Developing new drugs is notoriously slow and expensive. Leaders say 4 key strategies will speed up the process.

During the COVID-19 pandemic, innovations such as vaccines created in record time have challenged the health care industry to move more quickly.

Leaders in health care recently discussed solutions for speeding up the therapeutic process at Fortune‘s Brainstorm Health conference.

“I think the technology is moving faster than the clinical validation we can do for them. We have to ensure we don’t get left behind,” said Dr. Samit Hirawat, executive vice president and chief medical officer at Bristol Myers Squibb.

Hirawat appeared on a panel with Dr. Sanjiv Patel, president and CEO of Relay Therapeutics, and Dr. Harold Bernstein, chief medical officer and head of global clinical development at BioMarin.

They identified four key factors they believe will accelerate innovations in health.

1. Putting patients first

By focusing on diverse patients’ needs earlier in the process, biopharmaceutical companies can get better data that will help inform new drugs and treatments.

At BioMarin, which focuses on treatments for rare genetic diseases, the team only pursues research for treatments that would drastically improve patients’ lives.

“It’s only worth our time if we’re going to really transform patients’ lives. How they feel in the world, how they function, how they survive,” Bernstein said.

Bristol Myers Squibb has started speaking with patients and patient-advocacy groups before clinical trials to make the process more comfortable.

The company is also focused on expanding into underserved areas, Hirawat said.

“Engaging those extended communities we can gain more patients and increase the diversity of clinical trials,” he said.

2. Work with regulators

Working on new processes with regulators like the Food and Drug Administration is essential for accelerating innovation in health care, Hirawat emphasized.

“The technology is there. Can it be directly applied? No because we are behind in validation,” he said, adding that faster regulatory approval is key for innovation.

At Bristol, involving regulators earlier in the process is leading to more efficiency, he said.

Bernstein said a breakthrough in cystic fibrosis treatments shows how regulatory processes can be improved.

“When drugs were being developed for cystic fibrosis, you initially needed to run a trial for genetic mutations for the drug’s use and to get it paid for,” he said. “Recently, the FDA approved expansion based on a translational, in-vitro asset. Instead of running a trial at significant cost and delaying getting it to patients, we could simply show it, and get it cleared.”

Examples like this are “where we want to see the industry go,” he said.

3. Share knowledge

BioMarin is using partnerships to compile large datasets from other companies and academic studies, Bernstein said.

“Innovations can come from anywhere,” Hirawat said. “We want to work with other companies through acquisitions, mergers, collaborations, and discovery to get their perspectives.”

By combining data and research from multiple sources, researchers can reach conclusions more quickly and bring successful treatments to market.

“Other small companies or partnerships might give us some way of solving a problem,” he said.

4. Leverage machine learning and A.I.

Implementing A.I. is leading to advancements in genome sequencing, Patel said.

His firm, Relay Therapeutics, focuses on precision oncology and treatments for cancer based on the DNA of patients.

“In the past, lung cancer was lung cancer, and breast cancer was breast cancer,” Patel said. “Everyone received the same treatments. If you understand where your therapy will work you’ll have a much higher success rate.”

A.I. is helping craft treatments by studying pictures of proteins, and making connections between genetic factors that influence effective treatment and outcomes.

“Over the coming decades we’ll see these pieces get much better,” Patel said. “Hopefully it won’t take 10 to 12 years to develop a treatment, won’t cost $2 billion, and will lower the failure rate.”