US-12626794-B2 - System, server and method for predicting adverse events

US12626794B2US 12626794 B2US12626794 B2US 12626794B2US-12626794-B2

Abstract

A system includes a data collection engine, a plurality of items including radio-frequency identification chips, a plurality of third party data and insight sources, a plurality of interfaces, client devices, a server and method thereof for preventing suicide. The server includes trained machine learning models, business logic and attributes of a plurality of patient events. The data collection engine sends attributes of new patient events to the server. The server can predict an adverse event risk of the new patient events based upon the attributes of the new patient events utilizing the trained machine learning models.

Inventors

David Laborde

Assignees

BRAIN TRUST INNOVATIONS I, LLC

Dates

Publication Date: 20260512
Application Date: 20240326

Claims (9)

1 . A method for predicting a medical outcome associated with a new event, the method comprising: storing a plurality of past events, each of the plurality of past events including a plurality of patient attributes and a quantifiable medical outcome; training a neural network model (NNM) to generate the trained model, wherein the training of the NNM includes: performing pre-processing on the plurality of patient attributes for each of the plurality of past events to generate a plurality of input data sets; dividing the plurality of past events into a first set of training data and a second set of validation data; iteratively performing a machine learning algorithm (MLA) to update synaptic weights of the NNM based upon the training data; and validating the NNM based upon the second set of validation data; receiving a plurality of input attributes of the new event; performing pre-processing on the plurality of input attributes to generate an input data set; generating an output value from the trained model based upon the input data set; classifying the output value into an adverse event category to predict the medical outcome; and alerting and assigning an allocation of appropriate clinical resources to a patient based upon an optimization algorithm in accordance with the adverse event category of the new event and attributes of available clinical resources, wherein one or more of the plurality of input attributes of the new event includes handover related data.
2 . The method of claim 1 , wherein the performing of the pre-processing further includes using a trained Self-Organizing Map (SOM) to generate the input data set.
3 . The method of claim 1 , wherein one or more of the plurality of input attributes of the new event includes data from an RFID tag and a hospital information system.
4 . The method of claim 1 , wherein the handover related data includes handover metrics of a facility associated with the new event, the handover metrics including intervals between handovers.
5 . The method of claim 1 , wherein the event is a patient being treated by a doctor associated with a facility, the method further comprising: associating the output value of the event with the facility; and predicting an outcome of the facility by tracking a plurality of output values for different events associated with the facility.
6 . The method of claim 5 , wherein the output value of each event is a probability of an adverse event, the method further comprises determining the output value continuously at predetermined time periods to determine when the rate of increase of the probability of the adverse event output values is greater than a predetermined threshold.
7 . The method of claim 1 , wherein the receiving of the plurality of input attributes of the new event further includes receiving one or more messages including a patient identification and location information associated with a first RFID tag and a medical professional identification and location information associated with a second RFID tag from a data collection engine (DCE).
8 . A method for predicting a medical outcome associated with a new event, the method comprising: training a Self-Organizing Map (SOM) to generate a trained model, wherein the trained model includes a plurality of network nodes arranged in a grid or lattice and in fixed topological positions, an input layer with a plurality of input nodes representing input attributes of past patient events, wherein each of the plurality of input nodes is connected to all of the plurality of network nodes by a plurality of synaptic weights, wherein the training of the SOM includes: initializing values of the plurality of synaptic weights to random values, randomly selecting one past patient event and determining which of the plurality of network nodes is a best matching unit (BMU) according to a discriminant function, wherein the discriminant function is a Euclidean Distance; and iteratively calculating a neighborhood radius associated with the BMU to determine neighboring network nodes for updating, and updating values of synoptic weights for neighboring network nodes within the calculated neighborhood radius for a fixed number of iterations to generate the trained model; receiving a plurality of input attributes of the new event; performing pre-processing on the plurality of input attributes to generate an input data set; generating an output value from the trained model based upon the input data set; classifying the output value into an adverse event category to predict a medical outcome; and alerting and assigning an allocation of appropriate clinical resources to a patient based upon an optimization algorithm in accordance with the adverse event category of the new event and attributes of available clinical resources, wherein: one or more of the plurality of input attributes of the new event includes handover related data; wherein: the receiving of the plurality of input attributes of the new event further comprises receiving a plurality of new events, each including a plurality of input attributes; the generating of the output value and classifying the output value further includes generating a graphical image including input attributes for each of the plurality of new events; the method further comprises receiving a graphical display request from a remote client device and transmitting the graphical image to the remote client device as a response; and the graphical image is a cluster diagram including a plurality of clusters of new events having a similar characteristic.
9 . An application specific integrated circuit (ASIC) for a neural network model (NNM), the ASIC comprising: a plurality of neurons organized in an array, wherein the plurality of neurons includes: an input layer including input neurons which provide input data set signals associated with an event to the NNM, wherein the event is a quantifiable medical outcome, and the plurality of input data set signals include a plurality of plurality of patient attributes associated with the quantifiable outcome; a hidden layer including hidden neurons; an output layer including output neurons which provide output data set signals; and a plurality of synaptic circuits, each synaptic circuit including a memory for storing a synaptic weight, where each of the hidden neurons and output neurons includes an activation function, the activation function is one of: (1) the sigmoid function ƒ(x)=1/(1+e −x ); (2) the hyperbolic tangent function ƒ(x)=(e 2x −1)/(e 2x +1); and (3) a linear function ƒ(x)=x, wherein x is a summation of input neurons biased by the synoptic weights, wherein values of the synaptic weights are obtained by training the NNM, the training of the NNM includes: performing pre-processing on the plurality of input attributes for each of a plurality of past events to generate a plurality of input data sets; dividing the plurality of past events into a first set of training data and a second set of validation data; iteratively performing a machine learning algorithm (MLA) to update values of the synaptic weights of the NNM based upon the training data; and validating the NNM based upon the second set of validation data.

Description

CROSS-REFERENCE TO RELATED APPLICATIONS The present application is a continuation of U.S. patent application Ser. No. 18/150,749 filed on Jan. 5, 2023, which is a continuation of U.S. patent application Ser. No. 16/852,397 filed on Apr. 17, 2020 now U.S. Pat. No. 11,756,660, which is a continuation of U.S. patent application Ser. No. 16/578,942 filed on Sep. 23, 2019 now U.S. Pat. No. 10,628,739, which is a continuation of U.S. patent application Ser. No. 16/231,832 filed on Dec. 24, 2018 now U.S. Pat. No. 10,482,377, which is a continuation of U.S. patent application Ser. No. 16/012,088 filed on Jun. 19, 2018, which is a continuation of U.S. patent application Ser. No. 15/934,966 filed on Mar. 24, 2018 now U.S. Pat. No. 10,026,506, which is a continuation-in-part of U.S. patent application Ser. No. 15/704,494 filed on Sep. 14, 2017 now U.S. Pat. No. 9,928,342, which is a continuation-in-part of U.S. patent application Ser. No. 15/592,116 filed on May 10, 2017 now U.S. Pat. No. 9,848,827, which is a continuation of U.S. patent application Ser. No. 15/390,695 filed on Dec. 26, 2016 now U.S. Pat. No. 9,679,108, which is a continuation of U.S. patent application Ser. No. 15/004,535 filed on Jan. 22, 2016 now U.S. Pat. No. 9,569,589, which claims the benefit of U.S. Provisional Patent Application No. 62/113,356 filed on Feb. 6, 2015, the contents all of which are incorporated herein by reference. TECHNICAL FIELD The technical field generally relates to a system including a client device, data input sources and a server device. BACKGROUND A radio-frequency Identification (RFID) chip can transmit information to a reader in response to an interrogation signal or polling request from the reader. The RFID chip can be incorporated in a tag (RFID tag) which is placed on a medical item such as a patient or doctor identification so that information can be passively captured. An RFID tag can be an active-type with its own power source, or a passive-type or battery-assisted passive type with no or limited power source. Both the passive-type and battery-assisted passive type will be referred to here as passive-type for sake of brevity. Placing an active-type RFID tag on some items may not be feasible due to financial considerations, weight, etc. On the other hand, placing a passive-type RFID tag on items may be more feasible; however, a power source will be needed to passively obtain information. Artificial Intelligence (AI) technologies such as machine learning and deep learning have become ever present due to technological advances in data storage and processing. Machine learning at its most basic is the practice of using algorithms to parse data, learn from it, and then make a determination or prediction about something in the world. So rather than hand-coding software routines with a specific set of instructions to accomplish a particular task, the machine is “trained” using large amounts of data and algorithms that give it the ability to learn how to perform the task. Deep learning involves neural networks inspired by our understanding of the biology of our brains all those interconnections between the neurons. But, unlike a biological brain where any neuron can connect to any other neuron within a certain physical distance, these artificial neural networks have discrete layers, connections, and directions of data propagation. SUMMARY Medical care for patients is often provided among multiple providers across different facilities in a hospital system and in a plurality of care settings that may have no affiliation. The transfer of essential information and the responsibility for care of the patient from one health care provider to another is an integral component of communication in health care. This critical transfer point will be referred to herein as a handoff. It is very important to identify, diagnose and treat medical, operations, and administrative issues that may not be easily apparent during medical care to enable better care coordination, quality improvement, care surveillance, monitoring, and clinical business intelligence. However, this may be a challenge due to the multiple healthcare facilities and providers involved in the patient care. A system that can identify which patient(s) within a ward full of patients are predicted to be at high risk for a preventable adverse event such as, for example, a medical error would be beneficial. It would be further beneficial if such a system can predict which specific adverse event is at high risk. With such a system, hospital leaders and managers can determine preventive intervention and obtain better situational awareness (i.e. alert on chart, hanging signs on door, higher level of unit/ward manager involvement/oversight, other, etc.). Within a hospital facility or across a hospital system, a system that can identify which facilities and hospital areas (wards, operating rooms, procedure suites or areas, clinics, etc.) within specific facilities are predicted to ha