Python Project: Stochastic Simulation of Patient Virus Population Dynamics And Treatment Regimens: Effect of Delaying Drugs, Using 2 Drugs, and Patient non-compliance

Links to Python Project: Stochastic Simulation of Patient Virus Population Dynamics: Effect of Delaying Drug Treatment on Patient Disease, Using 2 Drugs, and Patient non-Compliance

Stochastic Simulation of Patient Virus Population Dynamics: Effect of Delaying Drug Treatment on Patient Disease, Using 2 Drugs, Patient non-Compliance

We used the classes implemented in the previous similar problem set in this project list (ResistantVirus, TreatedPatient).

In this problem set, we examined the effect of delaying treatment on the ability of the drug to eradicate the virus population.

We needed multiple simulations to observe trends in the distributions of patient outcomes.

In the last part of the previous similar problem set, we ran a simulation that consisted of 150 time steps before adding drug guttagonol, followed by another 150 time steps.

Here we run the same simulation but this time for 300, 150, 75 and 0 time steps before adding guttagonol to the patient, and then ran another 150 time steps.

We used the same init params as problem set 8 (refer to that description).

For each case, we repeated the experiment for enough trials to gain reasonable insight into the expected result.

Rather than averaging the final virus population across different trials as in the previous problem set, we used pylab's hist() function to plot a histogram of the final virus populations under each case for each trial.

The x-axis of the histogram was the final virus population values, and the y-axis the number of trials belonging to each histogram bin.

We had to justify the number of trials we ran for each case in order to obtain a reasonable distribution.

We filled in the function simulationDelayedTreatment (numTrials) to perform this simulation. With a delay of 300 and 150 cases, "0-5%" were cured, for 75, "6-15%", for 0, "86-100%", so the earlier the drug was applied the more likely for this virus it was to be cured.

We then examined these same things with 2 drugs and found it cured the patient more quickly when administered early for this virus.

We also discussed how to model a patient that refuses to take medication. You make a TreatedPatient class that stochastically does not take its medication, and ir predicted less patients would be cured.