The right indicator for the job: different levels of rigor may be appropriate for the development of quality indicators. Comment on Stelfox and Straus.