Common errors in epidemiological measurements, their effects on numerator and denominator data and their avoidance