How to facilitate (or discourage) community-based research: recommendations based on a Canadian survey