It's no less than 12 hours but ideally right before the next dose for venlafaxine and ODV (its active metabolite) plasma levels. The reference value where most will respond is when the ven+odv plasma level is between 195 and 400 ng/mL and risk of toxicity is when levels are over 800.
Venlafaxine has a dose-response relationship for MDD. There's evidence that you can push up the venlafaxine dose above 375mg/day up to 600mg/day if plasma levels are low, and above those doses dopamine reuptake is also blocked in some patients. There's a recent study of this for desvenlafaxine as well.
This is probably related to p-glycoprotein, which kicks drugs like venlafaxine/desvenlafaxine (as well as other SSRI/SNRIs) out of the CNS rather than metabolization by CYP P450 enzymes in the liver (for venlafaxine it's CYP2D6 but for ODV it's UGT1A1 for ODV, a glucuronidation phase II pathway and not phase I/CYP P450 pathway, and one that doesn't come up on typical pharmacokinetic genetic tests).
The gene test for p-glycoprotein is probably the most useful genetic test as patients were
3 times more likely to remit from depression with an NNT=3. This was followed up by a
second study for desvenlafaxine which suggests a dose-response relationship above 50mg up to 200mg whereas the PDR states that there's no evidence of more benefit above 50mg/day.
This is an even bigger deal if the patient is on another medication that is a strong p-glycoprotein inducer (most likely the case if pt isn't responding) or inhibitor. Really wish there was a way to get genetic testing for ABCB1 & ABCC1 in the US, which are the genes that code for p-glycoprotein.