US-20260128149-A1 - SEMAGLUTIDE DOSAGE MANAGEMENT

US20260128149A1US 20260128149 A1US20260128149 A1US 20260128149A1US-20260128149-A1

Abstract

Systems and methods for determining effectiveness of semaglutide for a patient. One embodiment is a system for analyzing dosage effectiveness. The system is configured to receive health data for a population of patients, extract metrics for sex, sleep apnea, hypertension, and prescription history from the health data, and train a predictive model based on the metrics. The system is configured to identify a patient and a semaglutide dosage, and operate the predictive model to predict a likelihood of the semaglutide dosage accomplishing a selected amount of weight loss for the patient during a time period. In an event the likelihood is below a threshold, the system is configured to recommend that the dosage not be prescribed to the patient. In an event the likelihood is above the threshold, the system is configured to recommend that the dosage be prescribed to the patient.

Inventors

Matthew Levy
Elizabeth Cirulli Rogers

Assignees

HELIX, INC.

Dates

Publication Date: 20260507
Application Date: 20241226

Claims (20)

1 . A system for analyzing dosage effectiveness, the system comprising: an interface configured to receive health data for a population of patients; and a controller configured to extract metrics for sex, sleep apnea, hypertension, and prescription history from the health data, and to train a predictive model based on the metrics; the controller further configured to identify a patient and a semaglutide dosage, and to operate the predictive model to predict a likelihood of the semaglutide dosage accomplishing a selected amount of weight loss for the patient during a time period; in an event the likelihood for the semaglutide dosage is below a threshold, the controller is further configured to recommend that the semaglutide dosage not be prescribed to the patient; in an event the likelihood for the semaglutide dosage is above the threshold, the controller is further configured to recommend that the semaglutide dosage be prescribed to the patient.
2 . The system of claim 1 wherein: the controller is further configured, in the event that the likelihood for the semaglutide dosage is above the threshold, to identify a lower dosage of semaglutide, and to operate the predictive model to predict a likelihood of the lower dosage accomplishing the selected amount of weight loss for the patient during the time period, in an event the likelihood for the lower dosage is below the threshold, the controller is further configured to recommend that the lower dosage not be prescribed to the patient; and in an event the likelihood for the lower dosage is above the threshold, the controller is further configured to recommend that the lower dosage be prescribed to the patient.
3 . The system of claim 1 wherein: the metrics further comprise a polygenic score for Body Mass Index (BMI), concurrent use of a non-Glucagon-like peptide-1 (GLP-1) weight loss drug with semaglutide, and previous use of another GLP-1 weight loss drug within a year immediately prior to prescription of semaglutide.
4 . The system of claim 1 wherein: the selected amount of weight loss is at least ten percent of body weight, and the threshold is between fifty and ninety-nine percent.
5 . The system of claim 1 wherein: the controller is further configured to train multiple predictive models, each of the predictive models trained using metrics specific to one or more demographics in categories selected from the group consisting of: sex, ancestry, age, and Body Mass Index (BMI); and the controller is further configured to select one of the predictive models based on demographics of the patient.
6 . The system of claim 1 wherein: the controller is further configured to train the predictive model using logistic regression.
7 . The system of claim 1 wherein: the controller is further configured to exclude metrics for patients of the population that have type two diabetes, prior to training the predictive model.
8 . A method for analyzing dosage effectiveness, the method comprising: receiving health data for a population of patients; extracting metrics for sex, sleep apnea, hypertension, and prescription history from the health data; training a predictive model based on the metrics; identifying a patient and a semaglutide dosage; operating the predictive model to predict a likelihood of the semaglutide dosage accomplishing a selected amount of weight loss for the patient during a time period; in an event the likelihood for the semaglutide dosage is below a threshold, recommending that the semaglutide dosage not be prescribed to the patient; and in an event the likelihood for the semaglutide dosage is above the threshold, recommending that the semaglutide dosage be prescribed to the patient.
9 . The method of claim 8 further comprising: in the event that the likelihood for the semaglutide dosage is above the threshold: identifying a lower dosage of semaglutide; operating the predictive model to predict a likelihood of the lower dosage accomplishing the selected amount of weight loss for the patient during the time period; in an event the likelihood for the lower dosage is below the threshold, recommending that the lower dosage not be prescribed to the patient; and in an event the likelihood for the lower dosage is above the threshold, recommending that the lower semaglutide dosage be prescribed to the patient.
10 . The method of claim 8 wherein: the metrics further comprise a polygenic score for Body Mass Index (BMI), concurrent use of a non-Glucagon-like peptide-1 (GLP-1) weight loss drug with semaglutide, and previous use of another GLP-1 weight loss drug within a year immediately prior to prescription of semaglutide.
11 . The method of claim 8 wherein: the selected amount of weight loss is at least ten percent of body weight, and the threshold is between fifty and ninety-nine percent.
12 . The method of claim 8 further comprising: training multiple predictive models, each of the predictive models trained using metrics specific to one or more demographics in categories selected from the group consisting of: sex, ancestry, age, and Body Mass Index (BMI); and selecting one of the predictive models based on demographics of the patient.
13 . The method of claim 8 wherein: training the predictive model comprises using logistic regression.
14 . The method of claim 8 further comprising: excluding metrics for patients of the population that have type two diabetes, prior to training the predictive model.
15 . A non-transitory computer readable medium embodying programmed instructions which, when executed by a processor, are operable for performing a method for analyzing dosage effectiveness, the method comprising: receiving health data for a population of patients; extracting metrics for sex, sleep apnea, hypertension, and prescription history from the health data; training a predictive model based on the metrics; identifying a patient and a semaglutide dosage; operating the predictive model to predict a likelihood of the semaglutide dosage accomplishing a selected amount of weight loss for the patient during a time period; in an event the likelihood for the semaglutide dosage is below a threshold, recommending that the semaglutide dosage not be prescribed to the patient; and in an event the likelihood for the semaglutide dosage is above the threshold, recommending that the semaglutide dosage be prescribed to the patient.
16 . A method for administering semaglutide, the method comprising: identifying a patient; selecting a dosage of semaglutide for the patient; operating a predictive model trained upon health data for a population using metrics of sex, sleep apnea, hypertension, and prescription history to predict a likelihood of the dosage accomplishing a loss of at least ten percent body mass for the patient during a time period of one year; in an event the likelihood is below a threshold, preventing administration of the dosage to the patient; and in an event the likelihood is above the threshold, administering the dosage to the patient.
17 . The method of claim 16 wherein: the threshold is between fifty and ninety-nine percent.
18 . The method of claim 16 wherein: operating the predictive model comprises operating a logistic regression model.
19 . The method of claim 16 wherein: the predictive model is further trained upon a polygenic score for Body Mass Index (BMI), concurrent use of a non-Glucagon-like peptide-1 (GLP-1) weight loss drug with semaglutide, and previous use of another GLP-1 weight loss drug within a year immediately prior to prescription of semaglutide.
20 . The method of claim 16 wherein: the population comprises a population of patients, excluding patients having type two diabetes.

Description

RELATED APPLICATIONS This non-provisional application claims priority to U.S. provisional application 63/717,784, filed on Nov. 7, 2024, which is incorporated herein by reference as if fully provided herein. FIELD The disclosure relates to the field of health care, and in particular to controlling a dosage of semaglutide administered to patients. BACKGROUND Semaglutide is a widely used pharmaceutical that has multiple applications, including for weight control. However, it is not uncommon for semaglutide to be ineffective in driving weight loss for certain patients, or for patients to discontinue semaglutide use due to undesirable side effects, such as bloating or nausea. Healthcare providers therefore continue to seek out new, robust solutions that enhance the ability to provide semaglutide to patients within a population in an efficacious manner. SUMMARY Embodiments described herein utilize predictive models trained on specific metrics of population data to anticipate the effectiveness of semaglutide dosages upon specific patients. This results in insights which may be used to determine whether to initiate, adjust, or discontinue a default semaglutide dosage for a patient. One embodiment is a system for analyzing dosage effectiveness. The system includes an interface configured to receive health data for a population of patients, and a controller configured to extract metrics for sex, sleep apnea, hypertension, and prescription history from the health data, and to train a predictive model based on the metrics. The controller is further configured to identify a patient and a semaglutide dosage, and to operate the predictive model to predict a likelihood of the semaglutide dosage accomplishing a selected amount of weight loss for the patient during a time period. In an event the likelihood for the semaglutide dosage is below a threshold, the controller is configured to recommend that the semaglutide dosage not be prescribed to the patient. In an event the likelihood for the semaglutide dosage is above the threshold, the controller is configured to recommend that the semaglutide dosage be prescribed to the patient. A further embodiment is a method for analyzing dosage effectiveness. The method includes receiving health data for a population of patients, extracting metrics for sex, sleep apnea, hypertension, and prescription history from the health data, and training a predictive model based on the metrics. The method further includes identifying a patient and a semaglutide dosage, and operating the predictive model to predict a likelihood of the semaglutide dosage accomplishing a selected amount of weight loss for the patient during a time period. In an event the likelihood for the semaglutide dosage is below a threshold, the method includes recommending that the semaglutide dosage not be prescribed to the patient. In an event the likelihood for the semaglutide dosage is above the threshold, the method includes recommending that the semaglutide dosage be prescribed to the patient. A further embodiment is a non-transitory computer readable medium embodying programmed instructions which, when executed by a processor, are operable for performing a method for analyzing dosage effectiveness. The method includes receiving health data for a population of patients, extracting metrics for sex, sleep apnea, hypertension, and prescription history from the health data, and training a predictive model based on the metrics. The method further includes identifying a patient and a semaglutide dosage, and operating the predictive model to predict a likelihood of the semaglutide dosage accomplishing a selected amount of weight loss for the patient during a time period. In an event the likelihood for the semaglutide dosage is below a threshold, the method includes recommending that the semaglutide dosage not be prescribed to the patient. In an event the likelihood for the semaglutide dosage is above the threshold, the method includes recommending that the semaglutide dosage be prescribed to the patient. A further embodiment is a method for administering semaglutide. The method includes identifying a patient, selecting a dosage of semaglutide for the patient, and operating a predictive model trained upon health data for a population using metrics of sex, sleep apnea, hypertension, and prescription history to predict a likelihood of the dosage accomplishing a loss of at least ten percent body mass for the patient during a time period of one year. In an event the likelihood is below a threshold, the method includes preventing administration of the dosage to the patient. In an event the likelihood is above the threshold, the method includes administering the dosage to the patient. Other illustrative embodiments (e.g., methods and computer-readable media relating to the foregoing embodiments) may be described below. The features, functions, and advantages that have been discussed can be achieved independently in various embodiments