Predictive risk assessment in patient and health modeling
A patient's health is modeled through multiple scenarios. A base model incorporates rules that govern a response by a human being to one or more diseases, as well as a relation between health metrics and the diseases. Health attributes of the patient are obtained. From the base model and the health attributes, a patient model is generated. The patient model is modeled under different parameters to generate health metrics. From an analysis of the parameters and the resulting health metrics, occurrence probabilities for each of the sets of parameters are determined. Risks are identified, which indicate a likelihood of the patient transitioning from the initial state to an adverse outcome such as a diseased state. A report provides a diagnosis of the patient and one or more remedies/interventions that are predicted, based on the plural sets of parameters and resulting health metrics, to avoid or prevent an adverse outcome.
1 . A computer-implemented method of providing a digital computer simulation of health of a human patient, comprising:
obtaining in computer memory, by a processor, a base model of a human being, the base model being a mathematical model incorporating rules that govern 1) a response by the human being to a disease, and 2) a relation between health metrics and the disease, the health metrics including a resistance metric indicating a measure of resistance to the disease;
obtaining in computer memory, by the processor, initial health attributes of the human patient, the initial health attributes including at least one of a genetic trait, a blood lipid profile, and a representation of a medical imaging report;
responsively generating in computer memory, by the processor, an initial patient model from the base model and the initial health attributes of the human patient, the initial patient model incorporating a patient-specific representation of the rules of the base model and a representation of the initial health attributes, the patient-specific representation of the rules including rules that govern response by the patient to a certain disease;
obtaining in computer memory, by the processor, updated health attributes of the human patient, the updated health attributes being collected from the human patient subsequent to collection of the initial health attributes;
responsively generating in computer memory, by the processor, an updated patient model from the base model and the updated health attributes of the human patient, the updated patient model incorporating the patient-specific representation the rules and a representation of the updated health attributes;
generating in computer memory, by the processor, plural sets of parameters, each of the plural sets of parameters indicating respective variables having a causal relation to the health metrics;
automatically generating a computer precision medicine evaluation of the patient by executing the updated patient model using a deterministic model, said executing being performed by the processor and producing: (a) respective simulation of the health metrics of the updated patient model under the plural sets of parameters over a given simulated time period;
(b) plural resistance metric results based on the deterministic model, each of the plural resistance metric results indicating a change in the resistance metric under a respective set of parameters over the given simulated time period, and
(c) an occurrence probability of the updated patient model transitioning from an initial state to each of a plurality of successive states during the given simulated time period, each of the successive states corresponding to a respective one of the plural sets of parameters, wherein the produced respective simulations, plural resistance metric results, and occurence probability support the computer precision medicine evaluation of the patient with respect to at least the certain disease;
automatically determining a risk, by the processor, based on the produced simulations, the plural resistance metric results, and the occurrence probability of the generated computer precision medicine evaluation, the risk being of the updated patient model transitioning from the initial state to a diseased state, the diseased state representing the patient being afflicted by the certain disease;
automatically determining, by the processor, a diagnosis and a remedcy, the remedy being determined by 1) identifying a subset of the health metrics that are negatively correlated with the diseased state, and 2) identifying the remedy as one of the respective variables associated with the identified subset of health metrics, the remedy identifying a modification to the updated patient model that reduces the determined risk;
generating as clinically-actionable output, by the processor, a map indicating 1) the processor determined diagnosis of the patient based on the determined risk, and 2) the determined remedy, said generating output being in a manner controlling computer diagnosis of patient-specific outcomes; and
responsive to the clinically-actionable output, treating the patient in accordance with the remedy by administering to the patient one or more of: a medication, physical therapy, and other theraby.
2 . The method of claim 1 , wherein the updated health attributes in computer memory are based on input by a plurality of medical service parties.
3 . The method of claim 2 , further comprising:
enabling computer network access to a patient file by the plurality of medical service parties; and
generating the updated health attributes in computer memory based on modification of the patient file by the plurality of medical service parties.
4 . The method of claim 1 , wherein the updated health attributes are based on input by at least one medical service party over a duration of at least one year.
5 . The method of claim 1 , wherein the updated health attributes are based on a periodic input by at least one medical service party.
6 . The method of claim 1 , further comprising:
determining authorization status of a medical service party;
selectively enabling computer network access to a patient file by the medical service party based on the authorization status; and
generating the updated health attributes in computer memory based on modification of the patient file by the medical service party.
7 . The method of claim 6 , wherein the patient file is associated with the human patient via a patient identifier (ID).
8 . The method of claim 1 , wherein the base model incorporates rules that govern a response by the human being to at least one of immune reactivations, autoimmunity, preexisting conditions and environmental parameters.
9 . The method of claim 1 , further comprising:
identifying, by the processor, a correspondence between a subset of the updated health attributes, the produced plural resistance metric results, and the determined risk; and
determining, by the processor, the diagnosis based on the identified correspondence, the diagnosis indicating the subset of updated health attributes that exceed a threshold correspondence with the determined risk.
10 . The method of claim 1 , further comprising:
modifying, by the processor, the updated patient model in computer memory to incorporate the modification.
11 . The method of claim 10 , further comprising reporting the remedy to a user.
12 . The method of claim 10 , wherein determining the remedy includes the processor:
generating an additional set of parameters in computer memory, the additional set of parameters indicating a health intervention, the health intervention including at least one of a medication, a therapy, and a diet;
responsively modeling the health metrics of the patient model in computer memory under the additional set of parameters to generate a resistance metric result; and
automatically identifying the remedy based on the health metrics associated with the health intervention.
13 . The method of claim 1 , wherein:
the map relates the diseased state to 1) corresponding instances of the plural sets of parameters and 2) the occurrence probability of the updated patient model transitioning from the initial state to the plurality of successive states; and
further comprising the processor determining, based on the map, at least one risk for at least one of the successive states of the updated patient model in computer memory, the at least one risk defining a probability of an outcome of the updated patient model including the diseased state.
14 . The method of claim 1 , further comprising the processor generating in computer memory a lookup table cross-referencing states of the updated patient model to corresponding ones of the risk.
15 . The method of claim 14 , wherein determining the risk includes the processor accessing the lookup table in computer memory using information on a given state of the patient or the updated patient model.
16 . The method of claim 14 , further comprising the processor:
responsively analyzing a state of the patient; and
wherein determining the risk includes accessing the lookup table in computer memory using information on the state of the patient.
17 . The method of claim 14 , wherein the lookup table cross-references the states of the updated patient model to the corresponding ones of the risk and at least one remedy.
18 . The method of claim 1 , wherein the generating the updated patient model in computer memory includes the processor:
generating a dependency graph having a plurality of nodes representing the health attributes; and configuring dependencies between the plurality of nodes based on the rules.
19 . A non-transitory computer-readable medium comprising instructions that, when executed by a computer, provide a digital computer simulation of health of a human patient including causing the computer to:
obtain in computer memory, by a processor, a base model of a human being, the base model being a mathematical model incorporating rules that govern 1) a response by the human being to a disease, and 2) a relation between health metrics and the disease, the health metrics including a resistance metric indicating a measure of resistance to the disease;
obtain in computer memory, by the processor, initial health attributes of a human patient, the health attributes including at least one of a genetic trait, a blood lipid profile, and a representation of a medical imaging report;
responsively generate in computer memory, by the processor, an initial patient model from the base model and the initial health attributes of the human patient, the initial patient model incorporating a patient-specific representation of the rules of the base model and a representation of the initial health attributes, the patient-specific representation of the rules of the base model including rules that govern response by the patient to a certain disease;
obtain in computer memory, by the processor, updated health attributes of the human patient, the updated health attributes being collected from the human patient subsequent to collection of the initial health attributes;
responsively generate in computer memory, by the processor, an updated patient model from the base model and the updated health attributes of the human patient, the updated patient model incorporating the patient-specific representation of the rules and a representation of the updated health attributes;
generate in computer memory, by the processor, plural sets of parameters, each of the plural sets of parameters indicating respective variables having a causal relation to the health metrics;
automatically generate a computer precision medicine evaluation of the patient by executing the updated patient model using a deterministic model, said executing being performed by the processor and producing: (a) respective simulations of the health metrics of the updated patient model under the plural sets of parameters over a given simulated time period,
(b) plural resistance metric results based on the deterministic model, each of the plural resistance metric results indicating a change in the resistance metric under a respective set of parameters over the given simulated time period, and
(c) an occurrence probability of the updated patient model transitioning from an initial state to each of a plurality of successive states during the given simulated time period, each of the successive states corresponding to a respective one of the plural sets of parameters, wherein the produced respective simulations, plural resistance metric results, and the occurrence probability support the computer precision medicine evaluation of the patient with respect to at least the certain disease;
automatically determine a risk, by the processor, based on the produced simulations, the plural resistance metric results, and the occurrence probability, the risk being of the updated patient model transitioning from the initial state to a diseased state, the diseased state representing the patient being afflicted by the certain disease;
automatically determine, by the processor, a diagnosis and a remedy, the remedy being determined by 1) identifying a subset of the health metrics that are negatively correlated with the diseased state, and 2) identifying the remedy as one of the respective variables associated with the identified subset of health metrics, the remedy identifying a modification to the updated patient model that reduces the determined risk;
generate as clinically-actionable output, by the processor, a map indicating 1) the processor determined diagnosis of the patient based on the determined risk, and 2) the determined remedy, said generating output being in a manner controlling computer diagnosis of patient-specific outcomes; and
responsive to the clinically-actionable output, provide treatment to the patient in accordance with the remedy by administering to the patient one or more of: a medication, physical therapy, and other therapy.