I looked up the abstract for this study on JAMA because as a well-trained licensed acupuncturist I find the conclusion absurd.
The group administering treatment were family physician acupuncturists. I have no idea what the education and licensing requirements are in Australia as of this writing, but there's a big difference between the 300 hours required for MD's here in the U.S. and the 4,000 hours of training I received.
Additionally, since acupuncture has a strong effect locally and systemically on reducing inflammation, increasing blood flow and nerve activity in avascular areas, I have consistently seen improvement in knee pain, often to my amazement, with only a few treatments. Of course, if there is a mechanical problem, acupuncture will not help. My point is, the study is weak because
a) the practitioners are not doing it right, and
b) the specific points used are not cited in the abstract- I generally apply a protocol using electroacupuncture

I have found, much to my chagrin, that often people in the media just to a conclusion based on the conclusion without reading and analyzing the research's design- the meat of the matter.