This applies also when the problem is already there. You can ask yourself ‘How can I fix it?’ but it’s a lot more useful to ask ‘How can I prevent it?’. In the latter, you’re more like to look at the cause of the problem, instead of its effect and you’re able to solve it much more effectively.
I encountered this in the example described at form follows function, where a circle would be perceived as a coin. I tried many things to counter-act this: bigger negative feedback on touch, distinctive sounds, longer restart time and some more. But in the end these didn’t do much to solve the problem. What did is taking a hard look at the circle itself and why it was perceived as a coin (or something to collect). Realizing it was because of a false affordance and turning that around to be an actual effective affordance is what solved the issue.