The Case-Control Approach Can be More Powerful for Matched Pair Observational Studies When the Outcome is Rare