From $p=2q+1$, we get $q={\large{\frac{p-1}{2}}}$
At the start of your attempt, you made a serious error . . .
From $q={\large{\frac{p-1}{2}}}$, it follows that
$$q^2=\left({\small{\frac{1}{4}}}\right)(p-1)^2$$
but it doesn't follow that
$$(q!)^2=\left({\small{\frac{1}{4}}}\right)\Bigl((p-1)!\Bigr)^2$$
You made another error when you suggested that
$$\left(\frac{p!}{p}\right)^2$$
"is clearly divisible by $p$".$\;$It's definitely not divisible by $p$.
Starting over, you can argue as follows . . .
Applying Wilson's Theorem, we get
\begin{align*}
&(p-1)! \equiv -1\;(\text{mod}\;p)\\[4pt]
\implies\;&
\bigl((1)(p-1)\bigr)
\bigl((2)(p-2)\bigr)
\cdots
\bigl((q)(p-q)\bigr)
\equiv -1\;(\text{mod}\;p)
\\[4pt]
\implies\;&
\bigl((1)(-1)\bigr)
\bigl((2)(-2)\bigr)
\cdots
\bigl((q)(-q)\bigr)
\equiv -1\;(\text{mod}\;p)
\\[4pt]
\implies\;&
(-1)^q(q!)^2
\equiv -1\;(\text{mod}\;p)
\\[4pt]
\implies\;&
(-1)^{2q}(q!)^2
\equiv -(-1)^q\;(\text{mod}\;p)
\\[4pt]
\implies\;&
(q!)^2
\equiv -(-1)^q\;(\text{mod}\;p)
\\[4pt]
\implies\;&
(q!)^2+(-1)^q
\equiv 0\;(\text{mod}\;p)
\\[4pt]
\end{align*}