Computer simulation as a method for evaluating the copper sulfate test for hemoglobin in diverse populations.

Abstract

The copper sulfate hemoglobin screening method was separately evaluated for three local population groups which do not conform to accepted normal standards of hemoglobin and/or plasma specific gravity (SG). This was carried out by a computer which simulated the distribution of plasma specific gravity and hemoglobin levels in random samples of blood donors. Simulations of 1,000 were carried out for males and for females in each of the three main population groups in Natal (Caucasians, Negroes and Asians). These were compared with a simulation of hypothetical normal subjects whose hemoglobin and plasma SG conformed to accepted standards. Variations in accuracy, relating to differences in plasma specific gravity and in hemoglobin, were demonstrated. Acceptances of donors who should be rejected (Type II errors) were most common in Negroes, but did not occur in the normal group. Type II errors did not occur in males with hemoglobin levels below 12.5 gm/dl or in females under 11.5 gm/dl. Erroneous rejections of prospective donors who, in fact, satisfied the criterion of acceptance (Type I errors) occurred in all groups, but were fewest in the normal group.