DETAILED ACTION
This Action is Responsive to Claims filed 02/18/2026.
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and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114.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 02/18/2026 has been entered.
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
Claims 1, 8-10, and 18-19 have been amended. Claims 1-20 are currently pending.
Response to Amendment
The amendments to Claims 8, 9, and 18 have overcome the Objections to minor informalities.
Response to Arguments
Applicant's arguments, see Pages 8-14, filed 02/18/2026, regarding the 35 U.S.C. 101 Rejection of Claims 1-20 have been fully considered but they are not persuasive.
The interpretation of the claim limitations is as follows: The “determining…” step is a generic axiom of generic machine learning functionality. A trained model makes a predictive output based on an input. This limitation, at broadest, merely points to the machine learning model in question performing generic, normal function, and amounts to post-solution activity.
The training data on which the model is trained (“generated by…”) is recited highly generically, and there is no structure or implementation precluding a human mind with or without the aid of pen and paper from performing the claimed steps of “identifying…”, “identifying…”, “determining…”, and “generating…”. These abstract idea mental process steps used to manipulate or refine the training data are then applied in the “trained based...on the training data,” step. This limitation amounts to instructions to apply the preceding set of algorithmic data manipulation steps. The next “determining…” takes the generic predictive output of the model and a qualifying criteria and determines a priority score. There is no structure or implementation precluding a human mind with or without the aid of pen and paper from formulating this priority score based on the generic predictive output and criteria. The allocating…” step amounts to instructions to apply the priority score ascertained in the previous step. The Examiner contends, under the broadest reasonable interpretation of the claims, that the alleged improvement to the functioning of a computer or other technological field is a direct result of training the model on the manipulation data, and using the output of said model in the formulation of a priority score. Without the algorithmic set of data manipulation steps, or the formulation of the priority score, the improvement is not realized at the ”allocating…” step. Per MPEP 2016.05(a), the specific improvement cannot come from the abstract idea mental process(es).
The Applicant misapplies the recent Desjardins decision in appearing to assume a claim merely reciting the adjustment of parameters is eligible subject matter. The claim of Desjardins has a clear connection between a non-abstract additional element within the claim, a recitation in the Specification to that non-abstract additional element, and the specific improvement realized by said additional element. The improvements cited by the Applicant on Pages [0025] and [0028] of the instant specification are recited highly generally, and are rooted in abstract idea mental process steps taken within the claim(s). This is not analogous to the Desjardins decision, or Example 47, which recites structure or implementation precluding a human mind from performing the steps or realizing the improvements to network security. Seethe updated 101 Rejection below.
Applicant’s arguments, see Pages 14-15, filed 02/18/2026, regarding the 35 U.S.C. 103 Rejection(s) of Claims 1-20 have been considered but are moot because the new ground of rejection does not rely on any reference applied in the prior rejection of record for any teaching or matter specifically challenged in the argument.
Claim Rejections - 35 USC § 101
The text of those sections of Title 35, U.S. Code not included in this action can be found in a prior Office action.
Claims 1-20 rejected under 35 U.S.C. 101 because the claimed invention is directed to a judicial exception (i.e., a law of nature, a natural phenomenon, or an abstract idea) without significantly more; and because the claims as a whole, considering all claim elements both individually and in combination, do not amount to significantly more than the abstract idea, see Alice Corporation Pty. Ltd. v. CLS Bank International, et al, 573 U.S. (2014). In determining whether the claims are subject matter eligible, the Examiner applies the 2019 USPTO Patent Eligibility Guidelines. (2019 Revised Patent Subject Matter Eligibility Guidance, 84 Fed. Reg. 50, Jan. 7, 2019.)
Step 1:
Claims 1-9 recite a computer-implemented method, which falls under the statutory category of a process. Claims 10-18 recite a system, which falls under the statutory category of a machine. Claims 19 and 20 recite one or more non-transitory computer-readable media storing processor-executable instructions, which falls under the statutory category of a manufacture.
Step 2A – Prong 1:
Claim 1 recites an abstract idea, law of nature, or natural phenomenon. The limitations of “identifying a retrospective event within a defined retrospective period of historical data,”, “identifying, a plurality of prospective event associated with the retrospective event within a defined prospective period of the historical data,”, “determining a binary ground-truth for the retrospective event based at least in part on the prospective event and a threshold measure, wherein the threshold measure is based on a percentile of a value distribution data object associated with a plurality of prospective events within the defined prospective period”, “and generating the training data for the machine learning model based at least in part on the retrospective event and the binary ground truth…”, and “determining, by the one or more processors, a prospective priority score for the predictive input entity based at least in part on a prospective qualifying criteria satisfaction predictive output and the predictive output;” under the broadest reasonable interpretation, cover a mental process including an observation, evaluation, judgment or opinion that could be performed in the human mind or with the aid of pencil and paper.
Identifying a retrospective and prospective event is practically performed within the human mind or with the aid of pen and paper. Determining a binary ground-truth is practically performed within the human mind or with the aid of pen and paper. Generating generic training data based on the identified event and determined ground-truth is practically performed within the human mind or with the aid of pen and paper. Determining a priority score based on outputs is practically performed within the human mind or with the aid of pen and paper.
Step 2A – Prong 2:
The additional elements of claim 1 do not integrate the abstract idea into a judicial exception. The claim recites the additional elements “A computer-implemented method”, “input entity”, “one or more processors”, “a computing resource”, “a programmatic agent”, and “output” are recognized as generic computer components recited at a high level of generality. Although they have and execute instructions to perform the abstract idea itself, this also does not serve to integrate the abstract idea into a practical application as it merely amounts to instructions to "apply it." (See MPEP 2106.04(d)(2) indicating mere instructions to apply an abstract idea does not amount to integrating the abstract idea into a practical application).
The additional elements of “a machine learning model”, “predictive output”, “training data”, “retrospective event”, “historical data”, “prospective event”, “binary ground-truth”, and “priority score” are recognized as non-generic computer components, but are recited at a high level of generality and are found to generally link the abstract idea to a particular technological environment or field of use (See MPEP 2106.05(h)).
The additional elements recited in the limitation “…trained using training data generated by”, “…wherein the machine leaning model is trained based at least in part on the training data,” and “and allocating, by the one or more processors, a computing resource to a programmatic agent based on the prospective priority score.” are found to be mere instructions to apply the abstract idea of determining feature criteria and curating training data accordingly to arrange data by priority (see MPEP 2106.05(f) indicating mere instructions to apply an abstract idea does not amount to integrating the abstract idea into a practical application).
The additional elements recited in the limitation “determining, by one or more processors and using a machine learning model, a predictive output for a predictive input entity…” is found to be pre- or post-extra-solution activity steps (See MPEP 2106.05(g)(iii) first list).
Step 2B:
The only limitation on the performance of the described method is a limitation reciting “A computer-implemented method”, “input entity”, “one or more processors”, “a computing resource”, “a programmatic agent”, and “output” These elements are insufficient to transform a judicial exception to a patentable invention because the recited elements are considered insignificant extra-solution activity (generic computer system, processing resources, links the judicial exception to a particular, respective, technological environment). The claim thus recites computing components only at a high-level of generality such that it amounts to no more than mere instructions to apply the exception using generic computer components; mere instructions to apply an exception using a generic computer component cannot provide an inventive concept (see MPEP 2106.05(f)).
The additional elements of “a machine learning model”, “predictive output”, “training data”, “retrospective event”, “historical data”, “prospective event”, “binary ground-truth”, and “priority score” are recognized as non-generic computer components, but are recited at a high level of generality and are found to generally link the abstract idea to a particular technological environment or field of use (See MPEP 2106.05(h)).
The additional elements recited in the limitation “…trained using training data generated by”, “…wherein the machine leaning model is trained based at least in part on the training data,” and “and allocating, by the one or more processors, a computing resource to a programmatic agent based on the prospective priority score.” are found to be mere instructions to apply the abstract idea (See MPEP 2106.05(f) indicating mere instructions to apply an abstract idea does not recite significantly more).
The additional elements recited in the limitation “determining, by one or more processors and using a machine learning model, a predictive output for a predictive input entity…” is found to be well-understood, routine, or conventional activity (See MPEP 2106.05(d)(II)(iv) third list).
Taken alone or in ordered combination, these additional elements do not amount to significantly more than the above-identified abstract idea. There is no indication that the combination of elements improves the functioning of a computer or improves any other technology. Their collective functions merely provide conventional computer implementation.
For the reasons above, claim 1 is rejected as being directed to non-patentable subject matter under §101. This rejection applies equally to independent claims 10 and 19.
Claim 10 recites similar limitations to claim 1, with the exception of “A system comprising: one or more processors; and one or more memories storing processor-executable instructions that, when executed by the one or more processors, cause the one or more processors to perform operations comprising:” (generic computer components).
Claim 19 recites similar limitations to claim 1, with the exception of “One or more non-transitory computer-readable media storing processor-executable instructions that, when executed by one or more processors, cause the one or more processors to perform operations comprising:” (generic computer components).
Dependent Claims:
Claim 2 (claims 11 and 20) recites additional elements “the prospective qualifying criteria satisfaction predictive output describes a likelihood of coordination of benefits with respect to the predictive input entity; and the predictive output describes a likelihood of claim filing with respect to the predictive input entity.” found to generally link the abstract idea to a particular technological environment or field of use (See MPEP 2106.05(h)).
Claim 3 (claim 12) recites additional elements “wherein the prospective priority score describes an investigation priority of the predictive input entity.” found to generally link the abstract idea to a particular technological environment or field of use (See MPEP 2106.05(h)).
Claim 4 (claim 13) recites abstract idea mental process steps “determining a high-utility subset of a plurality of retrospective events within the historical data based at least in part on the threshold measure, wherein the training data is generated based at least in part on the high-utility subset of the plurality of retrospective events.”
Claim 5 (claim 14) recites abstract idea mental process steps “wherein the threshold measure is determined based at least in part on a statistical distribution of event valuations associated with the plurality of retrospective events.”
Claim 6 (claim 15) recites abstract idea mental process steps “determining a predictive input channel of a plurality of predictive input channels that is associated with the predictive input entity;” and “and determining the prospective qualifying criteria satisfaction predictive output based at least in part on the predictive input channel.”
Claim 7 (claim 16) recites additional elements “wherein the plurality of predictive input channels comprises one or more of a model-based prospective channel, a rule-based prospective channel, a model-based real-time channel, or a rule-based real time channel.” found to generally link the abstract idea to a particular technological environment or field of use (See MPEP 2106.05(h)).
Claim 8 (claim 17) recites abstract idea mental process steps “assigning the predictive input entity to the programmatic agent based at least in part on the prospective priority score,” The additional element of “causing the programmatic agent to process the predictive input entity.” has been evaluated under Steps 2A – Rong 2 and 2B and found to be instructions to apply said abstract idea mental process steps (See MPEP 2106.05(f)).
Claim 9 (claim 18) recites abstract idea mental process steps “generating an investigation agent user interface for the programmatic agent that comprises one or more graphical elements respectively reflective of one or more investigation queue features of a related subset associated with the programmatic agent.”
Claim Rejections - 35 USC § 103
The text of those sections of Title 35, U.S. Code not included in this action can be found in a prior Office action.
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.
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.
Claim(s) 1-20 is/are rejected under 35 U.S.C. 103 as being unpatentable over Limoli et al. (US 11,031,111 B1), hereinafter Limoli; Chimmad et al. (Assessment of Healthcare Claims Rejection Risk Using Machine Learning, 2017), hereinafter Chimmad; and Maisog et al. (Using massive health insurance claims data to predict very high-cost claimants: a machine learning approach, 2019), hereinafter Maisog.
In regards to claim 1: The present invention recites: “A computer-implemented method comprising: determining, by one or more processors and using a machine learning model, a predictive output for a predictive input entity wherein the machine learning model is trained using training data generated by:” Limoli teaches “For purpose of illustration and not limitation, the prioritization module in performing the experience-driven statistical inference can be further configured to prioritize the retrieved information relating to a clinical context of a patient visit to a clinical context by fitting a non-parametric machine learning model to estimate the priority of observed data points in the retrieved information and unobserved data points in the retrieved information from clinical context information based on patterns from historical physician behaviors and contextually important patient outcomes.” (Column 2, Lines 28-37, mapping to the use of machine learning, since a predictive input, predictive output, and training with training data is a high-level description of all machine learning).
“identifying a retrospective event within a defined retrospective period of historical data,” Limoli teaches "In the third stage of the experience-driven statistical inference process 9, the prioritization engine 4 can direct a selected model 16 (e.g., gradient boosting regressor, neural network, etc.) to input the resulting processed features to estimate the priority 17 of various real clinical data points. The estimate produced by the model 16 can be compared against a composite priority index 25, comprising historical physician behaviors and patient outcomes 12" (Column 8, Lines 14-21). Limoli also teaches "a variety of machine learning techniques can be used for the selected model 16. For purposes of illustration and not limitation, the selected model 16 can be a neural network. According to an exemplary embodiment, the model 16 can first split the preprocessed features into an 80% sample used to train the model 16 and a 20% sample used to evaluate the model 16. An initial model (e.g., having 3 fully connected layers, 100 nodes each, leaky ReLU activation function) can be constructed with initial weights drawn from a random normal distribution. In the prediction stage, the 80% sample can be fed through the neural network, which produces an estimated priority 17 through a series of matrix multiplication operations using the model weights and nonlinear transformations through a selected activation function (e.g., ReLU) for each layer of the network. The estimated priority 17 can be compared to the actual priority for that visit to calculate a model error (e.g., using the mean absolute difference). The actual priority can be constructed for each visit by combining the historical conditional frequency for that observation with historical user event data. If the model error is not close to zero, the model weights can be adjusted in the direction that would decrease the priority error using stochastic gradient descent and backpropagation” (Column 8, Lines 42-48) (mapping the historical data/range to “retrospective events/period”).
“identifying, a plurality of prospective event associated with the retrospective event within a defined prospective period of the historical data,” Limoli teaches “the prioritization engine 4 can create a data table where each row represents one historical visit and each column represents a historical attribute related to that visit (e.g., provider specialty, diagnoses made prior to time of visit, etc.). The prioritization engine 4 can merge in data related to actions taken during that visit as new columns. Such merged data can serve as target variables (i.e., variables that will be predicted in a live setting)" (Column 7, Lines 53-61).
“determining, by the one or more processors, a prospective priority score for the predictive input entity based at least in part on a prospective qualifying criteria satisfaction predictive output and the predictive output;” Limoli teaches “Next, the prioritization engine 4 can sum each column in the guideline actions table across all relevant rows to determine which actions are most relevant to the context of the patient visit. Next, the prioritization engine 4 can increase the priority score in proportion to the sum. (Column 9, Lines 65-67 and Column 10, Lines 1-3, guidelines impact the priority score, as per the mapping above) and “the prioritization engine 4 can create a data table where each row represents one historical visit and each column represents a historical attribute related to that visit (e.g., provider specialty, diagnoses made prior to time of visit, etc.). The prioritization engine 4 can merge in data related to actions taken during that visit as new columns. Such merged data can serve as target variables (i.e., variables that will be predicted in a live setting)" (Column 7, Lines 54-61). As well as "In the third stage of the experience driven statistical inference process 9, the prioritization engine 4 can direct a selected model 16 (e.g., gradient boosting regressor, neural network, etc.) to input the resulting processed features to estimate the priority 17 of various real clinical data points. The estimate produced by the model 16 can be compared against a composite priority index 25, comprising historical physician behaviors and patient outcomes 12" (Column 8, Lines 13-21).
Limoli fails to explicitly teach:
“determining a binary ground-truth for the retrospective event based at least in part on the prospective event and a threshold measure…” While Limoli teaches “the prioritization engine 4 can direct a selected model 16 (e.g., gradient boosting regressor, neural network, etc.) to input the resulting processed features to estimate the priority 17 of various real clinical data points. The estimate produced by the model 16 can be compared against a composite priority index 25, comprising historical physician behaviors and patient outcomes 12.” (Column 8, Lines 14-21), which may reasonably read on ground-truth(s), Limoli fails to explicitly teach a binary ground-truth and a “threshold measure.” However, Chimmad, in a similar field of health- or health insurance-related data prediction, teaches “Three different Binary Classification algorithms (Classification Trees, SVM and Neural Networks) were used to classify the claims into High risk or Low risk for Denial. Training data sets were prepared by adding a status field of 100 (accepted) for all the claims with reason number (RSN) 100 and 200 (denied) for the remaining claims.” (Page 3, left column, 100 and 200 as options maps to a “binary” ground-truth label). Assigning of this accepted/denied label is determined by one or more thresholds regarding a specific data instance (See Figures 2 and 5, for example).
Chimmad, in a similar field of health- or health insurance-related data prediction, can save providers significant sums of money (Introduction). It would have been obvious to one of ordinary skill in the art at the time of the Applicant’s filing to take elements of the health data prioritization model of Limoli with the binary classification structure of Chimmad when considering insurance claims or similar data.
“and generating the training data for the machine learning model based at least in part on the retrospective event and the binary ground truth, wherein the machine leaning model is trained based at least in part on the training data,” Limoli teaches “Limoli teaches “In some embodiments, the priority index 25 can be continuous over the range [0, I], with higher values indicating higher priority, and can be constructed by non-linearly combining historical physician behaviors and patient outcomes 12 based on medical expertise.” (Column 8, Lines 32-37).” Additionally, Chimmad teaches “CR #1A: After preparing the dataset of size (# 13000 (10000 accepted and 3000 denied claim instances) with 3 major attributes – days of service, total charge amount and total paid amount, we randomly sampled 75% of this dataset and trained this subset using a Classification Tree (CART) algorithm using the R ML package. This algorithm classifies the dataset of mixed mode variables by recursively partitioning the claims data space and fitting a simple prediction model within each partition, which are graphically represented as a decision tree [12].” (Page 3, left column, Sections CR #1A, #1B, 1#C and subsequent sections for training data being created and used to train different models.
“and allocating, by the one or more processors, a computing resource to a programmatic agent based on the prospective priority score.” Limoli teaches “According to some embodiments, based on a set of inputs (e.g., physician specialty, past behavior of user, past behavior of similar users, clinical guidelines, patient profile, previous outcomes, monitored dialogs between clinician and patient, voice input from a clinician, etc.), the disclosed systems and methods can be configured to assign values across multiple dimensions, including urgency and relevancy, to the patient framework which can then be applied to the patients' clinical data. According to some embodiments, these assigned values can be used to create a rigorous methodology for surfacing the most important and/or relevant pieces of information to physicians, thereby enabling increased and/or improved expeditious data review and better treatment decision making than currently available 60 EHR systems" (Column 4, Lines 46-60).
Chimmad also teaches “We report on a comprehensive classification engine using multiple ML algorithms (Binary Classification Trees, Neural Networks (NN), Support Vector machines (SVM) and Naive Bayes Classification, to predict whether a given claim is likely to involve rework (i.e., has a high risk of rejection or denial) via binary classification" (Section I, Paragraph 4); "ML experiments were conducted using three different approaches to feature engineering as follows: (1) CARC Codes (RSNl) included in the dataset as a single 'Status' feature without one-hot coding; (2) same as above but also converting the CARC codes as eight distinct features using one-hot (or Dummy) coding and (3) engineering 4 additional (synthetic) features added to the dataset" (Section IlI, Paragraph 1); and "Three different Binary Classification algorithms (Classification Trees, SVM and Neural Networks) were used to classify the claims into High risk or Low risk for Denial. Training data sets were prepared by adding a status field of 100 (accepted) for all the claims with reason number (RSN) 100 and 200 (denied) for the remaining claims” (Section Ill, Paragraph 2).
While the combination of Limoli and Chimmad teaches the above, the combination fails to explicitly teach the particular data type recited in “…wherein the threshold measure is based on a percentile of a value distribution data object associated with a plurality of prospective events within the defined prospective period” However, Maisog, in a similar field of endeavor of claimant prediction analysis, teaches throughout their disclosure that prediction using threshold percentages of claimant value/spending (tope 1%, 5%, exceeded certain dollar amounts, etc.) would have been well-known in the art prior to the Applicant’s filing date (Table 1, Pages 7-9, at least).
Maisog highlights the importance of analyzing the claim data of high-cost claimants and discusses the accuracy of their own, as well as state of the art methods predicting off of this metric, as well as others (Abstract/Introduction, Table 1). It would have been obvious to one of ordinary skill in the art at the time of the Applicant’s filing to apply known methods and utilize known prediction metrics such claimant value/spending percentiles and/or thresholds in a combination of Limoli and Chimmad in order to leverage the known benefits of using such metrics.
In regards to claim 2: The present invention claims “wherein the prospective qualifying criteria satisfaction predictive output describes a likelihood of coordination of benefits with respect to the predictive input entity;” However, Chimmad teaches "The analyst has prepared the training dataset by identifying thousands of medical claims, matching each Medical Claim document from a Provider to a Payer with the corresponding Remittance Advice document generated by the Payer to the Provider in response. This later document contains the Reason Code descriptions shown in Table 1" (Section II-A, Paragraph 2). Chimmad Table 1 recites "This care may be covered by another payer per coordination of benefits"
“and the predictive output describes a likelihood of claim filing with respect to the predictive input entity.” Chimmad teaches "We built an ensemble of ML algorithms to classify claims against denial risk, using a dataset of 10000 accepted claims and 3000 denied claims from a prominent health insurance companies in the U. S. Days of service is the number of days elapsed starting from the date of service and payment data" (Section II-A, Paragraph 1).
In regards to claim 3: The present invention claims “wherein the prospective priority score describes an investigation priority of the predictive input entity.” Limoli teaches “For each metric, the prioritization engine 4 can extract and codify conditions for when the metric is applicable can be extracted by evaluating that metric's quantitative value 29 in the context of the visiting patient (e.g., based on the required age and diagnoses). The prioritization engine 4 can identify the key metric from the text and map to the appropriate code (e.g., treatment, diagnostics, and medications). In some embodiments, during runtime, the prioritization engine 4 can check the patient visit details against the conditions for each metric. For each matching metric, the prioritization engine 4 can increase the priority scores for the relevant codes" (Column 10, Lines 27-38).
In regards to claim 4: The present invention claims “determining a high-utility subset of a plurality of retrospective events within the historical data based at least in part on the threshold measure, wherein the training data is generated based at least in part on the high-utility subset of the plurality of retrospective events.” Limoli teaches the use of historical (retrospective) data (Column 5, Lines 20-25), Column 8, Lines 32-41 goes into the use of said historical data, which has a priority assigned to it, to train the model to make predictions of real and gap observations.
In regards to claim 5: The present invention claims “wherein the threshold measure is determined based at least in part on a statistical distribution of event valuations associated with the plurality of retrospective events.” Chimmad teaches “Claims with days of service greater 140 were denied. Repeating the simulation CR #1A several times, we found that on an average, claims with days of service greater than 140 were denied and needed rework, which is consistent with Figure 1 and justified by the regression tree below. Further, it can be seen (Figure 2) that Total Charged Amount on the claim is the key deciding feature, followed by the Total Amount Paid and the Duration of service and claim submission cycle (Days). It stands to reason that claims involving excessive charge amounts or long delays in processing are flagged for rework.” (Page 3, right column, see also Figure 3). See also how Maisog details basing predictions off of claimant valuation/spending would have been well-known in the art at the time of the Applicant’s filing.
In regards to claim 6: (As per Applicant’s Specification [0030], an “input channel” is being interpreted as a data repository) The present invention claims “determining a predictive input channel of a plurality of predictive input channels that is associated with the predictive input entity;” Limoli Column 4, Lines 42-45 teaches that the input data can come from multiple different sources (input channels), Column 5, Lines 1-28 also go into details pertaining to various data sources.
“and determining the prospective qualifying criteria satisfaction predictive output based at least in part on the predictive input channel.” Limoli teaches “In response to determining that iterative adjustment of the machine learning model is complete, the prioritization module can determine priority estimates for observed and unobserved data points in real time to be used to perform the machine interpretation of clinical guidelines.” (Column 2, Lines 51-55).
In regards to claim 7: The present invention claims “wherein the plurality of predictive input channels comprises one or more of a model-based prospective channel, a rule-based prospective channel, a model-based real-time channel, or a rule- based real time channel.” Limoli teaches “The prioritization module can also iteratively adjust the machine learning model using a loss calculated by comparing the priority estimates against a priority index comprising historical physician behaviors (model-based prospective) and patient outcomes. In response to determining that iterative adjustment of the machine learning model is complete, the prioritization module can determine priority estimates for observed and unobserved data points in real time (real-time) to be used to perform the machine interpretation of clinical guidelines. (rule-based)” (Column 2, Lines 47-55).
In regards to claim 8: The present invention claims “further comprising: assigning the predictive input entity to the programmatic agent based at least in part on the prospective priority score, and causing the programmatic agent to process the predictive input entity.” Limoli teaches “According to some embodiments, based on a set of inputs (e.g., physician specialty, past behavior of user, past behavior of similar users, clinical guidelines, patient profile, previous outcomes, monitored dialogs between clinician and patient, voice input from a clinician, etc.), the disclosed systems and methods can be configured to assign values across multiple dimensions, including urgency and relevancy, to the patient framework which can then be applied to the patients' clinical data. According to some embodiments, these assigned values can be used to create a rigorous methodology for surfacing the most important and/or relevant pieces of information to physicians, thereby enabling increased and/or improved expeditious data review and better treatment decision making than currently available 60 EHR systems" (Column 4, Lines 46-60).
Chimmad also teaches “We report on a comprehensive classification engine using multiple ML algorithms (Binary Classification Trees, Neural Networks (NN), Support Vector machines (SVM) and Naive Bayes Classification, to predict whether a given claim is likely to involve rework (i.e., has a high risk of rejection or denial) via binary classification" (Section I, Paragraph 4); "ML experiments were conducted using three different approaches to feature engineering as follows: (1) CARC Codes (RSNl) included in the dataset as a single 'Status' feature without one-hot coding; (2) same as above but also converting the CARC codes as eight distinct features using one-hot (or Dummy) coding and (3) engineering 4 additional (synthetic) features added to the dataset" (Section IlI, Paragraph 1); and "Three different Binary Classification algorithms (Classification Trees, SVM and Neural Networks) were used to classify the claims into High risk or Low risk for Denial. Training data sets were prepared by adding a status field of 100 (accepted) for all the claims with reason number (RSN) 100 and 200 (denied) for the remaining claims” (Section Ill, Paragraph 2).
In regards to claim 9: The present invention claims “generating an investigation agent user interface for the programmatic agent that comprises one or more graphical elements respectively reflective of one or more investigation queue features of a related subset associated with the programmatic agent.” Limoli teaches "FIG. 5 depicts an exemplary embodiment of the context-driven interface 5 that serves as both the user-facing manifestation of the prioritized data and a user-action capturing mechanism to generate data about physician practice patterns" (Column 10, Lines 39-43).
In regards to claims 10-18: Claims 10-18 recite similar limitations to claims 1-9, with the exception of “A system comprising: one or more processors; and one or more memories storing processor-executable instructions that, when executed by the one or more processors, cause the one or more processors to perform operations comprising:” therefore, both sets of claims are similarly rejected.
In regards to claims 19-20: Claims 19-20 recites similar limitations to claims 1-2, with the exception of “One or more non-transitory computer-readable media storing processor-executable instructions that, when executed by one or more processors, cause the one or more processors to perform operations comprising:” of claim 19, therefore, both claims are similarly rejected.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to GRIFFIN T BEAN whose telephone number is (703)756-1473. The examiner can normally be reached M - F 7:30 - 4:30.
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, Li Zhen can be reached at (571) 272-3768. 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.
/GRIFFIN TANNER BEAN/ Examiner, Art Unit 2121
/Li B. Zhen/ Supervisory Patent Examiner, Art Unit 2121
Prosecution Insights