A sample size calculation was performed assuming that a clinically significant difference between the groups was for 33% more forms to include all of the critical outcome measures. It was estimatedthat 73 referrals would be required to detect a 33% difference