Words jumped in output using hebrew

Mixing L-R and R-L languages creates a lot of potentially ambiguous situations. Pierre MacKay's article on TeX-XeT from way back when addressed some of these. What's happening is that it's treating the HCDI-WG (and the later English text) as being part of the surrounding Hebrew (so L-R embedded in R-L) rather than the Hebrew being embedded in the surrounding English. I don't have everything installed to be able to test this (or even run your example—I'm getting a missing font error that I'm too lazy to resolve), but you should be able to enclose your Hebrew in R{...} to get your desired result.

Edited to include following (untested) MWE:

documentclass{book}
usepackage[utf8x]{inputenc}
usepackage[hebrew,russian,english]{babel}

begin{document}
begin{itemize}
item
  textbf{Form}: WG
  textbf{Contributor}: Hila Gendler Shalev, Tel-Aviv University
  textbf{Citation}: 
  Gendler-Shalev, H. (2005). 
  R{התאמת שאלון  הורים} 
  HCDI-WG 
  R{לעברית} 
    [The adaptation of HCDI-WG parent questionnaire to Hebrew] [Master's Thesis, Tel Aviv University, Tev Aviv, Israel].
end{itemize}
    
end{document}