US 11,894,117 B1 · Discovering context-specific complexity and utilization sequences

Abstract

Systems, methods, and computer-readable media are provided for patient case and care complexity characterization, and detecting matches of an individual patient's record with collections of other patients' records, based on serial, longitudinal patterns, for facilitating efficient health services utilization, implementing programs to reduce complexity, preventive medicine, and risk management in health care. In an embodiment, time series are formed by electronically representing information pertaining to successive longitudinal episodes of health services utilization and the circumstances in which the episodes were incurred; calculating time-series K-nearest-neighbor clusters and distances for each combination; determining the cluster to which a given candidate patient complexity record is nearest, and prescribing one or more interventions specific to hazards that are characteristic of trajectories that are members of that cluster, or that are deemed to be relevant to mitigating those hazards, thereby preventing the adverse outcomes and subsequent excess utilization that are prevalent in that cluster.

Claims31

Dependent claims

A method, comprising:

for a candidate patient, determining, by a computer processor, a set of health records stored on one or more data stores distributed across multiple Physical locations for a population of reference patients having conditions, multiple episodes, or encounters over a period of time similar to the candidate patient, wherein each reference patient health record forms a reference set of clinical event sequences;
receiving, by the computer processor, candidate patient information associated with the candidate patient;
creating a candidate patient data frame based on the received candidate patient information;
and
determining, by the computer processor, an inter-event pattern present in the set of health records of the reference patients based on:
a comparison of the candidate patient data frame and one or more of the reference sets of clinical event sequences,
a matching set of clinical event sequences, and
a set of timeseries trajectory clusters determined by:
(a) performing, on at least one of the one or more data stores, a database scan that uses a sparse-matrix approach to decompose an original search lattice into sub-lattices that are processed independently in distributed memory associated with multiple parallel processors;
(b) computing and applying a multi-cluster model;
and
(c) extracting one or more cluster assignments,
wherein each timeseries trajectory cluster has a centroid.

10 more elements

One or more non-transitory media having instructions that when executed by one or more processors cause a plurality of operations, the operations comprising:

for a candidate patient, determining, by a computer processor, a set of health records stored on one or more data stores distributed across multiple physical locations for a population of reference patients having conditions, multiple episodes, or encounters over a period of time similar to the candidate patient, wherein each reference patient health record forms a reference set of clinical event sequences;
receiving, by the computer processor, candidate patient information associated with the candidate patient;
creating a candidate patient data frame based on the received candidate patient information;
and
determining, by the computer processor, an inter-event pattern present in the set of health records of the reference patients based on:
a comparison of the candidate patient data frame and one or more of the reference sets of clinical event sequences,
a matching set of clinical event sequences, and
a set of timeseries trajectory clusters determined by:
(a) performing, on at least one of the one or more data stores, a database scan that uses a sparse-matrix approach to decompose an original search lattice into sub-lattices that are processed independently in distributed memory associated with multiple parallel processors;
(b) computing and applying a multi-cluster model;
and
(c) extracting one or more cluster assignments,
wherein each timeseries trajectory cluster has a centroid.

10 more elements

Dependencies are inferred from the claim text.

Discovering context-specific complexity and utilization sequences

Life of this patent

Discovering context-specific complexity and utilization sequences

Life of this patent

Abstract

Claims31