the normal way is to improve the coupling between bias and secondary. If this is not enough and you still need a few more volts to get below 60, you can also increase the pre-load resistor. 

Thanks - do you mean increase output preload, the bias preload, or both? (i.e. R7, R10, or R11 in rdr195)?

I mean increase the output preload by reducing R11 say to 10k

Thanks, that dropped it by about 2V which should be just enough to sneak in below the 60V requirement, though more margin would be nice.

What can be done to the transformer to improve the performance? I wanted to re-use the same design for 240V with a few component changes, but it looks like I'll be violating 60V on the open output again. I'm using a  PQ20/20 core with 71 turns on the primary, followed by 27 turns for the bias, followed by 36 turns on the secondary, which is triple insulated (this was based on the design recommendation from PIXIs, the PQ20/20 was selected because it fit the available footprint a little better). The core was gapped for AL = 250nH/T^2.

The ideal case for the bias is if it is wind simultaneously with the secondary or sandwiched with the secondary winding layers. Another thing you can try is to put a 51v or 56v zener across the output and just check how much power it is consuming during ovp, if not much and you can get away with a halfwatt or 1 watt zener. The zener is only supposed to eat the excess energy that is peak charging the output cap and the value should be above the ovp threshold.