# US 2020/0286599 A1 · Experience Engine-Method and Apparatus of Learning from Similar Patients

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

- Number: US 2020/0286599 A1
- Document ID: US20200286599A1
- Status: Published application
- Kind: A1
- Application number: 15/758,988
- Filed: 2016-09-09
- Published: 2020-09-10
- Earliest claimed: 2015-09-11
- Assignees: Navya Network Inc.
- Applicants: Navya Network Inc.
- Inventors: Naresh Ramarajan; Gitika Srivastava
- Web page: https://patentgrep.com/patent/US20200286599A1

## CPC

- G16H 50/70: for mining of medical data, e.g. analysing previous cases of other patients
- G06F 16/20: of structured data, e.g. relational data
- G06F 16/22: Indexing; Data structures therefor; Storage structures
- G16H 10/60: for patient-specific data, e.g. for electronic patient records
- G16H 20/00: ICT specially adapted for therapies or health-improving plans, e.g. for handling prescriptions, for steering therapy or for monitoring patient compliance
- G16H 50/20: for computer-aided diagnosis, e.g. based on medical expert systems
- G16H 70/20: relating to practices or guidelines
- Y02A 90/10: Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
- 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)
- Y02A: TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE

## Abstract

The present solution covers identifying a recommended treatment for a patient based on records of similar patients, wherein the similarities are non-obvious and non-linear. The solution generates a similarity map that minimizes the variance of elements records among a curated group of patients, and this similarity map is used to find the patients who are most similar to an untreated patient.

## Claims

### Claim 1

A method of selecting a treatment for a patient based on a group of patients falling within a predetermined distance of a similarity map of data elements, the method comprising:
storing, in a database, records for patients, each record including (i) a first plurality of data elements regarding at least one medical condition and treatment of its respective patient and (ii) a second plurality of data elements, wherein each data element in the second plurality corresponds to an additional medical condition of the respective patient that has been inferred from data elements in the first plurality of data elements;
receiving, by a server system, (i) a context for treating an untreated patient and (ii) a record of the untreated patient;
identifying, by the server system, a group of patients, wherein each patient in the group received treatment in the same context as the untreated patient;
generating, by the server system, a similarity map that minimizes the variance of data elements among the group of identified patients;
determining, by the server system, similarity distances between the untreated patient and each patient in the group by applying the similarity map to data elements from the record of the untreated patient against data elements for each patient in the group;
selecting, by the server system, patients from the group that are within a predetermined similarity distance of the untreated patient; and
identifying, by the server system, a treatment from the treatments received by the selected patients to recommend to the untreated patient.

### Claim 2 (depends on 1)

The method of claim 1, wherein storing the records for patients comprises:
storing the first plurality of data elements for the record of the patient, the first plurality of data elements including conditions about the patient, observations about the patient, test results about the patient, clinical findings of the patient, obtained imaging about the patient, pathological findings of the patient, prior diagnosis, planned diagnostic interventions, prior therapeutic interventions, planned therapeutic interventions, morbidities, side effects of administered therapies, name of an expert physician, name of a hospital, and any combination thereof.

### Claim 3 (depends on 1)

The method of claim 1, wherein storing the records for patients comprises:
storing, for the record of the patient, the second plurality of data elements as (i) logical combinations of terms using Boolean operators, (ii) nested hierarchies, or (iii) directed acyclical graphs.

### Claim 4 (depends on 1)

The method of claim 1, wherein storing the second plurality of data elements for records for patients comprises:
retrieving, by the server system, inference definitions from storage; and
applying, by the server system, the inference definitions to the first plurality of data elements for a record to derive interconnected terms that are stored as the second plurality of data elements.

### Claim 5 (depends on 1)

The method of claim 1, wherein receiving the context for treating an untreated patient comprises:
receiving a clinical diagnosis, an objective for treatment, a procedural timing, or any combination thereof

### Claim 6 (depends on 5)

The method of claim 5, wherein the objective for treatment is curative, control, or palliative.

### Claim 7 (depends on 5)

The method of claim 5, wherein the procedural timing for treatment is primary, adjuvant, or neoadjuvant.

### Claim 8 (depends on 1)

The method of claim 1, wherein identifying the group of patients comprises:
identifying, by the server system, a group of patients, wherein each patient in the group has a same value for a data element in a respective record of the patient.

### Claim 9 (depends on 8)

The method of claim 8, wherein each patient in the group received a treatment recommendation from the same expert.

### Claim 10 (depends on 8)

The method of claim 8, wherein identifying the group of patients comprises:
constraining, by the server system, the group of patients to patients with values of additional data elements in common until a number of patients in the group reaches a minimum size.

### Claim 11 (depends on 1)

The method of claim 1, wherein generating the similarity map comprises:
identifying, by the server system, a set of data elements to use in the similarity map by performing relevant component analysis on data elements for the group of patients to learn non-linear similarities between data elements.

### Claim 12 (depends on 1)

The method of claim 1, wherein generating the similarity map comprises:
generating a locus of the group of patients, the locus representing average values across data elements for the group; and
summing a function of distances between the locus and each patient in the group of patients.

### Claim 13 (depends on 1)

The method of claim 1, wherein identifying the treatment from the treatments received by the selected patients to recommend to the untreated patient comprises:
selecting the treatment most common among the selected patients.

### Claim 14 (depends on 1)

The method of claim 1, wherein identifying the treatment from the treatments received by the selected patients to recommend to the untreated patient comprises:
selecting the treatment at random among the treatments received by the selected patients.
