US 11,657,901 B2
System and method for prediction-model-based display of distributions of health outcome information for patient populations in geographical areas
Amir Abdolahi, Waltham, MA (US); Cecilia Meijer, Cambridge, MA (US); Eran Simhon, Boston, MA (US); Gertjan Laurens Schuurkamp, Utrecht (NL); Reza Sharifi Sedeh, Malden, MA (US); and Jordan Lento, Alpharetta, GA (US)
Assigned to Koninklijke Philips N.V., Eindhoven (NL)
Filed by KONINKLIJKE PHILIPS N.V., Eindhoven (NL)
Filed on Nov. 26, 2018, as Appl. No. 16/199,701.
Claims priority of provisional application 62/613,817, filed on Jan. 5, 2018.
Prior Publication US 2019/0213302 A1, Jul. 11, 2019
Int. Cl. G16B 45/00 (2019.01); G16H 50/30 (2018.01); G16H 50/70 (2018.01); G16B 40/00 (2019.01); G16H 50/80 (2018.01)
CPC G16B 45/00 (2019.02) [G16B 40/00 (2019.02); G16H 50/30 (2018.01); G16H 50/70 (2018.01); G16H 50/80 (2018.01)] 11 Claims
 
1. A system configured to display distributions of predicted health outcome information for patient populations in geographical areas by generating prediction models trained on demographic, social, and prior health outcome information of the patient populations, so that a medical care provider system can better geographically align resources based on the distributions of predicted health outcome information for the patient populations in the geographical areas compared to an existing alignment of the resources, the system comprising:
an electronic storage medium storing machine readable instructions configured to cause one or more hardware processors to:
obtain, by the one or more hardware processors, demographic, social, and prior health outcome information for a patient population in a geographical area, wherein the demographic and social information relates to one or more of economics, neighborhood environments, education, health insurance coverage, or social interactions of the patient population, wherein the prior health outcome information indicates one or more of medical conditions experienced by the patient population, treatments received by the patient population, or results of the treatments on the medical conditions for the patient population, and wherein the demographic, social, and prior health outcome information is, at least in part, sourced from one or more sensors, and wherein the one or more sensors are coupled to one or more wearable devices configured to track physiological characteristics;
train, by the one or more hardware processors, a prediction model comprising one or more neural networks based on the demographic, social, and prior health outcome information, wherein:
the one or more neural networks each comprise a collection of neural units, with each neural unit of a neural network connected with other neural units of the neural network, and with such connections being enforcing or inhibitory in their effect on an activation state of connected neural units, the neural network including multiple layers of neural units where a signal path traverses from front layers to back layers;
the demographic, social, and prior health outcome information is inputted to the prediction model to cause a relative influence of each feature of the demographic, social, and prior health outcome information relative to other features to be determined by the one or more one or more neural networks, and
the demographic, social, and prior health outcome information is retrieved from a non-transitory computer readable storage medium;
output, by the one or more hardware processors, weighted features of the demographic and social information that are predictive of health outcomes for the patient population based on the relative influence of each feature from the prediction model; and
display, by the one or more hardware processors, a distribution of predicted health outcome information for the patient population in the geographical area based on the weighted features, the displaying of the distribution of predicted health outcome information comprising simultaneous display of a plurality of fields and/or views of a graphical user interface indicating information related to:
criteria received from a user used to define the patient population,
a number of patients in the patient population,
an indication of relative influence of individual weighted features on the distribution, and
health outcome risk indicators for medical conditions in specific regions of the geographical area.