Holy...okay. 3000 non-human primates might not be the correct number, but even 23 primates dead in this experiment is horrifying from a science perspective.
I worked in a neuroscience lab that did work with one (1) rhesus monkey. His name was Beckett. Beckett had better healthcare than a university undergraduate. Hell, it took less hoops to get some experimental studies involving undergrad volunteers up and running than it was to do one experiment with Beckett wherein he got to sit in a chair and eat grapes.
See, the paperwork regarding experiments with animals was intense. I spent three days filling out the forms and going back and forth with the ethics board to let us do experiments on terminal mice wherein they'd peacefully be put under anesthesia and never wake up from surgery. As in we were putting a sick mouse to sleep and had to be 100% sure it would not be in any pain or distress. Screwing up and causing a mouse harm or killing it outside of these exact bounds? You'd be lucky if you ever got to work with a mouse again.
Not only that, but since an animal was expected to die in the experiment, we had to report how many we needed in the experiment and then how many actually died. And if the latter number was higher, you better believe the ethics board would be on our case demanding to know what the actual hell happened.
(There's a plaque, by the way, in the labs, in remembrance of all the animals whose lives we ask for in the pursuit of science. It's right there when you walk in.)
The paperwork and restrictions get more and more complex and strict the higher up the "evolutionary chain" you go up. If you MUST use an animal model, you use the lowest one on the chain that can meet your requirements. As in, why use frogs if you can use zebrafish? Why use fish if you can use fruitflies? Why use a living creature at all if your computer simulation is good enough? And if you can't give a good answer, your proposal should be denied.
Back to the monkeys and Beckett. Monkeys, or "non-human primates" as they're often known as, are usually at the top of this chain. Ideally, you treat them like humans who are unable to consent for themselves. Which means they need advocates for them. Beckett pretty much had a vet assigned just to him. If he so much as had a sniffle, the vet had the power to veto any and all experiments he'd be involved in until he was feeling better.
This was important because if your lab had a monkey and that monkey died for reasons other than "natural causes" that the ethics board did not okay? There was going to be an investigation and if they didn't like what they found, not only would your lab never get another monkey again, but there was a damn good chance you'd be barred from working with any and ALL animals.
So from this perspective, "23 monkeys dead" should be read a very specific kind of shock and disgust. One dead should have made any kind of ethics board hit the breaks on this shit and start asking pointed questions as to "why".
(Oh, and in case you're wondering, there is an answer to the question of "what happens to monkeys who can't be used in research anymore?"
They're sent to the farm.
No, really. There are legit primate sanctuaries specifically for retired research primates. Where they get to live out the rest of their days in peace and safety in as close to a natural environment as possible. Beckett was a cranky old guy, and when I left the lab he was still there, but the plan had been for him to retire to one of those. I hope he got there and ate as many grapes as he could sucker the researchers into giving him as he could.)