On the Bias Arising from Relative Time Lag in COVID-19 Case Fatality Rate Estimation