# US 10,769,242 B2 · System and method for data mining very large drugs and clinical effects databases

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 10,769,242 B2
- Document ID: US10769242B2
- Status: Granted
- Kind: B2
- Application number: 15/541,435
- Filed: 2016-01-05
- Published: 2018-01-04
- Granted: 2020-09-08
- Earliest claimed: 2015-01-05
- Assignees: Children's Hospital Medical Center
- Applicants: Cincinnati Children's Hospital Medical Center
- Inventors: Bruce Aronow; Mayur Sarangdhar; Scott Carl Tabar; Anil Goud Jegga
- Web page: https://patentgrep.com/patent/US10769242B2

## CPC

- G16H 10/60: for patient-specific data, e.g. for electronic patient records
- G06F 16/2465: {Query processing support for facilitating data mining operations in structured databases}
- G16H 50/70: for mining of medical data, e.g. analysing previous cases of other patients
- G16H 70/40: relating to drugs, e.g. their side effects or intended usage
- G16H: HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

A current system allows data-driven hypothesis generation to identify therapeutic candidates for a disease phenotype treatment by identifying drugs and clinical indications associated with lower occurrences of disease-associated phenotype(s) by a drug/drug class. A current system may include a pharmaceutical hierarchical ontology; a phenotype hierarchical ontology; a record database comprising clinical event records; a database mining engine; and a mapping engine. The database mining engine may iteratively progress through a portion of the pharmacological hierarchical ontology and phenotype hierarchical ontology to iteratively select pairs of cohort entries from each ontology; and for each pair of cohort entries, query the clinical record database for matching records. The mapping engine may map each pair of cohort entries into a matrix comprising a drug-event cell for each pair and apply a value thereto representing the number of database records matching items returned by the database mining engine for the corresponding cohort entries.

## Claims

### Claim 1

A system for transforming cumulative clinical adverse events data into a visual representation for improved understanding of differential clinical outcomes, comprising:
a pharmacological hierarchical ontology of drug identifications;
a phenotype hierarchical ontology of clinical adverse effects and clinical indications;
a clinical record database comprising records containing clinical event information, each record including (a) demographic information for a patient experiencing a clinical event, (b) clinical indication information for the patient experiencing the clinical event, (c) drug identification information associated with the clinical event, (d) adverse reaction information associated with the clinical event, and (e) patient outcome information;
a database mining engine configured to,
iteratively progress through at least a selected portion of each of the pharmacological hierarchical ontology and the phenotype hierarchical ontology to iteratively select respective pairs of cohort entries from each of the pharmacological hierarchical ontology and the phenotype hierarchical ontology, and
for each pair of cohort entries, query the clinical record database for the database records matching the items in the pair of cohort entries; and
a mapping engine configured to,
map each pair of cohort entries into a data matrix comprising a cell for each pair of cohort entries,
apply a value associated with a drug-drug-event interaction to each cell which is, at least in part a function of a number of expected occurrences of the drug-drug-event interaction and a number of actual occurrences of the drug-drug-event interaction returned by the mapping engine, and
provide a visual representation of the data matrix including each mapped pair of cohort entries and each applied value.

### Claim 2 (depends on 1)

The system of claim 1, wherein the mapping engine is configured to apply a normalized per “x” number of patients value to each cell corresponding to a number of drug-event records returned by the database mining engine.

### Claim 3 (depends on 1)

The system of claim 1, wherein the mapping engine is configured to apply a relative risk value to each cell, wherein the relative risk value pertains, at least in part, to the relative risk of the adverse effects and clinical indications entry from the corresponding pair of cohort entries being associated with the drug identification from the corresponding pair of cohort entries.

### Claim 4 (depends on 3)

The system of claim 3, wherein the mapping engine is configured to calculate the relative risk value ‘RR’ to each cell based upon the equation
RR
=
(
a
a
+
b
)
(
c
c
+
d
)
where ‘a’ corresponds to the number of records in the database found by the database mining engine in which the drug identification from the corresponding pair of cohort entries is present and the adverse effects and clinical indications entry from the corresponding pair of cohort entries is also present;
where ‘b’ corresponds to the number of records in the database found by the database mining engine in which the drug identification from the corresponding pair of cohort entries is present and the adverse effects and clinical indications entry from the corresponding pair of cohort entries is not present;
where ‘c’ corresponds to the number of records in the database found by the database mining engine in which the drug identification from the corresponding pair of cohort entries is not present and the adverse effects and clinical indications entry from the corresponding pair of cohort entries is present; and
where ‘d’ corresponds to the number of records in the database found by the database mining engine in which the drug identification from the corresponding pair of cohort entries is not present and the adverse effects and clinical indications entry from the corresponding pair of cohort entries is also not present.

### Claim 5 (depends on 4)

The system of claim 4, wherein providing the visual representation includes applying the following visual indications to the cell based upon the calculated relative risk value ‘RR’:
a first visual indication corresponding to a minimal risk if RR is calculated to be less than or equal to 1;
a second visual indication corresponding to a moderate risk if RR is calculated to be greater than 1 and less than or equal to 2; and
a third visual indication corresponding to a high risk if RR is calculated to be greater than 2.

### Claim 6 (depends on 1)

The system of claim 1, wherein the mapping engine is configured to apply a normalized risk value to each cell, wherein the normalized risk value pertains, at least in part, to the relative risk of the adverse effects and clinical indications entry from the corresponding pair of cohort entries being associated with the drug identification from the corresponding pair of cohort entries.

### Claim 7 (depends on 6)

The system of claim 6, wherein the mapping engine is configured to calculate the normalized risk value ‘IC’ to each cell based upon the equation,
IC=log 2((AE Count*T)/(Unique Patients*AE Total))
where ‘T’ corresponds to the total number of records in the database;
where ‘AE Count’ corresponds the number of records in the database found by the database mining engine in which the drug identification from the corresponding pair of cohort entries is present and the adverse effects and clinical indications entry from the corresponding pair of cohort entries is also present;
where ‘AE Total’ corresponds the number of records in the database found by the database mining engine in which the adverse effects and clinical indications entry from the corresponding pair of cohort entries is present; and
where Unique Patients corresponds the number of records in the database found by the database mining engine in which the drug identification from the corresponding pair of cohort entries is present.

### Claim 8 (depends on 7)

The system of claim 7, wherein providing the visual representation includes applying the following visual indications to the cell based upon the calculated normalized risk value ‘IC’:
a first visual indication corresponding to drug-events database records being less than expected if IC is calculated to be less than 0;
a second visual indication corresponding to drug-events database records being more than expected if IC is calculated to be greater than 0;
a third visual indication corresponding to drug-events being as expected if IC is calculated to be equal to 0.

### Claim 9 (depends on 1)

The system of claim 1, wherein the configured to apply providing the visual representation includes applying a color to the cell representing the applied value for the cell.

### Claim 10 (depends on 9)

The system of claim 9, wherein
a first color corresponds to risk of the adverse effects and clinical indications entry from the corresponding pair of cohort entries being associated with the drug identification from the corresponding pair of cohort entries being relatively low; and
a second color corresponds to risk of the adverse effects and clinical indications entry from the corresponding pair of cohort entries being associated with the drug identification from the corresponding pair of cohort entries being relatively high.

### Claim 11 (depends on 1)

The system of claim 1, wherein:
the database mining engine is configured to,
iteratively progress through at least the selected portion of each of the pharmacological hierarchical ontology and the phenotype hierarchical ontology to iteratively select respective sets of cohort entries, each set including at least a first and second entry from the pharmacological hierarchical ontology and each set including at least an entry from the phenotype hierarchical ontology, and
for each set of cohort entries, query the clinical record database for the database records matching the items in the set of cohort entries; and
the mapping engine configured to,
map each set of cohort entries into a data matrix comprising a cell for each set of cohort entries, and
apply the drug-drug-event interaction value to each cell which is, at least in part a function of the number of database records matching a number of items returned by the database mining engine for the corresponding cohort entries.

### Claim 12 (depends on 11)

The system of claim 11, wherein the drug-drug-event interaction value applied to each cell is ‘Ω’ and ‘Ω’ is calculated from the following equation,
Ω=log 2((n111+alpha)/(E111+alpha));
where ‘n111’ corresponds to the number of actual occurrences of the drug-drug-event interaction returned by the mapping engine;
where ‘E111’ corresponds to the number of expected occurrences of the drug-drug-event interaction; and
where ‘alpha’ corresponds to a tuning parameter.

### Claim 13 (depends on 12)

The system of claim 12, wherein providing the visual representation includes applying the following visual indications to the cell based upon the calculated drug-drug-event interaction value ‘Ω’:
a first visual indication corresponding to drug-drug-events being less than expected if Ω is calculated to be less than 0;
a second visual indication corresponding to drug-drug-events being more than expected if Ω is calculated to be greater than 0;
a third visual indication corresponding to drug-drug-events being as expected if Ω is calculated to be equal to 0.

### Claim 14 (depends on 1)

The system of claim 1, wherein providing the visual representation includes applying a color to the cell representing the applied value for the cell.

### Claim 15 (depends on 1)

The system of claim 1, wherein the database mining engine is configured to a selected subgroup of at least one of the pharmacological hierarchy and the phenotype hierarchy.

### Claim 16 (depends on 15)

The system of claim 15, wherein the subgroup pertains to criteria for at least one field in the clinical record database.

### Claim 17 (depends on 15)

The system of claim 15, wherein the subgroup pertains to at least one demographic criteria in the clinical record database.

### Claim 18 (depends on 15)

The system of claim 15, wherein the subgroup pertains to at least one clinical information criteria in the clinical record database.

### Claim 19 (depends on 15)

The system of claim 15, wherein the subgroup pertains to at least one drug identification criteria in the clinical record database.

### Claim 20 (depends on 15)

The system of claim 15, wherein the subgroup pertains to at least one adverse reaction criteria in the clinical record database.

### Claim 21 (depends on 15)

The system of claim 15, wherein the subgroup pertains to at least one patient outcome criteria in the clinical record database.

### Claim 22 (depends on 15)

The system of claim 15, wherein the subgroup pertains to two more of at least one demographic criteria, at least one clinical information criteria, at least one drug identification criteria, at least one adverse reaction criteria and at least one patient outcome criteria.

### Claim 23 (depends on 15)

The system of claim 15, wherein the mapping engine is configured to display a comparison of values in the cells associated with the subgroup versus cells associated with a global representation of the database records.

### Claim 24 (depends on 1)

The system of claim 1, wherein the pharmacological hierarchical ontology of drug identifications is ordered under a multitude of parent concepts, and a multitude of generic pharmacological concepts under the parent concepts.

### Claim 25 (depends on 1)

The system of claim 1, wherein the drug identifications in the pharmacological hierarchical ontology include a plurality of drug entries, wherein each drug entry includes a plurality of names taken from a group consisting of: clinical names for the drug, commercial brand names for the drug, molecular names for the drug, foreign names for the drug, spelling variants for the drug, and mis-spellings for the drug.

### Claim 26 (depends on 1)

The system of claim 1, wherein the phenotype hierarchical ontology of clinical adverse effects and clinical indications is ordered under a multitude of hierarchical MedRA concepts.

### Claim 27 (depends on 1)

The system of claim 1, wherein the visual representation is provided in at least one of the following visual formats:
a heat map;
bar chart;
circular plot;
a correlation plot; and
a tag cloud.

### Claim 28 (depends on 1)

The system of claim 1, wherein providing the visual representation includes specifically highlighting those of the plurality of cells that show at least one of under-represented adverse reaction information and over-represented adverse reaction information.

### Claim 29 (depends on 1)

The system of claim 1, wherein providing the visual representation includes producing the plurality of cells in the form of an output that can be displayed on a user's computer.

### Claim 30

A system for mining cumulative clinical adverse events data for improved understanding of differential clinical outcomes, comprising:
a pharmacological hierarchical ontology of drug identifications;
a phenotype hierarchical ontology of clinical adverse effects and clinical indications;
a clinical record database comprising records containing clinical event information, each record including (a) clinical indication information for the patient experiencing the clinical event, (b) drug identification information associated with the clinical event, and (c) adverse reaction information associated with the clinical event;
a database mining engine configured to (i) iteratively progress through at least a selected portion of each of the pharmacological hierarchical ontology and the phenotype hierarchical ontology to iteratively select respective pairs of cohort entries from each of the pharmacological hierarchical ontology and the phenotype hierarchical ontology, and (ii) for each pair of cohort entries, query the clinical record database for the database records matching the items in the pair of cohort entries; and
a mapping engine configured to (i) map each pair of cohort entries into a data matrix comprising a cell for each pair of cohort entries, (ii) apply a first value associated with a drug-drug-event interaction to each cell which is, at least in part a function of a number of expected occurrences of the drug-drug-event interaction and a number of actual occurrences of the drug-drug-event interaction returned by the mapping engine, and (iii) provide a visual representation of the data matrix including each mapped pair of cohort entries and each applied first value.

### Claim 31 (depends on 30)

The system of claim 30, wherein the mapping engine is configured to apply a second value to each cell which is, at least in part a function of the number of drug-event database records matching a number of items returned by the database mining engine for the corresponding cohort entries.

### Claim 32 (depends on 30)

The system of claim 30, wherein the clinical record database further includes (d) and (e) patient outcome information.