The gold standard of evidence for medical treatments, the randomised controlled trial, particularly if double blind, and meta-analyses thereof is not part of the replication crisis. The replication crisis refers to life sciences research and sociological and psychological research which use less statistically robust experimental designs. There are other problems in medicine, specifically bias due to funding, and non-publication of negative studies. The former isn't going away. The only entities which can fund large randomised trials on a regular basis are large pharmaceutical or device companies. That boat has sailed. The latter is improving a bit due to study registration, but is still an issue.
It is true that many treatments don't have the kind of strong evidence one would like. There are multiple reasons for this. There are a few cases where a treatment has been shown to be ineffective, yet it is still widely offered. An example of this is orthopaedic surgery for back pain, or cardiac stents in certain circumstances. Usually, the reason for this discrepancy is obvious - the intervention is well compensated, and widely performed prior to evidence showing it isn't effective, so doctors just want to keep on doing it. The other reason for the discrepancy can be psychosocial factors. The use of antibiotics for children with upper respiratory tract infections is an example. Parents can be very demanding, and not surprisingly, there is a strong incentive to just prescribe antibiotics and get them out of the door.
I will not go into conflicts of interest here. It's a big topic. There have definitely been inappropriate relationships between pharmaceutical companies and doctors. This has become less of a problem due to regulation, but it is not eradicated. There is still a conflict of interest problem for hardware, eg pacemakers or hip replacements. This is why for a procedure like this, a second opinion is recommended.
Again some relevant history - there was an evidenced based medicine movement in the early 2000s that became popular. The gist of this was that applying high quality evidence to medical practice was paramount. The systematic review become much more prominent. This movement has died off, but the concept is still going strong. The problem was that this high quality evidence is often not available. The majority of systematic reviews concluded there was inadequate evidence. The second problem is that we want to move towards more precise tailored therapy for each patient, and by their very nature, randomised trials can only report average affects in large groups. They are thus well suited to relatively common problems with a clear diagnosis and relatively simple therapeutic interventions. Much of medicine isn't like this.
So although medicine strives to be evidence based, this is in many cases an aspiration rather than the reality. Nevertheless, it is important to remember that most doctors want to use the best evidence available.
As a consumer, the pragmatic approach I would take is as follows:
1A. Am I getting sound advice (part 1)? Ask the doctor for the evidence that what they propose works. This is the most efficient way. If they are vague or unhelpful in the consultation, you can ask them to send you something later. If they don't want to do that, you need to look elsewhere. I know this may or may not be practical depending on what is wrong and where you live. The best case scenario is to find a doctor you trust, and not worry about the rest so much.
1B. Am I getting sound advice (part 2)? Understand the incentive structure. When you go to a gastroenterologist for example, they get paid a lot more to do a colonoscopy than to talk to you. Unsurprisingly, they will tend to want to do a colonoscopy. This is why in my opinion, primary care is so important and the cornerstone of a good health system. What I mean is that you want to see a doctor who isn't a specialist first, who decides if sending you to a specialist is required. When you go direct to the specialist, you are at risk of getting biased advice, non-holistic care and unnecessary procedures.
2. Does the therapy make a difference? Know that effect size is important. Many treatments might be proven to make a difference, but the size of the difference is important. You should always be asking how much of a difference will this make? Do I want to have toxic chemotherapy that will extend my life a mean of 42 days? This is also inherently related to the value proposition of a therapy. You might want a chance at 42 days, the doctor might think you do or do not, so it is best to ask.
3. What level of side-effects can I tolerate? Educate yourself about side effects of drugs. It isn't feasible for a doctor to go over every side effect of a medication. They will tell you about the common problems, and also any problems which are severe and not extremely rare, and they should also tell you about side effects that may be particularly relevant to you eg if it makes you dizzy and you are a tight rope walker. Know that in placebo controlled trials, the placeboes do not have zero reported side effects, so that merely taking a pill can cause issues. Ask for detailed resources about side effects that you can take home and look at, but do not treat them all as equally likely. Decisions about whether a drug has too many side-effects or not are very individual.
The final thing to say is that some medical systems like to get in the way of actually looking after people, and the USA is a good example of that. This is again a whole other topic.