"ichi2" used not to rate a P in the edict edition. For years I've changed ichi1 to ichi2 when the term or form was obviously uncommon. There are about 100 of them.
The old ichi/news/spec tagging is a mess, and sorely needs revising.

I'm not sure what the reason for the downgrade is. We don't usually amend/drop (non-spec) freq tags unless there's an obvious error in the data. They're not based on Google n-grams. Plenty of ichi1-tagged terms have fewer n-grams than this. 
This doesn't meet the spec1/spec2 threshold.

@@ -5,0 +6 @@
+<ke_pri>ichi2</ke_pri>
@@ -8 +8,0 @@
-<ke_pri>spec1</ke_pri>
@@ -13 +12,0 @@
-<ke_pri>ichi2</ke_pri>
@@ -20 +18,0 @@
-<re_pri>spec1</re_pri>

Downgrade. It's still in the ichiman collection.

@@ -8 +8 @@
-<ke_pri>spec2</ke_pri>
+<ke_pri>spec1</ke_pri>
@@ -13 +13 @@
-<ke_pri>ichi1</ke_pri>
+<ke_pri>ichi2</ke_pri>
@@ -17 +17 @@
-<re_pri>ichi1</re_pri>
+<re_pri>ichi2</re_pri>
@@ -20 +20 @@
-<re_pri>spec2</re_pri>
+<re_pri>spec1</re_pri>

gg5
Glossing as an adjective

Google N-gram Corpus Counts
╭─ーーーー─┬─────────┬───────╮
│ 煩忙　　 │     351 │  0.1% │ - rK (kokugos)
│ 繁忙　　 │ 325,820 │ 99.8% │
│ はんぼう │     453 │  0.1% │
╰─ーーーー─┴─────────┴───────╯
╭─ーーーー─┬─────────╮
│ 繁忙　　 │ 325,820 │
│ 繁忙の　 │   4,164 │
│ 繁忙に　 │   3,190 │
│ 繁忙な　 │   2,730 │
│ 繁忙を　 │   2,124 │
│ 繁忙と　 │     744 │
│ 繁忙が　 │     513 │
│ 繁忙さ　 │     465 │
│ 繁忙は　 │     216 │
│ 繁忙して │      73 │
│ 繁忙する │      54 │
╰─ーーーー─┴─────────╯

I suppose the [ichi1] tag should be downgraded to [ichi2] or removed altogether

@@ -4,4 +3,0 @@
-<k_ele>
-<keb>煩忙</keb>
-<ke_pri>ichi1</ke_pri>
-</k_ele>
@@ -12,0 +9,5 @@
+</k_ele>
+<k_ele>
+<keb>煩忙</keb>
+<ke_inf>&rK;</ke_inf>
+<ke_pri>ichi1</ke_pri>
@@ -24 +24,0 @@
-<gloss>pressure of business</gloss>
@@ -25,0 +26,3 @@
+<gloss>high-stress</gloss>
+<gloss>hectic</gloss>
+<gloss>pressed (with business)</gloss>

@@ -18,0 +19 @@
+<re_pri>spec2</re_pri>