SQL question in a high volume web app environment

Question

Hi,
We have a high volume website used in a hospital setting. When a doc selects a patient, we have to delete all previous patient history records for the patient from a certain table and insert new records retrieved from a web svc called at runtime during patient selection.
Our 2 options are:
Option1:
1. Call Web svc, and always insert the results into pt_history, even if this creates duplicates (which it normally shouldn’t, if we only call every 24 hours)
2. Only pt claims_history records that are less than 24 hours old from app method calls
3. Nightly batch job cleans out records older than 24 hours
Option2:
1. Delete all existing history recs for a patient each time before calling web svc
2. Call web svc, and always insert the results into pt_history (no chance of duplicates here)
3. No change to existing methods
4. No nightly batch job
I'm leaning to option 1 to avoid contention\locks\deadlocks on the table. Any other solutions out there?
Also any insights into which may be a better option this being a high volume transaction (several 10s of calls every sec) and table could grow to be oretty large (2 million recs and more)
Thanks,
JGP

Answer 1

1) Why do you need to "delete all previous patient history records"? That seems contrary to the point of having patient history records...
We want to keep these records only for 24 hrs. Older than 24 hrs is not needed.
2) If you are just using this table to cache results from the web service call, could you use a temporary table?
We wan to cache across all sessions for 24 hrs. So temp table would not work.
Thanks,
JGP