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 .
Election/Restrictions
Applicant's election with traverse of Claims 12-14 in the reply filed on January 22nd 2026 is acknowledged. The traversal is on the grounds that there would not be a serious search burden as the distinguishing element between the species do not shift the claimed subject matter into a different field of search. This is not found persuasive because the species represent different statistical analysis methods, that would require a distinct search into varying optimization methods fields. Although the species are linked by the generic claim 8, they pose distinguishing element that show no overlap in the art. Each species would require a distinct prior art to read on each method, and no one art or search would be able to encompass all three, hence the search burden.
Claims 9-11 are withdrawn from further consideration pursuant to 37 CFR 1.142(b), as being drawn to nonelected methods of using average similarity and ratio of biased medical topics to calculate the first index, with Claim 8 being the generic or linking claim. Applicant timely traversed election requirement in the reply filed on January 22nd 2026.
The requirement is still deemed proper and is therefore made FINAL.
Claim Status
Claims 1-8, and 12-15 are currently pending and under exam herein.
Claims 9-11 are withdrawn.
Claim 1-8 and 12-15 are rejected.
Claims 3-8 and 12-14 are objected to.
Priority
Acknowledgment is made of applicant's claim for foreign priority based on an application filed in Taiwanese Application No. 111108870 filed on March 10th 2022. Receipt is acknowledged of certified copies of papers required by 37 CFR 1.55. Thus, the effective filing date of Claims 1-8 and 12-15 are March 10th 2022.
Information Disclosure Statement
The information disclosure statement (IDS) filed on 02/24/2023 has been received. The submission is in compliance with the provisions of 37 CFR 1.97. Accordingly, the information disclosure statement is being considered by the examiner.
Drawings
The Drawings filed on 06/17/2022 are accepted.
Specification
The disclosure is objected to because of the following informalities: incorrect data in table.
Paragraph [0032] states that medical record #N indicates that the patient has not had an unexpected hospitalization event as shown in Table 1, however, Table 1 indicates that medical record #N has had an unexpected hospitalization event
Appropriate correction is required.
Claim Objections
Claim 13 objected to because of the following informalities: “the firs statistical value” in the last line should read “the first statistical value”. Appropriate correction is required.
Claim Rejections - 35 USC § 112
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claim 8 and 12-14 is rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention. Claim 8 recites the limitation "comparing the first index and the second index to select the first topic number from the first topic number and the second topic number". It is unclear whether the first use of the phrase “first topic number” is representing the same thing as the second use of the phrase “the first topic number”. The dependent claims 12-14 are rejected as well, for they do not resolve the indefiniteness of the independent claim 8.
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 1-8 and 12-15 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
In accordance with MPEP § 2106, claims found to recite statutory subject matter (Step 1: YES) are then analyzed to determine if the claims recite any concepts that equate to an abstract idea, law of nature or natural phenomenon (Step 2A, Prong 1). In the instant application, the claims recite the following limitations that equate to an abstract idea/law of nature/natural phenomenon:
Claim 1 recites a method of checking drug interactions. The method starts by generating a drug combination set (3 drug combinations) from medical records, in which the first drug combination and the second drug combination share a first drug, while the first drug combination and third drug combination share a second drug (Abstract idea; mental process). Next, the method generates odds ratios (3 odds ratios) that correlates each drug combination with a hospitalization event based on mathematical counts and Equation (1) in the Specification of the instant application (Abstract idea; mathematical concepts and/or mental process). And based off of these odds ratios, the method then generates a first fraction (risk combination fraction) and a second fraction (normal combination fraction) based on Equation (2) and Equation (3) in the Specification (Abstract Idea; mathematical concepts and/or mental process). Lastly, the method outputs the first drug combinations in response to the first odds ratio being greater than a first threshold, the sum of the fractions (first and second) being greater than a second threshold, and the quotient of the fractions being less than a third threshold (Abstract idea; mathematical process and/or mental math). To summarized, the claim limitations are merely collecting patient data, and attempting to find high risk drug combinations that may lead to hospitalization events through statistical analysis and mathematical relations, which constitutes as an abstract idea. In addition, the mathematical equations and relations are simple enough to be also performed in the human mind, or with pen and paper as shown in the Specification of the instant application, such that these limitations would also fall under mental process in abstract ideas. Please see MPEP § 2106.04(a)(2) for more details.
Claim 2 recites the method of generating the first fraction, in claim 1, corresponding to the first drug. The method starts by comparing the second odds ratio to a risk threshold, and marking the second drug combination if the ratio is greater than the risk threshold (Abstract idea; mathematical concepts and/or mental process). The method then goes on to generate a third fraction, which is equivalent to (the number of drug combinations comprising the first drug but not the second drug, and is marked in the drug combination set) divided by (the number of drug combinations comprising the first drug but not the second drug), which is also the verbal equivalent of Equation (2) in the Specification of the instant application (Abstract Idea; mathematical concepts and/or mental process). Finally, the first fraction can then be obtained by subtracting the third fraction from 1, the verbal equivalent of Equation (3) in the Specification of the instant application (Abstract Idea; mathematical concepts and/or mental process). The limitations above are merely using mathematical relations to calculate fractions pertaining to the drug sets, and in the broadest reasonable interpretation, is simple enough to also be performed in the human mind. Hence, the claim limitations constitute an abstract idea.
Claim 3 recites the method of generating the drug combination set in claim 1. The method generates a “first unique drug combination set” by initially generating K (just a number) topic vectors with a latent Dirichlet allocation (LDA) model where each topic vector contains probability distributions of all drug combinations (Abstract Idea; mathematical concept and/or mental process). Topic modeling is just a natural language processing technique that can process text datasets to product a summary set of terms that represent themes. LDA modeling is a specific topic modeling approach, that assumes words that occur together are likely part of similar topics and generates topic distributions (lists of keywords with respective probabilities) based on word frequency and co-occurrences. In the broadest reasonable interpretation, the process of LDA modeling utilizes mathematical concepts to calculate probabilities of correlation, and in the simplest of scenarios could be done in the human mind. After generating the K number of topic vectors, the method goes on to select the drug combinations with the max probabilities in the first topic vector, to generate a “first important drug combination set” (Abstract Idea; mathematical concepts and/or mental process). Finally, the method uses the “first important drug combination set” to form the “first unique drug combination set” (Abstract Idea; mental process).
Claim 4 further defines the method of forming the “first unique drug combination set” from “first important drug combination set” based in claim 3. The method does this by deleting drug combinations that are in the “first important drug combination set” and a “second important drug combination set” generated from the second topic vector (Abstract Idea; mathematical concept and/or mental process).
Claim 5 further defines the method of generating a drug combination set based in claim 3, by generating a “first stable drug combination set”. The method first repeatedly generates a plurality of unique drug combination sets based on the methods in claim 3 and 4 above (Abstract Idea: Mathematical process and/or mental process). Then, the method generates the “first stable drug combination set” with drug combinations that show up a selected number of times (pass a threshold) in the plurality of unique drug combination sets (Abstract Idea; mathematical concepts and/or mental process). Finally, the method uses the “first stable drug combination set” to form the drug combination set.
Claim 6 further defines the method of generating a drug combination set based on the “first stable drug combination set” in claim 5. The method starts by generating a plurality of medical record vectors with LDA modeling, where each medical record vector comprises of probability distribution of the K topics (Abstract Idea; mathematical concepts and/or mental process). Then, the method determines a medical record set corresponding to the first topic based on the probability distribution (Abstract Idea; mathematical process and/or mental process). Next, the method calculates a ratio of at least one medical record to the medical record set, where the at least one medical record represent a medical record that has at least one drug combination in the “first stable drug combination set” (Abstract Idea: mathematical process and/or mental process). Finally, the drug combination set is generated from the “first stable drug combination set” if the ratio is above a threshold (Abstract Idea; mathematical concept and/or mental process).
Claim 7 further defines the method of determining a medical record set corresponding to the first topic, in claim 6. The method does this by picking the medical records in the medical record vectors that have a max probability in the first topic (Abstract Idea; mathematical concepts and/or mental process)
Claim 8 adds onto the method of claim 3, by generating a first index corresponding to the first topic number and a second index corresponding to a second topic number based on the medical records and LDA model (Abstract Idea; mathematical concept). Then the indexes are compared before selecting one index as K, the first topic number in claim 3 (Abstract Idea; mathematical concepts and/or mental process).
Claim 12 further defines the method in claim 8, with a method to generate the first index. The method starts by generating a plurality of medical record vectors based on the medical records and the first topic number using the LDA model, wherein that the medical record vectors comprise a probability distribution of the K topics (Abstract Idea; mathematical concept and/or mental process). Then the method divides the medical records into K groups according to the probability distribution, where the K groups correspond to the K topics (Abstract Idea; mathematical concept and/or mental process). Next, the method calculates a first statistical value representing the inter-group distances and a second statistical value representing the intra-group distances (Abstract Idea; mathematical concepts and/or mental process). Finally, the method calculates a ratio of the first statistical value to the second statistical value as the first index (Abstract Idea; mathematical concepts and/or mental process).
Claim 13 further defines the method of calculating inter-group distances in claim 12. The method starts by calculating a plurality of distances between the K topic vectors (Abstract Idea; mathematical concepts and/or mental process). Then, it adds all the distances together to get the first statistical value representing the inter-group distances (Abstract Idea; mathematical concepts and/or mental process).
Claim 14 further defines the method of calculating intra-group distances in claim 12. The method starts by calculating a plurality of distances between elements in the first group to generate a sum of intra-group distances corresponding to the first group (Abstract Idea; mathematical concepts and/or mental process). Then the method calculates a second sum of intra-group distance corresponding to a second group (Abstract Idea; mathematical concepts and/or mental process). Finally, the method adds the first sum of intra-group distances to the second sum of intragroup distances to obtain the second statistical value representing the intra-group distances (Abstract Idea; mathematical concepts and/or mental process).
Claim 15 an electronic device that is capable of checking drug interactions. The device starts by generating a drug combination set (3 drug combinations) from medical records, in which the first drug combination and the second drug combination share a first drug, while the first drug combination and third drug combination share a second drug (Abstract idea; mental process). Next, the device generates odds ratios (3 odds ratios) that correlates each drug combination with a hospitalization event based on mathematical counts and Equation (1) in the Specification of the instant application (Abstract idea; mathematical concepts and/or mental process). And based off of these odds ratios, the method then generates a first fraction (risk combination fraction) and a second fraction (normal combination fraction) based on Equation (2) and Equation (3) in the Specification (Abstract Idea; mathematical concepts and/or mental process). Lastly, the method outputs the first drug combinations in response to the first odds ratio being greater than a first threshold, the sum of the fractions (first and second) being greater than a second threshold, and the quotient of the fractions being less than a third threshold (Abstract idea; mathematical process and/or mental math). To summarized, the claim limitations are merely collecting patient data, and attempting to find high risk drug combinations that may lead to hospitalization events through statistical analysis and mathematical relations, which constitutes as an abstract idea. In addition, the mathematical equations and relations are simple enough to be also performed in the human mind, or with pen and paper as shown in the Specification of the instant application, such that these limitations would also fall under mental process in abstract ideas. Please see MPEP § 2106.04(a)(2) for more details.
The limitations regarding generating high risk drug combinations based on patient medical records utilize probabilities, statistical analysis, and mathematical concepts to analyze and correlate data, making them a mathematical concept. In addition, based on the broadest reasonable interpretation, there is no additional limit on the data being process, such that the mathematical calculations cannot be performed in the human mind, hence these limitations also constitute a mental process. These recitations are similar to the concepts of collecting information, analyzing it and displaying certain results of the collection and analysis in Electric Power Group, LLC, v. Alstom (830 F.3d 1350, 119 USPQ2d 1739 (Fed. Cir. 2016)), organizing and manipulating information through mathematical correlations in Digitech Image Techs., LLC v Electronics for Imaging, Inc. (758 F.3d 1344, 111 U.S.P.Q.2d 1717 (Fed. Cir. 2014)) and comparing information regarding a sample or test to a control or target data in Univ. of Utah Research Found. v. Ambry Genetics Corp. (774 F.3d 755, 113 U.S.P.Q.2d 1241 (Fed. Cir. 2014)) and Association for Molecular Pathology v. USPTO (689 F.3d 1303, 103 U.S.P.Q.2d 1681 (Fed. Cir. 2012)) that the courts have identified as concepts that can be practically performed in the human mind or mathematical relationships. Therefore, these limitations fall under the “Mental process” and “Mathematical concepts” groupings of abstract ideas. While claims 15 recite performing some aspects of the analysis with “an electronic device”, there are no additional limitations that indicate that this device requires anything other than carrying out the recited mental process or mathematical concept in a generic computer environment. Merely reciting that a mental process is being performed in a generic computer environment does not preclude the steps from being performed practically in the human mind or with pen and paper as claimed. If a claim limitation, under its broadest reasonable interpretation, covers performance of the limitation in the mind but for the recitation of generic computer components, then if falls within the “Mental processes” grouping of abstract ideas. As such, claims 1-8 and 12-15 are an abstract idea (Step 2A, Prong 1: YES).
Claims found to recite a judicial exception under Step 2A, Prong 1 are then further analyzed to determine if the claims as a whole integrate the recited judicial exception into a practical application or not (Step 2A, Prong 2). This judicial exception is not integrated into a practical application because the claims do not recite an additional element that reflects an improvement to technology or applies or uses the recited judicial exception in some other meaningful way. Rather, the instant claims recite additional elements that amount to mere instructions to implement the abstract idea in a generic computing environment. Specifically, the claims recite the following additional elements:
Claim 1 recites obtaining a plurality of medical records, wherein at least one of the plurality of medical records indicates whether a patient taking a first drug combination has a hospitalization event
Claim 1 recites outputting the first drug combination in response to a number of thresholds being met
Claims 15 recites an electronic device with a transceiver and a processor coupled to the transceiver
Claim 15 recites obtain, through transceiver, a plurality of medical records, wherein at least one of the plurality of medical records indicates whether a patient taking a first drug combination has a hospitalization event
Claim 15 recites outputting, through the transceiver, the first drug combination in response to a number of thresholds being met
There are no limitations that indicate that the claimed method of obtaining and outputting data along with the electronic device with a transceiver and processor require anything other than generic computing systems. As such, these limitations equate to mere instructions to implement the abstract idea on a generic computer that the courts have stated does not render an abstract idea eligible in Alice Corp., 573 U.S. at 223, 110 USPQ2d at 1983. See also 573 U.S. at 224, 110 USPQ2d at 1984. In general, linking the use of an abstract idea to a particular technological environment, such as a computer, does not integrate the abstract idea into a practical application based on MPEP 2106.05(h). As such, claims 1-8 and 12-15 are directed to an abstract idea as the additional elements do not integrate the judicial exceptions into a practical application (Step 2A, Prong 2: NO).
Claims found to be directed to a judicial exception are then further evaluated to determine if the claims recite an inventive concept that provides significantly more than the judicial exception itself (Step 2B). In claims 1-8 and 12-15, there is no additional element or limitation that would indicate anything other than carrying out the mathematical analysis on a generic computer. According to MPEP 2106.05(d), courts have held computer‐implemented processes not to be significantly more than an abstract idea (and thus ineligible) where the claim as a whole amount to nothing more than generic computer functions merely used to implement an abstract idea, such as an idea that could be done by a human analog (i.e., by hand or by merely thinking). The additional elements do not comprise an inventive concept when considered individually or as an ordered combination that transforms the claimed judicial exception into a patent-eligible application of the judicial exception. Therefore, the claims do not amount to significantly more than the judicial exception itself (Step 2B: No). As such, claims 1-8 and 12-15 are not patent eligible.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
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.
Claims 1-2 and 15 are rejected under 35 U.S.C. 103 as being unpatentable over Iyer et al. (Journal of American Medical Information Association Vol 21 Issue 2 Pgs. 353-362, Published Oct 24, 2013) further in view of Iyer et al. (CS 341 Report Pg 1-10 Published 2012), referred to Iyer CS et al. from now on. The limitations of the instant claim are italicized below.
With respect to claim 1, Iyer et al. teaches a method to identify significant drug-drug-event associations from electronic health records (EHRs) (Pg 353 left col para 2, A method of checking drug interaction). Iyer et al. utilized the Stanford Translational Research Integrated Database Environment (STRIDE) dataset that comprised of 9 million unstructured clinical notes that included both inpatient and outpatient notes (Pg 354 left col para 2, obtaining a plurality of medical records). The dataset also included the drugs that patients were taking along with adverse events that patients experienced (Pg 354 left col para 6, wherein at least one of the plurality of medical records indicates whether a patient taking a first drug combination has a hospitalization event). Iyer et al. then goes on to extract the drug combination and events from the medical record through annotation of the EHRs (Pg 354 right col para 2). Examples of the drug-drug-event combinations can be seen in Table 2 where drug combinations like Potassium Chloride and Lisinopril, Potassium Chloride and Spironolactone, and Lisinopril and Glipizide can be seen with their associated adverse event and count occurrences (Pg 360 left col Table 2, generating a drug combination set according to the plurality of medical records, wherein the drug combination set comprises the first drug combination, a second drug combination, and a third drug combination, wherein the first drug combination and the second drug combination both comprise a first drug, and the first drug combination and the third drug combination both comprise a second drug). Next, Iyer et al. goes on to compute association scores for a drug-drug-event tuple using standard methods to measure the disproportionality of the adverse event between the group exposed to the drugs and the comparison groups who has taken one drug or no drug (Pg 354 right col para 4). Iyer et al. does this by constructing a 2x2 contingency table (figure 2 on Pg 355) and then calculating the odds ratios (OR) for each drug combination from the contingency table (Pg 354 right col para 6, generating a first odds ratio between the first drug combination and the hospitalization event, a second odds ratio between the second drug combination and the hospitalization event, and a third odds ratio between the third drug combination and the hospitalization even according to the plurality of medical records). Iyer et al. also cleans up the drug combinations by removing tuples for which the event is an indication for either drug individually (Pg 356 left col para 2) and interactions that had less than 100 patients (353 left col para 5). Finally, Iyer et al. used a three-step threshold strategy to find the true representative interactions with an unadjusted odds ratio (UOR0.25) threshold of 4.7, a EB05 threshold of 1.5, and an adjusted odds ratio (AOR0.25) threshold of 1.1 (Pg 357 left col para 4, outputting the first drug combination in response to the first odds ratio being greater than a first threshold, a sum of the first fraction and the second fraction being greater than a second threshold, and a quotient of the first fraction and the second fraction being less than a third threshold).
Regarding claim 15, Iyer et al. teaches a method to identify significant drug-drug-event associations from electronic health records (EHRs) carried out on a computer which inherently comprises of a transceiver and processor (Pg 353 left col para 2, An electronic device of checking drug interaction, comprising a transceiver and a processor). Iyer et al. implies that the computer is able to receive datasets like the Stanford Translational Research Integrated Database Environment (STRIDE) dataset that comprised of 9 million unstructured clinical notes that included both inpatient and outpatient notes (Pg 354 left col para 2, obtain, through the transceiver, a plurality of medical records). The dataset also included the drugs that patients were taking along with adverse events that patients experienced (Pg 354 left col para 6, wherein at least one of the plurality of medical records indicates whether a patient taking a first drug combination has a hospitalization event). Iyer et al. then goes on to extract the drug combination and events from the medical record through annotation of the EHRs (Pg 354 right col para 2). Examples of the drug-drug-event combinations can be seen in Table 2 where drug combinations like Potassium Chloride and Lisinopril, Potassium Chloride and Spironolactone, and Lisinopril and Glipizide can be seen with their associated adverse event and count occurrences (Pg 360 left col Table 2, generating a drug combination set according to the plurality of medical records, wherein the drug combination set comprises the first drug combination, a second drug combination, and a third drug combination, wherein the first drug combination and the second drug combination both comprise a first drug, and the first drug combination and the third drug combination both comprise a second drug). Next, Iyer et al. goes on to compute association scores for a drug-drug-event tuple using standard methods to measure the disproportionality of the adverse event between the group exposed to the drugs and the comparison groups who has taken one drug or no drug (Pg 354 right col para 4). Iyer et al. does this by constructing a 2x2 contingency table (figure 2 on Pg 355) and then calculating the odds ratios (OR) for each drug combination from the contingency table (Pg 354 right col para 6, generating a first odds ratio between the first drug combination and the hospitalization event, a second odds ratio between the second drug combination and the hospitalization event, and a third odds ratio between the third drug combination and the hospitalization even according to the plurality of medical records). Iyer et al. also cleans up the drug combinations by removing tuples for which the event is an indication for either drug individually (Pg 356 left col para 2) and interactions that had less than 100 patients (353 left col para 5). Finally, Iyer et al. used a three-step threshold strategy to find the true representative interactions with an unadjusted odds ratio (UOR0.25) threshold of 4.7, a EB05 threshold of 1.5, and an adjusted odds ratio (AOR0.25) threshold of 1.1 (Pg 357 left col para 4, outputting the first drug combination in response to the first odds ratio being greater than a first threshold, a sum of the first fraction and the second fraction being greater than a second threshold, and a quotient of the first fraction and the second fraction being less than a third threshold).
However, Iyer et al. fails to teach the generation of a first fraction corresponding to the first drug and a second fraction corresponding to the second drug in claim 1. However, this concept of conditional probability was well known in the art at the time of the effective filing date of the invention and has been successfully demonstrated by Iyer CS et al.
Concerning claim 1, Iyer CS et al. also proposes a method to mine EHRs for associations between combinations of drugs and adverse events. Iyer CS et al. does this by using three 2x2 contingency tables to calculate the odds ratios and p-value for the drug combinations (Pg 4 3.2 Map Reduce Method para 2). The first table and its odds ratios measures counts for both drugs vs. the event against not taking both drugs vs. event, which is representative of the first odds ratio, second odds ratio, and third odds ratio (Pg 4 3.2 Map Reduce Method para 2). While the second table measures counts for taking one drug (drug 1) and not the other (drug 2) vs. the event against not taking both drugs vs. the event, which is representative of the first fraction corresponding to the first drug (Pg 4 3.2 Map Reduce Method para 2, generating a first fraction corresponding to the first drug, according to the second odds ratio, wherein the first fraction is negatively correlated with the second odds ratio). Lastly, the third table measure vice versa, the counts for taking drug 2 and not drug 1 vs. the event against not taking both drugs vs. the event, which is representative of the second fraction corresponding to the second drug (Pg 4 3.2 Map Reduce Method para 2, generating a second fraction corresponding to the second drug according to the third odds ratio, wherein the second fraction is negatively correlated with the third odds ratio and the first fraction is greater than or equal to the second fraction). Iyer CS et al. then gives examples of the fractions/percentages of the drug combinations associated with hypoglycemia and what the drugs are usually prescribed for (Pg 6, 4.2.1.2 Validation via the Facts and Comparisons database).
Regarding claim 2, Iyer CS et al. discloses the generation of first and second fraction correlating to the first and second drug respectively in a drug combination as taught above. Specifically, Iyer CS et al. teaches the generation of fractions that pertain to just the presence of one drug to the adverse event, to ensure that no one single drug was the main cause of the adverse event, and that it was truly the combination of the two drugs that contributed to the adverse event (Pg 4 3.2 Map Reduce Method para 2). Iyer CS et al. teaches marking drug combinations as significant or not significant based on threshold parameters as seen in Fig 4.1. (Pg 5 Fig 4.1, marking the second drug combination in response to the second odds ratio being greater than a risk threshold). Iyer CS et al. then goes on to calculate a fraction by counting all the drug combinations that contain the first drug that led to an adverse event divide by all the drug combinations that contained the first drug (Pg 6 Probability Equations, P[Hypo | Glyburide only] = 6.6%, equal to the number of drug combinations comprising the first drug but not comprising the second drug and being marked in the drug combination set, divided by the number of drug combinations comprising the first drug but not comprising the second drug in the drug combination set). Although Iyer CS et al. does not explicitly teach the generation of the first fraction by subtracting a third fraction from the number 1, it would have been obvious to try to ensure all the probabilities added up to 1 (calculating the first fraction according to the third fraction, wherein a sum of the first fraction and third fraction is equal to 1).
With respect to claim 15, Iyer CS et al. also proposes a method to mine EHRs for associations between combinations of drugs and adverse events. Iyer CS et al. does this by using three 2x2 contingency tables to calculate the odds ratios and p-value for the drug combinations (Pg 4 3.2 Map Reduce Method para 2). The first table and its odds ratios measures counts for both drugs vs. the event against not taking both drugs vs. event, which is representative of the first odds ratio, second odds ratio, and third odds ratio (Pg 4 3.2 Map Reduce Method para 2). While the second table measures counts for taking one drug (drug 1) and not the other (drug 2) vs. the event against not taking both drugs vs. the event, which is representative of the first fraction corresponding to the first drug (Pg 4 3.2 Map Reduce Method para 2, generating a first fraction corresponding to the first drug, according to the second odds ratio, wherein the first fraction is negatively correlated with the second odds ratio). Lastly, the third table measure vice versa, the counts for taking drug 2 and not drug 1 vs. the event against not taking both drugs vs. the event, which is representative of the second fraction corresponding to the second drug (Pg 4 3.2 Map Reduce Method para 2, generating a second fraction corresponding to the second drug according to the third odds ratio, wherein the second fraction is negatively correlated with the third odds ratio and the first fraction is greater than or equal to the second fraction). Iyer CS et al. then gives examples of the fractions/percentages of the drug combinations associated with hypoglycemia and what the drugs are usually prescribed for (Pg 6, 4.2.1.2 Validation via the Facts and Comparisons database).
It would have been prima facie obvious to one of ordinary skill in the art at the time of the effective filing date of the invention to implement the counts and conditional fractions of Iyer CS et al. with the method of Iyer et al. to fine tune the predictions of the drug combinations and screen for drugs combinations that were only risky when combined together. The concept of conditional probabilities and fraction were well known before the effective filing date of the instant application, and one of ordinary skill in the art would have been motivated to incorporate the conditional probabilities of Iyer CS et al. to further ascertain that the drug combination leading to the adverse event were accurate, and that no single drug was the cause for the adverse event. In addition, one of skill in the art before the effective filing date of the claimed invention would have a reasonable expectation of success at incorporating the fractions of Iyer CS et al. into the method of Iyer et al. as they were processing the same data and had the same contingency tables. It would have been equivalent to further calculations using well-established mathematical equations of conditional probabilities.
Allowable Subject Matter
Claims 3-8 and 12-14 are currently free from the prior art.
The following is an examiner’s statement over the prior art:
Claims 3-8 and 12-14 pertain to the generation of drug combinations through topic modeling in the form of a latent Dirichlet allocation (LDA) model. In the most basic sense, topic modeling works to categorize singular words in a document into topics of similar themes. LDA modeling is a specific type of topic modeling that defines each topic by a probability distribution over words, where words that often appear together are likely in the same topic. The instant application generates K topic vectors through a LDA model, where each topic comprises a probability distribution of all drug combinations (claim 3). Then the instant application screens the drug combinations by topic vectors, picking a plurality of drug combinations with a range of max probabilities in each vector to generate an important drug combination set for each vector (claim 3). Then, the important drug combination sets are screened out for duplicate drug combinations that appear in different vectors to create unique drug combination sets (claim 4). And the process is repeatedly executed before a stable drug combination set is picked out based on certain drug combinations appearing more than a certain amount of times (claim 5). Lastly, the instant application makes sure there are enough medical records displaying certain drug combination in the topic vector (claim 7) and picking the drug combinations that have a certain ratio of medical records to create the final drug combination set (claim 6). The instant application then goes backwards to specify a method of picking the optimal number for K, which represents the number of topics for the LDA model, from a first index number and a second index number (claim 8). The instant application implies that after an initial run of the LDA model with an initial K, the medical records should then be split into K groups (claim 12). Next, the inter-group distances between the groups are calculated and summed to get a first statistical value (claim 13). Then, the intra-group distances between groups are calculated and summed to get a second statistical value. Finally, the ratio of the first statistical value to the second statistical value would be the first index (claim 12).
The closest prior art that reads upon claims 3-8, and 12-14 are Bisgin et al. (BMC Bioinformatics Vol 12 Article S11, Published Oct 18 2011), Huang et al. (IEEE Access Vol 7 Pgs. 125280-125299, Published 2019), and Park et al. (Journal of Biomedical Informatics Vol 75 Pgs. 35-47, Published Nov 2017).
With respect to claim 3, Bisgin et al. teaches a method to group drugs with similar safety concerns and/or therapeutic use together based on topic modeling of drug labels (Pg 1 Method para). Unfortunately, Bisgin et al. did not utilize medical health records as specified in claim 3, but drug labels. Specifically, Bisgin et al. utilized LDA modeling to calculate the probability distribution of drugs over 100 topics as shown in Figure 1 (Pg 4 left col para 3, Pg 3 Figure 1). However, due to the way that words were defined in the model, Bisgin et al. only calculated probability distributions for a single drug, rather than a drug combination. Afterwards, Bisgin et al. assigns each drug to a single topic with the maximal topic conditional probability, to form a set of drugs correlating to each topic (Pg 3 Figure 1). Bisgin et al. did not pick a plurality of drugs with a range of max probabilities for each topic as specified in claim 3.
Concerning claim 4, because Bisgin et al. did not pick a plurality of drugs with a range of max probabilities for each topic and rather just assigned each drug to a single topic based on the max probability, there are no duplicates of each drug across topics.
Regarding claim 5, Bisgin et al. does not teach the repeated execution of the screening process to generate a stable set of drugs.
With regards to claim 6-7, because Bisgin et al. did not use medical records in their analysis, Bisgin et al. does not teach the use of generating a plurality of medical record vectors and screening the vectors to generate the final drug set.
Concerning claim 8 and 12-14, Bisgin et al. did not further specify how the topic number was determined.
With respect to claim 3, Huang et al. teaches a method to mine hidden medication patterns in electronic medical record (EMR) text (Pg 125280 Abstract). Huang et al. utilized LDA modeling, medication names as words, and EMR text for a patient as documents, to find latent medication patterns as topics (Pg 125285 right col 4) LDA Mining). However, similar to Bisgin et al., Huang et al. only extracts singular medications, not combinations of medications.
Regarding claims 4-7, Huang et al. does not explicitly teach selecting medications to form medication sets based on their maximal probabilities in a topic.
Concerning claim 8 and 12-14, Huang et al. does teach picking the optimal number of patterns/topics through repetition and representation to ensure that the patterns are distinct and there are no overlaps in drugs across patterns (Pg 125289 left col 2) LDA-Based Medication Clustering). However, Huang et al. does not teach calculating inter-group and intra-group distances for comparison.
With respect to claim 3 Park et al. teaches a disease-medicine pattern model (DMPM) based on LDA modeling to extract patterns from prescriptions and labeled diseases (Pg 38 right col para 5).
Regarding claims 4-7, Park et al. does not explicitly teach selecting medications to form medication sets based on their maximal probabilities in a topic.
Concerning claim 8 and 12-14, Park et al. teaches an expectation-maximization (EM) algorithm to determine the K topic number (Pg 39 right col para 2). Park et al. specifically used the predictive log-likelihood to predict K=15 (Pg 41 left col para 3). Unfortunately, Park et al. also did not use inter-group and intra-group distances to evaluate K.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
Bisgin, H., Liu, Z., Fang, H. et al. Mining FDA drug labels using an unsupervised learning technique - topic modeling. BMC Bioinformatics 12 (Suppl 10), S11 (2011).
H. Huang et al., "Discovering Medication Patterns for High-Complexity Drug-Using Diseases Through Electronic Medical Records," in IEEE Access, vol. 7, pp. 125280-125299, 2019
S. Park et al., Journal of Biomedical Informatics 75 (2017) Pgs. 35–47
LePendu P, Iyer SV, Bauer-Mehren A, et al. Pharmacovigilance using clinical notes. Clin Pharmacol Ther. 2013;93(6):547-555.
Gan J, Qi Y. Selection of the Optimal Number of Topics for LDA Topic Model-Taking Patent Policy Analysis as an Example. Entropy (Basel). 2021 Oct 3;23(10):1301.
Burke et al. US 2019/0005019 A1 Contextual Pharmacovigilance System
Dey et al. WO2019/171187 A1, Published Sep 12 2019, Adverse drug reaction analysis
Any inquiry concerning this communication or earlier communications from the examiner should be directed to WENYU YANG whose telephone number is (571) 272-0035. The examiner can normally be reached 8:00am - 5:00 pm.
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, Olivia Wise can be reached at (571) 272-2249. 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.
/W.Y./Examiner, Art Unit 1685 /OLIVIA M. WISE/Supervisory Patent Examiner, Art Unit 1685
Prosecution Insights