Progression analytics system

Patent metadata

Field	Value
Publication number	US-2016004840-A1
Application number	US-201514853377-A
Country	US
Kind code	A1
Filing date	Sep 14, 2015
Priority date	Mar 15, 2013
Publication date	Jan 7, 2016
Grant date	—

Abstract

A method of identifying insights related to the occurrence of an adverse health outcome of interest, comprises extracting electronic clinical data associated with historical healthcare encounters. The method also comprises defining patient groups based upon similar data patterns present in the extracted electronic clinical data wherein the patient groups have varying likelihood for the adverse health outcome. Still further, the method comprises deriving hypothesized etiological explanations for why one or more patient groups have higher likelihood when compared to other patient groups. Optionally, the method comprises identifying clinical interventions that are intended to reduce the likelihood of the adverse outcome for certain patient groups.

First claim

What is claimed is:

1 . A computer-implemented method of evaluating outcomes, comprising:
identifying a patient care-related outcome of interest; extracting electronic clinical data associated with historical healthcare encounters for a plurality of patients, by:
including in the plurality of patients, a first subset of patients that experienced the outcome of interest; and
including in the plurality of patients, a second subset of patients that did not experience the outcome of interest;
determining patient groups among the plurality of patients for which electronic clinical data is extracted, with the aid of a computer processor that executes a program, by:
defining each patient group by grouping together those patients having a similar data pattern present in their corresponding extracted electronic clinical data; and
selecting the data patterns such that the defined patient groups differentiate from one another in terms of a likelihood of the outcome of interest, consequences associated with the outcome of interest or both; and
deriving hypothesized etiological explanations for why one or more patient groups have different likelihoods of the outcome of interest, consequences associated with the outcome of interest or both, when compared to other defined patient groups, by comparing defined patient groups and identifying different likelihoods of the outcome of interest.

2 . The method of claim 1 , wherein identifying a patient care-related outcome of interest comprises selecting an adverse outcome of interest, and further comprising:
selecting a clinical intervention for a select patient group wherein the identified clinical intervention is directed to decrease the likelihood of the adverse outcome of interest or decrease the consequences of the adverse outcome of interest, or both, for the select patient group.

3 . The method of claim 2 , wherein identifying a clinical intervention comprises selecting a clinical intervention that is directed to prevent patients from entering higher-likelihood patient groups, prevent patients from entering higher-consequence patient groups, lower the likelihood of the adverse outcome for patients in higher-likelihood patient groups, lower the consequences of the adverse outcome for patients in higher-consequence patient groups, or any combination of the foregoing.

4 . The method of claim 1 , wherein identifying a patient care-related outcome of interest comprises selecting a favorable outcome of interest, and further comprising:
selecting a clinical intervention for a select patient group wherein the identified clinical intervention is directed to increase the likelihood of the favorable outcome of interest or increase the consequences of the favorable outcome of interest, or both, for the select patient group.

5 . The method of claim 4 , wherein selecting a clinical intervention comprises selecting a clinical intervention that is directed to assist patients in entering higher-likelihood patient groups, assist patients in entering higher-consequence patient groups, increase the likelihood of the favorable outcome for patients in lower-likelihood patient groups, increase the consequences of the favorable outcome for patients in lower-consequence patient groups, or any combination of the foregoing.

6 . The method according to claim 1 , wherein defining patient groups comprises creating at least one patient group as including at least one patient from the first subset of patients and at least one patient from the second subset of patients.

7 . The method according to claim 1 , wherein extracting electronic clinical data comprises extracting likelihood variables or consequence variables or both, wherein the likelihood variables define variables associated with a patient's likelihood of having the outcome of interest and consequence variables define variables associated with a patient's consequences associated with the outcome of interest.

8 . The method of claim 7 , wherein:
extracting likelihood variables comprises generating the likelihood variables as a result of reconciliation of likelihood factors identified by outcome-specific etiological models with available patient data; and extracting consequence variables comprises generating the consequence variables as a result of reconciliation of consequence factors identified by outcome-specific etiological models with available patient data.

9 . The method according to claim 7 , wherein grouping together those patients having a similar data pattern present in their corresponding extracted electronic clinical data comprises defining similar data patterns based upon the data values of a subset of the likelihood variables, consequence variables or both, such that the patient groups are defined in terms of variables and not in terms of whether or not the patient has experienced the outcome of interest.

10 . The method according to claim 9 further comprising:
converting the subset of the likelihood variables, consequence variables or both, into discrete measures having a fixed number of value options;
wherein grouping together those patients having a similar data pattern comprise grouping together those patients having the same data values associated with the discrete measures.

11 . The method according to claim 1 , wherein grouping together those patients having a similar data pattern present in their corresponding extracted electronic clinical data comprises grouping together those patients having a data pattern that includes both non-temporal and temporal data patterns.

12 . The method according to claim 1 , wherein grouping together those patients having a similar data pattern present in their corresponding extracted electronic clinical data comprises identifying a common trajectory associated with at least one defined patient group where the trajectory represents a data pattern across a time history of changes in the physiological state of patients, occurrences of events that patients experience, or a combination of states and events over time.

13 . The method according to claim 1 , wherein grouping together those patients having a similar data pattern present in their corresponding extracted electronic clinical data comprises defining a data pattern by defining values of one or more static variables that do not change during the course of a hospital encounter and/or defining a pattern across a time history of changes in the physiological state of patients, occurrences of events that the patients experience, or both.

14 . The method according to claim 1 , wherein defining patient groups comprises developing an outcome likelihood scoring algorithm that characterizes the likelihood of the outcome of interest as a function of likelihood variables derived from extracted electronic clinical data and defining the patient groups based on data patterns in the likelihood variables employed in the scoring algorithm.

15 . The method of claim 14 , further comprising configuring the outcome likelihood scoring algorithm to define the likelihood of the outcome of interest in terms of baseline and dynamic likelihoods.

16 . The method according to claim 1 , wherein defining patient groups comprises developing an outcome consequence scoring algorithm that characterizes the consequences associated with the outcome of interest as a function of consequence variables derived from extracted electronic clinical data and defining the patient groups based on data patterns in the consequence variables employed in the scoring algorithm.

17 . The method according to claim 1 , wherei

Assignees

• Battelle Memorial Institute

Inventors

• Rust Steven W
• Haber Daniel

Classifications

• G06F19/3443 Primary

  Physics · mapped topic

• G16Z99/00

  Subject matter not provided for in other main groups of this subclass · CPC title

• G06Q50/22

  Social work or social welfare, e.g. community support activities or counselling services · CPC title

• G16H50/70 Primary

  for mining of medical data, e.g. analysing previous cases of other patients · CPC title