# Making Uncertainties Plannable and Processes Opitimized

Optimize Clinical Studies

A pharmaceutical company needs patients who are willing to participate in a study for a new drug. And so the uncertainties already begin with the first question: Will enough suitable people be able to be enrolled in the participating clinics during the planned study period? There is a lot of stochasticity in a process like this – because not only is the enrollment of people in clinical trials to some degree left to chance, but the arrival of samples for the study is also stochastic.

#### Goal: Predict Workload in the Laboratory

The »Art of Guessing«, the basis of stochastics, accordingly characterizes the work in this environment. On behalf of a pharmaceutical company, the team at Fraunhofer ITWM is looking for mathematical solutions to make the entire process, up to and including personnel scheduling, more predictable. If there were no uncertainties, the researchers could calculate the arrival of samples individually and determine the workload generated by the clinical trial.

The daily workload of those who process the samples is often uncertain because it depends on the number of incoming samples, some of which have to be processed within a certain period of time due to their shelf life. The more accurately it is predicted how many samples will arrive and when, the better the workload in the laboratory can be planned. The team led by Dr. Sandy Heydrich and Dr. Heiner Ackermann aims to use a mathematical model to provide answers to many questions, such as: Is the current staff sufficient for the workload that will arise?

#### Make Every Day More Predictable

The project partner's goal is optimized and robust staff scheduling based on likely incoming samples. To this end, the researchers are developing a simulation tool that provides a forecast of the number of sample arrivals for each study day. At the heart of the model is a Monte Carlo simulation to estimate the number of incoming samples at the production site for each day in the study time horizon.Different influencing parameters are taken into account, but also different scenarios are represented – such as the fluctuating number of patients. The variable level of detail of the model optimally supports the planning process in every phase.