The Hospital Pass

As a kid I played Gaelic football. I was a pretty average player, especially on the ball, so I typically found it best to move it along quickly. One day, during tryouts for the school team, I made a terrible pass. In terms of accuracy, and selection of player (the best one on the team), it was fine. I aimed it high, and he had to jump for it, but that wasn't a problem in itself. The problem was that when he landed he was swarmed by three of the opposing team. After a struggle, the whistle was blown.

The coach later labelled my pass a "hospital pass". The three that swarmed my teammate hadn't come from afar, in fact, they were already crowded around him; this was why I chose to aim high in the first place. It was a pass that could put my teammate in the hospital: it landed him in a position where he would immediately have to fight for his life.

I didn't get selected for the team. Many a reader will be thinking "of course not" because this lesson about when to pass the ball is pretty damned obvious. Sorry for wasting your time.

Make your decision fast because there are no second chances.

The common sense of a good pass is preached far and wide. It's called a "hand-off", "setting someone up for success", "user-friendliness", or "intuitive design", depending on the domain. We want to make life easier for the people that we work with. We want to be understood. We are embarrassed to be seen as incompetent in public. Make the pass anyway. This not that. No, sorry, I mean "it depends". There's a real nuance to trace out here, and I'm going to keep writing until I've laid down all my evidence.

Algorithms to Live By

Thoroughly enjoyable book, I reckon it's great for overthinkers, and a bit dangerous for someone like me who tends to enjoy moving fast. This is where I first learned of "the secretary problem". It would have worked really well in my football career:

1. Estimate how long I am likely to keep possession of the ball.
2. Spend one third of that time noncommittally assessing options.
3. Spend the next two thirds of the time actively looking for the better pass to make.
4. Pass the ball as soon as the better option appears.

Great! Done. Dusted. Make your decision fast because there are no second chances. It's an appropriate mindset for Gaelic football. Less so for its original example of hiring a secretary. You have both more and less time than you think. So hurry up and take it slow. Don't overthink it. You're only making a decision about the most complicated thing in the known universe (a person and their abilities). Basically, read "The Inner Game of Tennis", realise there is no combination of words you can read that will substitute for lived experience, and I'll stop haranguing you with obtuse one-liners.

What I really want to do is bring greater awareness to the unglamorous truth of teamwork, so that we might be able to pay enough attention to the right details to get good at it. Hence the hospital pass, which I propose as the down-to-earth successor to the "hail Mary pass". The Inner Game of Tennis is big on perception over judgement. To achieve this, it often employs mirrors. Before making a pass, look to see if your teammate is in a favourable position. If they're not, think how you'd fare in their position. If you reckon you'd muscle through the difficulty anyway, play the ball. Don't think about whether they're more or less skilled than you, that's judgement creeping in and it'll fuck up the calculation. Just play the ball like things will happen exactly as you think they could.

Innovative Medical Treatments

Innovative medical treatments do not and cannot make it to the market through the will and vision of a single person pushing it from start to finish. My claim to fame is the Brain Spine Interface: a brain implant connected via software to a spinal cord implant that enables people with paraplegia to once again take steps through their own volition. Enabling participants of that study to take their newly installed neuroprosthetic and use it at home is both a moral imperative and a heinous hospital pass. On paper the risk analysis doesn't go very well, it is much safer to keep them in a wheelchair. Yet imagine yourself in the chair and you'll take any chance you're given to get back out of it again. The exercise becomes getting the maths to line up with what the heart already knows. "First, do no harm" is a lie. It's always been a lie.

Last summer I began working on Qualyscan, Terapet's solution for particle therapy verification. The goal was to present it to the world. The video opens with the line "what if you could see what happens inside the target during particle therapy". A thought experiment. I didn't know exactly how Qualyscan could be used to maximum potential, I just knew I had to pass the ball to somebody in a better position than me. The only certainty was the guilt I felt: yes, Qualyscan would cause real pain through the disruption of existing workflows, it could increase the workload, it could create uncertain moments through the creation of conflicting data. I passed the ball and waited for the whistle to be blown.

Can we use it in-vivo?

It felt like we had skipped a step, but too late, we'd passed the ball and the field was running with it. At PTCOG in Deauville I got the chance to play with every treatment planning system on the market and the rationale became clearer to me then. Stepping into the cockpit of particle therapy was a pretty wild experience. Once beam angles are locked down, the medical physicist moves on to writing the objective: dose targets for the tumour, dose limits for the organs at risk, robustness considerations and different ways of even quantifying "dose": deciding whether to stick with a flat relative biological effectiveness conversion, or factor in the linear energy transfer at the same time. They play a hundred trade-offs against one another, in the hopes of steering the beam to achieve the best possible outcome, and all of this in a windowless cockpit.

So how would I imagine myself muscling through all of this? With difficulty, of course, and that's before I've added dynamic arc, multi-ion, and FLASH therapy to the mix. Decision paralysis feels imminent. The only way out is to test and measure. Qualyscan feels fundamentally different in this mix: an output signal, not an input parameter. Will I know exactly what a PET image of a good irradiation looks like? I won't, at least not at first, but give me something to look at so that there's a chance that I can form an intuition.

So the time for deployment of the Qualyscan In-Vivo is now, not after we've figured all the use cases out. There are brilliant people on the other end ready to take this tool and run with it, I'm sure of it. It's what I'd do in their position. I have, more or less, been in their position. That brings me back to .NeuroRestore, and scaling their mission across 100 centres.

We spent a lot of time figuring out how to streamline the process of functional mapping, particularly for our walking trials. If you can't figure out the spatiotemporal patterns of stimulation that can elicit, synergistically, the muscle contractions required to take steps, you're not going anywhere. The process requires you to build an intuition and explore a large problem space while dealing with daily performance fluctuations. The idea that a machine learning solution exists to cope with all this complexity and find the optimal stimulation pattern is conceptually very appealing. How else could the therapy ever feasibly reach the 100-centre scale? You would need an army of highly adaptable, highly intelligent people with an interest in neural engineering to deal with it. And that's the thing, those people exist. If there's something that makes the engineer at .NeuroRestore unique, it's the fact that they did it first, when the question of whether it was even possible still lingered. For the rest, the assurance that it can be done is enough. Onward Medical is preparing ArcIM for the market now, with the initial indication being blood pressure. We know that epidural spinal cord stimulation is safe. We know it can be used to treat chronic pain. We'll know soon that it's effective to treat orthostatic hypotension. For the rest? Walking, anti-spasticity, bladder control, trunk stability and sexual function are possibilities begging for greater exploration. The ArcIM platform can unlock all of that, even if the official indication isn't there yet. I think the medical world is ready to take this ball and run with it. ◆