They agree this is going to cause a lot of random ripples, which they are asking Merrin to handle, but they're very confident she can manage that, and also within thirty minutes she's going to have vastly finer levers for staying on top of the situation. 

...That's actually a lot of the justification.

Is this potentially going to cause some inconvenient dysregulation of other vital signs and electrolytes on the scale of an hour or two? Yes. Is the electrolyte imbalance - still evolving unpredictably as the patient’s gut regains circulation and more of the seawater contents hanging out there and in the local-circulation vicinity are fully absorbed into the systemic circulation - at least as much of a problem on the scale of four hours, let alone the full 12-16 hours before they’re actually done with the weird stuff? Also yes.

But - the experts and the markets think - it's going to significantly boost the efficiency of Merrin's later levers for fixing that.

The basic problem is blood viscosity. Hypothermia by itself already changes how the blood behaves on a chemical level, making it thicker - and thus harder to shove through an external filter circuit.

And they don’t just need the ECMO to work for at least four hours without clotting off or putting intractable load on a struggling heart - the interim-stabilization protocol taking shape also calls for the state-of-the-art dialysis system, not because the patient is in kidney failure, but because it’s a direct rather than indirect lever for Merrin to regulate the patient’s electrolyte balance to within a narrow set of parameters.

The narrow parameters are partly about survival, minimizing the chance that the patient deteriorates too far to reverse before they reach the real hospital facility at all, but really they’re mostly about optimizing the eventual neurological outcome.

(Markets are currently calling 80% odds of some degree of brain damage, and (on the hypothetical where they use the public version of the neuroprotective protocol, since while like a third of the experts involved know about the secret one, these are in fact still public markets) 55% odds of enough damage that the patient won’t recover enough to live independently, and 35% odds that he won’t even be walking or speaking after this.)

Anyway, they think they’re trading a ~3% increase in short-term risk of death in transit (and a ~10% chance of enough deterioration that the patient’s brain takes significant more irreversible damage before they can start the neuroprotective protocol), for…a less certain but much larger 15-25% decrease in the risk of moderate-to-severe brain damage overall. 

Based on their current best (though still rather sketchy) extrapolation of the patient’s preferences-over-outcomes-involving-tradeoffs, this is a tradeoff he would choose. And the markets really do have a lot of faith in Merrin’s ability to do damage-mitigating troubleshooting even if something does go seriously wrong.

Updated market predictions:

- Survival to hospital: down by well under a percentage point, still at 94%.

- Odds of a major pre-hospital-arrival complication causing significant additional brain damage: up to 33%

- Final-prognosis prediction on, controlling for 'patients survives rewarming and recovers', at least mild brain damage: down to 72%.

- Prediction on moderate brain damage: down to 44%

- On severe brain damage: down to 27%.

(This takes well under 30 seconds to express in Baseline.)

Ugh. Okay. Those are sufficiently better odds that it's worth the enormous hassle. People having a lot of faith in her to do really hard things well is SO SCARY but Merrin will absolutely do her best? 

5 minutes to go! 1350 ml of IV fluids now have the correct electrolyte concentration - the medicopter comes with a machine for prepping that, they're not doing it by hand - and it should be at the exact right temperature and ready to administer in another minute or two. 

...Turbulence is kind of bad right now. 

[meta complaint to be deleted later that there is NO DISCORD ACCESS ON THIS FLIGHT WIFI it's terrible]

Someone else placed a peripheral IV in the patient's arm while Merrin was doing...something...there's a few minutes in the middle there that are sort of a blur. Merrin really wants central line access, but putting sharp things near major veins is risky when one is subject to frequent sudden acceleration forces in random directions. While Merrin is technically, certed for it even in 'non-ideal conditions', and the training sims for that involved conditions less ideal than these, it's still not her favorite sort of adventure - and the way to do it safely, or at least more safely, is to go slowly, which means two minutes even if literally everything goes perfectly.

...She makes the call herself to just go ahead and start the fluids now, taking the risk that they'll hit the patient's core circulation more slowly and less predictably, but at least they'll have most of the initial fluid bolus in by the time she's starting her next procedures. 

She watches the patient's vital signs like a hawk. 

Her guess - and the market predictions - were right; the patient's temperature is still dropping, he's down to 22.5 now. 

But, for all that, he looks surprisingly okay. Well. Objectively speaking he looks awful, but taking into account the sequence of events before they picked him up, he looks less worrying than she would have expected. Despite a numerical blood pressure reading that is, one, all over the place, and two, still consistently under 60 systolic, if she looks at just his face and neck, the waxy bloodless look is clearing away and he looks much less like a dead or dying person, and more like someone who is admittedly very cold and very much in shock but is getting bloodflow. 

Also Vellis is now shaving his head to finish placing the EEG electrodes and he has SO MUCH HAIR. There's a little handheld vacuum for the purpose of things like 'not getting hair in your sterile field' but there's SO MUCH of it, and Vellis is trying to solo the sensor setup plus cleanup, because Nettir is busy getting another arterial blood gas sample. The vibration and turbulence are not helping, and his unusual pale hair color means it's kind of hard to see, but there sure is hair everywhere now. 

Well, when they LAND and the medtechs from the ship team run in to collect their patients, maybe Merrin can STEAL one of them and make them CLEAN UP THE HAIR, because she wants Vellis on placing a central line while Merris does the more complicated setup. 

Merrin is registering a strong mental prediction that his perfusion, at least to core organs, is actually improving. To the point that she sort of wonders if they're having a sensor calibration issue - the equipment and the setup protocol are supposed to avoid that but they did set it up really fast under far from ideal condition - or if maybe the software that adjusts from radial-artery pressure to calculate brain perfusion pressure is getting borked by all the motion artifact and the noise-smoothing is actually filtering out some of the real signal. 

- you know what she's going to go ahead and register that out loud as an observation. Along with her observation that this patient at least looks less worrying than her generic sense of what-to-expect-given-these-circumstances.

She wonders (out loud, or rather subvocalizing into her microphone to avoid distracting her team) if, in addition to his age and health, it's relevant that he's more physically fit than the patients usually modeled in sims, and that's the source of her ongoing background feeling of pleasant-surprise here? Like, it's just definitely the case that if someone regularly does very intense physical exercise, pushing their cardiovascular system to its limits on a daily basis, then the compensatory mechanisms there have more reserve.

Also temperature-acclimatization is a thing - regularly exposing a given physiological system to not-actually-dangerous stressors will shift the baseline state of the control system, up-regulating or down-regulating various responses to adverse conditions, and it sounds like this guy hasn't just been on a boat today, but for weeks or months, and maybe that means regularly getting soaked in cold water, and so the current situation, on a physiological and even biochemical level, is at least less far outside the parameters his body is accustomed to. 

(Also she's going to get a manual cuff-inflating blood pressure from his other arm, it's simple and low-tech but sometimes, and now seems like one of those times, that's exactly what you want.) 

Merrin is not in fact the first person to consider this hypothesis, but her independently confirming it - especially given that she's inferring it mostly from direct sense-observations of the patient plus years of trained intuition, rather than via a more abstract model of the system - is informative. 

And the physiological indicators that they hoped would improve with some rapid IV fluid administration are, in fact, improving, to a degree that's on the higher end of the predicted range. 

Diagnostic prediction markets are updating that the patient's odds of surviving transport-and-rewarming, on the current protocol that involves not rewarming him as soon as they can, did briefly dip after they made the call to treat the dehydration fast rather than carefully, but they're now back up, and settling at 94%.  

Could be a lot worse! Merrin will take that. 

(She is also, now that she's had a few minutes to absorb the situation - and, honestly, it also helps that she's had some positive updates - feeling significantly calmer about the whole superheated insane medical directive situation.

As in: it's bizarre and upsetting (and really confusing, in a way that leaves her with an itchy open-mental-loop of feeling like she's missing some sort of key context on the patient or the broader situation and that might matter), but it is not, in fact, the terrifying black-and-white scenario her mind immediately jumped to, where– where whether she, personally, is Good Enough will end up causally responsible for the patient's survival versus True Death.

She has no context on why he made the decision to refuse cryo (and is mostly not thinking about that because, whatever the potential answers, they're probably ALSO upsetting and she does not need to be wasting any of her limited metacognition on not-immediately-relevant questions with predictably upsetting answers) - but people don't make that decision lightly. Meaning that, even if he survives this, she shouldn't actually expect he would change his mind. He...might, in some weird edge cases, though her brain is annoyingly doing the thing where it FLAILS at her rather than answering when she tries to query it for an actual numerical probability estimate. But. Low??? 

Her counterfactual impact on this situation...was already, always, a narrow slice of this pie. Both because she's one person in a huge team effort mostly made up of people WAY SMARTER AND COOLER than her, where her only real advantage is being the one physically here first and with lots of sim time on the work that will buy the Very Serious People time to figure out clever solutions - and because, obviously, if she weren't available then Exception Handling would just have sent someone else. They sent her because she seemed like the best person to bet on given what they knew at the time, but the next-best person isn't actually much worse. (Honestly, performance-wise they might be better, and a lot of the reasoning is about the cost; Exception Handling knows that Merrin won't need as much recovery time as usual after a scenario like this.) 

And her counterfactual impact isn't actually on a True Death, because...absent some weird edge-case scenario, that was going to happen anyway, in forty or fifty years, and it's STILL really weird that Kalorm apparently prefers PERMANENTLY CEASING TO EXIST over...whatever he even thinks is bad about cryopreservation??? ...but it's his life and his choice and he, presumably, had endorsed reasons for it. 

So. By her efforts, here, she can buy him a higher chance of those fifty years. Which - is real, and important, and Merrin is still really very motivated by it - but it doesn't make sense to panic over it. Not that it made sense to panic before either, but it's helpful to draw those lines around what her actual, realistic influence is, here. 

(The ridiculously well-funded prediction markets are freaking her out in a different way, which is less amenable to calming down with that whole line of thought, but Merrin at least has practice on handling that.) 

90 seconds to landing. They were flying higher to avoid some of the turbulence, so expect more of that on the descent. 

Patient status: 

...Yeah there's something going on with either the sensor calibration/data-analysis-software, or maybe differential blood pressure in the lower vs upper arm, or it could literally be a difference between the two sides of the patient's body, you get that sometimes for reasons that are rarely explicable. Manual reading is 68/42. 

Latest blood gas result is up! (They still don't have an at-all-reliable continuous O2 sat reading, even though, when Merrin absentmindedly checks, there is now a faintly palpable radial pulse; under non-moving-helicopter conditions they could probably get a reading, on an earlobe if not a finger, but the sensor is trying to pick up a very weak signal against the backdrop of rather a lot of noise, and it's trying its best to give them numbers anyway but nobody is actually putting any trust in them.) 

Anyway. Latest results are almost not horrifying! This is actually really impressive given just how horrifying the initial readings were! 

(The normal range for the subparts of this test are: 10.5-13.5 kilopascal (kPa) for the arterial partial pressure of oxygen, 5.1-5.6 kPa for carbon dioxide, 22-28 milliequivalents per liter (mEq/L) for bicarbonate, the main acid-base buffer involved in the kidney-metabolic-side regulation of blood pH. Because the respiratory CO2-concentration and metabolic bicarbonate-concentration sides of the pH equation can compensate for each other, as long as only one side of the equation is affected by a given medical problem, the body usually succeeds at staying within an impressively tight pH range, centered on a pH of 7.4, with only a 0.02% range of variation in either direction.

The patient's initial blood pH of 6.95 was an entire 6% outside those delicately-maintained parameters, because both sides of the equation were severely affected; lactic acid is, as the name hints, an acid, and the large quantities of it dumped into the patient's bloodstream after those long minutes of desperate anaerobic metabolism were enough to exhaust and exceed the existing bicarbonate-buffer; the measured bicarb level from the first blood draw was down to 3.9 mEq/L. And meanwhile his blood CO2 was up to almost 18 kPa, triple the usual upper limit.

Oxygenation doesn't directly affect pH (though the indirect effect, via cells deprived of their primary fuel source and falling back on anaerobic metabolism, is one that happens fast enough) but it's still one of the most important metrics here, and his initial reading was 3.5 kPA, a third of the usual value.)

Anyway. The sky-high lactate is taking longer to clear - that requires metabolic work, much of it done by the liver - but, since they knew they couldn't afford to wait for that, the patient received some IV bicarbonate solution almost immediately once they had IV access and test results (this was a sufficiently predictable side-task that Merrin was neither consulted nor noticed it happening, because she assumed it would). And the blood CO2 concentration is now down to only about 10 kPa, which is still abnormal but it's not even double the limit. His measured bicarbonate is now 10 and his blood pH is now back above 7.2, still abnormal but back within the 'generally considered compatible with life' range; since blood pH affects approximately all metabolic pathways, this is kind of important. 

Partial pressure of O2 is up to 6.3 kPa. The temporary semi-stabilization protocol now being set, customized to all the details calls for a slightly-but-not-extremely lower range than the normal range at a normal body temperature, aiming for (roughly, they're actually calibrating mainly on measured O2 saturation on the assumption that they will AT SOME POINT have this as a continuous measurement) between 5.5 and 6.5 kPa, so...they're actually pretty much there? 

(Though of course the eventual protocol will be led from more direct brain-perfusion sensor data that they do not currently have, due to the inadvisability of poking sharp things near important organs while in a moving helicopter in a storm; all the current results are at best an estimate of that, and a lossier one than usual given how far outside usual-parameters the patient's body is currently operating.) 

Eeeeeeeeeeeeee!!!! Things continue to improve faster than Merrin's (intuitive) median estimate and this is the sort of thing that makes a Merrin really happy! 

(She has a very pointless urge to hug the patient about it, which would neither help nor be noticed or appreciated by him in any way.) 

Aaaaaaaaaaaaaaand touchdown on the helipad on the ship!

As soon as they're firmly secured down - ship staff are stationed nearby ready to manage the equipment for that - which should be in about sixty seconds, they'll open the side door and there will be a briefly stressful 5-10 min period of, like, eight additional people inside their not-actually-that-large medicopter, but soon after that, the other four patients will be elsewhere and there will be BREATHING SPACE to set up all the fancy additional equipment that will, hopefully, let them keep the patient from deteriorating any further before they reach Default. 

That being said, it would - if this seems realistic and the personnel actually on the scene don't predict that the distraction will impede their performance - be really appreciated if they start that setup process right away rather than waiting for the transfer to be done. Especially the more invasive sensors. 

(It is possible that a large number of people betting on the Diagnostic and Treatment Planning markets are currently very frustrated about the limited suite of sensor data coming in for their statistical models to chew on.) 

- yeah Merrin is also feeling that way and she's already kicked out everyone except Vellis and told them to drop the plastic-sheet 'room' separator, so that the instant the medicopter is confirmed secured against unexpected sliding around on a wet ship deck, she can start - carefully, not rushing, but definitely efficiently - placing the handful of invasive sensors that will hang out in the venous sinuses of the patient's brain and tell her how bad the damage looks right now.

(Everything else keeps going slightly-but-noticeable better than she was expecting, but the helpful factors she was speculating about - youth, excellent cardiovascular fitness, even being physiologically accustomed to random cold-water dunks - are not actually going to be protective against anoxic brain damage. Merrin is in SUSPENSE over here.) 

Even the large oceangoing ship is...not incredibly like being on solid ground? There are vibrations conveyed even through the medicopter's internal gyroscopic stabilization system, and there's definitely an ongoing rhythmic swinging-swaying motion from the waves. 

Rhythmic motion is fine, Merrin has trained for this and she can compensate for it on a basic motor-memory level. Vibrations are inconvenient mainly because they interfere (not a lot, but some) with the internal accelerometer sensors she's using to have any idea where the probes she's inserting have gotten to, and add some visually annoying static to the real-time ultrasound imagery that Vellis is getting for her. 

It's fine. Everything is fine. She will just go about 20% slower and make sure to time anything delicate to the moment in the sway-back-and-forth motion when they're moving the least, but she's done this like sixty times in sims and honestly most of them were meaner than this. 

It takes them about four minutes, and another minute after that for the sensors to calibrate and start filling an entire additional screen with readings. 

....Wow those are not great readings, though. 

They're not!!! Merrin doesn't like it!!!!! Those are badwronggross numbers!!!!

There are a lot of abnormal cell-damage readings, some of them ones specific to reperfusion damage, and the ones that have a normal reading of 'approximately None, definitely undetectable' are not just reading as 'barely above the detection threshold'. Some of them are, like, four times higher than it. 

It's in many ways not-exactly-comparable to the situation that Merrin faced two and a half years ago. For one, it's a redesigned and significantly better sensor, one that is both more sensitive and measures a lot more cell-damage factors. Last time, they were guessing and - maybe to some extent - hoping for the best. This time, they'll know a lot more of it from the beginning. 

Inconveniently, if unsurprisingly, there isn't a lot of human trial data on what the 'abnormal' sensor readings mean. 

It was (eventually, once involved parties found a way to introduce the new advance in medical tech without it looking suspicious or leaking any information about a still-secret medical case) a genuinely important step forward. Making it public took a while, but universal adoption happened almost immediately after that, and the basic sensor use-case went from 'obscure experimental lab tech' to available in most hospitals on the planet within a month.

However. It's nearly always been used to better inform the medical prediction markets on whether and when to shift from a lifesaving-recovery focused treatment plan to cryopreservation (or, occasionally, the reverse). This is incredibly valuable, but: most patients are not in the situation that this patient is, and most patients have advance directives that lean pretty conservative on risking irreversible brain damage.

Not everyone, people do vary, but it just hasn't been very long, and while there are now a few hundred at-least-vaguely-relevant case studies to refer to, they are mostly not actually very analogous. On average, dath ilani are rather unlikely to end up suffering this sort of mishap (the case Merrin dealt with in Harkanam was itself a rare and surprising event even leaving out the patient's identity and importance to Civilization.)

A large fraction of the documented use-cases for the sensor and related protocols were for stroke patients: usually elderly, usually aware that their bodies were headed toward the slow but inevitable cascade of gradually failing physiological control systems, and thus likely to have updated their medical advance directives accordingly. However, the local damage of a stroke, and resulting dysfunctional-physiological-compensation-cascade, is actually a pretty different situation from the widespread diffuse damage of global anoxic brain injury. 

A much smaller cohort of patients involves traumatic brain injuries - because accidents do, sometimes, happen, even when most people are being very careful. That patient population tends younger and healthier, and correspondingly tends to have a higher-functioning baseline they could return to with a full recovery (and medical directives that take this into account, and insurance that will pay out higher prices per QALY saved). But major head trauma doesn't tend to happen in isolation, and additional traumatic injuries or major blood loss add new confounders; also, even in controlled rat studies, the cascades that happen as the body tries to respond to the injury are actually pretty different. 

There is some data on patients who suffered an (also usually age-related) heart attack with cardiac arrest and suspected anoxic brain damage - somewhat more comparable to the current patient - but this is rare to begin with, hearts are a much better-understood organ with correspondingly better preventative medicine, and most patients at risk of a major heart attack either go for cryo anyway, or have a pacemaker and closer monitoring. 

In terms of near-drowning incidents, involving at least some degree of hypothermia, where the sensor and protocol were used and good documentation exists, it's mostly a tiny and really upsetting dataset of young kids. (Even in dath ilan, kids are the most likely demographic to end up in water-related incidents.) Most of those don't also involve this degree of hypothermia. The best and most comparable data on the protocol is from this source, and - it looks fairly positive, actually, in that population? But young children are in many ways a unique population. Cellular regeneration mechanisms just work better, neuroplasticity is far higher, and of course the cardiovascular system is nearly always in excellent shape. And of course it's harder to know what the counterfactual looks like, for a three-year-old after a 'full recovery' from a near-drowning-related potential brain injury. Three-year-olds' cognitive test scores are already imperfectly predictive of adult results to begin with, and besides, this entire setup has only been in testing for like a year. 

And even aside from the dearth of endpoint measures on patients who displayed concerning early signs – since those are exactly the patients who, even if their testaments hadn't just called for cryo in any case with that degree of uncertainty and risk of a bad outcome, very often would request that given specific reasons to update toward a poor prognosis – but even apart from that, none of those are the current patient's situation. 

There are just not very many adult near-drownings period. This is, of course, a good thing. It means that the water-safety training taught to all children is mostly adequate, and people are following it. But it also means that there are exactly three examples from the entire last 14 months that are even vaguely analogous. The data is better than nothing, but it approximately informs them that submersion time of 4 or 7 minutes respectively, at age 21 or 17 respectively, followed by immediate medical treatment at a specialist hospital, is a good sign. Being trapped under the ice on a frozen lake for forty-three minutes, while also living in a Quiet City a long trip from the nearest large hospital, and also being almost fifty, is not so much a good sign. It is pretty unclear how to interpolate from this to whatever territory lies in the middle between those extremes. 

All of the other, more prevalent use cases show a different pattern of damage-signals than other cases. They're the closest vaguely-comparable scenarios, but of course, as it is said, Reality can be such a huge asshole about what counts as having tested your code inside a domain that's close enough.