Special Feature: The Cost Effectiveness of Pap Smear Screening
By Kenneth L. Noller, MD
I am very worried about the future of cervical cytology screening. The Pap smear is the single best cancer-screening tool ever developed. It has resulted in approximately a 70% reduction in cervical cancer in all countries in which it has been used regularly. It is not perfect. A single smear probably misses almost as much disease as it picks up. However, cervical neoplasia is such a slowly progressing lesion that even if it is missed the first time, it will certainly be discovered if the woman is screened as few as several times a decade. An ever-higher proportion of cases of cervical cancer in the United States are among those few women who are never or rarely screened.
Despite this remarkable record of success, more and more techniques are being developed for the purpose of improving cervical cytology sensitivity. The introduction of the endocervical brush was one of the first advances. Unlike the more recent techniques, the brush was relatively inexpensive. The more recent changes (liquid-based cytology and HPV-DNA testing) will add tremendously to the cost of Pap smear screening if we continue our current practice of screening every women every year.
Why are we so convinced that annual cervical cytology screening is a good idea? Of course the best answer is that it has been demonstrated to work. Nonetheless, there is little evidence in the literature that screening needs to be annual, and a growing body of evidence that less frequent screening is just as effective and more cost efficient. Recently, 2 articles appeared in the Journal of the American Medical Association that used advanced modeling techniques to examine the cost effectiveness of various methods of evaluating ASCUS Pap smears and the possible role of HPV-DNA testing in screening.1,2
It is often difficult for us clinicians to read cost-effectiveness articles, as the mathematical modeling techniques are quite difficult to understand. Perhaps the best way to approach reading articles such as these is first to identify the underlying hypothesis, then to look at the assumptions made by the authors in the methods section, and finally to read the comments. Often, many of the details in the methods section as well as the results section are meaningless, clinically. Rather, it is the interpretation of the results from the model that is most important. Both of these recent articles have done a good job of developing their model.
One thing should be very obvious to any clinician who reads a cost-efficiency paper: The assumptions made by the authors are always open to criticism. Much like a case-control study where it is always possible to criticize the controls, the assumption made by the authors concerning costs, frequency of abnormalities, etc, may not be in concert with our own practices. Those who carefully read the literature before attempting to apply modeling techniques usually wind up with generally acceptable results. I believe the JAMA article that reviewed the various methods of management of ASCUS Pap smears represents such an article. While I can find things in it with which I disagree, overall it is based on excellent clinical information and its results are probably true.
But back to my main point. I am very worried about the future of Pap smear screening if we continue to believe that every woman needs a Pap smear every year. When the test only cost $8 or $10, such an approach was reasonable. Now, when a Pap smear using liquid-based cytology can easily be billed at $50-100 and if HPV testing is indicated an additional $100 added, will third-party payers continue to allow us to screen every year?
I think the answer is surely "No." Rather, third-party payers are going to demand that the whole area of Pap smear screening be examined carefully, and that an entirely new system of screening be developed. This might mean less frequent screening, but with better techniques. It might mean fewer colposcopic examinations for women with minimally atypical smears, and hopefully will result in much less "treatment" of LGSIL lesions in young women. While there is no official position as yet, the American Cancer Society is currently examining the whole area of cervical cytology screening and should be ready to report within the next few months. I would be amazed if we do not see major changes in their new recommendations compared to our general practice of annual screening.
Perhaps it will be easier for all of us to change our practice if we remember that there never was a scientific basis for annual Pap smear screening. Rather, at least in the United States, annual screening developed more as a result of the availability of oral contraceptives in the 1960s when a woman could not obtain a refill for her OC’s without a Pap smear than from data. While yearly cytology reduces cervical cancer, so does biennial or triennial screening as demonstrated in many published trials.
1. Mandelblatt J, et al. JAMA. 2000;287(18):2372-2381.
2. Kim JJ, et al. JAMA. 2000;287(18):2382-2390.