How would you palliate a metastatic lesion abutting a joint with an associated effusion?

I don't know that there is data for this - not that I could find.

Bone metastases themselves cause pain due to multiple factors - mass effect, inflammation, and microenvironment changes. I'd guess the effusion is potentially due to the existence of the metastases (an inflammatory reaction), rather than having concern that there are tumor cells in there. In fact, I presume that there would be some cancer cells in the fluid. 

The radiation is going to act on the metastases itself - decreasing the size of it, causing microenvironment changes, and decreasing inflammation. I would think that it is possible that this may help with the effusion. I wouldn't necessarily spare the effusion, but I would not make it a point to include it purposefully. 

I am interested in other's opinions on this.

I have never worried about this, but the closest similar scenario would be a pleural metastasis with a malignant pleural effusion. You treat the cancer that is causing the symptom (which is the pleural met that is say, invading the rib, or spine, or chest wall, or lung or bronchial tree, etc.) and not the entire pleural cavity. 

Similarly, in the scenario mentioned above, focus on what you're palliating, not on what is not causing the patient's symptoms. If doing 'standard' palliative doses, can come back as necessary.

If somehow was say an oligometastatic case with a joint effusion and you're going for durable local control, could consider tapping the effusion and sending for cytology similar to a pleural effusion and then purposefully treat the joint space, as the main reason we never treat the entire pleura (and pleural effusion is an exclusion criteria from every oligomet trial) is due to toxicity concerns.