I don't think the x-ref is needed.

@@ -6 +5,0 @@
-<ke_pri>spec1</ke_pri>
@@ -22 +20,0 @@
-<xref type="see" seq="1009330">どんな・1</xref>

I can't see any problem with adding sK to the 如何 forms. Some may get rK instead of the kokugos have them.
In WWWJDIC the kanji forms shunted back for uk entries.

どういう	9246980	98.2%
どう言う	160947	1.7%
如何いう	3333	0.0%
如何言う	1687	0.0%

Please see longer thoughts on どうしようもない.  I cherry-picked this one additional entry to look at, and it's the only other one I will touch as part of this question.

Question is: can we start using [sK] on 如何/どう entries (other than どう itself) that don't have exceptionally high occurrences of the kanji form (maybe a kotowaza or something might lean on old kanji, I don't know).

Comment is: On any dictionary app that displays the first kanji form prominently (even with a [uk] marking), we have a lot of entries with 如何 displayed(jmdictdb's own advanced search results, and jisho.org are two examples).  This seems to me like a kanji form that a learner at an appropriate level can internalize, and not have spelled out in every entry using どう.  Conversely, it seems like a distraction for any non-advanced learner.

@@ -9,0 +10 @@
+<ke_inf>&sK;</ke_inf>
@@ -12,0 +14 @@
+<ke_inf>&sK;</ke_inf>

reverso

I don't think "somehow" "why" etc. are really 
correct.

@@ -22,4 +21,0 @@
-<gloss>somehow</gloss>
-<gloss>how</gloss>
-<gloss>in what way</gloss>
-<gloss>why</gloss>
@@ -26,0 +23,2 @@
+<gloss>what sort of</gloss>
+<gloss>what</gloss>