I don't have the ins and outs of your case. I'm a student, not an obsterician, or neonatologist. That said, I do know delivering a 25 weeker is insanely risky even if perfect care is provided, for both his immediate outcomes and avoiding long-term disability. (80% of babies born before 26w have some form of disability e.g low IQ, blindness, etc [1]).
No sane obstetrician is going to deliver based on a family opinion like that - it is a true last resort. If you were wrong, or less lucky, as horrible as it is to say, your son would not be with us and the surgeon would be out of a job with the "what if" on their conscious (which trust me, all obstericians have anyway).
When your son needed to be delivered, it sounded like he was delivered. Frankly no neonatal respiratory distress at 22 weeks sounds insanely lucky to me; normally before 35 weeks it's a concern.
I can't say if this was a failure of the medical professionals in charge of your care or not, but from the information you've given us it doesn't sound like it. Tens, maybe hundreds of patients will have presented identically to your partner with a far less severe diagnosis. Getting their babies out early could have left them permanently disabled or dead, on the offchance it was percreta/accreta/increta: especially in the absence of definitive imaging, which they did attempt to perform. Medicine is not perfect by a long shot. Managing uncertainty is one of the many skills of senior doctors.
Medicine isn't as simple as a lot of people in this thread think it to be.
DOI: Medical Student
[1]: https://www.tommys.org/pregnancy-information/premature-birth...
In our cases, her condition was deteriorating at a frightening rate. She could barely speak. Under such stress one must also consider the possibility her body giving up and aborting the pregnancy. Doctors had put cameras up and down and ultra sound etc. They were trying to keep her going as long as possible to give baby the best chance. I definitely would not advocate for anyone to make rash decisions in such delicate situations, and of course, my post omitted myriad details for brevity.
The cause was later determined to be the result of a D&C for retained placenta that was performed following the birth of our first child a few years earlier.
Medical training and experience can reduce the impact, but not overcome it. It is a reason why someone can't entirely outsource their heath to the medical system, and something that I think people miss when they set medical expectations.
From the perspective of the medical system, it is a numbers game, where the optimal number of deaths is non-zero. More time and effort spent in one area means less in another. Saving one life might cost two others.
Because of this, I think it is most useful to discuss ways to increase medical efficiency. Better data transfer, better collation, better processes. AI/big data review has a place in this, but we should be aware of it's limitations, particularly on the data it doesn't capture.
The most dramatic was in 2009. My wife was admitted to emergency at 22 weeks pregnant with severe cramping, nausea, and severe IBS like symptoms. She couldn't keep food down, leading to daily stomach pumping and eventually IV nutrition.
The episodes started every few days but grew more intense and frequent. After 8 days, the first doctor seemed lost. When I asked about a second opinion, he lectured me that it was "his call to make, not mine." That killed any trust - I asked him not to return.
The next doctor, an eccentric but brilliant GI specialist, ordered new tests. I obsessively googled every result (partly to cope, partly to help). When he suggested a motility drug, my research found a case of spontaneous abortion linked to it. We declined.
Three weeks in, I noticed something odd in her lab work - different blood counts from different locations. More late-night googling led me to placental percreta, where the placenta grows through the uterus and attaches to other organs. Several ultrasounds proved inconclusive.
With daily episodes now and my wife deteriorating, we insisted on an emergency C-section despite warnings about our baby's lung development. During the procedure, even with an epidural, she went into extreme pain. Her eyes rolled back, the surgical team's demeanor changed instantly, and I was told to leave the OR.
Waiting outside for who knows how long. I knew the risks. Some noise down the hallway, a few nurses called me over. They were wheeling a humidicrib down the hall to the ICU. Little champs whole leg was no longer than my pinky finger. They took him into the ICU and sent me to the hospital room to wait.
The OB arrived an hour or so later and explained she was alive and they had to remove something. He took out a photo and and showed how the placenta had indeed broken through her uterus and attached to the intestine. He then pointed to a particular point and explained how it would have ruptured at any moment. The decision to push for the C-section likely saved both lives. Our son, never developed lung issues. Lots of luck.
This experience showed me how even amazing doctors, overwhelmed with cases and data, can miss crucial connections. In 2009, they didn't have the tools to correlate data the way I could with google search-fu.
AI diagnosis tools will save lives - they can make these connections faster and more reliably than either doctors or desperate husbands with search engines. And it will just improve over time.