From my standpoint, I'd prefer maybe not to look toward a phase-out but toward an agreed approach on developing what the acceptable risk is for the product and then taking steps to mitigate that risk.
Maybe in the three-year period we can re-evaluate how the measurements of these risk mitigation strategies worked. At that point, if there is still a problem or if there is new science, we can have another conversation around whether the value proposition of the risk is great enough.