As reported in the Scientific American, new research has shown that it doesn't help make any "risk assessment" more accurate. On the contrary it was reported to possibly cause hurt - to rely on a formulaic way to predict the risk of a suicide. Many have debated using different forms of psychometric tools or predictive actuarial risk assessment tools, but ultimately, when someone actually takes their life, the clinician is left in shock and dismay.
Clinicians with experience will tell you that if people wanted to die, there was no way in hell that you could do or say anything that could prevent them from topping themselves. Rather, why not unpack and show these individuals the context of their suicidality and help them understand why they wanted to die.
Here's the original article - https://www.scientificamerican.com/article/suicide-risk-assessment-doesnt-work/
When is the right time, the right time? What if I terminate too early? What if I'm just holding onto my client? Consider a different perspective.