An economic model for comparing alternative policies for cervical cytologic smear screening

In this paper a mathematical model is developed to determine the probability that a truly negative cervical cytologic smear will be correctly identified by a system of screening policies that uses one or more cytotechnologists to independently and sequentially prescreen such smears for the detection of cervical cancer. This is an extension of previous work that modeled the probability of detecting a truly positive smear under the same set of policies. In this system any positive reading by a cytotechnologist causes a slide to be rescreened by a pathologist, and if all cytotechnologists declare a slide to be negative, the slide is placed in a pool for random selection of slides to be rescreened by the pathologist. The policy of single screening by a cytotechnologist, with subsequent 10% rescreening of the negative slides, as suggested by the Clinical Laboratory Improvement Amendments of 1988, is thus embedded in the policies that are modeled. In addition, a cost model is developed that takes into account the cost of screening a slide by a cytotechnologist, of rescreening a slide by a pathologist, of a false-positive reading and of a false-negative reading. This cost model can be used to determine which policy is optimal for the parameters that pertain to a specific situation. Examples are presented to illustrate the use of the cost model. The results of a computer simulation model are also presented to validate the mathematical results and to display the variability of total cost.