DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Status of the Claims
The status of the claims as of the response filed 5/12/2026 is as follows: Claims 2-3, 7-8, 13-14, and 18-19 are cancelled, and all previously given rejections for these claims are considered moot. Claims 1, 12, and 20 are currently amended. Claims 4-5, 15-16, and 21-26 are as previously presented. Claims 6, 9-11 and 17 are original. Claims 27-28 are new. Claims 1, 4-6, 9-12, 15-17, and 20-28 are currently pending in the application and have been considered below.
Response to Amendment
Rejection Under 35 USC 101
The claims have been amended but the 35 USC 101 rejections are upheld.
Rejection Under 35 USC 103
The amendments made to the claims introduce limitations that are not fully addressed in the previous office action, and thus the corresponding 35 USC 103 rejections are withdrawn. However, Examiner will consider the amended claims in light of an updated prior art search and address their patentability with respect to prior art below.
Response to Arguments
Rejection Under 35 USC 101
On pages 11-12 of the response filed 5/12/2026 Applicant argues that the amended claims “cannot be classified as managing personal behavior or interactions between people” because “it is impractical to consider that that [sic] one or more clinicians can analyze large quantities of data from multiple data sources and perform initial fitting/training of a predictive model using the categorized data to determine an outcome dataset with operating parameters for the model that specify how the model transforms inputs to outputs, and further fit/train the predictive model using the standardized population data and initial opening [sic] parameters as suggested by the Office.” Applicant’s arguments are fully considered, but are not persuasive. Examiner first notes that the claims as presently drafted do not require such vast quantities of data as Applicant appears to assert; “a plurality of data sources” could include as few as two data sources, “population data for a plurality of first individuals” could include data about as few as two patients, etc. Regardless, Examiner notes that the practicality or scale of one or more human actors performing the certain method of organizing human activity is not a consideration under this grouping of abstract idea; to fall into this category, the claim merely need recite limitations that describe certain methods of organizing human activity like managing personal behavior, relationships, or interactions between people. The purpose of the instant invention is to identify a patient population to target for clinical interventions or outreach, which Examiner maintains falls into this grouping because it describes mathematically-based patient population identification operations that personnel at a hospital, insurance company, or other interested clinical institution could perform. Examiner maintains that clinical personnel managing their personal behavior and interactions with colleagues and/or patients would be capable of organizing training data from various sources, fitting a predictive model to the training data at least twice, and calculating various metrics about patients in a population to stratify the patients and identify those with metrics above a threshold who would most likely benefit from enrollment into care management programs. Accordingly, Examiner maintains that the instant claims recite certain methods of organizing human activity.
On page 12 Applicant argues that the claims do not recite a mathematical concept, but “are merely based on or involve, a mathematical concept,” submitting that “the pending claims relate to identifying patients for enrollment into care interventions and do not set forth or describe any particular mathematical relationships, calculations, formulas, or equations using words of mathematical symbols.” Applicant points to Example 39 as an example of a claim found not to recite mathematical concepts, and Example 47 as an example of a claim that did recite mathematical concepts. Applicant’s arguments are fully considered, but are not persuasive. The eligible claim of Example 39 recited collecting a set of digital facial images from a database, applying one or more transformations to each digital facial image, creating a first training set comprising the collected set of digital facial images, the modified set of digital facial images, and a set of digital non-facial images, training a neural network in a first stage using the first training set, creating a second training set for a second stage of training comprising the first training set and digital non-facial images that are incorrectly detected as facial images after the first stage of training, and training the neural network in a second stage using the second training set. These operations are based on mathematical concepts (e.g. training a neural network is based on capturing mathematical relationships in a node-based architecture), but the claims themselves did not recite any specific formulas or calculations. In contrast, the instant claims do recite mathematical calculations, including at least standardizing population data to determine training data by determining a plurality of covariates, a plurality of impact factor datasets, and past population engagement in care interventions based on the population data; performing initial training of the HTE model to determine an HTE outcome dataset comprising model object parameters and a set of HTE model hyperparameters specific to the plurality of covariates and impact factor datasets that are configured to transform the population data to produce treatment effects for the first plurality of individuals; combining output information with additional metrics to generate combined strategic stratification metrics associated with a plurality of second individuals; and determining the plurality of second individuals for enrolling into the care management interventions based on comparing the combined strategic stratification metrics with one or more strategic stratification threshold values. These operations align with the examples provided in Example 47 as well as in MPEP 2106.04(a)(2)(I)(C) because they describe mathematical calculations used to determine numerical covariates, operating parameters, hyperparameters, etc., transform data from one value to another, make comparisons between data values, etc. Accordingly, Examiner maintains that the instant claims do recite mathematical concepts rather than just involving mathematical concepts.
On page 13 Applicant argues that the instant claims “reflect a technical improvement” because they are “directed to intelligently determining for enrolling into the care management interventions to maximize impact.” Applicant asserts that [0001] of the specification describes a technical problem that has been solved by the instant invention, including “determining and identifying individuals that benefit from care management interventions, which requires analyzing the impact of a plurality of clinical impact factors… to determine which individuals benefit the most from care management interventions.” Applicant’s arguments are fully considered, but are not persuasive. Examiner respectfully disagrees that identifying patients that would benefit most from care interventions based on analysis of various types of clinical data is a technical problem; this challenge reflects a difficulty in business decision-making for clinical entities like hospitals, health insurance providers, etc. but the specification provides no specific technological drawbacks or issues with existing machine learning modeling techniques, computing technology, or any other technical field that the invention seeks to solve. Merely improving the business decision-making of clinical entities in identifying patients for appropriate outreach or intervention by applying high-level AI or ML models to this abstract field does not amount to an improvement in technology or technical solution to a technical problem. Examiner notes that improvements to the abstract idea itself do not provide an improvement to technology, as outlined in MPEP 2106.05(a): “It is important to note, the judicial exception alone cannot provide the improvement. The improvement can be provided by one or more additional elements.” See also 2106.05(a)(II): “it is important to keep in mind that an improvement in the abstract idea itself… is not an improvement in technology.”
On pages 13-15 Applicant analogizes the instant claims with those found eligible in Ex Parte Desjardins, at least because they “are directed to methods of determining whether to target individuals for care management interventions that are scalable, and optimize staffing requirements.” Applicant further submits that the claims solve the technical problem of identifying individuals that benefit from care management interventions “in a computationally efficient way by performing initial training of the care management HTE model to determine an HTE outcome dataset for the care management HTE model and generate the training data based on the plurality of covariates, the plurality of impact factor datasets, the past population engagement, and the HTE outcome dataset.” Applicant’s arguments are fully considered, but are not persuasive. In Desjardins, the claims reflected an improvement to how a machine learning model itself is trained and operates to address the technical problem of ‘catastrophic forgetting' encountered in continual learning systems, which was identified and explained as a technical problem in the specification. In contrast, the instant specification does not outline a specific technical problem in machine learning technology whose solution is reflected in the claims. As explained above, Examiner respectfully disagrees that identifying patients for care management interventions is a technical problem. Examiner further notes that the specification does not describe any specific technical benefits of the initial model training step and subsequent creation of the training dataset, and there is no mention at all of improvements to computational efficiency. The training and use of AI/ML models in the instant invention appear to be invoked as means with which to digitize and/or automate otherwise-abstract patient data analysis and care intervention management operations such that patient care management decisions may be made more quickly or at scale via digital infrastructure, rather than as an effort to improve any underlying machine learning model architecture or training methods as was the case in Desjardins. Examiner notes that per MPEP 2106.05(f)(2), “‘claiming the improved speed or efficiency inherent with applying the abstract idea on a computer' does not integrate a judicial exception into a practical application or provide an inventive concept.” Applicant has not provided any evidence that the implementation of the abstract idea with high-level machine learning models and computing components provides any technical improvements beyond the improved speed or efficiency inherent with applying patient care management operations in an automated/computing environment.
On page 15 Applicant argues that the claims should be found patent eligible under Step 2B because they are “novel and non-obvious,” and recite “a combination of features which go beyond what is well-understood, routine, and conventional,” specifically pointing to alleged deficiencies of the Gopal, Luo, and Winlo references to “disclose or suggest at least certain features from the independent claims.” Applicant’s arguments are fully considered, but are not persuasive. Applicant has not identified any specific additional elements or combination of additional elements that are believed to be unconventional, and appears to broadly assert that the cited prior art fails to teach or suggest some of the features of the invention and thus they amount to an unconventional combination. Examiner notes that issues of patentability over the prior art are a separate consideration to the question of eligibility under 35 USC 101; MPEP 2106.05(I) states that:
Although the courts often evaluate considerations such as the conventionality of an additional element in the eligibility analysis, the search for an inventive concept should not be confused with a novelty or non-obviousness determination. See Mayo, 566 U.S. at 91, 101 USPQ2d at 1973 (rejecting "the Government’s invitation to substitute §§ 102, 103, and 112 inquiries for the better established inquiry under § 101 "). As made clear by the courts, the "‘novelty’ of any element or steps in a process, or even of the process itself, is of no relevance in determining whether the subject matter of a claim falls within the § 101 categories of possibly patentable subject matter." Intellectual Ventures I v. Symantec Corp., 838 F.3d 1307, 1315, 120 USPQ2d 1353, 1358 (Fed. Cir. 2016) (quoting Diamond v. Diehr, 450 U.S. at 188–89, 209 USPQ at 9).
Accordingly, whether the claims are found to be novel and/or non-obvious over the prior art has no bearing on analysis of patent eligibility under 35 USC 101.
Further, the only additional elements beyond that abstract idea itself recited in the claims include a computing platform comprising one or more processors executing instructions stored on a non-transitory computer-readable medium to perform the receiving, standardizing, determining, performing, generating, training, inputting, combining, and providing steps, as well as specifying that the care management HTE model comprises one or more care management machine learning – artificial intelligence (ML – AI) models and that the information is provided for display on a care management computing device. These additional elements merely serve to automate steps that could otherwise occur via a human actor managing their personal behavior and/or interactions with others, and thus amount to instructions to apply the abstract idea using generic computer components (see MPEP 2106.05(f)). Additionally, the combination of a computing platform executing ML-AI models and displaying information at a computing device for the purpose of treatment/intervention planning is a well-understood, routine, and conventional combination, as evidenced by at least Figs. 1-5 of Winlo et al. (US 20190156955 A1); abstract, Fig. 7, & [0071] of Basu et al. (US 20210241907 A1); and abstract & Fig. 1 of Hasan et al. (US 20230352134 A1).
For the reasons outlined above, the 35 USC 101 rejections are upheld.
Rejection Under 35 USC 103
On page 16 Applicant argues that “Gopal is silent regarding the model object parameters being specific to the plurality of covariates and the plurality of impact factor datasets and are configured to transform the population data to produce treatment effects for the first plurality of individuals.” Applicant submits that “at best, Gopal describes a chi square value that is a statistical measure that represents relationships between various variables (e.g., signals and readmissions)” but “makes no mention of any statistical measure that is configured to transform population data to produce treatment effects.” Applicant’s arguments are fully considered, but are not persuasive. Figs 1 & 3 and paras. [0015] & [0020]-[0021] of Gopal describe training a predictive model using the prepared training data (i.e. including the covariates, impact factor datasets, and past population engagement data as explained in para. 26 of the non-final rejection mailed 2/20/2026). Predictive models such as decision trees, regression equations, and neural networks (the examples of model types trained by Gopal’s system in [0015]) are collections of learned mathematical/statistical relationships (i.e. model object parameters) that transform specific inputs into a desired output based on the specific associations learned from the training data, such that training of these types of models is considered to determine an outcome dataset comprising model object parameters that specify an operation of the model during implementation to transform input data (i.e. population data) into desired outputs (i.e. treatment effects) and are specific to the input training data containing the covariates and impact factor datasets. Accordingly, Examiner maintains the combination of Gopal, Luo, and Winlo in the updated 35 USC 103 rejections below.
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1, 4-6, 9-12, 15-17, and 20-28 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1
In the instant case, claims 1, 4-6, and 9-11 are directed to a method (i.e. a process), claims 12, 15-17, and 25-27 are directed to a platform (i.e. a machine), and claims 20-24 and 28 are directed to a non-transitory computer-readable medium (i.e. a manufacture). Thus, each of the claims falls within one of the four statutory categories. Nevertheless, the claims fall within the judicial exception of an abstract idea.
Step 2A – Prong 1
Independent claims 1, 12, and 20 recite steps that, under their broadest reasonable interpretations, cover certain methods of organizing human activity, e.g. managing personal behavior, relationships, or interactions between people. Specifically, claim 12 (as representative) recites:
An enterprise computing platform, comprising: one or more processors; and a non-transitory computer-readable medium having processor-executable instructions stored thereon, wherein the processor-executable instructions, when executed by the one or more processors, facilitate:
receiving, from a plurality of data sources, population data for a plurality of first individuals;
standardizing the population data to determine training data for a care management heterogeneous treatment effect (HTE) model, wherein standardizing the population data comprises:
determining, based on the population data, a plurality of covariates, a plurality of impact factor datasets for the plurality of first individuals, and past population engagement in care management interventions for the plurality of first individuals;
performing initial training of the care management HTE model using the plurality of covariates, the plurality of impact factor datasets, and the past population engagement to determine an HTE outcome dataset for the care management HTE model, wherein the HTE outcome dataset comprises model object parameters and a set of HTE model hyperparameters that specify an operation of the care management HTE model during implementation, wherein the model object parameters are specific to the plurality of covariates and the plurality of impact factor datasets and are configured to transform the population data to produce treatment effects for the first plurality of individuals; and
generating the training data based on the plurality of covariates, the plurality of impact factor datasets, the past population engagement, and the HTE outcome dataset;
training the care management HTE model using the training data, wherein the care management HTE model comprises one or more care management machine learning - artificial intelligence (ML - Al) models;
determining a plurality of second individuals for the care management interventions based on using the trained care management HTE model, wherein determining the plurality of second individuals for the care management interventions comprises:
inputting a plurality of new impact factors and a plurality of new covariate datasets into the one or more care management ML – AI models to determine output information for the plurality of second individuals;
combining the output information with additional metrics to generate combined strategic stratification metrics associated with the plurality of second individuals; and
determining the plurality of second individuals for enrolling into the care management interventions based on comparing the combined strategic stratification metrics with one or more strategic stratification threshold values; and
providing, for display on a care management computing device, information indicating the plurality of second individuals for the care management interventions.
But for the recitation of generic computer components like a processor executing instructions stored in a non-transitory computer-readable medium and a care management computing device, the italicized functions, when considered as a whole, describe treatment effectiveness determination and patient population identification operations that could otherwise be achieved by human actors (e.g. a clinician, researcher, or administrator) managing their personal behavior and interactions with others (e.g. colleagues or patients). For example, a clinician could look up population data from a plurality of data sources (studies, case reports, written resources, colleagues’ experiences, etc.), standardize the population data by categorizing the data into different categories, perform initial fitting/training of a predictive model using the categorized data to determine an outcome dataset with operating parameters specifically gleaned for the model that specify how the model transforms inputs to outputs (e.g. covariates in a regression equation, cutoff thresholds for a decision tree, etc.), and further fit/train the predictive model using the standardized population data and initial operating parameters. The clinician could then use the fitted/trained model to determine which new patients may benefit from certain care management interventions by inputting new patient data into the model and receiving predictions output from the model, considering the model output with other known patient metrics to calculate combined stratification metrics, and compare the stratification metrics with one or more thresholds to identify the patients most likely to benefit from enrollment into certain care interventions. The clinician could finally visually indicate (e.g. in a report, graph, or other visual means) the identified patients so that they may be contacted for enrollment into the indicated care interventions. Thus, claim 12 recites an abstract idea in the form of a certain method of organizing human activity. Claims 1 and 20 recite substantially similar subject matter as claim 12 and are found to recite an abstract idea under the same analysis.
The independent claims also recite steps that can be considered to recite an abstract idea in the form of mathematical concepts in addition to certain methods of organizing human activity. Such steps include:
standardizing the population data to determine training data for a care management heterogeneous treatment effect (HTE) model, wherein standardizing the population data comprises:
determining, based on the population data, a plurality of covariates, a plurality of impact factor datasets for the plurality of first individuals, and past population engagement in care management interventions for the plurality of first individuals;
performing initial training of the care management HTE model using the plurality of covariates, the plurality of impact factor datasets, and the past population engagement to determine an HTE outcome dataset for the care management HTE model, wherein the HTE outcome dataset comprises model object parameters and a set of HTE model hyperparameters that specify an operation of the care management HTE model during implementation, wherein the model object parameters are specific to the plurality of covariates and the plurality of impact factor datasets and are configured to transform the population data to produce treatment effects for the first plurality of individuals; and
generating the training data based on the plurality of covariates, the plurality of impact factor datasets, the past population engagement, and the HTE outcome dataset;
training the care management HTE model using the training data, wherein the care management HTE model comprises one or more care management machine learning - artificial intelligence (ML - Al) models;
determining a plurality of second individuals for the care management interventions based on using the trained care management HTE model, wherein determining the plurality of second individuals for the care management interventions comprises:
inputting a plurality of new impact factors and a plurality of new covariate datasets into the one or more care management ML – AI models to determine output information for the plurality of second individuals;
combining the output information with additional metrics to generate combined strategic stratification metrics associated with the plurality of second individuals; and
determining the plurality of second individuals for enrolling into the care management interventions based on comparing the combined strategic stratification metrics with one or more strategic stratification threshold values.
These steps describe a process for determining covariates and other data from a dataset, training a mathematical predictive model by identifying model object parameters and hyperparameters, using the trained mathematical model to predict outputs, using the predicted outputs to calculate additional metrics, and comparing the additional metrics to thresholds to identify patients for enrollment into care interventions, which amounts to a mathematically-based data analysis process and thus fits in the “mathematical concepts” grouping of abstract idea. Claims 1 and 20 recite substantially similar subject matter as claim 12 and are found to recite an abstract idea under the same analysis.
Dependent claims 4-6, 9-11, 15-17, and 21-28 inherit the limitations that recite an abstract idea from their dependence on claims 1, 12, or 20, and thus these claims also recite an abstract idea under the Step 2A – Prong 1 analysis. In addition, claims 4-6, 9-11, 15-17, and 21-28 recite additional limitations that further describe the abstract ideas identified in the independent claims. Specifically, claim 21 recites that the care management HTE model comprises a plurality of treatment effect model parameter values, and that the set of HTE model hyperparameters are associated treatment effect model parameter values from the plurality of treatment effect model parameter values that are above an accuracy threshold. These limitations merely further describe mathematical workings of the model in an abstract way such that they also describe an abstract idea. Claims 4, 15, and 22 further describe the types of data that fit into the impact factor datasets, each of which are types of data that a human actor would be capable of accessing, categorizing, and evaluating via mathematical operations. Claims 5, 16, and 23 further describe using impact factor datasets and covariates to train the model, which a human actor could accomplish by using such data types to fit/train a predictive model via mathematical processes as indicated for the independent claims above. Claims 6, 17, and 24 recite determining past population outcomes indicating post-engagement clinical and/or financial healthcare outcomes for the first individuals and utilizing such data in the model training step, which a human actor could achieve by making determinations about post-engagement outcomes based on the population data and using such outcome data to fit/train the predictive model as indicated for the independent claims above. Claims 9-11 and 25-28 specify various types of health plans that the patient populations may be enrolled in; a human actor would be capable of obtaining and evaluating data from patients enrolled in these types of plans.
However, recitation of an abstract idea is not the end of the analysis. Each of the claims must be analyzed for additional elements that indicate the abstract idea is integrated into a practical application to determine whether the claim is considered to be “directed to” an abstract idea.
Step 2A – Prong 2
The judicial exception is not integrated into a practical application. In particular, independent claims 1, 12, and 20 do not include additional elements that integrate the abstract idea into a practical application. The additional elements of claims 1, 12, and 20 include a computing platform comprising one or more processors executing instructions stored on a non-transitory computer-readable medium to perform the receiving, standardizing, determining, performing, generating, training, inputting, combining, comparing, providing, etc. steps, as well as specifying that care management HTE model comprises one or more care management machine learning – artificial intelligence (ML – AI) models and that the information is provided for display on a care management computing device. These additional elements, when considered in the context of each claim as a whole, merely serve to automate steps that could occur via a human actor managing their personal behavior and/or interactions with others (as described above), and thus amount to instructions to “apply” the abstract idea using generic computer components (see MPEP 2106.05(f)). For example, use of the computing platform to perform the various steps merely digitizes/automates the otherwise-abstract steps of receiving population data, standardizing the population data by making determinations about the data and performing initial fitting/training of a predictive model, performing further fitting/training of the model, determining a plurality of second individuals by using the model to predict outcomes, combine the predicted outcomes with additional metrics, and compare the stratification metrics to a threshold, and finally providing information indicating the plurality of second individuals such that they take place in a computerized environment. Specifying that the HTE model is an AI-ML model merely utilizes the high-level concept of artificial intelligence or machine learning as a means to digitize/automate the otherwise-abstract steps of fitting/training and using a predictive model. Specifying that the information indicating the plurality of second individuals is provided for display at a care management computing device again merely invokes a high-level computing device as a means with which to digitize the output of information from a predictive model such that it occurs in a computerized environment. Accordingly, these additional elements are merely invoked as tools with which to digitize/automate the otherwise abstract functions of the invention, and claims 1, 12, and 20 as a whole are each directed to an abstract idea without integration into a practical application.
The judicial exception recited in dependent claims 4-6, 9-11, 15-17, and 21-28 is also not integrated into a practical application under a similar analysis as above. Claims 4-6, 9-11, 15-17, and 21-28 are performed with the same additional elements introduced in the independent claims, without introducing any new additional elements of their own, and accordingly also amount to mere instructions to apply the abstract idea.
Accordingly, the additional elements of claims 1, 4-6, 9-12, 15-17, and 20-28 do not integrate the abstract idea into a practical application because they do not impose any meaningful limits on practicing the abstract idea. Claims 1, 4-6, 9-12, 15-17, and 20-28 are directed to an abstract idea.
Step 2B
The claims do not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to integration of the abstract idea into a practical application, the additional elements of a computing platform comprising processors and a non-transitory computer-readable medium, a specifically ML-AI HTE model, and a care management computing device for performing the receiving, standardizing, determining, performing, generating, training, inputting, combining, comparing, providing for display, etc. steps of the invention amount to mere instructions to apply the exception using generic computer components. As evidence of the generic nature of the above recited additional elements, Examiner notes the following portions of Applicant’s specification:
[0039], noting various examples of known computing elements that may be embody the computing platform, such as computing devices, computing platforms, systems, servers, engines, software functions, applications, etc.
[0040], noting various examples of known computing devices that may be embody the care management computing device, such as a desktop, laptop, table, mobile device, smart watch, IoT device, etc.
[0042], noting a generic arrangement of an exemplary computing device within the system.
[0057], noting the ML-AI models “may be any type of ML-AI model (e.g., unsupervised, supervised, and/or deep learning)” and providing several examples of known ML-AI model types such as XGBoost regression/classifier, causal forest, EconML, etc.
These disclosures do not indicate that the elements of the invention are particular machines and instead provide generic, high-level examples of known computer hardware and ML-AI model types, such that one of ordinary skill in the art would understand that any generic computing platform, ML-AI models, and computing device could be used to implement the invention. Further, the combination of these additional elements is not expanded upon in the specification as a unique arrangement and as such relies on the knowledge of one of ordinary skill in the art to understand the combination of components within a computer system as a well-known and generic combination for automating an abstract idea that could otherwise be performed as a certain method of organizing human activity and thus do not provide an inventive concept. Additionally, the combination of a computing platform executing ML-AI models and displaying information at a computing device for the purpose of treatment/intervention planning is a well-understood, routine, and conventional combination, as evidenced by at least Figs. 1-5 of Winlo et al. (US 20190156955 A1); abstract, Fig. 7, & [0071] of Basu et al. (US 20210241907 A1); and abstract & Fig. 1 of Hasan et al. (US 20230352134 A1).
Analyzing these additional elements as an ordered combination adds nothing that is not already present when considering the elements individually; the overall effect of the computer platform, ML-AI models, and computing device in combination is to digitize and/or automate a treatment effectiveness and patient population identification operation that could otherwise be achieved as a certain method of organizing human activity. Thus, when considered as a whole and in combination, claims 1, 4-6, 9-12, 15-17, and 20-28 are not patent eligible.
Claim Rejections - 35 USC § 103
This application currently names joint inventors. In considering patentability of the claims the examiner presumes that the subject matter of the various claims was commonly owned as of the effective filing date of the claimed invention(s) absent any evidence to the contrary. Applicant is advised of the obligation under 37 CFR 1.56 to point out the inventor and effective filing dates of each claim that was not commonly owned as of the effective filing date of the later invention in order for the examiner to consider the applicability of 35 U.S.C. 102(b)(2)(C) for any potential 35 U.S.C. 102(a)(2) prior art against the later invention.
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
Claims 1, 4-6, 9, 12, 15-17, 20-25, and 28 are rejected under 35 U.S.C. 103 as being unpatentable over Gopal et al. (US 20160358282 A1) in view of Luo (Reference U on the PTO-892 mailed 2/20/2026) and Winlo et al. (US 20190156955 A1).
Claims 1, 12, and 20
Gopal teaches a method, comprising:
receiving, by a computing platform and from a plurality of data sources, population data for a plurality of first individuals (Gopal Fig. 1, [0007], [0015], noting a computerized system obtains historical data for a patient population from a variety of sources, e.g. claims data, clinical/health program participation data, consumer data, etc.);
standardizing, by the computing platform, the population data to determine training data for a care management heterogeneous treatment effect (HTE) model (Gopal Figs. 1 & 3, [0015], [0019]-[0020], noting the obtained historical population data is cleansed, mined, and otherwise prepared for use in training a predictive model), wherein standardizing the population data comprises:
determining, based on the population data, a plurality of covariates, a plurality of impact factor datasets for the plurality of first individuals, and past population engagement in care management interventions for the plurality of first individuals (Gopal Figs. 2 & 4, [0015], [0018]-[0021], noting various types of patient factors from the obtained historical population data are used to train the predictive model, including covariates (e.g. demographic data like age and gender, diagnosis, medications, etc.), impact factors (e.g. hospital admit count, comorbidity index or count, medication count, CMS risk score, etc.), and past population engagement in care management interventions (e.g. clinical/health program participation data));
performing initial training of the care management HTE model using the plurality of covariates, the plurality of impact factor datasets, and the past population engagement to determine an HTE outcome dataset for the care management HTE model, wherein the HTE outcome dataset comprises model object parameters (Gopal Figs 1 & 3, [0015], [0020]-[0021], noting a predictive model is trained using the prepared training data, i.e. including the covariates, impact factor datasets, and past population engagement data as explained above. Predictive models such as decision trees, regression equations, and neural networks (the examples of model types trained by the system in [0015]) are collections of learned mathematical/ statistical relationships (i.e. model object parameters) that transform specific inputs into a desired output based on the specific associations learned from the training data, such that training of these types of models is considered to determine an outcome dataset comprising model object parameters that specify an operation of the model during implementation to transform input data (i.e. population data) into desired outputs (i.e. treatment effects) and are specific to the input training data containing the covariates and impact factor datasets); and
generating the training data based on the plurality of covariates, the plurality of impact factor datasets, the past population engagement, and the HTE outcome dataset (Gopal Fig. 3, [0020], noting prepared population data (i.e. including the covariates, impact factor datasets, and past population engagement data as explained above) is used to generate a validation dataset that is used to tune the trained model (which would include the statistical relationships learned from the initial training operation));
training, by the computing platform, the care management HTE model using the training data, wherein the care management HTE model comprises one or more care management machine learning - artificial intelligence (ML - Al) models (Gopal Figs 1 & 3, [0015], [0020]-[0021], noting a predictive model is trained and tuned using the prepared training data, including tuning (i.e. retraining) of the model with the validation dataset as in [0020]; the predictive model can be used to make care management intervention enrollment decisions as in [0026] and is thus considered equivalent to a care management HTE model comprising one or more care management ML-AI models in accordance with Applicant’s definition of such models in para. [0057] of the specification as “any type of ML – AI model (e.g., unsupervised, supervised, and/or deep learning) that can be used to determine (e.g., identify) individuals to be enrolled into care management interventions”);
determining, by the computing platform, a plurality of second individuals for the care management interventions based on using the trained care management HTE model (Gopal Figs. 1 & 6A-B, [0026]-[0027], [0031], noting new patient data is applied to the trained model to determine which patients should be selected for or enrolled in certain interventions), wherein determining the plurality of second individuals for the care management interventions comprises:
inputting a plurality of new impact factors and a plurality of new covariate datasets into the one or more care management ML – AI models to determine output information for the plurality of second individuals (Gopal Figs. 6A-B, [0023], [0026]-[0031], noting new patient profiles with various factors (i.e. impact factors and covariates) may be input to the trained model to obtain a readmission risk output);
combining the output information with additional metrics (Gopal Figs. 6A-B, [0026]-[0029], noting patient readmission risks (i.e. the output information) are evaluated in combination with other filtering or selection criteria (i.e. additional metrics) to determine which patients should be enrolled in the care management interventions); and
determining the plurality of second individuals for enrolling into the care management interventions based on comparing the output with one or more strategic stratification threshold values (Gopal Figs. 6A-B, [0026]-[0029], noting patient readmission risks (i.e. the output information) are compared to a threshold to determine which patients should be enrolled in the care management interventions); and
providing, by the computing platform and (Gopal [0031], claim 15, noting a daily referral list of the identified patients may be generated for review by a computer user, i.e. provided on a computer device).
In summary, Gopal teaches a computerized method of training, tuning, and using a predictive model to determine patient populations to undergo care management interventions. The predictive model of the system can capture and learn statistical relationships between inputs and a desired output such that the system is considered to learn model object parameters that specify an operation of the care management HTE model during implementation. However, Gopal is silent regarding determining HTE model hyperparameters. Additionally, in Gopal a readmission risk score output from the model may be compared with a threshold and then considered in combination with additional filtering criteria (i.e. other metrics) to identify patients for care interventions, but the risk score is not explicitly combined with additional metrics prior to comparison to generate combined strategic stratification metrics that are then compared to the strategic stratification threshold. Finally, in Gopal a list of the identified patients may be generated and provided for review by a computer user, indicating some manner of computerized user interface to output the identified patients. However, the reference does not specify that there is any visual display of information about the identified patients, and thus fails to explicitly disclose providing, by the computing platform and for display on a care management computing device, information indicating the plurality of second individuals for the care management interventions.
However, Luo teaches that machine learning models include both ordinary and hyper parameters that govern operation of the model, and that hyperparameters may be automatically tuned or selected based on measurements of accuracy associated with each potential combination of hyperparameters for a given algorithm (Luo abstract, Sections 1.2-1.3, Section 2.1). It would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to modify the known methods of training a machine learning model of Gopal to include determining both ordinary and hyper parameters of the model as in Luo in order to make the machine learning methods accessible to layman users such as clinicians while skipping the manual and labor-intensive process of selecting an effective algorithm and/or combination of hyperparameter values (as suggested by Luo section 1.3).
Additionally, Winlo teaches an analogous computerized method for identifying target patient populations for care management interventions (Winlo abstract) in which a calculated risk score is combined with other calculated metrics for a patient to generate combined stratification metrics that are then compared with one or more thresholds to identify patients for enrollment into a care intervention (Winlo [0085], noting “the selection module 230 may identify candidate members as target members based on various combinations of the risk, benefit, and participation scores exceeding a certain threshold value” and “the selection module 230 may identify candidate members as target members based on the aggregate of the risk, participation and benefit scores exceeding a certain threshold value,” showing that a risk score (i.e. analogous to the readmission risk score of Gopal) may be combined with other metrics for comparison to one or more stratification thresholds) and where information about the identified target patient populations may be provided to a care management computing device for display (Winlo [0034], [0087], noting graphical presentation of a user interface at a client device including information about target members). It would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to modify the threshold-based patient risk score stratification methods of the combination to include aggregation of additional patient metrics prior to comparison to a threshold score as in Winlo in order to improve upon crude, single-data-type patient identification cutoffs and consider additional important patient-level metrics that impact the success of intervention programs so that human and computer resources are more efficiently utilized in reaching out to targeted patients that are actually most likely to engage with and benefit from care interventions (as suggested by Winlo [0004]). It further would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to modify the generation of a patient list as in the combination such that it is graphically displayed for visual review as in Winlo in order to allow users such as clinicians to actually get useful, actionable information about which patients to target for intervention presented in a graphical manner (as suggested by Winlo [0034] & [0087]).
Regarding claim 12, Gopal in view of Luo and Winlo teaches an enterprise computing platform, comprising: one or more processors; and a non-transitory computer-readable medium having processor-executable instructions stored thereon, wherein the processor-executable instructions, when executed by the one or more processors, facilitate (Gopal [0023], claims 8 & 15, noting a computerized system performing the method, such as a server or computer executing programming instructions) the method of claim 1, as explained above.
Regarding claim 20, Gopal in view of Luo and Winlo teaches a non-transitory computer-readable medium having processor-executable instructions stored thereon, wherein the processor-executable instructions, when executed, facilitate (Gopal [0023], claims 8 & 15, noting a computerized system performing the method, such as a server or computer executing programming instructions) the method of claim 1, as explained above.
Claims 4, 15, and 22
Gopal in view of Luo and Winlo teaches the method of claim 1, and the combination further teaches wherein the plurality of impact factor datasets comprises a plurality of impact factor metrics, wherein the plurality of impact factor metrics comprise fall risks, emergency room (ER) risks, medical adherence indicators, chronic condition counts, mental illness indicators, usage of durable medical equipment (DME), new onset of diseases, and/or drug safety indicators (Gopal Figs. 2 & 4, [0018]-[0021], noting various types of patient factors from the obtained historical population data are used to train the predictive model, including impact factors like CMS risk score or hospital admit or readmit count (equivalent to ER risks), comorbidity index or count (equivalent to chronic condition counts), medication count (equivalent to drug safety indicators), etc.)).
Claims 15 and 22 recite substantially similar subject matter as claim 4, and are also rejected as above.
Claims 5, 16, and 23
Gopal in view of Luo and Winlo teaches the method of claim 1, and the combination further teaches wherein training the care management HTE model using the training data comprises: using the plurality of impact factor datasets and the plurality of covariates to train the one or more care management ML - AI models, wherein the plurality of impact factor datasets and the plurality of covariates are features for the one or more care management ML - AI models (Gopal Figs. 1-4, [0015], [0018]-[0021], noting the various patient factors from the obtained historical population data (i.e. the covariates and impact factor datasets, as explained above) are used to train the predictive model and result in the discovery of statistical relationships between the input variables that are selected as predictor features in the validated model).
Claims 16 and 23 recite substantially similar subject matter as claim 5, and are also rejected as above.
Claims 6, 17, and 24
Gopal in view of Luo and Winlo teaches the method of claim 5, and the combination further teaches tracking and utilizing intervention participation data (Gopal) as well as clinical outcome data (Gopal Fig. 2, [0018], noting length of stay) in training the predictive model. Thus, Gopal in view of Winlo teaches wherein standardizing the population data further comprises: determining past population outcomes for the care management interventions for the plurality of first individuals, wherein the past population outcomes indicate (Gopal Fig. 2, [0018]-[0021], noting clinical outcome data such as length of stay is used to train the predictive model; see also [0015], [0030], noting the system tracks and utilizes intervention participation data in training the predictive model).
In summary, the present combination teaches tracking and utilizing intervention participation data as well as clinical outcome data in training the predictive model, but it does not appear to specify that the clinical outcomes are post-engagement clinical outcomes resulting from undergoing certain interventions. Accordingly, the present combination fails to explicitly disclose wherein the past population outcomes indicate post-engagement clinical and/or financial healthcare outcomes for the plurality of first individuals after undergoing the care management interventions. However, Winlo further teaches analyzing historical member data including post-engagement clinical or financial outcomes of a member population after undergoing care management interventions to train a predictive model (Winlo [0055]-[0057]). It would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to modify the unconnected past population outcomes and participation data of the combination to specifically include evaluation of past population outcomes indicating post-engagement clinical and/or financial healthcare outcomes resulting from certain interventions as in Winlo in order to train the model to quantify the benefit that each intervention may provide to each member so that the targeting of patient populations most likely to benefit from care management interventions is improved (as suggested by Winlo [0056]).
Claims 17 and 24 recite substantially similar subject matter as claim 6, and are also rejected as above.
Claims 9, 25, and 28
Gopal in view of Luo and Winlo teaches the method of claim 1, and the combination further teaches wherein the plurality of first individuals and the plurality of second individuals are enrolled into MEDICARE (Gopal Fig. 3, [0019], noting Medicare claims as the member population).
Claims 25 and 28 recite substantially similar subject matter as claim 9, and are also rejected as above.
Claim 21
Gopal in view of Luo and Winlo teaches the non-transitory computer-readable medium of claim 20, and the combination further teaches wherein the care management HTE model comprises a plurality of treatment effect model parameter values, and wherein the set of HTE model hyperparameters are associated treatment effect model parameter values from the plurality of treatment effect model parameter values that are above an accuracy threshold (Gopal Figs 1 & 3, [0015], [0020]-[0021], noting the predictive model is trained using the prepared training data; predictive models such as decision trees, regression equations, and neural networks (the examples of model types trained by the system in [0015]) are collections of learned mathematical/statistical relationships that transform specific inputs into a desired output, such that these types of models are considered to comprise a plurality of treatment effect model parameter values. See also Luo abstract, Sections 1.2-1.3, Section 2.1, noting trained machine learning models include a plurality of ordinary and hyper parameters that are learned, and the combination of ordinary and hyper parameters correlated with the highest accuracy measures are selected for use (considered equivalent to Applicant’s disclosure of the “best” or “most accurate” parameters being determined in para. [0058] of the specification)).
Claims 10-11 and 26-27 are rejected under 35 U.S.C. 103 as being unpatentable over Gopal, Luo, and Winlo as applied to claims 1 or 12 above, and further in view of Chandra et al. (US 20190172564 A1).
Claims 10 and 26
Gopal in view of Luo and Winlo teaches the method of claim 1, and the combination further teaches that the system may be utilized by a health benefits provider with a covered patient-member population, e.g. Medicare (Gopal [0015], [0019]). However, the present combination fails to explicitly disclose wherein the plurality of first individuals and the plurality of second individuals are enrolled into MEDICAID. However, Chandra teaches an analogous predictive model training pipeline that utilizes data from patients enrolled in Medicaid (Chandra [0101], [0122]). It would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to modify the covered patient-member populations of the combination to include patients enrolled in Medicaid as in Chandra in order to train and utilize a predictive model specific to that known type of member-population.
Claim 26 recites substantially similar subject matter as claim 10, and is also rejected as above.
Claims 11 and 27
Gopal in view of Luo and Winlo teaches the method of claim 1, and the combination further teaches that the system may be utilized by a health benefits provider with a covered patient-member population, e.g. Medicare (Gopal [0015], [0019]). However, the present combination fails to explicitly disclose wherein the plurality of first individuals and the plurality of second individuals are enrolled into a commercial plan. However, Chandra teaches an analogous predictive model training pipeline that utilizes data from patients enrolled in commercial plans (Chandra [0101], [0122]). It would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to modify the covered patient-member populations of the combination to include patients enrolled in commercial plans as in Chandra in order to train and utilize a predictive model specific to that known type of member-population.
Claim 27 recites substantially similar subject matter as claim 11, and is also rejected as above.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure. DuBard et al. (Reference U on the accompanying PTO-892) describes methods of determining predictive model variables to target patients for health care intervention based on impactability.
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to KAREN A HRANEK whose telephone number is (571)272-1679. The examiner can normally be reached M-F 8:00-4:00 ET.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Shahid Merchant can be reached on 571-270-1360. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/KAREN A HRANEK/ Primary Examiner, Art Unit 3684
Prosecution Insights