Although the supplied scale factor for PGA supplied for this issue/enhancement is 1.0, it actually should be 1.2 for structural engineering applications. Also would need to be changed in nshmp-apps. Confirm with @amakdisi and/or @srezaeian
Unless users are clamoring for a "max-direction PGA", I'm not sure it makes sense to implement the 1.2 factor for PGA. As far as I know, PGA isn't being widely used for structural engineering applications - it's primarily used for geotechnical applications, where the geometric mean is maintained.
That being said, there was a BSSC meeting earlier this year where there was some discussion about engineers using PGA as a proxy for a "zero-period" spectral acceleration for certain structural applications, although I don't recall a consensus being reached on that issue... @nluco may be able to weigh in?