Data is always good.  I used my Clarostat 240C power decade box to vary the load from 10ma to the point where regulation was lost.   Input voltage was 13.0VDC and the resulting current was verified with a Keithley 2000 DMM. 

Short version:   After adjusting the output to 300VDC  the output voltage remained between 291 and 298V from 10 ma to 110 ma loads,  which translates to a range of 3 to 33 watts into the load.   At 100ma the heatsink temperature was 40C and the input current was 2.56A for input power of 33.28 watts.   Output voltage at 100ma was 292VDC or 29.2 watts,   resulting in efficiency of 29.2/33.28 = 88%.    Well within the expected range and in my view, this is the maximum working limit.

At 120ma load current the output was down to 280VDC for 33.6 watts output but the input current was 3.54A or 46 watts,  resulting in efficiency of only 73%.  This was reflected in a rise in the heatsink temperature to 54C.   Subjectively, it also "smelled hot".

I've had failures running these things at higher voltages so 300V and 100ma are my practical limits, but that's a substantial amount of regulated HV power!   The output is clean but if you find the need for more, normal filtering and shielding practices apply.  The worst problem I had was in powering a light aircraft receiver with a 75kHz IF - which is right at the switching frequency!   That required shielding and physical separation to eliminate magnetic coupling.    A light load of a few ma helps with regulation at low current and you really want a bleeder resistor anyhow (these things can bite!).

A final tip - make sure any RF loads are well bypassed - it doesn't take much RF getting back into the board to pop the IC.   I keep UC3843s and FETs on hand for those occasions.

73, Bob W9RAN