Several thoughts -
(a) 5 deg bank - relevant to Vmca rather than any consideration of climb performance. In the absence of any information to the contrary in the POH, one should presume that Vmca was determined with 5 deg into the operating engine
(b) You can turn an aircraft any which way ... provided you can keep min control speed Vmca is VERY bank dependent. If you bank the wrong way, Vmca goes up, significantly .. eg as I recall the B52 goes up something like 30-35 knots for 5 deg into the failed engine.
(c) Either way never get below V2 If Vtoss is not too distant from Vmca then, if conditions replicate Vmca, you risk going over on your back if you turn to the dead engine at Vtoss. Rarely are you given the actual data in a civil POH.
(d) best climb (minimum descent in many light twins at weight and DH) typically is around 2-3 deg into the operating engine and approximates nil sideslip .. this is true of just about any multiengine aircraft. Any other bank angle increases sideslip, drag, and down you go faster .. ergo OEI turns in light twins are a risky business if the level performance is marginal (which it near invariably is ..)
(e) turn into the dead engine CAREFULLY to achieve a tighter turn radius I've given it some thought but it still doesn't make too much sense .. I'm obviously missing some critical point in the suggestion ?
(f) Incr in bank means more drag not from wings to level to a few degrees into the operating engine ... if one wanted to nitpick ..