Reproducing an antenna system to get the most accurate analysis of the antenna can be somewhat difficult to do and to be sure that you're really reproducing a similar system to the HT itself. If you haven't already, read my post about the NanoVNA to get a broader scope of info on what I do to analyze antennas.
There's no point optimizing an HT antenna setup on a perfect ground plane free of obstructions etc to see what the antenna can do ideally as it's rarely in an ideal situation in normal use. You want to set up something that's as close to an actual HT as possible. I've tried a variety of things from just holding the coax and keeping it at a normal distance from my head/body, I tried a metal box, and I've settled on a pretty typical 3" wide tin can as that has a pretty similar profile to an HT in size. I've also heard of folks gutting a dead radio and installing coax into it which is something I'm going to try next time I have a head HT on my hands.
Currently what I do is put a 1M pigtail on the NanoVNA and screw it into the can and while holding the can I do the calibration using the calibration kit that came with the NanoVNA so that the antenna system is calibrated. I then attach my antennas one at a time and hold the can near my face and in free air and compare the readings. I decided to try holding the can about a foot away from my face as I would normally hold an HT for my readings and re-analyzed all of my antennas.
When I re-analyzed my antennas I got slightly different results than my previous metal box jig which is good, although some were decidedly worse, including some that I've seen measurements from other folks saying they were very low SWR in their testing so I still don't think I've gotten a good setup yet and I'm going to keep tinkering with it. I'm not looking for "better" numbers, I want accurate ones and am just not sure how to get those consistently.