( 1/3 L), spirit 12 grams (4 cL), wine 12 grams (12 cL) and cider 12 grams ( 1/3 L). Information on smoking habits was collected with a set of standardized questions and the data was expressed as the amounts of cigarettes per day. Habitual physical activity including both the number and total time used for physical exercises can also be measured from an uer. Coffee consumption was assessed with a set of standardized questions and expressed as the intake of standard servings of coffee (cups) per day.

“The measuring values obtained from measuring can e.g. subsequently be used to define scores index measure, for example, for low risk (=0), medium risk (=1) and high risk (=2) categories for each individual risk factor following captured patterns on health-related risk assessment in relation to alcohol consumption, smoking, BMI status and physical activity. Herein, the variables can e.g. be categorized into three ordinal levels to yield increased statistical power as compared to previously used prior art dichotomous classification. For alcohol consumption the score measures can e.g. be defined as follows: 0=no consumption; 1=alcohol consumption between 1-14 (men) or 1-7 (women) standard drinks per week; 2=alcohol consumption exceeding 14 drinks (men) or 7 drinks (women) per week. For smoking 0=no smoking, 1=1-19 cigarettes per day, 2=≥20 cigarettes per day; for BMI 0=BMI<25; 1=BMI≥25 and <30; 2=BMI≥30. For physical activity 0 represents those with physical activity over 4 hours per week; 1=those with physical activity between 0.5 and 4 hours per week and 2=those with physical activity less than 30 min/week. The sum of these score measures can e.g. be used to measure a total number of risk factors, with higher score measures (e.g. maximum=8) indicating an unhealthier lifestyle. Serum liver enzymes (ALT and GGT) can e.g. be measured using a standard clinical chemical methods, such as on an Abbott Architect clinical chemistry analyzer of the manufacturer (Abbott Laboratories, Abbott Park, Ill., USA). High-sensitivity CRP, a biomarker of inflammation, can e.g. be determined using a latex immunoassay (Sentinel Diagnostics, Milan, Italy) with the Abbott Architect c8000 clinical chemistry analyzer. Lipid profiles can e.g. include determinations of total cholesterol, high-density lipoprotein-associated cholesterol (HDL), low-density lipoprotein (LDL) and total triglycerides using standard enzymatic methods. All laboratory or telematics-based measurements can be subjects to cut-offs for the normal limits of the different markers e.g. as follows: ALT (50 U/L men; 35 U/L women), GGT (60 U/L men; 40 U/L women), CRP (3.0 mg/L), cholesterol (5 mmol/L), HDL cholesterol (1.0 mmol/L men, 1.2 mmol/L women), LDL cholesterol (3.0 mmol/L), triglycerides (1.7 mmol/L).

“The characteristics can e.g. be measured using pattern recognition, machine-learning or applied analysis structures of variance (ANOVA) e.g. with polynomial contrasts to reveal possible trends across increasing risk score categories. The distribution of abnormal biomarker levels across the risk categories can e.g. be analyzed by chi-square measurements for trend. Binary logistic regression can e.g. be used to estimate the odds ratios (ORs) of abnormal biomarker levels associated with the risk score categories, adjusting for age and coffee consumption, as these factors are known to potentially associate with abnormal biomarker levels and showed association in univariate analysis. All factors can e.g. be inputted simultaneously into the multivariable modelling and/or machine-learning structure. Potential multicollinearity among the covariates can e.g. be measured by generating the Variance Inflation Factor (VIF). Correlations between the risk scores and various biomarkers can be measured using Spearman’s rank correlation coefficients. For significance, a defined threshold p-value can be implemented e.g. considering p-value <0.05 statistically significant.”

There is additional background information. Please visit full patent to read further.”

In addition to the background information obtained for this patent application, NewsRx journalists also obtained the inventors’ summary information for this patent application: “It is an object of the invention to provide an automated optimized risk assessment and risk measuring system, operatable in the field of life and health risks, thereby providing the technical basis for automated risk assessment and risk-transfer (including automated underwriting (UW)). The system should allow for systematic capturing, measuring, quantifying, and forward-looking generating of appropriate risk and risk accumulation measures of risk-transfers and risk-transfer portfolios associated with risk exposures of living individuals (humans) based on physical measuring parameter values and data, i.e. the impact of a possibly occurring physical event in a defined future time window. In the present invention, this includes measuring and rating the score measures (“resilience score”) of the life or health of a person. It is a further object of the present invention to propose a processor-driven system or platform providing an automated digital channel for automatically concluding and dynamically adapting risk-transfers between a risk-transfer service user and a risk-transfer service provider, which does not exhibit the disadvantages of the known systems. In particular, it is an object of the present invention to provide an inventive technical teaching for automation which is easily integratable in other processes, productions chains or risk assessment and measuring systems, e.g. by appropriate APIs. This invention also aims at providing an automated system to enhance risk assessment of individuals as applicants for a risk-transfer or among life risk-transfer systems using predictive processing efficiency and accuracy.

“According to the present invention, these objects are achieved particularly through the features of the independent claims. In addition, further advantageous embodiments follow from the dependent claims and the description.

“According to the present invention, the above-mentioned objects are particularly achieved by the inventive, automated, digital measuring system for measuring relative occurrence frequencies or more specifically, a measurable occurrence probability of a medical event and/or health event and/or life risk event (given by as a measuring value of risk measurands or score index measurands) of living individuals for a future measurement time window, wherein the relative risk measurands provide a measure for the frequency of occurrences of specific medical events having impacting consequences in specified ranges to the living individual within a defined cohort of living individuals relative to a randomized cohort of living individuals, in that the system comprises a defined parametrization for capturing risk shape pattern for living individual, the parametrization of the risk shape pattern comprising at least lifestyle factor values measuring physical activity and/or sleep and/or nutrition and/or mental wellbeing and/or substance use and/or environmental conditions, in that the system captures a multitude of risk shape pattern for living individuals at least by means of said lifestyle factor values, wherein the captured multitude of risk shape pattern are clustered by the system, each cluster defining a prototype of risk shape pattern assigned to a dedicated relative risk measurand, in that a newly captured risk shape pattern of a living individual is mapped by the system to one of the prototypes of risk shape pattern based on measured lifestyle factor values associated with the living individual, wherein the dedicated relative risk measurand assigned to the risk shape pattern is outputted as resilience score value of the living individual. The invention has inter alia the advantage that it provides an increasingly optimized risk assessment improving the technical accuracy, flexibility and transparency of such automated system. Further, the system allows to provide and dynamically adapt clustering and rating based on prototypes of risk shapes, and thus providing an automated underwriting risk measurement and rating with high technical accuracy. The invention also has, inter alia, the advantage that the system can be modified or adapted to produce relevant individual (consumer) facing scores to match a specific development in a given market (e.g. relative risk rating, health age etc.). The inventive system also has the advantage, that it allows a dimensionality reduction to rely on relevant measuring parameters that can improve the prediction power of the modeling structure without having redundancies. The data dimension can e.g. be reduced by the system itself, using integrated feature selection techniques and feature extraction namely, Correlation-Based Feature Selection (CFS) and Principal Components Analysis (PCA). Machine learning algorithms, namely Multiple Linear Regression, Artificial Neural Network, REPTree and Random Tree classifiers were implemented on the dataset to predict the risk level of applicants.

“In an embodiment variant, the parametrization of the risk shape pattern further comprises clinical factor values measuring (i) build factor values comprising a measured height and/or weight factor value, and/or (ii) lipids factor values comprising a measured total cholesterol factor value and/or a high-density lipoprotein factor value and/or a triglycerides factor value, and/or (iii) blood pressure factor values comprising a measured systolic and diastolic blood pressure factor value, and/or (iv) glucose metabolism factor values comprising a measured fasting/non-fasting glucose factor value and/or glycated hemoglobin (hemoglobin A1c) factor value and/or diabetes status, and/or (v) liver function factor values comprising a measured gamma-glutamyltransferase (GGT) factor value and/or an alanine transaminase (ALT) factor value and/or an aspartate transaminase (AST) factor value and/or an alkaline phosphatase (Alk Phos), and/or (vi) family history of diabetes and circulatory disorders factor values. The parametrization of the risk shape pattern can e.g. further comprise clinical factor values measuring (i) a measured calcium score factor value, and/or (ii) a measured C-reactive protein factor value, and/or (iii) a measured heart rate variability factor value. This embodiment variant has, inter alia, the same advantages as the previous one, however, allows to even improve further the technical accuracy, flexibility and transparency of the automated system.

“In another embodiment variant, in case of defined threshold values are exceeded by at least one of the captured factor values, the parametrization of the risk shape pattern can e.g. be further extended by the input measures comprising (i) a measured waist circumference factor value, if a threshold value of one of the build factor values is exceeded, and/or (ii) a measured apolipoproteins factor value, if a threshold value of at least one of the lipid factor values is exceeded, and/or (iii) a measured relationship and/or relative diagnosis age of family history factor value, if a threshold value of at least one of the family history factor values is exceeded, and/or (iv) a factor value indicating certain diabetes sub-types, if a threshold value of at least one of the glucose metabolism factor values is exceeded, and/or (iv) a measured sport-driven physical activity measures value, if a threshold value of at least one of the physical activity factor values is exceeded, and/or (v) a measured activity intensity qualifier factor value, if a threshold value of at least one of the physical activity factor values is exceeded, and/or (vi) a measured binge drinking value, if a threshold value of a drinking factor value is exceeded, and/or (vii) measured factor values based on an additional popular screening questionnaires, if a threshold value indicating mental wellness is exceeded. The sport-driven physical activity measures can e.g. comprise factor values at least indicating cycling and/or swimming activity of the individual. This embodiment variant has, inter alia, the same advantages as the previous one, however, allows to even improve further the technical accuracy, flexibility and transparency of the automated system.

“In still another embodiment variant, the system comprises a machine-based, automated simulation structure modelling and capturing interactive effects between any of the factors of the parametrization of the risk shape pattern, the factors providing the input measures to the system. The machine-based, automated simulation structure can e.g. be machine-learning based. This embodiment variant has, inter alia, the same advantages as the previous one, however, allows the system to be further optimized by reducing the used measuring parameters to the minimal needed number. Thus, this embodiment variant, inter alia, allows to optimize data processing time and efficiency, due to a minimized set of required input measuring values.”

There is additional summary information. Please visit full patent to read further.”

The claims supplied by the inventors are:

“1. A measuring system for measuring relative occurrence frequencies or measurable occurrence probabilities of a medical event and/or health event and/or life risk event associated with living individuals in a measurement future time window as relative risk measurand values based on measured physical parameter values by sensory and measuring devices, wherein the relative risk measurands providing a measure for the frequency of physical occurrences of specific medical events and/or health events and/or life risk events having impacting consequences in specified ranges to the living individual within a defined cohort of living individuals relative to a randomized cohort of living individuals, the measuring system comprising: processing circuitry configured to perform a defined parametrization for capturing risk shape pattern for living individual, the parametrization of the risk shape pattern comprising at least lifestyle factor values measuring physical activity and/or sleep and/or nutrition and/or mental wellbeing and/or substance use and/or environmental conditions, wherein the risk shape pattern further comprises clinical factor values measuring (i) build factor values comprising a measured height and/or weight factor value, and/or (ii) lipids factor values comprising a measured total cholesterol factor value and/or a high-density lipoprotein factor value and/or a triglycerides factor value, and/or (iii) blood pressure factor values comprising a measured systolic and diastolic blood pressure factor value, and/or (iv) glucose metabolism factor values comprising a measured fasting/non-fasting glucose factor value and/or glycated hemoglobin (hemoglobin A1c) factor value and/or diabetes status, and/or (v) liver function factor values comprising a measured gamma-glutamyl-transferase (GGT) factor value and/or an alanine transaminase (ALT) factor value and/or an aspartate transaminase (AST) factor value and/or an alkaline phosphatase (Alk Phos), and/or (vi) family history of diabetes and circulatory disorders factor values, perform a dimensionality reduction for reducing the used number of parameters by an integrated feature selection and feature extraction structure, wherein the system comprises a correlation-based feature selection structure for generating subsets of attributes by selecting a subset of features containing highly correlated features with a class of features, but uncorrelated to each other based on generated correlation values comprising at least Pearson’s correlation coefficient, minimum description length (MDL), symmetrical uncertainty, and relief, and wherein the system comprises a principal components analysis (PCA) structure for unsupervised linear feature extraction reducing the size of the data by extracting features having most information, capture a multitude of risk shape pattern for living individuals at least by said lifestyle factor values, wherein the captured multitude of risk shape pattern are clustered by the system, each cluster defining a prototype of risk shape pattern assigned to a dedicated relative risk measurand, wherein a newly captured risk shape pattern of a living individual is mapped by the system to one of the prototypes of risk shape pattern based on measured lifestyle factor values associated with the living individual, and wherein the dedicated relative risk measurand assigned to the risk shape pattern is outputted as resilience score value of the living individual.

“2. The measuring system for measuring relative risk measurands of a living individual according to claim 1, wherein the parametrization of the risk shape pattern further comprises clinical factor values measuring (i) a measured calcium score factor value, and/or (ii) a measured C-reactive protein factor value, and/or (iii) a measured heart rate variability factor value.

“3. The measuring system for measuring relative risk measurands of a living individual according to claim 1, wherein, in case of defined threshold values are exceeded by at least one of the captured factor values, the parametrization of the risk shape pattern is further extended by the input measures comprising (i) a measured waist circumference factor value, if a threshold value of one of the build factor values is exceeded, and/or (ii) a measured apolipoproteins factor value, if a threshold value of at least one of the lipid factor values is exceeded, and/or (iii) a measured relationship and/or relative diagnosis age of family history factor value, if a threshold value of at least one of the family history factor values is exceeded, and/or (iv) a factor value indicating certain diabetes sub-types, if a threshold value of at least one of the glucose metabolism factor values is exceeded, and/or (iv) a measured sport-driven physical activity measures value, if a threshold value of at least one of the physical activity factor values is exceeded, and/or (v) a measured activity intensity qualifier factor value, if a threshold value of at least one of the physical activity factor values is exceeded, and/or (vi) a measured binge drinking value, if a threshold value of a drinking factor value is exceeded, and/or (vii) measured factor values based on an additional popular screening questionnaires, if a threshold value indicating mental wellness is exceeded.

“4. The measuring system for measuring relative risk measurands of a living individual according to claim 3, wherein the sport-driven physical activity measures comprise factor values at least indicating cycling and/or swimming activity of the individual.

“5. The measuring system for measuring relative risk measurands of a living individual according to claim 1, wherein the system comprises a machine-learning based simulation structure modelling and capturing interactive effects between any of the factors of the parametrization of the risk shape pattern, wherein the factors provide the input measures to the system.

“6. A digital, cloud-based marketplace platform providing automated, risk underwriting and risk assessment for health and/or life risks by configuring, launching and processing of customized firs-tier and/or second-tier risk-transfer products for risk-exposed living individuals as first units and carriers/brokers as second units, wherein an automated risk-transfer placement is provided by the digital platform in a digital environment by a first online channel comprising a parameter-driven, rule-based underwriting process for creating or participating at risk-transfer structures by a pricing and underwriting engine, wherein an automated claim handling is provided by the platform by a claim triage and handling engine as a second online channel, and wherein an automated accounting is provided by the platform by a balance sheet provision and management engine and policy administration engine as a third online channel, wherein the digital platform comprises the measuring system for measuring relative risk measurands of living individuals based on physical parameter values measured by sensory and measuring devices as resilience score according to claim 1 and each digital service of the platform by a contribution measure to an extended resilience score of the living individuals purchasing the risk-transfer products on the digital marketplace platform and benefiting from the digital services of the digital platform, wherein the platform comprises a dimensionality reduction for reducing the used number of parameters by an integrated feature selection and feature extraction structure, wherein the platform comprises a correlation-based feature selection structure for generating subsets of attributes by selecting a subset of features containing highly correlated features with a class of features, but uncorrelated to each other based on generated correlation values comprising at least Pearson’s correlation coefficient, minimum description length (MDL), symmetrical uncertainty, and relief, and wherein the platform comprises a principal components analysis (PCA) structure for unsupervised linear feature extraction reducing the size of the data by extracting features having most information, wherein the extended resilience score provides a measure based on the measured current health status of the living individuals and/or the measured probability to purchase risk-transfer cover and/or the measured probability to start or keep behavior for improving their health status, wherein the measuring of the extended resilience score encompasses different type of risks at least comprising mortality risks and/or morbidity risks and/or longevity risks together with the probability to claim for a risk-transfer benefit and/or the measured evolving health status of the living individuals, and wherein the contribution measure to the extended resilience score is measured by assessing the variance of an individual’s extended resilience score by changing first individual’s parameters at least comprising adding or omitting a specific risk-transfer cover and/or triggering start or maintenance of a nutrition program.

“7. The digital, cloud-based marketplace platform according to claim 6, wherein detected changes in the extended resilience score triggers dynamically changes in pricing parameters and/or benefit parameters of one or more risk-transfer or financial products.

“8. The digital, cloud-based marketplace platform according to claim 7, wherein dynamically detected changes in the resilience score trigger at least dynamic assessment and/or reassessment and/or repricing by the digital platform at inception and throughout a duration of a risk-transfer and/or of the user relationship established by the digital platform.”

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

URL and more information on this patent application, see: ANGELAKOPOULOU, Aspasia; DUCKER, Michael; MARTIN, Alan; RIX, Douglas; SCHOONBEE, John. Digital System For Automated Measuring Of Relative Risk Measurands And Scores Of Living Individuals And Method Thereof. U.S. Patent Application Number 20240145096, filed September 27, 2023 and posted May 2, 2024. Patent URL (for desktop use only): https://ppubs.uspto.gov/pubwebapp/external.html?q=(20240145096)&db=US-PGPUB&type=ids

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