Could someone please tell me how i can go about calculating NNT (number needed to treat) or NNH (number needed to harm), with continuous data? I am a student in the health profession and am delivering a Grand Rounds seminar.

To reflect the benefit of a certain drug to a patient population and to determine the risk or probability of a patient either developing an intended beneficial drug effect or an adverse reaction, we calculate an event rate vs a placebo, which is the Absolute Risk Reduction --> (n/N placebo group) - (n/N drug group), and then calculate the inverse to be the NNT or NNH --> 1/[(n/N placebo group) - (n/N drug group)]. Usually, i have no problem calculating this, however this time my study only reported it's results as continuous data (i.e. as numerical change on a scaled score (ex. -4.8 change on ABC scale, or +2.5 change on DEF scale), without any accompanying info in the results about # patients reporting the change.

My question is, is there a way that i can calculate NNT and NNH with a the continuous data numerical changes in scales, or can i use the confidence interval or something to calculate it? I would really love it if someone could please help me out with this, as i'm so confused and really could use some clarification.

Thanks in advance. Hopefully someone will respond...