problem with cointegration in Heterogeneous panel

[amirpooya]

Hi every one.
I've been having a question from one month. at least I understand I couldn't answer it without help! here is my question.
I estimate unbalanced panel with 6 period & 20 cross-section.
I used im-pesaran & shim also fisher ADF for testing unit root. 5 of my variables are I(1) & one is I(0)!
now I couldn't run pedroni or fisher for testing cointegration among them. because Eviews give me error: insufficient number of observation.
I can use just Kao test. & it confirm long-run relationship . Is it true? Can I trust to this result?
I will be very appreciate it if anybody can help me! please help me!