The MSHA document you posted says he was making a phase to phase measurement and goes on to specifically stress best practices should be phase to ground.
In any case, your point is taken. A DMM was involved in an injury. The MSHA determined that the meter did not meet it's IP67 rating, but did not say that was the cause of the failure. I wonder what the UL report found?
Unfortunately for us rubbernecking meter geeks, we don't have a full record to examine nor was there any EEVBlog-style teardown and video to look at. There's certainly a lot we don't know, including what the supposed nonconformities were in the meter. I wasn't able to find much more information at the time. I only brought it up then as an example of disreputable practices by manufacturers and now as an admittedly rare example of meter-related injuries. Your statement that arc-flash incidents with a substandard meter as the primary cause are rare is probably true, but I'd say that a major factor in that is that people working on high-energy systems are unlikely to be using crap meters. There was some chatter on other forums that I remember from that time where some electricians were essentially saying they would walk of the job before testing a 995 volt system with anything other than a Fluke while others were claiming that a DMM was not the proper tool for the job at all, regardless of brand. I suspect the latter might be true--there are specialized medium and high voltage instruments for this purpose and an underground mine could be a CAT-IV environment which would rule out almost every meter I know of except the Brymens that are CAT IV/1000V (but not MSHA or IP67....) I don't know if the OP is still around, but I think the question he might want to ask is "do I even want to be having this discussion?"
As for mistakes by the tech or the mine operator, there may have been some, but that would be speculative unless someone has more info than I do. Perhaps a DMM was the wrong tool or PPE was required, in which case the employer did not properly train or equip the electrician. Perhaps the test leads were not appropriate, although I'm assuming from the MSHA statement that the meter failed internally. The statement by the MSHA regarding phase-to-phase seems unfortunate to me, making a messy legal case even messier if not for the quick dismissal on other grounds. I have seen this statement before about not doing phase-to-phase measurements, but this is not correct no matter who says it and anyone with experience with 3-phase systems testing could tell you why. I realize that phase-to-ground will typically be a lower voltage and might be less susceptible to certain transients, but either your meter is rated for the voltage or it isn't.
In some cases, phase-to-ground testing with a DMM will not spot certain problems. This is why Fluke sells those very expensive 4-channel Scopemeters--they have high voltage probes and can measure all 3 phases at once, either relative to a common reference or not. However, if you cannot directly measure the phase angle between your three readings like you can with a scope, your DMM cannot tell you what your phase-to-phase voltage is nor can you calculate it. This can result in some difficult to understand issues.
For an unlikely but easy-to-understand example, consider a 480V 3-phase wye supply with a ground-referenced neutral feeding a large isolated delta-wound motor through a remote contactor. Now imagine the motor is not running and you want to know why. Two of the wires from the contactor to the motor have gone open due to physical damage of some sort. First you measure all 3 phases to ground in front of the contactor, you get ~277VAC for each. Next you activate the motor and measure the phases after the contactor and again you get ~277VAC each time. Finally, you crawl down some tunnel and measure at the motors junction box and again you get ~277VAC on each phase but the motor doesn't run. You tell your boss to order a $20K motor and....