Computational Modeling Can Bring Better, Cheaper Medicines to a Pharmacy Near You

Check out the two parts of my blog series on the impact of the research that I do in computational modeling for pharmaceutical device design and manufacturing. See my previous blog post for my reflection on communicating my science and engineering to others.

Part 1: Reducing Pharmaceutical Production Costs Using Continuous Manufacturing and Computational Modeling

Part 2: Computational Modeling to Design Pharmaceuticals that Require Fewer Dosages for Same Treatment

Reducing Pharmaceutical Production Costs Using Continuous Manufacturing and Computational Modeling

Continuous manufacturing is the typical industrial practice characterized by a flow of materials through an assembly line often operating 24/7 to produce a product. I have been familiar with continuous manufacturing most of my life: my dad builds tires in 12-hour rotating shifts at a Goodyear factory, and I spent a summer as a chemical engineering intern in continuous manufacturing for refining oil into fuels. A striking exception to the rule of using continuous manufacturing to produce chemical products is the pharmaceutical industry, which uses batch manufacturing. Like bakeries that bake many batches of different pastries using regular or jumbo-sized kitchen equipment like mixers and ovens to accommodate the varied tastes of many customers, the pharmaceutical industry uses jumbo-sized organic chemistry lab techniques for producing batches of specialty chemicals with a range of demands. These techniques do not necessarily scale well to larger production. The differences between batch and continuous manufacturing can be illustrated by sub sandwiches. Subway’s highly successful marketing campaign for the $5 foot long has made them the go-to restaurant for affordable, made-to-order sandwiches. Subway uses an assembly line to create customized sub sandwiches in bulk (continuous manufacturing). One of my friends on a tight budget did an experiment in the economics of creating a sandwich at home (batch manufacturing) with quality on par with Subway. His question: is it cost effective to make your own sandwich versus buying Subway’s foot long? His answer was a resounding no. Not only did it cost a lot for the groceries, but he had to go shopping, prepare the veggies, and think of ways to eat all the groceries before they spoiled or eat the same type of sandwich repeatedly. (A similar case study is available online.) Compared to continuous manufacturing the batch method of manufacturing a sandwich has more expensive raw ingredients, can create a lot of waste, and takes more time. Also, product quality may vary widely between batches. The same things are typical of the pharmaceutical industry. The Novartis-MIT Center for Continuous Manufacturing (CCM) focuses on shifting the pharmaceutical industry to the more cost effective practice of continuous manufacturing (for more on the CCM click here and here).

In my research in the CCM, I have developed a computational model to describe a chemical assembly line for making pharmaceuticals. We add medicines and flavors to a liquid mixture, which is spread on a conveyor belt. This thin liquid layer has to be carefully dried to form an edible film like Listerine Pocketpaks breath strips or Fruit Roll-Ups before it leaves the equipment, and the film must have uniform concentrations of medicine and precise physical properties. I use mathematical descriptions of the time-varying processes involved in drying the film (heating, evaporation, motion of water molecules, and movement along the conveyor belt) to aid in design and quality control of the system. We use computational tools to monitor changes to the input conditions and undesirable disturbances, quickly calculate how these changes should affect the output using my model, and make corrections to the temperatures or speeds within the equipment to ensure that the final product reliably meets its quality specifications, preventing waste of materials, time, and energy. Thus, fabrication of high quality pharmaceuticals through continuous manufacturing is enabled by computational science and applied mathematics and should lower production costs.

Computational Modeling to Design Pharmaceuticals that Require Fewer Dosages for Same Treatment

Controlled release drug delivery is a category of pharmaceutical dosage techniques where the level of medicine in a patient's body is sustained for a long period of time with a single dose. Examples of controlled release delivery include skin patches for nicotine or estrogen that work for days or weeks and intrauterine birth control devices that release a constant supply of hormones for years. These devices are wonderful improvements over traditional dosages forms like pills or injections that require repeated doses. The daily birth control pill is an example of a traditional dosage form that has life-altering consequences for a single forgotten dose. Controlled release drug delivery alleviates pressure on patients to remember every one of the frequent doses. For a chronically forgetful person like me, this is great news. Currently, controlled release drug delivery is only available for a few medicines and generally involves a material (usually plastic) that has to be removed from the skin or from inside the body after all the medicine has been released. I study biodegradable materials that decompose into natural products that are not harmful to the body and do not require removal. The hope is that many types of medications could be delivered with these materials to improve patient convenience for care of chronic conditions and mental illnesses with major consequences of forgetfulness such as schizophrenia and Alzheimer's disease.

I develop mathematical models to design how medicines are released from biodegradable polymer spheres that contain medicine. I use computational experiments to explore conditions that may give the optimal design. Thomas Edison was a prolific American inventor credited with the invention of the light bulb, phonograph, and motion picture camera. Just imagine Edison's impact if he had mathematical models or computer simulations that predicted the best design, saving him hundreds of experiments. Then he might have been able to be as productive as Nikola Tesla (see The Oatmeal for a striking, humorous comparison of Edison and Tesla).

My models combine the biodegradation chemical reactions and transport of the medicine as the spheres dissolve. Because many conditions change in an interdependent manner, the equations describing them cannot be solved through simple calculations. Instead, I use advanced computational recipes or algorithms for accurately approximating the solutions to equations in the model. More computational power enables more sophisticated algorithms, better spatial resolution, and increases in the length of time that can be simulated. These contribute to more accurate predictions for the timing of drug release, which in turn enable more realistic computational experiments to identify best designs for the spheres.