Researchers Submit Patent Application, “Method and Apparatus for Discovering a Sequence of Events Forming an Episode in a Set of Medical Records from…
Researchers Submit Patent Application, "Method and Apparatus for Discovering a Sequence of Events Forming an Episode in a Set of Medical Records from a Patient", for Approval (USPTO 20180082025)
By a
The patent's assignee is
News editors obtained the following quote from the background information supplied by the inventors: "The embodiments relate to a system and method to detect an episode of a disease or other medical condition, for example by applying machine learning methods. It has applications in the areas of healthcare profiling, healthcare monitoring and improvement, and even as an aid to selection of treatment.
"At the onset of certain medical conditions, a patient (or subject) can demonstrate several symptoms and be treated separately or jointly with different treatments. Also, several conditions may jointly affect a patient: this is known as co-morbidity. In this document, a disease/condition episode is defined as including any observable symptoms, payable investigative method, treatments, and medications associated with an instance of that disease/condition.
"For instance, a typical measles episode can consist of the following states: infection, incubation, non-specific/mild period, acute period, recovery, and communicable period. At different stages, a patient can demonstrate different symptoms and signs. The patient can be affected by high temperature, cough, headaches, and rash. The patient may be subject to blood tests and chest x-rays. Severe measles can have complications demonstrating other symptoms and signs which need to be dealt with during the patient's interaction with the healthcare system. All these need to be grouped into one episode to enable better healthcare statistics/profiling, improved treatment and also non-medical improvements such better payment management under medical insurance. Also, long term conditions can have acute episode, e.g. a bipolar disorder can have a mania episode, during which period all the interactions should be properly categorized for better patient care.
"When treating a patient within one particular episode, the disease progresses from one disease state (with particular symptoms) to another (with different symptoms). For instance, an episode of chest-infection can start with high fever, moving to persistent coughing, moving to asthmatic symptoms, moving to full recovery or LTC asthma. These are different states of a particular instance of the disease and may require different medication strategies.
"Detecting an episode of a certain disease (for example the start and end of an episode and/or the events that are attributable to that episode) is important in health care in general and in medical insurance industries in particular. The reasons are as follows: 1. In order to have well targeted studies in the medical domain, it is useful to differentiate co-morbid diseases. Such diseases can then be studied separately to avoid data 'pollution'. 2. Certain conditions may be caused by the treatment of prior conditions. Also it is equally important to understand the fluctuation of long term conditions. It is therefore necessary to clearly define and separate different episodes. 3. Medical claims (for example of the same episode) are normally grouped together for a single insurance payout. This also applies to countries providing non-contributory healthcare welfare, e.g.
"Episode defining or episode grouping is not a trivial task. Thus far, the dominant solution is based on expert opinions. This approach is, however, often prone to errors. 1. Expert opinions only provide ranges of values based on 'normal' conditions. For cases falling out the boundaries of such normal ranges, the episode definition is difficult. 2. Human conditions vary from individual to individual and even from time to time with respect to the same individual. The 'normal' range cannot cover all the inter- and intra-individual variances.
"The inventors have come to the realization that the inefficiency of current practice calls for a dynamic approach for episode grouping."
As a supplement to the background information on this patent application, NewsRx correspondents also obtained the inventors' summary information for this patent application: "Additional aspects and/or advantages will be set forth in part in the description which follows and, in part, will be apparent from the description, or may be learned by practice of the embodiments.
"According to an embodiment of a first aspect, there is provided an apparatus for discovering a sequence of events in a set of medical records from a patient, the sequence forming an episode of a medical condition, the apparatus comprising:
"a state transition learner, to parse published clinical guidelines and to extract probabilities of transition between a number of states of a medical condition as a state transition model;
"a clinical finding learner, to extract typical findings of the medical condition from domain knowledge as a finding model and to compute the probability of a particular finding for a particular state and to save the probabilities in an overall episode model including the state transition model and the finding model; and
"an episode grouper, to use the overall episode model, and the set of medical records to discover a sequence of events, to group the sequence of events into an episode of the medical condition, and to differentiate the medical condition from apparently similar medical conditions or co-morbidities, or both.
"These features allow discovery of an episode without necessarily involving a human expert and by using different knowledge sources which are publically available. They allow separation of a sequence of events associated with a particular condition from other events, and thus differentiation from other conditions which may be similar and from other symptoms occurring at the same time but for different causes (co-morbidities). The same features can even also allow preliminary identification of a condition, because once a match is made to an episode (or several candidate episodes), it is possible to narrow down the courses of disease development specific to that condition.
"An episode is a single (significant) occurrence/instance of a certain illness (which is used synonymously and interchangeably herein with medical condition or disease). It can be one occurrence of a longer series of occurrences. An episode is normally defined as the group of symptoms and signs from the onset of certain relevant symptoms and signs to the disappearance of the symptoms and signs. Each episode can have a number of states. Normally, a 'naive' episode grouper or simple episode grouper will split data to define an episode from the first interaction between a patient and the healthcare system until the patient is deceased or discharged for that condition.
"The purpose of an episode grouper according to the embodiments, and as explained in more detail later herein, can be to identify which findings (e.g. symptoms/signs, treatments and medicines) belong to one episode of certain disease. This can be important information for diagnostic assistance, analysis, insurance and quality assurance purposes.
"The term finding(s) or observation(s) is used to refer to observable patient interactions which are thus effectively medical findings. Hence the term can include medical intervention and medication, as well as complaints, symptoms and diagnosis as noted by a healthcare practitioner.
"Events are findings coming from a particular patient record which is to be grouped or classified, for example according to a learnt model, as described in more detail later. (An 'event' in an individual's medical record is an observable medical result or medical intervention that could have significant meaning in the course of diseases.) So, in short, events are documented individual instances of findings. Findings refer to the general model.
"The episode grouper can be to discover a sequence as a subset of events in the set of medical record that can be grouped into an episode, and to exclude remaining events as not forming part of the medical condition. This allows the separation of co-morbidities.
"The state transition learner can be arranged to derive the probability of transition between states from internet search results. This may be using internet search results (and, for example, co-occurrence of the states in the search results). The internet search results can be confined to one or more on-line medical publications to give better accuracy.
"The clinical learning finder can compute the probability of a particular finding for a particular state in the state transition model, preferably using a training data set, for instance a set of patients' records for the medical condition in question. In this way, each finding from the domain knowledge, such as from on-line medical protocols and guidelines, is mapped to the states in the state transition model based on the occurrence of the finding in each state. Of course if the finding does not occur in a particular state in the medical records (or the level is below a threshold), then the probability can be set to zero, and there is effectively no link between the finding and the state.
"Thus the clinical learning finder can compute the probability of a particular finding for a particular state using co-occurrence of the finding and the state in public data.
"The overall episode model can contain, for each of one or more medical conditions, links between the findings in the finding model for the medical condition and the states in the state transition model for that medical condition and probabilities associated with the links. Preferably, the model covers a wide range of medical conditions, and holds a separate, individual linked state transition and finding model for each of them.
"The overall episode model also preferably contains links to a leak term and probabilities associated with the links. Here, the leak term corresponds to a situation in which a finding is observed which is not relevant to any state in the state transition model for the medical condition.
"The episode grouper may match sequences of events in the set of medical records to the overall episode model and detect the best match between a sequence of events and the overall episode model.
"For example, the episode grouper is to match the powerset of the sequence of events in the set of medical records to the overall episode model, with the exception of the empty set and/or any sets including a number of events below a threshold. It is important that the events are kept in the correct order in the powerset, but events in the time line may be omitted (and for instance included later in an episode of another condition). The threshold may be 2 events, so that a 'sequence' of a single event is not assessed for a match with an episode. The various subsets of sequences can be assessed in any order, for example in random order or fully/partially in parallel.
"The episode grouper may match sequences of events to the overall episode model by calculating the probability of arriving at the sequence of events in the set of medical records for each sequence, based on a given initial state in the patient medical records, the state transition model and based on the probability of observable medical findings in the overall episode model corresponding to the events in the sequence of events.
"The episode grouper may process remaining events in the set of medical records once a sequence of events has been grouped into an episode, by matching sequences of remaining events. This allows other conditions to be identified on the basis of co-morbidities which might otherwise have been categorized as part of the same condition.
"According to an embodiment of a second aspect, there is provided a method for discovering a sequence of events in a set of medical records from a patient, the sequence forming an episode of a medical condition, the method comprising:
"parsing published clinical guidelines and extracting probabilities of transition between a number of states of a medical condition as a state transition model;
"extracting typical findings of the medical condition from domain knowledge as a finding model;
"computing the probability of a particular finding for a particular state;
"saving the probabilities in an overall episode model including the state transition model and the finding model; and
"using the overall episode model, and the set of medical records to discover a sequence of events, to group the sequence of events into an episode of the medical condition, and to differentiate the medical condition from apparently similar medical conditions or co-morbidities.
"According to an embodiment of a third aspect, there is provided a computer program which when executed on a computer apparatus carries out a method for discovering a sequence of events in a set of medical records from a patient, the sequence forming an episode of a medical condition, the method comprising:
"parsing published clinical guidelines and extracting probabilities of transition between a number of states of a medical condition as a state transition model;
"extracting typical findings of the medical condition from domain knowledge as a finding model;
"computing the probability of a particular finding for a particular state;
"saving the probabilities in an overall episode model including the state transition model and the finding model; and
"using the overall episode model, and the set of medical records to discover a sequence of events, to group the sequence of events into an episode of the medical condition, and to differentiate the medical condition from apparently similar medical conditions or co-morbidities.
"A method or computer program according to preferred embodiments can comprise any combination of the apparatus aspects, but without limitation to specific hardware. Methods or computer programs according to further embodiments can be described as computer-implemented in that they require processing and memory capability.
"The apparatus according to preferred embodiments may be described as configured or arranged to, or simply 'to' carry out certain functions. This configuration or arrangement could be by use of hardware or middleware or any other suitable system. In preferred embodiments, the configuration or arrangement is by software.
"Thus according to one aspect there is provided a program which, when loaded onto at least one computer configures the computer to become the apparatus according to any of the preceding apparatus definitions or any combination thereof.
"According to a further aspect there is provided a program which when loaded onto the at least one computer configures the at least one computer to carry out the method steps according to any of the preceding method definitions or any combination thereof.
"In general the computer may comprise the elements listed as being configured or arranged to provide the functions defined. For example this computer may include memory, processing, and a network interface.
"The embodiments can be implemented in digital electronic circuitry, or in computer hardware, firmware, software, or in combinations of them. The embodiments can be implemented as a computer program or computer program product, i.e., a computer program tangibly embodied in a non-transitory information carrier, e.g., in a machine-readable storage device, or in a propagated signal, for execution by, or to control the operation of, one or more hardware modules.
"A computer program can be in the form of a stand-alone program, a computer program portion or more than one computer program and can be written in any form of programming language, including compiled or interpreted languages, and it can be deployed in any form, including as a stand-alone program or as a module, component, subroutine, or other unit suitable for use in a data processing environment. A computer program can be deployed to be executed on one module or on multiple modules at one site or distributed across multiple sites and interconnected by a communication network.
"Method steps can be performed by one or more programmable processors executing a computer program to perform functions by operating on input data and generating output. Apparatus can be implemented as programmed hardware or as special purpose logic circuitry, including e.g., an FPGA (field programmable gate array) or an ASIC (application-specific integrated circuit).
"Processors suitable for the execution of a computer program include, by way of example, both general and special purpose microprocessors, and any one or more processors of any kind of digital computer. Generally, a processor will receive instructions and data from a read-only memory or a random access memory or both. The essential elements of a computer are a processor for executing instructions coupled to one or more memory devices for storing instructions and data.
"The description is in terms of particular embodiments. Other embodiments are within the scope of the following claims. For example, the steps can be performed in a different order and still achieve desirable results. Multiple test script versions can be edited and invoked as a unit without using object-oriented programming technology; for example, the elements of a script object can be organized in a structured database or a file system, and the operations described as being performed by the script object can be performed by a test control program.
"Elements have been described using the terms 'learner', 'grouper' etc. The skilled person will appreciate that such functional terms and their equivalents may refer to parts of the system that are spatially separate but combine to serve the function defined. Equally, the same physical parts of the system may provide two or more of the functions defined.
"For example, separately defined means may be implemented using the same memory and/or processor as appropriate."
For additional information on this patent application, see: HU, Bo;
Keywords for this news article include: Business, Software,
Our reports deliver fact-based news of research and discoveries from around the world. Copyright 2018, NewsRx LLC
Long Term Care Market to Witness a Pronounce Growth by 2023
LA’s former assistant fire chief was paid $1.4 million last year, topping city’s list of highest-paid retirees
Advisor News
Annuity News
Health/Employee Benefits News
Life Insurance News