Timing bias refers to any issues with the timing of the intervention that could impact the outcomes. For example, in the ALPS study amiodarone was no better than placebo in the treatment of out of hospital cardiac arrest. (Kudenchuk 2016) However, treatment was not given until an average of 19 minutes into the resuscitation, which might be too late for it to have any meaningful effect. Despite being a very well done trial, the conclusion could be wrong if it turns out that giving amiodarone earlier in the resuscitation is helpful.
Timing bias is a type of intervention bias.
Timing bias is different from lead time bias, which is a type of selection bias.
This post is part of a series of posts on bias in medical research. You can find the whole bias catalogue here.
You can find more evidence based medicine resources here.
References
Kudenchuk PJ et al. Amiodarone, Lidocaine, or Placeboe in Out-of-Hospital Cardiac Arrest. NEJM 2016.