Apparatus, method, and computer-readable medium for determining a drug for manufacture

1. An apparatus for determining a drug for manufacture, the apparatus being communicably coupled via a network to a manufacturing device, the apparatus comprising:

processing circuitry configured to

receive input data related to candidate drug programs, the input data describing, for each candidate drug program, a candidate drug, a disease indication, and a geo location associated with a development of the candidate drug,

acquire data from a database based upon the input data, wherein the acquired data comprises chronological data and qualitative data of historical drug programs related to each candidate drug program, the qualitative data describing characteristics of clinical trials associated with the historical drug programs,

generate, for each candidate drug program, at least one success model and at least one timeline model based on the acquired data from the database related to the historical drug programs, each of the at least one success model and the at least one timeline model being related to chronological events of the clinical trials associated with the historical drug programs, the chronological events being one or more milestone dates that define the clinical trials, wherein

each of the at least one success model is a logistic regression model including, as parameters thereof, a subset of characteristics of a respective clinical trial determined to positively affect a likelihood function defining the logistic regression model, and

each of the at least one timeline model is a proportional hazard model including, as parameters thereof, a subset of characteristics of a respective clinical trial determined to positively affect a likelihood function defining the proportional hazard model,

determine, from a composite of the at least one success model and a corresponding composite of the at least one timeline model for each candidate drug program, one or more outputs related to a corresponding chronological event for each of the candidate drug programs,

select, based upon the determined one or more outputs for each of the candidate drug programs, one of the candidate drug programs for manufacture, and

transmit, to the manufacturing device via the network, manufacturing information related to the manufacture of the candidate drug associated with the selected one of the candidate drug programs, the manufacturing device beginning the manufacture of the candidate drug upon receipt of the manufacturing information from the apparatus, wherein

the composite of the at least one success model predicts a probability that a respective candidate drug program will reach a particular milestone date of a respective clinical trial, and

the corresponding composite of the at least one timeline model predicts a time required for the respective candidate drug program to reach the particular milestone date of the respective clinical trial.

2. The apparatus according to claim 1 , wherein the processing circuitry is configured to select the selected one of the candidate drug programs for manufacture by

comparing, as the one or more outputs for the candidate chug programs, a success probability that each candidate drug program will reach the particular milestone and a time required for each candidate drug program to reach the particular milestone.

3. The apparatus according to claim 2 , wherein the processing circuitry is configured to select, based on the comparing and as the selected one of the candidate drug programs for manufacture, a candidate drug program that maximizes a success probability that a respective candidate drug program will reach a particular milestone.

4. The apparatus according to claim 1 , wherein coefficients of the parameters of the logistic regression model are optimized via a maximum likelihood estimator.

5. The apparatus according to claim 1 , wherein coefficients of the parameters of the proportional hazard model are optimized via a maximum likelihood estimator.

6. A method for determining a drug for manufacture, comprising:

receiving, by processing circuitry, input data related to candidate drug programs, the input data describing, for each candidate drug program, a candidate drug, a disease indication, and a geo location associated with a development of the candidate drug;

acquiring, by the processing circuitry, data from a database based upon the input data, wherein the acquired data comprises chronological data and qualitative data of historical drug programs related to each candidate drug program, the qualitative data describing characteristics of clinical trials associated with the historical drug programs;

generating, by the processing circuitry and for each candidate drug program, at least one success model and at least one timeline model based on the acquired data from the database related to the historical drug programs, each of the at least one success model and the at least one timeline model being related to chronological events of the clinical trials associated with the historical drug programs, the chronological events being one or more milestone dates that define the clinical trials, wherein

each of the at least one success model is a logistic regression model including, as parameters thereof, a subset of characteristics of a respective clinical trial determined to positively affect a likelihood function defining the logistic regression model, and

each of the at least one timeline model is a proportional hazard model including, as parameters thereof, a subset of characteristics of a respective clinical trial determined to positively affect a likelihood function defining the proportional hazard model;

determining, by the processing circuitry, from a composite of the at least one success model and a corresponding composite of the at least one timeline model for each candidate drug program, one or more outputs related to a corresponding chronological event for each of the candidate drug programs;

selecting, by the processing circuitry, based upon the determined one or more outputs for each of the candidate drug programs, one of the candidate drug programs for manufacture; and

transmitting, by the processing circuitry, to a manufacturing device via a network, manufacturing information related to the manufacture of the candidate drug associated with the selected one of the candidate drug programs, the manufacturing device beginning the manufacture of the candidate drug upon receipt of the manufacturing information, wherein

the composite of the at least one success model predicts a probability that a respective candidate drug program will reach a particular milestone date of a respective clinical trial, and

the corresponding composite of the at least one timeline model predicts a time required for the respective candidate drug program to reach the particular milestone date of the respective clinical trial.

7. The method according to claim 6 , wherein the selecting, by the processing circuitry, further comprises

comparing, as the one or more outputs for the candidate drug programs, a success probability that each candidate drug program will reach the particular milestone and a time required for each candidate drug program to reach the particular milestone.

8. The method according to claim 7 , wherein the selecting selects, based on the comparing as the selected one of the candidate drug programs for manufacture, a candidate drug program that maximizes a success probability that a respective candidate drug program will reach a particular milestone.

9. The method according to claim 6 , wherein coefficients of the parameters of the logistic regression model are optimized via a maximum likelihood estimator.

10. The method according to claim 6 , wherein coefficients of the parameters of the proportional hazard model are optimized via a maximum likelihood estimator.

11. A non-transitory computer-readable storage medium storing computer-readable instructions that, when executed by a computer, cause the computer to perform a method of determining a drug for manufacture, comprising:

receiving input data related to candidate drug programs, the input data describing a candidate drug, a disease indication, and a geo location associated with a development of the candidate drug;

acquiring data from a database based upon the input data, wherein the acquired data comprises chronological data and qualitative data of historical drug programs related to each candidate drug program, the qualitative data describing characteristics of clinical trials associated with the historical drug programs;

generating, for each candidate drug program, at least one success model and at least one timeline model based on the acquired data from the database related to the historical drug programs, each of the at least one success model and the at least one timeline model being related to chronological events of the clinical trials associated with the historical drug programs, the chronological events being one or more milestone dates that define the clinical trials, wherein

each of the at least one success model is a logistic regression model including, as parameters thereof, a subset of characteristics of a respective clinical trial determined to positively affect a likelihood function defining the logistic regression model, and

each of the at least one timeline model is a proportional hazard model including, as parameters thereof, a subset of characteristics of a respective clinical trial determined to positively affect a likelihood function defining the proportional hazard model;

determining, from a composite of the at least one success model and a corresponding composite of the at least one timeline model for each candidate drug program, one or more outputs related to a chronological event for each of the candidate drug programs;

selecting, based upon the determined one or more outputs for each of the candidate drug programs, one of the candidate drug programs for manufacture; and

transmitting, to a manufacturing device via a network, manufacturing information related to the manufacture of the candidate drug associated with the selected one of the candidate drug programs, the manufacturing device beginning the manufacture of the candidate drug upon receipt of the manufacturing information, wherein

the composite of the at least one success model predicts a probability that a respective candidate drug program will reach a particular milestone date of a respective clinical trial, and

the corresponding composite of the at least one timeline model predicts a time required for the respective candidate drug program to reach the particular milestone date of the respective clinical trial.

12. The non-transitory computer-readable storage medium according to claim 11 , wherein the selecting further comprises

comparing, as the one or more outputs for the candidate chug programs, a success probability that each candidate drug program will reach the particular milestone and a time required for each candidate drug program to reach the particular milestone.

13. The non-transitory computer-readable storage medium according to claim 12 , wherein the selecting selects, based on the comparing as the selected one of the candidate drug programs for manufacture, a candidate drug program that maximizes a success probability that a respective candidate drug program will reach a particular milestone.

14. The non-transitory computer-readable storage medium according to claim 11 , wherein coefficients of the parameters of the logistic regression model are optimized via a maximum likelihood estimator.

15. The non-transitory computer-readable storage medium according to claim 11 , wherein coefficients of the parameters of the proportional hazard model are optimized via a maximum likelihood estimator.