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 Claims
This Office Action is responsive to the applicant response filed August 20, 2025.
Claims 1-19 are currently pending and have been fully examined.
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 1-19 is/are 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.
Step 1
The claim(s) recite(s) subject matter within a statutory category as a process (claim 10), a machine (claim 1), and an article of manufacture (claim 19) which is recited as a method, system, and non-transitory computer readable medium that performs the steps and/or functions of:
generate a feature vector that includes a first plurality of values and a second plurality of values,
wherein the first plurality of values corresponds to a respective plurality of static variables that are knowable at a time a patient goes into labor, and
the second plurality of values corresponds to a respective plurality of dynamic variables that are associated with a particular time during labor, the second plurality of values includes at least a most recent cervical dilation value;
provide the feature vector to a trained machine learning model, wherein the trained machine learning model was trained using a plurality of labeled feature vectors associated with a respective plurality of patients associated with one or more known labor outcomes,
wherein each of the plurality of labeled feature vectors included values corresponding to the plurality of static variables and the plurality of dynamic variables associated with a respective patient and associated with a cervical dilation value in a range that includes the most recent cervical dilation value, and
each of the plurality of labeled feature vectors is associated with an indication of one or more unfavorable outcomes experienced by the respective patient;
receive, from the trained machine learning model, an output indicative of a risk that the patient will experience at least one of the one or more unfavorable outcomes; and
cause information indicative of the risk to be presented to a user to aid the user in determining whether to recommend intrapartum Cesarean delivery for the patient.
Step 2A: Prong 1
When taken individually and as a whole, the steps corresponds to concepts identified as abstract ideas by the courts, such as “certain methods of organizing human activity”, which are interactions between individuals that can include: fundamental economic principles or practices; commercial or legal interactions (including agreements in the form of contracts; legal obligations; advertising, marketing or sales activities or behaviors; business relations); and managing personal behavior or relationships or interactions between people (including social activities, teaching, and following rules or instructions).
The claim is directed to a system to perform the process of providing recommendations based on a patient’s risk, which is performed by the system performing the limitations underlined above.
These describe the system managing the behavior of an individual by providing rules or instructions based on a patient’s level of risk, and the patient’s level of risk is determined by generating feature vectors that include a first and second plurality of values, performing calculations generated by a statistical model that identifies a relationship between the patient’s data and the predicted outcomes. These concepts are used to generate the patient’s risk determination that is used as part of providing the rules or instructions to the user. These concepts relate to the process that a medical professional would follow in making a medical recommendation to recommend intrapartum Cesarean delivery for the patient and thus is considered to be a certain method of organizing human activity.
Step 2A: Prong 2
The claims do not include additional elements that are sufficient to be considered a practical application because the additional elements amount to: insignificant extra-solution activity (MPEP 2106.05(g)), generally linking the application of the abstract idea to a particular field of use or technological environment (2106.05(h)), or mere instructions to apply it with a computer (MPEP 2106.05(f)), as discussed below.
Mere Instructions to Apply the Abstract Idea Using a Computer
The steps reciting the use of computer components, such as providing the data to the trained machine learning model or receiving the output from the machine learning model, serve as mere instructions to apply the abstract idea using a computer. Mere instructions to apply the abstract idea using a computer are not sufficient to integrate the abstract idea into a practical application or amount to significantly more than the abstract idea (MPEP 2106.05(f)).
Step 2B
The claims also do not include additional elements that are sufficient to be considered a significantly more than the abstract idea because the additional elements amount to: insignificant extra-solution activity (MPEP 2106.05(g)), mere instructions to apply it with a computer (MPEP 2106.05(f)), generally linking the application of the abstract idea to a particular field of use or technological environment (MPEP 2106.05(h)), or a well-understood, routine, and conventional limitation (MPEP 2106.05(d)), as discussed below.
The steps addressed above in Step 2A: Prong 2, when considered again under Step 2B are not considered to make the claims amount to significantly more than the abstract idea because those steps, when considered additionally with regards to Step 2B, are still considered to be either insignificant extra-solution activity, mere instructions to apply an abstract idea with a computer, or generally linking the application of the abstract idea to a particular field of use or technological environment, which are types of limitations that are not sufficient to make the claims amount to significantly more than the abstract idea (MPEP 2106.05.I.A).
The steps recited as either being part of the abstract idea or insignificant extra-solution activity are all examples of at least one of: storing and retrieving data from a memory (retrieving data that is stored locally), sending and receiving data over a network (receiving data when that data is from a remote source), electronic recordkeeping, or performing repetitive calculations. All of those functions have been identified as well-understood, routine, and conventional functions of a generic computer that are not significantly more than the abstract idea when claimed broadly or as an extra-solution activity (MPEP 2106.05(d).II).
The recited computer components (e.g., the at least one hardware processor and the non-transitory computer readable medium) are all generically recited components (see specification, par. [0060]-[0061]). Commercially available components, generic computer components, and specially-programmed computer components performing the functions of a generic computer are not considered to be amount to significantly more than the abstract idea (MPEP 2106.05(b)).
When considered as a whole, the components do not provide anything that is not present when the component parts are considered individually. Using the broadest reasonable interpretation, the system as a whole is a system of general purpose computer components that analyze patient data and provide recommendations based on the patient’s determined risk level. This is a general purpose computer system performing the abstract idea and insignificant extra-solution activities through these generically described devices performing well-understood, routine, and conventional functions of a generic computer (MPEP 2106.05(d).II).
Dependent Claim Analysis
Claims 2-9 are ultimately dependent from Claim(s) 1 and includes all the limitations of Claim(s) 1. Therefore, claim(s) 2-9 recite the same abstract idea of certain methods of organizing human activity of claim 1.
Claim 2 recites narrowing details of the trained machine learning model, narrowing the additional element recited in the independent claim rather than presenting a further additional element.
Claims 3 and 7 both recite additional limitations that serve to further describe the mathematical concepts (baseline values, plotting outcomes on graphs using curves to represent average risk scores and risk scores for specific patients experiencing a specific outcome) involved that are used as part of the abstract idea. Therefore, these limitations are also part of the abstract idea. “Because a judicial exception is not eligible subject matter, Bilski, 561 U.S. at 601, 95 USPQ2d at 1005-06 (quoting Chakrabarty, 447 U.S. at 309, 206 USPQ at 197 (1980)), if there are no additional claim elements besides the judicial exception, or if the additional claim elements merely recite another judicial exception, that is insufficient to integrate the judicial exception into a practical application.” (MPEP 2106.04.II.A.2).
Claims 4-6 and 9 all recite additional limitations that serve to select by type or source the data to be manipulated by describing the types of data that are to be used when performing the analysis. Selecting by type or source the data to be manipulated is an insignificant extra-solution activity that is not sufficient to integrate the abstract idea into a practical application or amount to significantly more than the abstract idea (MPEP 2106.05(g)).
Claim 8 recites additional limitations that amount to performing an additional iteration of the steps performed as part of claim 1. This merely describes performing the steps again using newly acquired data and does not provide any unpredictable results. Therefore, these limitations are not sufficient to integrate the abstract idea into a practical application or amount to significantly more than the abstract idea.
Claims 10-18 are ultimately dependent from Claim(s) 10 and includes all the limitations of Claim(s) 10. Therefore, claim(s) 10-18 recite the same abstract idea of certain methods of organizing human activity of claim 10.
Claims 10-18 all recite additional limitations that are the same or substantially similar to the limitations of claims 2-9, respectively. Therefore, claims 10-18 are rejected under 101 for the same reasons as claims 2-9.
Claim Rejections - 35 USC § 103
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, 10, 19 are rejected under 35 U.S.C. 103 as being unpatentable over Basu (US PGPub 2018/0203978) in view of Cossler (US PGPub 2018/0182475).
Claim 1
Basu teaches a system for predicting a risk of one or more unfavorable labor outcomes in a patient, the system comprising:
at least one hardware processor [para. 25-26 teaches the uses of multiple wearable computing devices] that is programmed to:
generate a feature vector that includes a first plurality of values and a second plurality of values, wherein the first plurality of values corresponds to a respective plurality of static variables and the second plurality of values corresponds to a respective plurality of dynamic variables [para 23 teaches an approach “combining static features low-frequency, non-clinical data (weight, symptom reports) and continuous dynamic data from wearable physiological sensors…using a machine-learning approach…to predict a user’s risk of worsening symptoms…over a specific time period based on a combination of static data available at discharge and dynamic data collected by wearable sensors”];
provide the feature vector to a trained machine learning model, wherein the trained machine learning model was trained using a plurality of labeled feature vectors associated with a respective plurality of patients associated with one or more known labor outcomes, wherein each of the plurality of labeled feature vectors included values corresponding to the plurality of static variables and the plurality of dynamic variables associated with a respective patient [para 26 teaches the system includes a secondary computing device that includes a machine learning model that evaluates static and dynamic data; para 61 – “population data may be used as an input to machine learning model.. data pertaining to a plurality of patients undergoing similar treatment and monitoring may be included in population data…. Population data may include data from a plurality of previous heart failure patients, from a hospital’s internal records, published literature, cross-hospital data sets”; para 68-77 teaches data available for machine learning includes EMR data including thousands of features and hundreds of medical records],
and each of the plurality of labeled feature vectors is associated with an indication of one or more unfavorable outcomes experienced by the respective patient [para 102 – “Risk scores may also be affected by a lack of medication ingestion, as non-compliance with prescriptions is a strong indicator of negative outcomes”; paras 58, 79, 96 – the data processed includes medication data];
receive, from the trained machine learning model, an output indicative of a risk that the patient will experience at least one of the one or more unfavorable outcomes [para 23 - “combining static features low-frequency non-clinical data (weight, symptom reports) and continuous dynamic data from wearable physiological sensors to provide a continuous prediction of disease state… a machine learning model may be used to predict a user’s risk of worsening symptoms, DHF, and/or a readmission over a specific time period based on a combination of static data available at discharge and dynamic data collected by wearable sensors”, worsening symptoms being unfavorable outcomes; para 82 – “N machine-learning models are trained for each of the EMR-based models… the N EMR-based models are run on the EMR data for the patient; their outputs, along with the features from the device, are then used as the features for the final model that produces the risk score”; para 87 – “a machine learning model is generated based on the patient’s static data at time of discharge as well as EMR records for numerous other patients. The static data and potentially dynamic data from other patients and any dynamic data acquired while the patient was under clinical care may also inform the machine-learning model. A set of classifiers for the dynamic data may then be used to evaluate incoming dynamic data for the patient over time and to determine a risk score for decompensation”]; and
cause information indicative of the risk to be presented to a user to aid the user in determining whether to recommend intrapartum Cesarean delivery for the patient [para 94 – “the risk score may be presented to the user… the risk score may also be being presented to a clinician”].
Basu does not explicitly teach that:
the static variables are knowable at a time a patient goes into labor,
the dynamic variables are associated with a particular time during labor and includes at least a most recent cervical dilation value,
the dynamic variables are associated with a cervical dilation value in a range that includes the most recent cervical dilation value.
However, Cossler teaches static variables knowable at a time a patient goes into labor [para 53 – “the EMR/HER data can include any known and recorded medical and/or health information about current and past patients… the EMR/HER data can include information… including but not limited to demographic profile information for respective patients, information regarding past medical history (e.g., including past diagnosis, procedures, conditions, etc.), vital signs, progress notes, medications, immunization dates, allergies, laboratory report data, imaging studies, pathology report data, care plan information, insurance information, and the like”], and dynamic variables associated with a particular time during labor and a cervical dilation value within a range [para 61 teaches tracking a patient’s current (i.e., most recent) physiological state or condition, timing associated with the tracked physiological parameters, changes in a physiological status or state of a patient and including “changes in dilation”; para 80 teaches that key data points with respect to laboring patients includes an onset of labor and a change in the dilation of the cervix; para 137 – “different colors or color shades can be employed to indicate a characteristic of the information represented thereby. For example, different shades of yellow can be employed to distinguish between different degrees of dilation. As seen in GUI 600, as dilation increases from 2 cm to 9 cm (e.g., range of cervical dilation value), the shade of color associated with each cell including the dilation information intensifies].
Both Basu and Cossler are directed towards collecting and monitoring physiological conditions of a patient and applying machine learning to predict the risk of adverse outcomes. Thus, they are regarded to be analogous references directed towards solving similar problems even if different contexts (healthcare delivery and heart health). It would have been obvious before the effective filing date to one of ordinary skill in the art to modify the teachings of Basu to include the steps of include dynamic variables associated with a particular time during labor and including a cervix dilation value within a range, as taught by Cossler, because doing so further enhances the ability of Basu to combine static data with supplemental dynamic, real-time data to provide an accurate and continuously updating risk score which can be used to recommend treatment decisions [Basu, para 3].
Claims 10 and 19 recite limitations that are substantially similar to those of claim 1, the only difference being the statutory category of invention (claim 1 being a system, claim 10 being the corresponding method and claim 19 being the corresponding CRM claim). Thus, the same rejection applies.
Further, Basu teaches a non-transitory computer readable medium containing computer executable instructions [para 111-115 – logic machine may be configured to execute instructions… such instructions may be implemented to perform a task, implement a data type, transform the state of one or more components, achieve a technical effect, or otherwise arrive at a desired result… logic machine may include one or more processors configured to execute software instructions] for performing the claimed methodology.
Claims 2, 3, 11, 12 are rejected under 35 U.S.C. 103 as being unpatentable over Basu in view of Cossler, and further in view of Sullivan (US PGPub 2016/0135706).
Claim 2
Basu in view of Cossler teaches the system of claim 1, but does not explicitly teach wherein the trained machine learning model is a gradient boosting machine model comprising a plurality of decision trees.
However, Sullivan teaches the use of gradient boosting machine model comprising a plurality of decision trees [para 314-315 – “a machine learning classifier as described in further detail below can be trained on a large population, for example, a population that can range from several thousand to tens of thousands of patient records comprising electrophysiology, demographic and medical history information. The machine learning tool can include but is not limited to classification and regression tree decision models, such as random forest and gradient boosting, (e.g., implemented using R or any other statistical/mathematical programming language)…. An overview of how a random forest tool may be applied to a given dataset can illustrate how a classification tool may work in interpreting given parameters or metrics. A random forest is a collection of decision trees. A decision tree is a flow chart-like structure in which each node represents a test on a metric and each branch represents the outcome of the test. The tree culminates in a classification label, e.g., a decision taken at the end after computing each of the metrics. Each tree in a random forest tool gets a “vote” in classifying a given set of metrics. There are two components of randomness involved in the building of a random forest. First, at the creation of each tree, a random subsample of the total data set is selected to grow the tree. Second, at each node of the tree, a “splitter variable” is selected and the underlying patients are separated into two classes. For example, patients in one class (e.g., Response or occurrence of sudden cardiac arrest) can be separated from those in another class (e.g., Non-Response). The tree is grown with additional splitter variables until all terminal nodes (leaves) of the tree are purely one class or the other. The tree is “tested” against patient records that have been previously set aside. Each patient testing record traverses the tree, going down one branch or another depending on the metrics included in the record for each splitter variable. The patient testing record is assigned a predicted outcome based on where the record lands in the tree (a vote). The entire process may be repeated with new random divisions of the underlying dataset to produce additional trees and ultimately a “forest”. In each case, a different subset of patients can be used to build the tree and test its performance.”].
Basu and Sullivan both apply machine learning to collected physiological data of a patient to predict the risk of adverse outcomes. Thus, they are regarded to be analogous references directed towards solving similar problems. It would have been obvious before the effective filing date to one of ordinary skill in the art to modify the teachings of Basu to use gradient boosting machine model comprising a plurality of decision trees, as taught by Sullivan, because doing so further strengthens machine learning model of Basu to improve accuracy in determining the risk of outcome to the patient.
Claim 11 recites limitations that are substantially similar to those of claim 2, the only difference being the statutory category of invention. Thus, the same rejection applies.
Claim 3
Basu in view of Cossler in view of Sullivan teaches the system of claim 2, and Basu teaches wherein the at least one hardware processor is further programmed to:
receive, from a baseline machine learning model, a baseline output indicative of a risk that the patient will experience at least one of the one or more unfavorable outcomes based on variables knowable at the time the patient went into labor, wherein the baseline machine learning model was trained using a second plurality of labeled feature vectors associated with a respective plurality of patients associated with one or more known labor outcomes; wherein each of the second plurality of labeled feature vectors included values corresponding to the plurality of static variables associated with a respective patient and omitted any dynamic variables associated with the respective patient; and each of the plurality of labeled feature vectors is associated with an indication of one or more unfavorable outcomes experienced by the respective patient, and include the baseline output in the feature vector [para 23 - “combining static features low-frequency non-clinical data (weight, symptom reports) and continuous dynamic data from wearable physiological sensors to provide a continuous prediction of disease state… a machine learning model may be used to predict a user’s risk of worsening symptoms, DHF, and/or a readmission over a specific time period based on a combination of static data available at discharge and dynamic data collected by wearable sensors”, the static features low-frequency non-clinical data being variables knowable at the time the patient went into labor, worsening symptoms being unfavorable outcomes; para 62 – “Static data 320, dynamic data 330, periodic data 350, and population data 360 may be fed into machine-learning model 310. Initially, static data 320 and population data 360 may be weighted more heavily until a threshold amount of dynamic data 330 and periodic data 350 is acquired. Static data 320 and population data 360 may be used to establish initial sets of dynamic data classifiers 370 and periodic data classifiers 380. The determined classifiers may then be used to evaluate incoming dynamic data 330 and periodic data 350, respectively”, the various data including values associated with a patient; para 82 – “N machine-learning models are trained for each of the EMR-based models. Once these machine-learning models are trained, they are run on the cohort data R to produce outputs for each of the subjects i. These initial risk scores are then concatenated with the device-derived dynamic features X to form the training data for the final model. A production model for this approach is shown at 445 of FIG. 4E. The N EMR-based models are run on the EMR data for the patient; their outputs, along with the features from the device, are then used as the features for the final model that produces the risk score”, the initial risk score being the baseline output and each of the N machine-learning models being trained using cohort data; para 92 – “evaluation of the second set of data against the dynamic data classifiers comprises an evaluation of a fixed history of dynamic features against the dynamic data classifiers . A machine learning model may be configured as a discrete classifier that uses all of a user ' s first set of data as an input along with a fixed history (e.g., three days) of the user's second set of data , but does not make use of historical data from the second data set prior to that window”, dynamic data from prior to the window of fixed history not being included, or omitted dynamic data],
Claim 12 recites limitations that are substantially similar to those of claim 3, the only difference being the statutory category of invention. Thus, the same rejection applies.
Claims 4, 13 are rejected under 35 U.S.C. 103 as being unpatentable over Basu in view of Cossler in view of Sullivan in further view of Cantor (US PGPub 2009/0137925).
Claim 4
Basu in view of Cossler in view of Sullivan teaches the system of claim 3, wherein the trained machine learning model is trained to predict risk that the patient will experience at least one of the one or more unfavorable outcomes.
Basu does not explicitly teach that the predicted risk of experiencing an unfavorable outcome is based on data collected through 4 centimeters (cm) cervical dilation, and wherein the most recent cervical dilation value is a cervical dilation value that is at least 4 cm cervical dilation and less than 5 cm.
However, Cantor teaches cervical dilation values of 4-5 cm during labor [para 3 – “During pregnancy, the cervix remains lengthy and thick to protect the baby. During labor, the cervix effaces and dilates. Effacement refers to the thickness of a cervix and is measured by percentage thinned. A 0% effaced cervix is thick, and a 100% is very thin. Dilation refers to the opening of the cervix and is traditionally measured in centimeters. A cervix with 0 cm dilation is closed, a cervix is in early labor between 0-4 cm, a cervix in active labor is dilated between 4 and 8 cm, a cervix in transition is between 8-10 cm, and a fully dilated cervix is approximately 10 cm. Dilation and effacement work together to open the cervix to allow the baby to pass into the vagina. The state of the cervix is only one of many critical pieces of information considered when deciding how to safely deliver the infant.”].
Basu and Cantor both collect physiological data of a patient during medical treatment. Thus, they are regarded to be analogous references directed towards a similar field of endeavor. It would have been obvious before the effective filing date to one of ordinary skill in the art to modify the teachings of Basu to collect data through 4 to 5 centimeters (cm) cervical dilation, as taught by Cantor, since the state of the cervix is among the critical information considered when deciding how to safely deliver a patient [Cantor, para 3], which would enhance the ability of Basu to combine static data and records with supplemental dynamic, real-time data to provide an accurate and continuously updating risk score which can be used to recommend treatment decisions [Basu, para 3, 24].
Claim 13 recites limitations that are substantially similar to those of claim 4, the only difference being the statutory category of invention. Thus, the same rejection applies.
Claims 5, 14 are rejected under 35 U.S.C. 103 as being unpatentable over Basu in view of Cossler in further view of Frenz (US PGPub 2019/0029875).
Claim 5
Basu in view of Cossler teach the system of claim 1, but does not explicitly teach wherein the plurality of static variables includes variables corresponding to parity, a binary indication of whether the patient has previously delivered via Cesarean, and the patient's age.
However, Frenz teaches predictors used for developing a classification model, the predictors including age, parity and number of Casarean sections [para 32 – parameters can be in the form of yes/no statement, or in the form of a three-level classification; para 154-155, Table 3 – Age and Parity are Baseline data, as are Number of caesarean sections].
Basu and Frenz apply machine learning to collected physiological data of a patient to predict the risk of adverse outcomes. Thus, they are regarded to be analogous references directed towards solving similar problems. It would have been obvious before the effective filing date to one of ordinary skill in the art to modify the teachings of Basu to use age, parity and number of Casarean sections as predictors, as taught by Frenz, since incorporating more indicators as predictors within the classification model which would enhance the ability of Basu to combine static data and records with supplemental dynamic, real-time data to provide an accurate and continuously updating risk score which can be used to recommend treatment decisions [Basu, para 3, 24].
Claim 14 recites limitations that are substantially similar to those of claim 5, the only difference being the statutory category of invention. Thus, the same rejection applies.
Claims 6, 15 are rejected under 35 U.S.C. 103 as being unpatentable over Basu in view of Cossler in further view of Sharf (US PGPub 2004/0236193).
Claim 6
Basu in view of Cossler teaches the system of claim 1, but does not explicitly teach wherein the plurality of dynamic variables includes variables corresponding to cervical dilation, cervical effacement, and head station.
However, Sharf teaches using cervical dilation, effacement and head orientation as a dynamic variable measured during birth [para 4 – “It is common practice to monitor the progress of birth by measuring the degree of cervical dilation and monitoring contraction frequency (other parameters may be monitored as well, such as maternal heart rate)… Currently, the gold standard of measurement for dilation, head orientation, station, effacement and cervical consistency is the human hand. However, this measurement method is not only unsanitary, it is also intrusive and inaccurate. Since the accuracy and meaning of the measurements depends on the person measuring (and even for a same person a 1 cm error is considered normal), when shifts change, measurements change. In addition, it is hypothesized that the cervix dilation increases momentarily during contraction, and fetus head location and orientation changes.”; para 102 – “orientation information is used to determine the degree of cervical effacement and/or the degree and/or effect of contractions. Optionally, orientation information and/or location information is used to measure twisting of cervical os 108. Alternatively or additionally, orientation information is used to determine the passage of parts of fetus 102 past canal parts that have attached probes. Alternatively or additionally, orientation information is used to determine the orientation of head 104, for example to indicate twisting and turning of the head.”; para 125 - “the effacement and/or dilation of cervix 108 are tracked, for example, to generate a distortion map of cervix 108”].
Basu and Sharf apply machine learning to collected physiological data of a patient to predict the risk of adverse outcomes. Thus, they are regarded to be analogous references directed towards solving similar problems. It would have been obvious before the effective filing date to one of ordinary skill in the art to modify the teachings of Basu to use cervical dilation, effacement and head orientation as a dynamic variable measured during birth, as taught by Sharf, since incorporating more indicators as predictors within the classification model which would enhance the ability of Basu to combine static data and records with supplemental dynamic, real-time data to provide an accurate and continuously updating risk score which can be used to recommend treatment decisions [Basu, para 3, 24], and further since studying the distortion of a cervix using effacement and/or dilation and head orientation may indicate a fault in the cervix which might cause premature birth or danger to the mother or child, which can lead to immediate treatment [Sharf, para 125].
Claim 15 recites limitations that are substantially similar to those of claim 6, the only difference being the statutory category of invention. Thus, the same rejection applies.
Claims 7-8, 16-18 are rejected under 35 U.S.C. 103 as being unpatentable over Basu in view of Cossler in further view of Zhong (US PGPub 2017/0053552).
Claim 7
Basu in view of Cossler teaches the system of claim 1, but does not explicitly teach wherein the at least one hardware processor is further programmed to:
plot the outcome on a graph, wherein the graph includes a curve representing average risk scores for patients that did not experience unfavorable outcomes and a second curve representing risk scores for patients that experienced one or more unfavorable outcomes; and
cause the graph to be presented as the information indicative of the risk.
However, Zhong teaches plotting data on a chart or graph relative to a mean using curves, with favorable and unfavorable outcomes that visualize the risk [Figure 4 teaches a plot with data that “can be graphically presented using multiple swoosh curves with a highlighted line that indicates the median or mean, and side bands that indicate the interquartile range”, a plurality of curves that are “overlaid to show good versus bad outcomes corresponding to different variables”, with the median and interquartile ranges marked].
Basu and Zhong apply machine learning to collected physiological data of a patient to predict the risk of adverse outcomes. Thus, they are regarded to be analogous references directed towards solving similar problems. It would have been obvious before the effective filing date to one of ordinary skill in the art to modify the teachings of Basu to plot outcome data and results on a graph that includes curves representing average risk of patients who experienced favorable and unfavorable outcomes, as taught by Zhong, since providing a visual indication on the specific patient’s measured, recorded and monitored physiological readings relative previous patients having experienced both desirable and undesirable outcomes, would enhance the ability of Basu to combine static data and records with supplemental dynamic, real-time data to provide an accurate and continuously updating risk score which can be used to recommend treatment decisions [Basu, para 3, 24]
Claim 16 recites limitations that are substantially similar to those of claim 7, the only difference being the statutory category of invention. Thus, the same rejection applies.
Claim 8
Basu in view of Cossler in view of Zhong teaches the system of claim 7, and Basu teaches wherein the at least one hardware processor is further programmed to:
generate a second feature vector that includes the first plurality of values and a third plurality of values, wherein the third plurality of values corresponds to a respective plurality of dynamic variables that are associated with a second particular time during labor, including at least a most recent cervical dilation value [para 23 teaches an approach “combining static features low-frequency, non-clinical data (weight, symptom reports) and continuous dynamic data from wearable physiological sensors…using a machine-learning approach…to predict a user’s risk of worsening symptoms…over a specific time period based on a combination of static data available at discharge and dynamic data collected by wearable sensors”];
provide the second feature vector to a second trained machine learning model, wherein the second machine learning model was trained using a second plurality of labeled feature vectors associated with the respective plurality of patients associated with the one or more known labor outcomes [para 26 teaches the system includes a secondary computing device that includes a machine learning model that evaluates static and dynamic data; para 61 – “population data may be used as an input to machine learning model.. data pertaining to a plurality of patients undergoing similar treatment and monitoring may be included in population data…. Population data may include data from a plurality of previous heart failure patients, from a hospital’s internal records, published literature, cross-hospital data sets”; para 68-77 teaches data available for machine learning includes EMR data including thousands of features and hundreds of medical records],
wherein each of the second plurality of labeled feature vectors included values corresponding to the plurality of static variables and the plurality of dynamic variables associated with a respective patient and associated with a cervical dilation value in a second range that includes the most recent cervical dilation value included in the third plurality of values [para 26 teaches the system includes a secondary computing device that includes a machine learning model that evaluates static and dynamic data; para 61 – “population data may be used as an input to machine learning model.. data pertaining to a plurality of patients undergoing similar treatment and monitoring may be included in population data…. Population data may include data from a plurality of previous heart failure patients, from a hospital’s internal records, published literature, cross-hospital data sets”; para 68-77 teaches data available for machine learning includes EMR data including thousands of features and hundreds of medical records], and
each of the second plurality of labeled feature vectors is associated with an indication of one or more unfavorable outcomes experienced by the respective patient [para 102 – “Risk scores may also be affected by a lack of medication ingestion, as non-compliance with prescriptions is a strong indicator of negative outcomes”; paras 58, 79, 96 – the data processed includes medication data];
receive, from the second trained machine learning model, a second output indicative of an updated risk that the patient will experience at least one of the one or more unfavorable outcomes [para 87 – “A set of classifiers for the dynamic data may then be used to evaluate incoming dynamic data for the patient over time and to determine a risk score for decompensation”];
generate an updated graph by plotting the second outcome on the graph [para 116 teaches that a display subsystem may be used to present a visual representation of data]; and
cause the updated graph to be presented to the user to aid the user in determining whether to recommend intrapartum Cesarean delivery for the patient [para 93-94 – “the machine-learning model may assume that all of the dependence in previous time steps is captured in the previous model’s output. With this assumption, to predict a decompensation risk at time t, the machine-learning model may only look at dynamic data associated with time t, along with the model output from time t−1. The first set of data may be reincorporated at each time step, and/or may factor explicitly only into the initial output to and would then be implicitly carried forward in the model output…the risk score may be presented to the user… the risk score may also be being presented to a clinician”].
Basu teaches generating alerts when the measured quantities exceed simple thresholds [para 21] and collecting data that dynamically updates over time and using a machine learning model to indicate risk scores for the user based on an evaluation of the dynamically updating data [para 123-125]
Basu does not explicitly teach that:
the dynamic variables are associated with a second particular time during labor and includes at least a most recent cervical dilation value that exceeds an upper limit of the range associated with the trained model;
the dynamic variables are associated with a cervical dilation value in a second range that includes the most recent cervical dilation value included in the third plurality of values.
Cossler teaches:
monitoring cervical dilation value within a range to determine an indication of the stage of labor being at a second particular time different than a first particular time [para 61 teaches tracking a patient’s current (i.e., most recent) physiological state or condition, timing associated with the tracked physiological parameters, changes in a physiological status or state of a patient and including “changes in dilation”; para 80 teaches that key data points with respect to laboring patients includes an onset of labor and a change in the dilation of the cervix; para 137 – “different colors or color shades can be employed to indicate a characteristic of the information represented thereby. For example, different shades of yellow can be employed to distinguish between different degrees of dilation. As seen in GUI 600, as dilation increases from 2 cm to 9 cm (e.g., range of cervical dilation value), the shade of color associated with each cell including the dilation information intensifies… different colors can be employed to distinguish between different stages associated with labor (e.g., stage I is marked by light green, stage II marked by light purple) and different categories associated with labor (e.g., category I is marked by pink and category II is marked by red)”, the stage of labors representing different particular times during labor];
collecting data for a physiological condition exceeding an upper limit threshold [para 31 - AI system can further be configured to characterize events or conditions associated with a risk/severity level above a threshold level as being significant. In other implementation, the AI system can further determine a score to associate with a significant event or condition that reflects its level of risk/severity; para 76 – “with respect to laboring patients, significant events or conditions can include the key data and decision points noted above. Similarly, key data and decision points associated with treatment of a medical condition, performance of a surgical procedure and the like can also be considered significant events or conditions. In another example, a significant event or condition can include detection of a physiological parameter, parameter value, and/or combination of different physiological parameters/parameter values monitored for a patient (e.g., the mother and/or the infant in labor scenarios), that warrants further monitoring and/or requires a clinical response (e.g., performance of a medical procedure, change in dosage and/or provision of a pharmaceutical). According to this example, significant events/conditions can include the detection of a particular physiological parameter alone (e.g., detection of a pathogen present in the body), a change in a specific physiological parameter (e.g., decrease in blood pressure), detection of a value of a particular physiological parameter being above or below a set threshold and/or above or below the threshold for a defined duration of time (e.g., decrease in blood pressure below normal level for more than 1.0 minute), and the like. In another example related to patient data 104, a significant event or condition can include one or more physiological parameters and/or parameter values that correspond to a medical complication, or a defined physiological state the patient is experiencing (e.g., a parameter or combination of physiological parameters that reflect the patient has entered stage II of labor).”]; and
updating the graph by plotting new data and outcomes [para 93 – “the visual representation can further be updated in real tie to reflect new tracked events and conditions that arise… in implementations in which the AI response component determines or infers new relevant parameters for tracking, the AI response component can facilitate automatically tracking the new parameter and providing a corresponding visual representation of the new tracked parameter in the GUI].
Both Basu and Cossler are directed towards collecting and monitoring physiological conditions of a patient and applying machine learning to predict the risk of adverse outcomes. Thus, they are regarded to be analogous references directed towards solving similar problems even if different contexts (healthcare delivery and heart health). It would have been obvious before the effective filing date to one of ordinary skill in the art to modify the teachings of Basu to include the steps of include dynamic variables associated with a second particular time during labor and including a cervix dilation value that exceeds a range, as taught by Cossler, because doing so further enhances the ability of Basu to combine static data with supplemental dynamic, real-time data to provide an accurate and continuously updating risk score which can be used to recommend treatment decisions [Basu, para 3].
Claim 17 recites limitations that are substantially similar to those of claim 8, the only difference being the statutory category of invention. Thus, the same rejection applies.
Claims 9 and 18 are rejected under 35 U.S.C. 103 as being unpatentable over Basu in view of Cossler in view of Zhong in view of Cantor.
Claim 9
Basu in view of Cossler in view of Zhong teach the system of claim 8, but do not explicitly teach wherein the range associated with the trained machine learning model includes cervical dilation from about 4 cm to less than 5 cm and the second range associated with the second trained machine learning model includes cervical dilation from about 5 cm to less than 6 cm.
Cantor teaches cervical dilation values ranging from 4-5 cm and 5-6 cm during different stages of labor [para 3 – “During pregnancy, the cervix remains lengthy and thick to protect the baby. During labor, the cervix effaces and dilates. Effacement refers to the thickness of a cervix and is measured by percentage thinned. A 0% effaced cervix is thick, and a 100% is very thin. Dilation refers to the opening of the cervix and is traditionally measured in centimeters. A cervix with 0 cm dilation is closed, a cervix is in early labor between 0-4 cm, a cervix in active labor is dilated between 4 and 8 cm, a cervix in transition is between 8-10 cm, and a fully dilated cervix is approximately 10 cm. Dilation and effacement work together to open the cervix to allow the baby to pass into the vagina. The state of the cervix is only one of many critical pieces of information considered when deciding how to safely deliver the infant.”].
Basu and Cantor both collect physiological data of a patient during medical treatment. Thus, they are regarded to be analogous references directed towards a similar field of endeavor. It would have been obvious before the effective filing date to one of ordinary skill in the art to modify the teachings of Basu to collect data through 4 to 5 centimeters (cm) cervical dilation and 5 to 6 centimeters cervical dilation, as taught by Cantor, since the state of the cervix is among the critical information considered when deciding how to safely deliver a patient [Cantor, para 3], which would enhance the ability of Basu to combine static data and records with supplemental dynamic, real-time data to provide an accurate and continuously updating risk score which can be used to recommend treatment decisions [Basu, para 3, 24].
Claim 18 recites limitations that are substantially similar to those of claim 9, the only difference being the statutory category of invention. Thus, the same rejection applies.
Response to Arguments
Applicant’s arguments with respect to the prior art rejection of the claims 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.
Applicant's arguments filed August 20, 2025 have been fully considered but they are not persuasive.
Applicant argues that the pending claims recite meaningful limitations that sufficiently limit its practical application such that the claims do not seek to tie up any judicial exception such that others cannot practice it. Specifically, applicant argues the limitation of “generate a feature vector that includes a first plurality of values and a second plurality of values, wherein the first plurality of values corresponds to a respective plurality of static variables… and the second plurality of values corresponds to a respective plurality of dynamic variables that are associated with a particular time during labor, the second plurality of values includes at least a most recent cervical dilation value” and providing a feature vector to a certain type of trained machine learning model (trained with a plurality of labeled feature vectors associated with known labor outcomes) in claim 1 and another certain type of machine learning model trained using a second plurality of labeled feature vectors in claim 3.
This argument is not persuasive. The limitations argued by applicant are among the limitations that comprise the abstract idea recited in the claims, and thus would not provide a practical application, as only additional elements can integrate the abstract idea into a practical application. Furthermore, even if the claims do not attempt to tie up all embodiments of the judicial exception, as argued by applicant, the absence of complete preemption does not demonstrate that a claim is eligible. While preemption is the concern underlying the judicial exceptions, it is not a standalone test for determining eligibility. Rapid Litig. Mgmt. v. CellzDirect, Inc., 827 F.3d 1042, 1052, 119 USPQ2d 1370, 1376 (Fed. Cir. 2016). See MPEP 2106.04.
The claims do not provide any improvement to the functioning of a computer or to any other technology or technical field. The additional elements in combination (as well as individually) do not amount to an inventive concept, e.g., because they are not more than the non-conventional and non-generic arrangement of known, conventional elements.
Applicant argues that the finding that the pending claims are not directed to certain methods of organizing human activity because the claims do not recite any rules or instructions that a person or group of people are to follow to achieve some outcome.
This argument is not persuasive. MPEP 2106.04(a)(2)(II)(C) states that BASCOM is an example of claims that are found to organize human activities while not requiring humans to do anything. Other examples of managing personal behavior recited in a claim include: i. filtering content, BASCOM Global Internet v. AT&T Mobility, LLC, 827 F.3d 1341, 1345-46, 119 USPQ2d 1236, 1239 (Fed. Cir. 2016) (finding that filtering content was an abstract idea under step 2A).
Applicant argues that the pending claims do not recite a mathematical concept and at most, merely recite limitations that are based on or involve a mathematical concept, and thus, “mathematical concepts” is not an abstract idea grouping that applies to the claims.
This argument is persuasive. “Mathematical concepts” is not an abstract idea grouping that applies to the claims; however, the claims still recite an abstract idea per the “certain methods of organizing human activity” grouping.
Applicant argues that the subject matter of the pending claims is integrated into a practical application. Applicant points to MPEP 2106.04(d)(I) and paragraphs 3-9, 42, 46-47, 49-50, 67, 70 of the specification in asserting that the subject matter of the pending claims provides a technical solution to the various challenges associated with existing techniques and technologies.
This argument is not persuasive. None of the cited paragraphs point to any improvements to the functioning of the computer, technical field or underlying technology. Having “more accurate” predictions as a result of “more successful accounting for dynamic confounders” is not an improvement to the functioning of a computer, technical field or underlying technology. Providing additional or individual data to a model may improve the accuracy of the model, but does not constitute a technical improvement. It is noted that the modeling and predictors are all part of the recited abstract idea set forth in the independent claim. At best, applicant is identifying the novelty of the abstract idea. However, the technical solution is evaluated based on the additional elements, not the abstract idea. The trained machine learning model is the sole additional element recited in the independent claim, with the remainder of the claim being the abstract idea. The trained machine learning model is utilized in an “apply it” manner. The claim language does not provide any specificity regarding how the trained machine learning model operates to obtain the risk score; at most, the claim recites details regarding the data input and the data used in training the machine learning model, but no details regarding the actual operations performed by the trained machine learning model itself. It is not evident how the claims provide any improvement to the machine learning model beyond providing additional and newer/updated data, which is not an improvement to the model itself.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to PETER H CHOI whose telephone number is (469)295-9171. The examiner can normally be reached M-Th 9am-7pm.
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.
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.
/PETER H CHOI/Supervisory Patent Examiner, Art Unit 3681