The “blue sky” approach will use a system of hundreds of off-the-shelf sensors and mobile devices to provide real-time, around-the-clock disease symptom information to researchers and doctors. It will then leverage IBM’s machine learning capabilities to find unique connections between recorded symptoms and other clinical data, such as timing and dosing of medicine.

“The idea was how can we take the digital revolution, continuous monitoring, and use that to improve the quality of clinical trials and care delivery of therapeutics,” said Peter Bergethon, head of quantitative medicine for Pfizer. “We want to define the digital signature of a person, then characterize how they’re feeling and how they’re responding to medication.”

The research facility will be setup near IBM’s Watson Research Center in Yorktown Heights, NY. It will be designed to look like a regular house but underneath is a network of connected sensors that can collect a data treasure trove. The sensors will measure things like motor function, cognition, sleep and daily activities such as grooming, dressing, and eating.

Pfizer and IBM will rotate in as many as 200 participants, both those with Parkinson’s disease and control subjects who don’t have the neurological condition, who will live in the space for a period of time and produce reams information from these sensors. By receiving a constant flow of data from a real-life scenario and linked with a clinical assessment, doctors can have a better idea of how a patient is responding to a medicine and how his or her disease is progressing.

“Until now, health care has only been episodic. Your caregiver can only get data during that episode of interaction during your appointment,” said Ajay Royyuru, director of health care and life sciences research at IBM. “We could be changing that with continuous measurement and improved quantitative data. It takes personalized care to a whole new level.”

The research project will run over the next two to three years, said Royyuru. The end goal is multi-fold. For one, IBM wants to figure out exactly which sensors provide the right insight for Parkinson’s patients. This could then be put into patient’s home to provide real-time data to doctors so they can provide more holistic care rather than relying on a patient’s self-reported symptoms.

For Pfizer, finding this right array of sensors can help speed up clinical trials of new Parkinson’s drugs–or even find better ways to use existing treatments. The feedback from sensors, along with the power of cognitive computing, could highlight symptoms that indicate a patient should get a higher or lower dose of a medicine.

As for new medicines, Pfizer is already hoping to use this IBM-enabled setup to test its newest Parkinson’s drug that’s in the pipeline. Pfizer is planning to bring its compound currently known as PF-06649751 to a phase 3 trial in 2019 using this connected approach.

“We want to demonstrate the direct value of this drug and show how it improves movement,” said Bergethon. “We are focused on demonstrating with real world data that this drug helps not just while you’re at the doctor but all of the time.”

Ultimately, the research setup will also be used for research projects in other degenerative diseases to show how new treatments can improve daily wellbeing. Pfizer is hoping to expand this into diseases like Huntington’s disease, Alzheimer’s, chronic heart failure, and even oncology-related pain management.

“We’re looking for an infrastructure and model that can apply across the sphere of health care,” said Bergethon.