Thing is, though, that school trips to amusement parks are neither a recent nor local phenomenon, and neither are rapids rides. If there had been obvious and inherent safety problems with either, accidents such as this would have been a lot more common. Even if the odds of unsupervised children - even narrowed down to one particular age group - hurting themselves badly was one in a thousand, we'd see dozens if not hundreds of serious injuries every day. These rides have a capacity of a thousand people or more per hour, and there are hundreds of them out there. And parks are swarming with school kids during large parts of the season, while serious injuries are rare.
I guess that one rider standing up during a calm moment on a rapids ride brings the risk of serious injury up from "marginal" to "quite unlikely". In, say, 95 % of cases, it'd be fine. Up, swap places, sit down, no biggie. In the last few percent, the rider would be thrown against something by a sudden and violentr movement. In a small (but not minuscule) fraction of those cases, injury would occur. Most likely teeth being knocked out, bruises, a broken arm if unlucky. In a smaller fraction of the injury subset again, the rider would be thrown out of the boat. At that point, survival is a coin flip. Rapids rides are brutal like that. But in most cases, the odds of something going badly wrong is really small, even when safety procedures are broken. If standing up led to injury every time, there's no way in hell the ride would be allowed to operate.
The fault tree in the above example begins with a rider standing up (that would be the "top event"), and of the thousands of possible outcomes, only a few lead to injury. You probably can't reduce it much further without killing the capacity of the ride, and making it safer to stand up might have an encouraging effect (that is, the odds of the top event leading to disaster is reduced, but more top events occur since the risk is considered so low, leading to the overall number of disasters staying the same). What you could try to do is reducing the likelihood of the top event occurring, preferably without compromising other aspects (loading times, other safety aspects, ride enjoyability...).
However, at some point you have to consider the risk acceptable, and the actions taken to mitigate it sufficient. One in a million, one in ten million... It can never be zero, and the more you want to reduce it, the more it will cost. There has to be a cutoff point.
I think Drayton Manor just hit a perfect storm of unlikely factors. The girl shouldn't have stood up, but she did. Even though she did stand up, she shouldn't have fallen, but she did. Even though she did fall, she shouldn't have fallen out, but she did. And that's the disaster scenario. It normally wouldn't occur, and there were risk-reducing measures in place. But marginal odds are still means there is a possibility, and reduced risk is still risk.