Responding constructively to frustration

We all experience frustration. And ideally we develop ways to manage - to relieve the stress and to be constructive.

I get frustrated frequently as I move RB-190 down the path into clinical trials, particularly driven by the field's commitment to the status quo erecting barriers that seem to be getting wider and taller.

My primary coping mechanism is to focus on a project that needs to be completed - typically appearing relatively small and not very significant. But when enough of these small tasks are completed, they become in aggregate significant progress.

Focusing on manufacturing

So I turned recently to the ongoing manufacturing process. The obvious tasks are things including ordering the ingredients needed to make the drug and the excipients to turn the drug into tablets (excipients are inert chemicals that are proven safe that make the drug stick together as a tablet and also allow the tablet to dissolve quickly in your stomach, for example). The manufacturer takes care of that, just as they take care to clean the equipment fully to guarantee that the last drug they made won't get into RB-190 tablets.

One task needed is to design the labels that will go on the bottles of tablets. The bottles will contain tablets in one of three strengths: 40 mg, 160 mg or 320 mg. Other bottles will have tablets that look exactly the same as the 40mg, 160 mg and 320 mg tablets - but have no active drug (a placebo). During the trial we'll need to hand out the right bottles to the right people. It would not be acceptable to make a mistake by sending active drug at a high dose to a study participant who was receiving placebo. The data say RB-190 should be started low and gradually increased - not abruptly introduced (intentionally or accidentally) at a high dose.

So I spent a day developing a process, a way to code the bottles (including so that there were no characters that could be confused, such as 0 - O, 1 - I and 5 - S). I did it manually and then used perplexity.ai to do the same. That's where I got the idea to create codes without characters that can be confused. I shared it with the manufacturer. Task completed.

Integrating AI into our process

Last week I turned my attention to complete a task about which I was very excited. A company figured out how to use a commercially available large language model (a.k.a., AI platform) to read data files and draft a regulatory document called an IND (Investigational New Drug application). An IND is required to be allowed by FDA so that drug can be shipped to investigative sites and a clinical trial can commence.

I did some preliminary work with the company to test their system and it was pretty impressive. It wouldn't write a document that could be submitted to the FDA, but it seemed capable of producing a really good first draft. So I engaged with them to make the commitment.

With this task, I failed. We could not come to agreement on the terms of the contract.

The company was not willing to accept the amount of liability I required in the event that a breach of the contract were to result in loss of confidentiality of our trade secret information. Because they rely on a large AI platform, I needed to be confident that our information would not be stored by that 3rd party and could not be used to train its platform.

At first, it appeared they were resistant to assume liability for a breach of cybersecurity of the AI platform. In today's world, everyone is at risk of cyberattack and it made sense that they did not want to assume this liability. So I offered to carve that out, unless such a breach were the result of negligent or criminal acts. But that was not enough.

It seemed that this company decided it would not accept a contract without a very low cap on liability in order to protect itself. I understood. But if there were a breach of confidentiality that compromised the likelihood of success of Right Brain Bio, I couldn't tell my investors with a straight face that I allowed that risk to exist - to compromise our trade secrets - with no liability provision to protect their investment or the potential return on their investment. I needed to protect Right Brain Bio.

So unfortunately, I could not bring in this AI tool into our process. We'll just rely on old-fashioned human intelligence.

Shifting from frustration to task completion

But the key take-away is the importance of pushing my frustration aside and making sure it did not control me or slow me down. Instead, I responded constructively to the frustration by completing important tasks. And we remain on the path to start dosing in 2026. Apologies that I am not committing to a specific date or month, as there is still a wide range of start dates that we are working hard to make more specific - and sooner. Stay tuned!