Learners of a second language are often at a disadvantage because they expect adult level proficiency in a short period of time. They often forgo focussing on fluency in order to chase that proficiency. In the end they fail at both.
Some of the most unfortunate myths that persist about language learning really hurt. For example, if you ask the average person how many words of vocabulary they need to be "fluent", they often pick numbers like 2000. But that's the level of vocabulary for a 2 or 3 year old. Adult level proficiency requires upwards of 20,000 word families.
Basically, what I'm trying to get at is that English skills of people in their early twenties are guaranteed to suck unless they are using English almost as much as their native language. It takes native people 10-20 years to get a good level of proficiency. What we can hope to do is to refocus education on fluency rather than proficiency so that at least people can converse freely in a small area. Then over time they will gradually accumulate proficiency in other areas.
This opinion is unfortunately not widely shared in the teaching profession at large :-(
Also the instruction of both junior high school and high school English has changed. It is now (officially, anyway) taught using only English in the classroom. I left the year before it entered high school, so I never officially participated, but I got some of the students who went through the junior high program. Even in my very low level school, the fluency improved dramatically.
Finally, they began demoting teachers with poor language skill a few years ago. It's actually a bit unfortunate because some good teachers got demoted whereas poor teachers with good language skill got promoted. I suppose you can't have everything.
Anyway, they are improving over time. The one thing I should point out is that teaching language fluency is actually very difficult. Things like grammar translation is still considered the best technique in many universities. Even though Pimsleur demonstrated the effectiveness of spaced repetition in the 60's, virtually nobody builds lessons that way. If you look at the ideas of Stephen Krashen and techniques like TPRS, even though they are very effective for fluency, only the lunatic fringe utilize them.
I actually gave a demonstration of TPRS (and related skills like circling) to the Shizuoka board of education once. They seemed very excited and even said, "This is the way English should be taught", so I have some hope that they are looking in the right direction. But again, nobody gets this right. Even the most effective techniques are still quite barbaric IMHO. It's one of the reasons I'm sad to be out of language education. Lot's of important discoveries to be made ;-)