Assessing agreement between two measurement systems: an alternative to the limits of agreement approach