DETAILED CORRESPONDENCE
This is a non-final office action on merits in response to the arguments and/or amendments filed on 02/11/2026 and the request for continued examination filed on 03/11/2026.
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 claims
Amendments to claims 1, 3, 5-6, 10, and 20-21 are acknowledged and have been carefully considered. Claims 1-21 are pending and considered below.
Continued Examination Under 37 CFR 1.114
A request for continued examination under 37 CFR 1.114, including the fee set forth in 37 CFR 1.17(e), was filed in this application after final rejection. Since this application is eligible for continued examination under 37 CFR 1.114, and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114. Applicant's submission filed on 03/11/2026 has been entered.
Subject Matter Free of Art
Claims 1-21 include subject matter that is free of prior art. The cited prior art of record fails to expressly teach or suggest, either alone or in combination, the features found within independent claims 1, 10, 20, and 21.
For claims 1, 20, and 21, the cited prior art of record fails to expressly teach or suggest, either alone or in combination, selecting, from patient data processed using a hybrid machine learning model comprising a static portion and a dynamic portion to generate an initial efficacy score, a defined subset of features, wherein the patient data includes at least one feature not included in the defined subset of features and the defined subset of features corresponds to an underrepresented outcome class in training data used to train the hybrid machine learning model, including determining whether the patient has experienced an adverse event when using a different medication in the same class as the medication being evaluated; generating an adjustment factor by processing the defined subset of features using an adjustment component; and scaling the initial efficacy score based on the adjustment factor to generate a final efficacy score, wherein final efficacy scores generated using the adjustment component more accurately adjust for the underrepresented outcome class as compared to initial efficacy scores generated using the hybrid machine learning model without the adjustment component.
For claim 10, the cited prior art of record fails to expressly teach or suggest, either alone or in combination, training a hybrid machine learning model comprising a static portion and a dynamic portion based on at least a subset of patient data and a determined efficacy of a medication, including generating an initial efficacy score by processing the patient data using the hybrid machine learning model; selecting, from the patient data processed using the hybrid machine learning model to generate the initial efficacy score, a defined subset of features, wherein the patient data includes at least one feature not included in the defined subset of features and the defined subset of features corresponds to an underrepresented outcome class in training data used to train the hybrid machine learning model, including determining whether the patient has experienced an adverse event when using a different medication in the same class as the medication; generating an adjustment factor by processing the defined subset of features using an adjustment component; and scaling the initial efficacy score based on the adjustment factor to generate a final efficacy score, wherein final efficacy scores generated using the adjustment component more accurately adjust for the underrepresented outcome class as compared to initial efficacy scores generated using the hybrid machine learning model without the adjustment component.
The closest prior art of record includes 1) Bostic et al. (U.S. Patent Publication 2021/0202102 A1), referred to hereinafter as Bostic, 2) Shriberg et al. (U.S. Patent Publication 2022/0165371 A1), referred to hereinafter as Shriberg, and 3) Brandes et al. (U.S. Patent Publication 2020/0364520 A1), referred to hereinafter as Brandes.
Bostic teaches receiving patient data, identifying a medication with respect to a patient, using patient attributes and medication information as features input to machine learned models, training machine learned models using patient treatment outcome data, and evaluating the efficacy of drug treatment plans. Bostic further teaches medication machine learned models and models corresponding to different medications or classes of medications. However, Bostic fails to teach or suggest selecting, from patient data processed using a hybrid machine learning model to generate an initial efficacy score, a defined subset of features corresponding to an underrepresented outcome class in training data used to train the hybrid machine learning model, and processing that defined subset of features using an adjustment component to generate an adjustment factor for scaling the initial efficacy score to generate a final efficacy score that more accurately adjusts for the underrepresented outcome class as compared to an initial efficacy score generated without the adjustment component. Bostic further fails to teach or suggest determining whether the patient experienced an adverse event when using a different medication in the same class as the medication being evaluated.
Shriberg teaches processing patient data using machine learning models to generate health state predictions and scores, including generating an intermediate score, processing patient data, weighting and aggregating metrics, normalizing patient assessments, and adjusting model outputs to improve the accuracy of patient health assessments. However, Shriberg fails to teach or suggest selecting, from patient data processed using a hybrid machine learning model to generate an initial efficacy score, a defined subset of features corresponding to an underrepresented outcome class in training data used to train the hybrid machine learning model. Shriberg further fails to teach or suggest processing such a defined subset of features using an adjustment component to generate an adjustment factor and scaling the initial efficacy score based on the adjustment factor to generate a final efficacy score that more accurately adjusts for the underrepresented outcome class as compared to an initial efficacy score generated using the hybrid machine learning model without the adjustment component.
Brandes teaches a machine learning classifier trained using training data, determining an underrepresented class associated with insufficient representation in the training data, obtaining data relating to the underrepresented class, and selecting data sets by comparing features of the obtained data with features of the underrepresented class. Brandes further teaches adding selected data associated with the underrepresented class to the training data and retraining the classifier to improve prediction of rare or underrepresented classes. However, Brandes fails to teach or suggest selecting, from patient data processed using a hybrid machine learning model to generate an initial efficacy score, a defined subset of features corresponding to an underrepresented outcome class and processing that defined subset of features using an adjustment component to generate an adjustment factor. Brandes further fails to teach or suggest scaling the initial efficacy score based on such an adjustment factor to generate a final efficacy score that more accurately adjusts for the underrepresented outcome class as compared to an initial efficacy score generated using the hybrid machine learning model without the adjustment component.
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 1-21 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1
Under step 1, the analysis is based on MPEP 2106.03, and claims 1-19 and 21 are drawn to a method and claim 20 is drawn to a non-transitory computer-readable storage medium. Thus, each claim, on its face, is directed to one of the statutory categories (i.e., useful process, machine, manufacture, or composition of matter) of 35 U.S.C. §101.
Step 2A Prong One
Claim 1 recites the limitations of identifying a medication to be evaluated with respect to the patient; generating an initial efficacy score by processing the patient data describing the patient, wherein the initial efficacy score indicates predicted efficacy of the medication for the patient; selecting a defined subset of features from the patient data describing the patient that was processed to generate the initial efficacy score, wherein the patient data includes at least one feature not included in the defined subset of features and the defined subset of features corresponds to an underrepresented outcome class; determining whether the patient has experienced an adverse event when using a different medication in the same class as the medication; generating an adjustment factor by processing the defined subset of features from the patient data; and scaling the initial efficacy score based on the adjustment factor to generate a final efficacy score for the patient, wherein final efficacy scores generated using the adjustment component more accurately adjust for the underrepresented outcome class as compared to initial efficacy scores generated without the adjustment component. These limitations, as drafted, recite observations, evaluations, and judgments describing patient information, medication history, treatment outcomes, and predicted medication efficacy. Specifically, the claim encompasses identifying a medication for evaluation, reviewing, and evaluating patient information to generate a predicted efficacy for the medication, selecting patient features associated with an underrepresented outcome class, determining from the patient's medication history whether the patient experienced an adverse event when using another medication in the same medication class, generating an adjustment based on the selected patient information, and applying that adjustment to the initial efficacy score to determine a final predicted efficacy. These observations, evaluations, judgments, and calculations can practically be performed in the human mind or by a person using pen and paper. For example, a person could review patient information and medication history, identify relevant patient features and prior adverse events, assign an initial efficacy value and an adjustment value based on the information, and scale the initial efficacy value using the adjustment value to arrive at a final efficacy value. Although the claim recites performing steps using a hybrid machine learning model and an adjustment component, these computer components implement the evaluations, judgments, and calculations and are considered as additional elements under Step 2A, Prong Two. Accordingly, claim 1 recites a mental process and therefore recites an abstract idea.
Claim 10 recites the limitations of identifying a medication consumed by the patient; determining an efficacy of the medication based at least in part on whether the medication was successful in treating a disorder of the patient; selecting a defined subset of features from the patient data, wherein the patient data includes at least one feature not included in the defined subset of features and the defined subset of features corresponds to an underrepresented outcome class; determining whether the patient has experienced an adverse event when using a different medication in the same class of the medication; generating an adjustment factor by processing the defined subset of features from the patient data; and scaling the initial efficacy score based on the adjustment factor to generate a final efficacy score for the patient, wherein final efficacy scores generated using the adjustment component more accurately adjust for the underrepresented outcome class as compared to initial efficacy scores generated without the adjustment component. These limitations recite observations, evaluations, judgments, and calculations that describe patient information, medication history, treatment outcomes, and the importance of selected patient features. These evaluations, as drafted and under their broadest reasonable interpretation, describe steps that can practically be performed in the human mind or by a person using pen and paper. For example, these steps include reviewing patient information, identifying a medication previously taken by the patient, determining whether the medication successfully treated a disorder, identifying particular features of the patient information for consideration, and determining from the patient's medication history whether an adverse event occurred with another medication in the same medication class, generating an adjustment value based on the selected patient information, and applying that adjustment value to the initial efficacy value to determine a final efficacy value. Although the claim recites performing certain operations using a hybrid machine learning model and an adjustment component, these computer components implement the underlying evaluations, judgments, and calculations and are considered as additional elements under Step 2A, Prong Two. Accordingly, these limitations fall within the mental process grouping of abstract ideas.
Independent claims 20 and 21 recite identical or nearly identical steps with respect to claim 1 (and therefore also recite limitations that fall within this subject matter grouping of abstract ideas), and these claims are therefore determined to recite an abstract idea under the same analysis.
Under Step 2A Prong Two
The claimed limitations, as per claim 1, include:
receiving patient data describing a patient;
identifying a medication to be evaluated with respect to the patient;
generating an initial efficacy score by processing the patient data describing the patient using a hybrid machine learning model comprising a static portion and a dynamic portion, wherein the initial efficacy score indicates predicted efficacy of the medication for the patient;
selecting a defined subset of features from the patient data describing the patient that was processed using the hybrid machine learning model to generate the initial efficacy score, wherein the patient data includes at least one feature not included in the defined subset of features and the defined subset of features corresponds to an underrepresented outcome class in training data used to train the hybrid machine learning model, comprising
determining whether the patient has experienced an adverse event when using a different medication in a same class of the medication;
generating an adjustment factor by processing the defined subset of features, from the patient data, using an adjustment component;
scaling the initial efficacy score based on the adjustment factor to generate a final efficacy score for the patient, wherein final efficacy scores generating using the adjustment component more accurately adjust for the underrepresented outcome class, as compared to initial efficacy scores generated using the hybrid machine learning model without the adjustment component; and
providing the medication for the patient based at least in part on the final efficacy score.
The claimed limitations, as per claim 10, include:
receiving patient data describing a patient;
identifying a medication consumed by the patient;
determining an efficacy of the medication based at least in part on whether the medication was successful in treating a disorder of the patient;
training a hybrid machine learning model comprising a static portion and a dynamic portion based on at least a subset of the patient data and the determined efficacy of the medication, comprising:
generating an initial efficacy score by processing the patient data describing the patient using the hybrid machine learning model;
selecting a defined subset of features from the patient data describing the patient that was processed using the hybrid machine learning model to generate the initial efficacy score, wherein the patient data includes at least one feature not included in the defined subset of features and the defined subset of features correspond to an underrepresented outcome class in training data used to train the hybrid machine learning model, comprising
determining whether the patient has experienced an adverse event when usinq a different medication in a same class of the medication;
generating an adjustment factor by processing the defined subset of features, from the patient data, using an adjustment component;
scaling the initial efficacy score based on the adjustment factor to generate a final efficacy score for the patient, wherein final efficacy scores generating using the adjustment component more accurately adjust for the underrepresented outcome class, as compared to initial efficacy scores generated using the hybrid machine learning model without the adjustment component; and
deploying the hybrid machine learning model.
Examiner Note: underlined elements indicate additional elements of the claimed invention identified as performing the steps of the claimed invention.
The judicial exceptions expressed in claims 1 and 10 are not integrated into a practical application. The claims as a whole merely describe how to generally “apply” the concepts of evaluating patient and medication information, predicting medication efficacy, generating and adjusting efficacy scores based on selected patient features and prior adverse events, and determining a final efficacy score in a computer environment. The claimed computer components (i.e., using a hybrid machine learning model comprising a static portion and a dynamic portion (claim 1), using the hybrid machine learning model (claims 1 and 10), training data used to train the hybrid machine learning model (claims 1 and 10), using an adjustment component (claims 1 and 10), using the hybrid machine learning model without the adjustment component (claims 1 and 10), and training a hybrid machine learning model comprising a static portion and a dynamic portion based on at least a subset of the patient data and the determined efficacy of the medication (claim 10)) are recited at a high level of generality and are merely invoked as tools to perform the process of evaluating patient and medication information, generating an initial predicted efficacy, adjusting that predicted efficacy based on selected patient information associated with an underrepresented outcome class, and generating a final predicted efficacy (see MPEP 2106.05(f)). The claims do not recite a specific technological implementation of the hybrid machine learning model or adjustment component that meaningfully limits the judicial exception, and instead use these components to implement the recited evaluations, judgments, and calculations. Merely implementing the abstract idea using computer-based tools does not integrate the abstract idea into a practical application. Accordingly, when considered individually and in combination, these additional elements amount to mere instructions to apply the judicial exception using computer components and do not integrate the judicial exception into a practical application.
The judicial exception expressed in claims 1 and 10 is not integrated into a practical application. The claims recite the additional elements of receiving patient data describing a patient (claims 1 and 10); providing the medication for the patient based at least in part on the final efficacy score (claim 1); and deploying the hybrid machine learning model (claim 10). These limitations are recited at a high level of generality (i.e., as a general means of obtaining information for use in the analysis, applying the result of the analysis by providing a medication, and deploying the resulting machine learning model, respectively) and amount to mere data gathering and insignificant application of the results, which are forms of insignificant extra-solution activity (see MPEP 2106.05(g)). Specifically, receiving the patient data merely obtains the information for the subsequent evaluations and calculations that are performed, and providing the medication and deploying the hybrid machine learning model merely occur after the analysis without further specifying a specific technological manner of providing the medication or deploying the model that meaningfully limits the judicial exception. Accordingly, even when considered in combination, these additional elements constitute insignificant extra solution activity and do not integrate the abstract idea into a practical application.
Therefore, under step 2A, the claims are directed to the abstract idea, and require further analysis under Step 2B.
Under step 2B
Claims 1 and 10 do not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed with respect to Step 2A, the claims as a whole merely describe how to generally “apply” the concept of evaluating patient information to predict and adjust the efficacy of a medication for a patient (claim 1) and evaluating patient and medication information to determine medication efficacy and train a model for predicting and adjusting medication efficacy (claim 10) in a computer environment. The recited hybrid machine learning model, including the static and dynamic portions and adjustment component are used as computer-based tools for performing the recited evaluations, judgments, and calculations and do not provide an inventive concept beyond the abstract idea. Therefore, even when the additional elements are considered individually and as an ordered combination, nothing in the claims adds significantly more (i.e., an inventive concept) to the abstract idea.
For claims 1 and 10, under Step 2B, the additional elements of receiving patient data describing a patient (claims 1 and 10); providing the medication for the patient based at least in part on the final efficacy score (claim 1); and deploying the hybrid machine learning model (claim 10) have been evaluated. As noted in Electric Power Group, LLC v. Alstom S.A., 830 F.3d 1350, 1354, 119 USPQ2d 1739, 1742 (Fed. Cir. 2016), merely collecting information for analysis without a technological improvement does not add significantly more to an abstract idea. The recitation of receiving patient data amounts to no more than collecting information before performing the recited analysis and calculations, and therefore constitutes mere data gathering that does not provide an inventive concept. Additionally, as noted in In re Brown, 645 Fed. App'x 1014, 1016–17 (Fed. Cir. 2016), an activity that merely applies the result of an abstract process constitutes insignificant application of the judicial exception. In these claims, providing the medication for the patient based at least in part on the final efficacy score (claim 1) merely applies the result of the efficacy evaluation without specifying a particular medication, dosage, or specific treatment that meaningfully limits the abstract idea. Also, deploying the hybrid machine learning model (claim 10) merely applies the result of the training process without specifying any technological method of deployment that meaningfully limits the abstract idea. These limitations are recited at a high level of generality and merely supplement the analysis without providing an inventive concept. Accordingly, individually and in combination with the other additional elements, the recited data gathering and insignificant application do not amount to significantly more than the judicial exception. Therefore, claims 1 and 10 do not recite an inventive concept under Step 2B are not patient eligible.
Claims 6-9, 13, and 17-19 recite no further additional elements, and only further narrow the abstract idea. The previously identified additional elements, individually and as a combination, do not integrate the narrowed abstract idea into a practical application for reasons similar to those explained above, and do not amount to significantly more than the narrowed abstract idea for reasons similar to those explained above.
Claims 2-5, 11-12, 14-16 recite the additional elements of the static portion of the hybrid machine learning model (claims 2-3, 11-12), the dynamic portion of the hybrid machine learning model (claims 2-3, 11-12), and the hybrid machine learning model (claims 4-5, 12, 14-16). However, these additional element amount to implementing an abstract idea on generic computing components. As such, these additional elements, when considered individually or in combination with the previously identified additional elements, do not integrate the abstract idea into a practical application or amount to significantly more than the abstract idea.
Thus, as the dependent claims remain directed to a judicial exception, and as the additional elements of the claims do not amount to significantly more, the dependent claims are not patent eligible.
Therefore, the claims here fail to contain any additional element(s) or combination of additional elements that can be considered as significantly more and the claims are rejected under 35 U.S.C. 101 for lacking eligible subject matter.
Claim Rejections - 35 USC § 103
In view of Applicant’s amendments and arguments, the rejection of claims 1-21 under 35 U.S.C. 103 is withdrawn.
Response to Arguments
Applicant’s arguments and amendments, see Remarks/Amendments submitted 02/11/2026 with respect to the rejection of claims 1-21 have been carefully considered and are addressed below.
Claim Rejections - 35 USC § 101
Applicant's arguments have been considered but are not persuasive. Applicant states that the claims do not recite an abstract idea and that the recited hybrid machine learning model and adjustment component integrate any alleged judicial exception into a practical application by improving the accuracy and operation of the machine learning model. However, claims 1 and 10 recite the steps of evaluating patient and medication information, generating an initial efficacy score, selecting patient features associated with an underrepresented outcome class, determining whether the patient experienced an adverse event with another medication in the same medication class, generating an adjustment factor, and scaling the initial efficacy score using the adjustment factor to generate a final efficacy score. These limitations recite evaluations, judgments, and calculations that fall within the mental process grouping of abstract ideas. The recitation of a hybrid machine learning model and adjustment component does not alter the abstract analysis merely because those computer components are used to perform the recited evaluations and calculations.
The Examiner has also considered Applicant's statement regarding technological improvement, including that the adjustment component may improve efficacy score accuracy where training data is unbalanced or where attributes may be ignored by a trained model. Nevertheless, the claims do not recite a particular improvement to the operation of the machine learning model. Instead, the claims use the hybrid machine learning model to generate an initial efficacy score and separately use an adjustment component to generate an adjustment factor from selected patient features, and this is then used to scale the initial efficacy score. The recited improvement, that the resulting final efficacy scores accurately adjust for an underrepresented outcome class, describes an improvement in the accuracy of the result of the medication efficacy analysis (i.e., the abstract idea). The claims do not further recite a particular modification to the training algorithm or learning process of the hybrid machine learning model that causes the model to operate in a different way. Therefore, the stated improvement reflects improved performance of the abstract medication efficacy evaluation using the recited computer components as tools, instead of an improvement to computer functionality or machine learning technology itself.
The Examiner has further considered Applicant's statement regarding Desjardins, but the present claims are different from those recited in Desjardins. In Desjardins, the claimed limitations addressed catastrophic forgetting by modifying how the machine learning model was trained to learn new tasks while preserving previously learned knowledge. Although Applicant states that the claimed approach improves accuracy for underrepresented outcome classes in the present claims, the claims do not recite a modification to how the hybrid machine learning model learns or retains information. Accordingly, Desjardins does not establish that the claimed machine learning model and adjustment component integrate the judicial exception into a practical application.
Lastly, the Examiner evaluated the claims as a whole and has not treated the recited machine learning components as ineligible merely because they involve machine learning. Instead, the rejection considers the specific functions required by the claims, including the hybrid static and dynamic model, selection of features corresponding to an underrepresented outcome class, generation of the adjustment factor, and scaling of the initial efficacy score. When considered individually and as an ordered combination, these limitations use computer based components to implement the evaluation and adjustment of predicted medication efficacy, and the remaining limitations constitute data gathering or insignificant application of the analysis, as discussed in Step 2A, Prong Two and Step 2B. Accordingly, Applicant's arguments do not establish that the claims reflect a technological improvement or do not integrate the judicial exception into a practical application. The rejection of claims 1 and 10 under 35 U.S.C. § 101 is maintained.
Claim Rejections - 35 USC § 103
In view of Applicant’s amendments and arguments, the rejection of claims 1-21 under 35 U.S.C. 103 is withdrawn.
Conclusion
The prior art made of record and not relied upon is considered pertinent to Applicant's disclosure.
Bostic et al. (U.S. Patent Publication 2021/0202103 A1) teaches a method to simulate a patient’s current and future health states by analyzing relationships between patient data and historical data, enriching the data, and applying machine learning algorithms.
Valuck et al. (U.S. Patent Publication 2021/0043292 A1) teaches a computer system that generates therapeutic treatment recommendations for a patient by analyzing patient data using machine learning algorithms.
Langheier et al. (International Publication No. WO 2006072011 A2) teaches methods, systems, and computer programs for developing predictive models using clinical data to forecast medical outcomes and evaluate intervention strategies.
Purushothaman (U.S. publication 2021/0343384 A1) teaches an AI based system for managing autoimmune conditions using machine learning to actively monitor patients, predict symptom flares, and deliver treatment recommendations via patient, provider, and payer interface.
Griffin et al. (U.S. publication 2023/0178237 A1) teaches a machine learning based method that determines the efficacy of drug combinations for treating a specific morbidity by training on electronic medical records and generating personalized treatment proposals based on patient specific health information.
Dil Nahlieli (U.S. publication 2021/0249137 A1) teaches a method that maps patient specific medications and parameters to a data structure linking active ingredients, predicted medication induced outcomes, and risk factors, then computes aggregated risk scores for each outcome to assess patient specific mediation risks.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to KYRA R LAGOY whose telephone number is (703)756-1773. The examiner can normally be reached Monday - Friday, 8:00 am - 5:00 pm EST.
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, Kambiz Abdi can be reached at (571)272-6702. 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.
/K.R.L./Examiner, Art Unit 3685
/KAMBIZ ABDI/Supervisory Patent Examiner, Art Unit 3685
Prosecution Insights