Page 67 - Screening for Cervical Cancer: Systematic Evidence Review
P. 67
Chapter III. Results
the laboratory facility. At the University of North Carolina Healthcare System, the Hybrid
Capture II test kit cost $43 in February 2001. The clinical laboratory fees were $125 for Hybrid
Capture II testing and $160 for consensus PCR at the same time. As any other diagnostic
modality, increased demand should eventually increase capacity, in part through competition,
and lower costs. These market forces are not yet at work, and costs have not converged or
declined.
For cost modeling purposes this uncertainty about the expense of the test can easily be
handled by varying the model estimates of test cost. Adequate estimates of the cost of primary
care office-based cervical cancer screening and resulting evaluation and treatment relevant to the
32
US health care system are available. However, the remaining parameters required for cost-
effectiveness analysis hinder such estimates. Factors such as repeatability of HPV test results,
incidence and types of HPV by age category, rates of progression and regression of HPV, and
sensitivity and specificity for predicting outcomes over varied intervals of time are inadequately
defined for using HPV tests in screening.
No recent literature specific to the United States is available to model and compare
effectiveness or costs of screening or triage uses of HPV testing that have been proposed. The
most complete contemporary model to date, constructed based on the United Kingdom screening
system, considered adding HPV to Pap testing, replacing Pap testing with HPV testing, and
adding HPV testing to surveillance of low-grade lesions versus continuing conventional
surveillance. The authors concluded: The uncertainty, as expressed by the differences between
models, is so large, the results are inconclusive. Adding HPV testing to cervical cancer
screening may or may not improve the (cost) effectiveness of screening. There are relatively few
longitudinal HPV screening studies with enough time lapse between measurement points to
67