In the absence of cancer registry data, is it sensible to assess incidence using hospital separation records?

In the absence of cancer registry data, is it sensible to assess incidence using hospital separation records?