You will get a complete picture on sample size estimation from the following

Sample size estimation in epidemiologic studies

Article Sample size estimation in epidemiologic studies

Thank you Prof.Ajit kumar Roy , after I read the paper you recommended and found that I thought my problems in a comprehensive way.

Although I have two outcomes occured in sequnentially, I just need to figure it out one by one? For instance, in my study, population A has to have the outcome 1 and then the second outcome occured, so I use a proportion of outcome 2 as "p" to calculate a sample size firstly, then what should I do for the next?

Hi Esther Zhao,

Sample size calculation is based on the analysis you are planing to do and subsequently the parameter you are interested in to estimate with desired precision.

What type of analysis are you planing?

Would it be a survival analysis with multiple outcome? Something like a competing risk analysis?

Hi Sreenath Arekunnath Madathil,

Thank you for answering my question! I think you are right, it will be a survival analysis with multiple outcome and this is will be a open cohort without any intervention. As I mentioned above, population A has to have the outcome one and then the second outcome occured. Under this condition, how to decide a appropriate sample size based on two outcomes?

Would you like to give more support or details for answer? Thank you! I really do not have further ideas.

What would be your dependent variable of interest? Is it the time to outcome1 and time to outcome2? If yes, I think it is similar to accelerated failure time models. Not aware of simple equations to calculate for this analysis.

In which case simulations are the way to go. I would highly suggest you to meet a statistician for this one.

Hi Sreenath Arekunnath Madathil,

The dependent variables are both the time to outcomes and whether the outcomes would occur. However, maybe I prefer to more interested in whether the two outcomes will happend or not.

That would be kind of complicated....

Anyway, I think your suggestion will be helpful, sincerely thanks again for your help!

If the two outcomes are independent then estimate the sample size required for outcome 1 and 2 separately. Consider the larger sample size for your study there by your two outcomes will be achieved.

Hi Kandasamy Ravichandran,

Thank you for your help! While the problem is the two outcomes are not independent, if the outcome 1 did not happend the outcome 2 will not occur in anyone. Under this situation, what can I do to calculate the sample size?

Hope you will give me a hand!

You must consider that in spite of exist two hypotheses, the sample size have to achieve a minimum number of participants. Due to the foregoing I recommend you calculating both samples and using the bigger one in order to satisfy the research objectives.

Hi Víctor Hugo González Cárdenas ,

I think I may understand what you say, calculate sample size twice time and pick the biggger one as to be the more appropriate. Thank you for your answer!

For a proper answer more details like what is your two outcomes, who is your cohort etc. are needed. Ons simple solution is consider one of the outcome as primary and the other as secondary then calculate the sample size based on the primary outcome.