The argument you are making is basically identical to the old "guns don't kill people" argument which actually has been proven completely false. Similarly, minimal suicide prevention measures (like fences on bridges, or again lack of access to guns) while they can usually be circumvented by the extremely determined do usually prevent suicides.
How do you measure the efficacy of a fence on a bridge? By the number of suicides by jumping off a bridge or by looking at the suicide numbers in aggregate? Of course it would reduce suicide by jumping but is it actually reducing suicide generally or only a specific method?
England changed from coal gas to natural gas. That prevented one very common method, and it led to a drop in total suicide rates.
It took a while for method substitution to happen.
We also saw similar drops when catalytic convertors were added to cars in the UK.
One of the important parts of reducing access to means an methods is to cause people to switch to less lethal methods. Removing access to coproxamol (in the UK) saved lives because people switched to other meds. Any overdose is dangerous, but some overdoses are less likely to be lethal if medical attention is sought quickly.
England changed the quantities of paracetamol that people could buy. This link only talks about self-poisoning (so it doesn't address the method substitution) but it does talk about characteristics of some people who chose this method: did they go to buy the meds or did they use what was in the home? Were they able to buy large quantities or did the legislation work? What was the length of time between having the initial thought of wanting to overdose and then carrying out the act? http://cebmh.warne.ox.ac.uk/csr/resparacet.html
At the moment one of the strongest recommendations we can make for suicide prevention is to reduce access to means and methods, because that has clear evidence to support it.
You can hear Professor Nav Kapur talk about it here: https://www.youtube.com/watch?v=iWPEVhrWZS0&t=415s
The NCISH will probably have more information or links to research about method substitution: https://sites.manchester.ac.uk/ncish/
It's important that removing access to means and methods is not the only thing we do! It's important, but it's only a part of a package of suicide prevention measures.
> How do you measure the efficacy of a fence on a bridge? By the number of suicides by jumping off a bridge or by looking at the suicide numbers in aggregate?
You don't look at "suicides", you look at self inflicted deaths. You look at self inflicted deaths from people jumping from high places, and you look at the total number of self inflicted death in the area. So far we strongly think that fencing off places like multi-story car-parks saves life and reduces total numbers of self-inflicted deaths.