CN-122025160-A - System, method of operation, and predictive processor for predicting patient cessation of prescription diagnosis and treatment

CN122025160ACN 122025160 ACN122025160 ACN 122025160ACN-122025160-A

Abstract

A system, method of operation, and prediction processor for predicting patient cessation of prescription treatment are disclosed. The system includes a memory device storing a training data set including treatment data for a patient group and patient data, and patient data and prior treatment data for a patient, and further including an indication of whether the patient stopped prescribing treatment or a procedure, and a prediction processor receiving treatment data for the patient from the APD machine, storing the treatment data to the memory device, determining a probability that the patient will stop prescribing treatment by applying the patient data, the treatment data, and the prior treatment data to the at least one patient prediction model, and causing the probability to be displayed in a user interface on a clinician device.

Inventors

Christie Elizabeth. Garcia
Timothy Louis. Kudelka
JONATHAN ALAN HANDLER
Ian Janosh Kelaimen
Mark Anthony. Penny
Angelo A. Salto
Andrew Thomas. Gebhart
Ashkan Khozad
Richard. Scott Teisseire

Assignees

巴克斯特国际公司
巴克斯特医疗保健股份有限公司

Dates

Publication Date: 20260512
Application Date: 20201105
Priority Date: 20191105

Claims (12)

1. A system for predicting a target patient to stop prescription diagnosis administered by an automated peritoneal dialysis ("APD") machine, the system comprising: a memory device that stores: A training data set comprising treatment data and patient data for a group of patients, the training data set further comprising an indication as to whether the patient stopped treatment prescribed or programmed, At least one patient prediction model formed using the training data set, the at least one patient prediction model including at least inputs of (i) counts or frequencies of alerts generated by APD machines, (ii) information related to peritoneal dialysis cycles, (iii) patient blood pressure values, and (iv) patient weight values, and Patient data and prior diagnosis data of the target patient undergoing the prescribed treatment or procedure, and A prediction processor communicatively coupled to the memory device, the prediction processor configured to: receiving the diagnosis and treatment data of the target patient from the APD machine, Storing said diagnosis and treatment data received by the interface device to said memory device, Determining a probability that the target patient will cease the prescribed diagnosis by applying the patient data, the diagnosis data, and the prior diagnosis data of the target patient to the at least one patient prediction model, an Causing the probabilities to be displayed in a user interface on the clinician device.
2. The system of claim 1, wherein the prediction processor is further configured to: Identifying the most important parameters contributing to said probability, and Causing an indication of the most important parameter to be displayed within the user interface on the clinician device.
3. The system of claim 1 or 2, wherein the memory device comprises a data structure relating medical fluid delivery advice to a range of probabilities.
4. The system of claim 3, further comprising a guidance processor communicatively coupled to the memory device and configured to: Comparing the probability of the target patient to a range of the probabilities; Selecting at least one suggestion based on the comparison, and Causing the at least one suggestion to be displayed within the user interface on the clinician device.
5. A method of operation of a system for predicting a target patient to stop prescription diagnosis administered by an automated peritoneal dialysis ("APD") machine, the system comprising a memory device and a prediction processor communicatively coupled to the memory device, the method of operation comprising: the memory device stores: A training data set comprising treatment data and patient data for a group of patients, the training data set further comprising an indication as to whether the patient stopped treatment prescribed or programmed, At least one patient prediction model formed using the training data set, the at least one patient prediction model including at least inputs of (i) counts or frequencies of alerts generated by APD machines, (ii) information related to peritoneal dialysis cycles, (iii) patient blood pressure values, and (iv) patient weight values, and Patient data and prior diagnosis data of the target patient undergoing the prescribed treatment or procedure, and The prediction processor receives the treatment data for the target patient from the APD machine, The prediction processor stores the diagnosis and treatment data to the memory device, The prediction processor determines a probability that the target patient will cease the prescribed diagnosis by applying the patient data, the diagnosis data, and the prior diagnosis data of the target patient to the at least one patient prediction model, an The prediction processor causes the probability to be displayed within a user interface on a clinician device.
6. The method of operation of claim 5, wherein the method of operation further comprises: The prediction processor identifying the most important parameter contributing to the probability, and The prediction processor causes an indication of the most important parameter to be displayed within the user interface on the clinician device.
7. The method of operation of claim 5 or 6, wherein the memory device includes a data structure relating medical fluid delivery recommendations to a range of probabilities.
8. The method of operation of claim 7, wherein the system further comprises a guidance processor communicatively coupled to the memory device, wherein the method of operation further comprises: The guideline processor compares the probability of the target patient to a range of the probabilities; The guideline processor selects at least one suggestion based on the comparison, and The guideline processor causes the at least one suggestion to be displayed within the user interface on the clinician device.
9. A prediction processor for predicting a target patient to stop prescription diagnosis administered by an automated peritoneal dialysis ("APD") machine, the prediction processor configured to: Receiving, from an APD machine, treatment data for the target patient undergoing a prescribed treatment or procedure; The diagnosis and treatment data is stored to a memory device, Accessing at least one patient prediction model formed using a training dataset comprising at least inputs of (i) counts or frequencies of alerts generated by APD machines, (ii) information related to peritoneal dialysis cycles, (iii) patient blood pressure values, and (iv) patient weight values, the training dataset comprising treatment data and patient data for a patient group, the training dataset further comprising an indication of whether the patient stopped a prescribed treatment or procedure; determining a probability that the target patient will cease the prescribed therapy by applying patient data of the target patient, the therapy data, and prior therapy data to the at least one patient prediction model, and Causing the probabilities to be displayed in a user interface on the clinician device.
10. The prediction processor of claim 9, wherein the prediction processor is further configured to: Identifying the most important parameters contributing to said probability, and Causing an indication of the most important parameter to be displayed within the user interface on the clinician device.
11. The prediction processor of claim 9 or 10, wherein the memory device includes a data structure that relates medical fluid delivery recommendations to a range of probabilities.
12. The prediction processor of claim 11, wherein the prediction processor is further configured to: Comparing the probability of the target patient to a range of the probabilities; Selecting at least one suggestion based on the comparison, and Causing the at least one suggestion to be displayed within the user interface on the clinician device.

Description

System, method of operation, and predictive processor for predicting patient cessation of prescription diagnosis and treatment The present application is a divisional application of national application number 202080076094.4 (international application number PCT/US2020/059117, international application day 2020, month 11, 5, title of application "including medical fluid delivery systems for managing analysis of patient participation and compliance with diagnosis)". Technical Field The present invention relates to medical fluid data transfer systems for determining and/or predicting patient compliance, and in particular to systems, methods of operation, and prediction processors for predicting patient cessation of prescription clinics. Background Currently, it is almost impossible to have patients participate in tasks outside of the medical environment for long periods of time. Similar to starting fitness members or purchasing treadmills, many patients often start to be more aggressive. Initially, the patient is willing to participate in a self-administered medical treatment (e.g., a medical fluid delivery treatment). Typically, these medical treatments are performed in the patient's home and/or clinic. For medical fluid delivery diagnosis, patients must connect themselves to the medical fluid delivery machine (or container containing renal failure diagnosis fluid) to clean their blood against toxin accumulation. Portions of the diagnosis may include administrative tasks that the patient must perform, such as weighing their weight, measuring their blood pressure, and/or recording information related to their diagnosis. Clinicians often review patient recorded information to ensure that a diagnosis is prescribed. The clinician also reviews the recorded data to determine if adjustments to the diagnosis are needed. Over time, patients become less enthusiastic, as repeated diagnoses lose novelty, and become another routine duty. It is envisioned that patients would be more willing to participate in more exciting, relaxing or stimulating activities than to self-administer medications or to go to the clinic multiple times per week to receive the same treatment. As many patients continue to diagnose, they sometimes begin to omit the additional tasks attendant to the diagnosis. Omitting additional tasks and becoming less enthusiastic for diagnosis, it is possible to create gaps in the clinical supervision of ongoing diagnosis. As patients further reject the treatment, they may begin skipping the treatment, either making the treatment only for a portion of the prescribed time, or giving up the treatment altogether, with a health risk in such a treatment. Disclosure of Invention Disclosed herein is a medical fluid data transfer system for determining and/or predicting patient compliance. The medical fluid data transfer system is configured to improve patient compliance by tracking how the patient uses or otherwise interacts with a medical fluid delivery machine, such as an automated peritoneal dialysis ("APD") machine. In some embodiments, the medical fluid data transfer system analyzes data from the medical fluid delivery machine to determine compliance of the patient with one or more prescribed treatments or procedures. If patient compliance has fallen below a prescribed threshold or tends to fall below a threshold, the medical fluid data transfer system may operate one or more evidence-based algorithm models. As disclosed herein, the model is configured to provide suggestions to the clinician as to how to improve patient compliance with one or more prescribed treatments or procedures by addressing potential problems that the patient may be experiencing. In some embodiments, the medical fluid data transfer system may alternatively or additionally include one or more artificial intelligence ("AI") patient prediction models configured to identify patients at risk of stopping their diagnosis or being below a prescribed compliance threshold. The one or more AI patient prediction models are configured to determine a score of interest that indicates a probability that a given patient may end a diagnosis or be below a desired compliance threshold. The AI patient prediction model determines the score of interest by considering the treatment data from the medical fluid delivery machine and readily available patient information, as it relates to prescribed treatment. In this way, the AI patient prediction model is configured to accurately determine the risk of the patient using readily available data without having to access third party data or other medical data stored in the patient medical record. In addition to providing a score of interest, the example AI patient prediction models described herein are configured to provide visibility or insight as to how the score of interest is calculated. For example, the AI patient prediction model may provide a number of important reasons or attributes that contribute to the attentio