Monitoring Data Quality in Syndromic Surveillance: Learnings from a Resource Limited Setting