The new exams will mark a return to using English questions for reading and listening tests. So we will see a return to "discrete skill testing". This issue is hotly debated within the language-teaching community. The argument for it goes that it ensures clarity for children and that a test of listening, for example, should be just that, not a test of someone's ability to understand a written rubric in the target language. This argument has merit and it is the way that Asset Languages decided to go when they designed their assessment system.
I would argue against it. In the UK, because of our high level of public accountability and because of our systems of performance management, we teach strongly to the exam. In fact, this was the case long before league tables when there was a more informal type of accountability. The results of teaching to the test is the "backwash" effect. From half way through the final year before the exam we practise exam skills and work through past papers as far as we can. Text books are designed to fit with the assessment system and feature similar style assessment tasks.
The result of discrete skill testing is, therefore, that teachers will use a lot more English in lessons than they would ideally like to. Methodology will be compromised as we do lots of tests with exercises couched in English.
This happened post 1987 when GCSE first came into being and exams were soon re-designed to include more target language in questions and rubrics. True, this has been adjusted over the years so that we now see a greater emphasis on visuals and multiple choice.
So don't be surprised if there is a backlash against discrete skill testing sometime in the future as the pendulum swings back.
Anyway, discrete skill testing is what we've got, so we must guard carefully against over-using English in the classroom to ensure that our students get enough high quality target language input.