Notes:
When clinicians and patients collaboratively decide on a treatment, they need to understand how much better the discussed treatment is compared to an alternative. For categorical outcomes, metrics such as NNT (Number Needed to Treat) are commonly used. NNT and NNH (Number Needed to Harm) indicate the number of patients that need to receive the intervention for one additional positive (or negative) outcome compared to a control. However, for continuous outcomes, standardized mean difference (Cohen’s d) is typically used as the effect measure. Clinicians and patients may find it challenging to interpret the effect size based on Cohen’s d alone; thus, converting d to a risk difference or NNT can make it more understandable.

See Reference 4 for a systematic review using this method, or click here to access the methodology in Reference 2.

1. When the control event rate is not provided, the method by Kraemer & Kupfer ^[1] is used
2. When the control event rate is provided, the method by Furukawa & Leucht ^[2] is applied.