Predicting risk of serious gastrointestinal complications in patients taking nonsteroidal anti-inflammatory drugs

• US 6,560,541 B1
• Filed: 11/23/1999
• Issued: 05/06/2003
• Est. Priority Date: 04/23/1998
• Status: Expired due to Fees

• Current Assignee: Board of Trustees of the Leland Stanford Junior University (Stanford University)
• Original Assignee: Board of Trustees of the Leland Stanford Junior University (Stanford University)
• Inventors: Singh, Gurkirpal
• Primary Examiner(s): Allen, Marianne P.
• Assistant Examiner(s): Clow, Lori A.
• Application Number:
  US09/447,963
• Time in Patent Office: 1,260 Days
• Field of Search: 700/49, 700/90, 702/179, 702/19, 705/2, 707/104, 128/920, 128/923
• US Class Current: 702/19
• CPC Class Codes: G06F 17/18 for evaluating statistical data , e.g. average values, frequency distributions, probability functions, regression analysis forecasting specially adapted for a specific administrative, business or logistic context G06Q10/04
• CPC Class Codes: G16H 15/00 ICT specially adapted for medical reports, e.g. generation or transmission thereof
• CPC Class Codes: G16H 20/10 relating to drugs or medications, e.g. for ensuring correct administration to patients
• CPC Class Codes: G16H 50/30 for calculating health indices; for individual health risk assessment
• CPC Class Codes: G16H 50/50 for simulation or modelling of medical disorders
• CPC Class Codes: Y10S 128/92 Computer assisted medical diagnostics
• CPC Class Codes: Y10S 128/923 by comparison of patient data to other data

Abstract

A computer-implemented method for calculating estimated risk of serious gastrointestinal (GI) hospitalization of a patient taking nonsteroidal anti-inflammatory drugs includes the steps of obtaining patient values of a plurality of predictive factors, calculating an estimated risk from the patient values, and displaying the estimated risk, preferably on a monitor. The predictive factors may be age, global health status, proportion of time taking prednisone, occurrence of a previous GI side effect, and occurrence of a previous GI hospitalization. Preferably, the estimated risk is calculated from a Cox proportional hazard model, derived from a database of arthritis patients, using either continuous or categorical variables. The continuous model applies a standard Cox model to the predictive factors. The categorical model divides each predictive factor into categories, and assigns a point value to each predictive factor. The point values are added to derive a point total, which is converted to an estimated risk using a predetermined correlation between point total and estimated risk.

Claims

1. A computer-implemented method for determining an estimated risk R of serious gastrointestinal (GI) hospitalization of a patient taking non-steroidal anti-inflammatory drugs, said method comprising the steps of:

a) obtaining patient values of predictive factors x₁, . . . , x₅ wherein said predictive factors are age, baseline global health status, proportion of time taking prednisone, previous occurrence of a GI side effect and previous occurrence of GI hospitalization;

b) calculating said estimated risk R from a model applied to said patient values, wherein said model is a Cox proportional hazard model that is based on longitudinal data, wherein said estimated risk R is calculated according to R=1−

S(1)^B, wherein;

S(1) is a mean probability of not having a serious GI hospitalization within one year, and

B=e^A, A=L−

M, $M=\underset{i=1}{\overset{5}{\sum <br><br>}}<br><br>\text{\hspace{1em}<br><br>}<br><br>a_i<br><br>m_i,\text{\hspace{1em}<br><br>}<br><br>\mathrm{and}<br><br>\text{\hspace{1em}<br><br>}<br><br>L=\underset{i=1}{\overset{5}{\sum <br><br>}}<br><br>\text{\hspace{1em}<br><br>}<br><br>a_i<br><br>x_i,$

 

wherein m_i are a plurality of predetermined mean values of said predictive factors, and a_i are a plurality of Cox coefficients, wherein S(1) is 0.99045, wherein said plurality of predetermined mean values are m₁ is 56.82 years, m₂ is 41.04, m₃ is 0.374, m₄ is 0.263, and m₅ is 0.013, and wherein said plurality of Cox coefficients are a₁ is 0.050/years, a₂ is 0.010, a₃ is 1.109, a₄ is 0.373, and a₅ is 1.957; and

c) displaying said estimated risk R on a display wherein said estimated risk R is displayed as a quantitative or qualitative measure.

2. A computer-implemented method for determining an estimated risk R of serious gastrointestinal (GI) hospitalization of a patient taking non-steroidal anti-inflammatory drugs, said method comprising the steps of:

a) obtaining patient values of predictive factors x₁, . . . , x₅ wherein said predictive factors are age, baseline global health status, proportion of time taking prednisone, previous occurrence of a GI side effect and previous occurrence of GI hospitalization;

b) calculating said estimated risk R from a model applied to said patient values, wherein said model is a Cox proportional hazard model that is based on longitudinal data wherein said step of calculating said estimated risk R comprises the step of;

i) assigning a point value to each of said predictive factors, wherein said point value is related to said patient value of said predictive factor;

ii) adding said point values for each of said predictive factors to obtain a point total P; and

iii) converting said point total P to said estimated risk R, wherein said step of converting uses a correlation between said point values and said estimated risk R or a point total P—

risk table, wherein said R=1−

S(1)^B, wherein S(1) is 0.99049, B=e^A, A=L−

M, M is 2.7662, and L equals 0.237*P; and

c) displaying said estimated risk R on a display wherein said estimated risk R is displayed as a quantitative or qualitative measure.

3. A computer-implemented method for determining an estimated risk R of serious gastrointestinal (GI) hospitalization of a patient taking non-steroidal anti-inflammatory drugs, said method comprising the steps of:

a) obtaining patient values of predictive factors x₁, . . . , x₅ wherein said predictive factors are age, baseline global health status, proportion of time taking prednisone, previous occurrence of a GI side effect and previous occurrence of GI hospitalization, wherein said predictive factors are point values p₁, . . . , p₅ from said patient of said predictive factors x₁, . . . , x₅, wherein;

p₁ is;

0 when said age is below 20,

1 when said age is between 21 and 25,

2 when said age is between 26 and 30,

3 when said age is between 31 and 35,

4 when said age is between 36 and 40,

5 when said age is between 41 and 45,

6 when said age is between 46 and 50,

7 when said age is between 51 and 55,

8 when said age is between 56 and 60,

9 when said age is between 61 and 65,

10 when said age is between 66 and 70,

11 when said age is between 71 and 75,

12 when said age is between 76 and 80,

13 when said age is between 81 and 85, and

14 when said age is above 85;

p₂ is;

0 when said baseline global health status is between 91 and 100,

1 when said baseline global health status is between 51 and 90,

2 when said baseline global health status is between 36 and 50,

3 when said baseline global health status is between 11 and 35, and

4 when said baseline global health status is between 0 and 10;

p₃ is;

0 when said number of months taking prednisone is 0,

1 when said number of months taking prednisone is between 1 and 3,

2 when said number of months taking prednisone is between 4 and 6,

3 when said number of months taking prednisone is between 7 and 10, and

4 when said number of months taking prednisone is between 11 and 12;

p₄ is;

0 when said patient potential previous occurrence of a GI side effect is no, and

2 when said potential previous occurrence of a GI side effect is yes; and

p₅ is;

0 when said potential previous occurrence of GI hospitalization is no, and

8 when said potential previous occurrence of GI hospitalization is yes;

b) calculating said estimated risk R from a model applied to said patient values, wherein said model is a Cox proportional hazard model that is based on longitudinal data; and

c) displaying said estimated risk R on a display wherein said estimated risk R is displayed as a quantitative or qualitative measure.