In the case of the St Maartin incident, why did the crew use RNAV for the approach rather than use the VOR approach which is more accurate.
A VOR is not "more accurate". Lateral tracking on a LNAV is more accurate; it puts you on the CL (as the boys at Mildura found; they would have died had they done a VOR). It is true that in this case the VOR had a lower MDA, but as the report says, they only loaded the LNAV "just in case"; the ATIS indicated, and they briefed for, a visual approach. So good on them for that.
Interesting that following this near-miss, Westjet created (or had created) an RNP (AR). Now they're getting serious.
Therefore, there is a long visual flight segment following the MAP where the crew is required to manage the descent to the runway threshold in order to complete the landing (Figure 4). It is not common for WestJet pilots to fly long visual segments of an IFR approach such as that of the RNAV (GNSS) Rwy 10 at TNCM. Even less common are long visual segments over water and with the type of weather encountered during the occurrence approach.
I think the TSB got this bit wrong. Even though the chart depicts a Visual Segment, 700ft/2nm is basically on a 3° profile, so no different to any other NPA flown as a CDFA. With a vis limit of 3600m, the runway should have been visible almost immediately.