Have you had a read around to see how validity and reliability has been ensured by other authors who have done similar in their studies? Of the top of my head, there are a few things that you can maybe do.
1) Have a totally transparent, systematic approach to data collection from your sources.
2) Maintain an audit trail, documenting clearly the flow and processing of the data.
3) Member checking to ensure your approach is reliable - you draw the data using the prescribed approach, someone else (the more, the merrier) draws the data using the prescribed approach - any systematic differences?
4) Maybe you can demonstrate validity by having some kind of focus group with a sample of clinicians/researchers, who can reach a consensus on when whether the data for your research represents what it is supposed to.
To be honest, you haven't given very much information on your research - so it's a hard one to answer. Most important of all, read, read and read around.