This question is more complicated than it appears.
One of the biggest challenges with depression studies is that the placebo group always improves dramatically, too. Using your terms, "over 50%" of the placebo group would likely show improvements in their depression inventories.
This makes it very complicated to interpret the studies, because now you have to look for how much more the active treatment group improves relative to the placebo group.
This is a huge detail that gets abused a lot by anti-pharma people, who write headlines about how SSRIs are "barely better than placebo" and then ignore the actual statistics. Another common tactic is to try to reframe the thresholds in different terms like "effect size" and then pool studies together to try to show that the "effect size" is below some arbitrary threshold.
Another challenge is that placebo response has been getting stronger over the years and nobody really knows why. Some antidepressant studies have even been halted because, ironically, the placebo group improved so much that there numerically wasn't much room left for the active group to be statistically better given the sample size. This is less of a problem with very large scale studies where smaller margins can be shown to be more statistically significant, but those are expensive and rare.
There isn't really a question about whether or not they are effective for many patients in the world of empirical treatment. However, if you go digging through the internet you can find plenty of commentary trying to convince you they don't work. Sadly, I've had some close friends and family members delay SSRI treatment for years because they read too many of these studies, but when they finally gave in and did a trial it turned their life around. The drugs aren't perfect and don't work for every situation, but they do work for a lot of people.
In practice, the average person will fall more in the range of maybe 5-15 due to vague symptoms like "I don't sleep as well as I used to" triggering some of the points. The average depressed person who seeks treatment might fall in the range of 25-35.
Now imagine the placebo group goes in with an average score of 35 and improves to a score of 25 by the end of the test. The SSRI group improves to an average of 20 by the end of the test. Is this significant? Well, it depends on how many patients you have in the sample size.
That's the problem. There's only so much room in these scales for improvement, so when both groups improve a lot you need to have a larger sample size to get statistical significance. Getting a lot of patients in a study (hundreds) is very expensive, so it's only a small number of studies that can pull this off.
So the published results are over-sampling studies where the statistics happened to work and under-sampling studies where they didn't.
Is this really abusing anything? It seems like the correct conclusion to draw from this is we should be giving depressed people placebos instead of SSRIs
What this should tell you is that SSRIs are almost always worth trying, because you might be one of the lucky ones for whom they work great - even if chances are they won't. Psychiatric treatment always works on this basis anyway - try one drug for a few weeks, see how well it does, then decide to either adjust dose, or try something else. Most patients eventually find something that works for them, though a the minority that don't is also large, unfortunately.
> but when they finally gave in and did a trial it turned their life around
One might wonder: could this be, in potentially large part, the placebo effect? I suppose that a good answer to that question would be a very big deal.
There was a deluge of media a couple decades ago about Prozac and its dramatic effects.
A few years later, I read some report that the studies testing its efficacy had ambiguous conclusions.