Identifying risk of a medical event
First Claim
1. A method for identifying risk of a medical outcome based on an analysis of multiple risk factors, comprising:
- providing a reference database comprising a plurality of reference data points, each reference data point comprising a value for each of n risk-factor parameters, wherein n≧
2;
wherein each reference data point further comprises a corresponding outcome value; and
wherein the outcome value is determined by the prior occurrence or non-occurrence of a medical outcome;
mapping each reference data point as a map point in (n+1)-dimensional space, thereby producing a reference map;
wherein the position of each map point in n dimensions is determined by the values of the associated reference data point'"'"'s n risk-factor parameters; and
wherein the position of each map point in the (n+1)th dimension is determined by the outcome value for the associated reference data point;
applying an interpolation algorithm to the reference map to produce an interpolated (n+1)-dimensional map;
wherein the interpolated (n+1)-dimensional map comprises a surface that defines a function that maps values of the n risk-factor parameters to predicted outcome values; and
wherein each of the predicted outcome values represents at least one of a relative risk and a probability of occurrence of the medical outcome;
mapping a query data point, comprising values for each of the n risk-factor parameters, onto the interpolated (n+1)-dimensional map;
wherein the location of the query data point on the interpolated (n+1)-dimensional map is determined, at least in part, by the query data point'"'"'s values of the n risk-factor parameters;
determining a predicted outcome value for the query data point by mapping, with the function, the query data point'"'"'s values of the n risk-factor parameters to a point in the surface, the point in the surface being indicative of the predicted outcome value for the query data point; and
outputting at least one of the query data point'"'"'s predicted outcome value, its associated relative risk, and its associated probability, to an output device.
1 Assignment
0 Petitions
Accused Products
Abstract
Described is a method of determining the relative risk of an outcome based on an analysis of multiple risk factors. A graphical method is used to take values corresponding to risk parameters and an event outcome to produce a smoothed surface map representing relative risk over an entire space defined by n risk factors. Applying a query data point to the surface map permits the determination of the estimated outcome probability for the query data point, based on its location on the surface map. The method then reports a relative risk or other probability measure associated with the query data point. Also described is a method of analysis in which subpopulations previously identified as high-risk can be further analyzed with respect to risk posed by additional factors.
-
Citations
37 Claims
-
1. A method for identifying risk of a medical outcome based on an analysis of multiple risk factors, comprising:
-
providing a reference database comprising a plurality of reference data points, each reference data point comprising a value for each of n risk-factor parameters, wherein n≧
2;
wherein each reference data point further comprises a corresponding outcome value; and
wherein the outcome value is determined by the prior occurrence or non-occurrence of a medical outcome;
mapping each reference data point as a map point in (n+1)-dimensional space, thereby producing a reference map;
wherein the position of each map point in n dimensions is determined by the values of the associated reference data point'"'"'s n risk-factor parameters; and
wherein the position of each map point in the (n+1)th dimension is determined by the outcome value for the associated reference data point;
applying an interpolation algorithm to the reference map to produce an interpolated (n+1)-dimensional map;
wherein the interpolated (n+1)-dimensional map comprises a surface that defines a function that maps values of the n risk-factor parameters to predicted outcome values; and
wherein each of the predicted outcome values represents at least one of a relative risk and a probability of occurrence of the medical outcome;
mapping a query data point, comprising values for each of the n risk-factor parameters, onto the interpolated (n+1)-dimensional map;
wherein the location of the query data point on the interpolated (n+1)-dimensional map is determined, at least in part, by the query data point'"'"'s values of the n risk-factor parameters;
determining a predicted outcome value for the query data point by mapping, with the function, the query data point'"'"'s values of the n risk-factor parameters to a point in the surface, the point in the surface being indicative of the predicted outcome value for the query data point; and
outputting at least one of the query data point'"'"'s predicted outcome value, its associated relative risk, and its associated probability, to an output device. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36)
-
-
37. A method, of identifying or estimating risk of a medical outcome in a subpopulation of patients, comprising:
-
providing a first plurality of reference data points representing a population of patients, each reference data point comprising a value for each of n risk-factor parameters, wherein n≧
2;
wherein each reference data point further comprises a corresponding outcome value; and
wherein the outcome value is determined by the prior occurrence or non-occurrence of a medical outcome;
mapping each reference data point as a map point in (n+1)-dimensional space, thereby producing a reference map;
wherein the position of each map point in n dimensions is determined by the values of the associated reference point'"'"'s n risk-factor parameters; and
wherein the position of each map point in the (n+1)th dimension is determined by the outcome value for the associated reference data point;
applying an interpolation algorithm to the reference map to produce an interpolated (n+1)-dimensional map;
wherein the interpolated (n+1)-dimensional map comprises a surface that defines a function that maps values for the n risk-factor parameters to predicted outcome values; and
wherein each of the predicted outcome values represents at least one of a relative risk and a probability of occurrence of the medical outcome;
locating on the interpolated (n+1)-dimensional map a first subpopulation of reference data points corresponding to a first subpopulation of patients within the population of patients, the first subpopulation of reference data points representing an increased or decreased risk of the medical outcome relative to another subpopulation of patients within the population;
providing a plurality of subpopulation data points corresponding to the first subpopulation of patients, each subpopulation data point comprising a value for each of m risk-factor parameters, wherein m≧
2;
wherein each subpopulation data point further comprises a corresponding subpopulation outcome value; and
wherein the subpopulation outcome value is determined by the prior occurrence or non-occurrence of the medical outcome;
mapping each subpopulation data point as a map point in (m+1)-dimensional space, thereby producing a second reference map;
wherein the position of each map point in m dimensions is determined by the values of the associated subpopulation data point'"'"'s m risk-factor parameters; and
wherein the position of each map point in the (m+1)th dimension is determined by the subpopulation outcome value for the associated subpopulation data point;
applying an interpolation algorithm to the second reference map to produce an interpolated (m+1)-dimensional map;
wherein the interpolated (m+1)-dimensional map comprises a second surface that defines a second function that maps values for the m risk-factor parameters to subpopulation predicted outcome values; and
wherein each of the subpopulation predicted outcome values represents at least one of a subpopulation relative risk and a subpopulation probability of occurrence of the medical outcome; and
outputting at least one of a subpopulation predicted outcome value, a subpopulation relative risk, and a subpopulation probability of occurrence of the medical outcome, to an output device.
-
Specification