D r. Renshaw's thoughtful review article 1 reminds us of the major ambiguities that remain in assessing the vital issues of overall Papanicolaou (Pap) screening sensitivity, assessing comparative accuracy among individual laboratories, and assessing adherence to "the standard of practice." Despite the success of the Pap test in decreasing the rate of cervical carcinoma in the U.S. by ΟΎ 70%, public expectations of 100% effectiveness have never been achieved in any screened population reported to date. 2,3 In essence, potential public "disappointment" with cervical carcinoma screening reflects the gap between the capability of the Pap test to prevent cervical carcinoma and cervical carcinoma deaths and the expectations of the public. The unrealistically high expectations of the public are especially reflected in the legal system, in which lay juries actually serve as the ultimate arbiters of "the standard of practice." For example, a trial lawyer's newsletter from 1997 recently opined that a claim for medical malpractice against some health care provider almost certainly exists in any case in which a woman develops cervical carcinoma and undergoes a hysterectomy or dies, unless the woman utterly failed to obtain even periodic Pap smears. 4 The expectation expressed here is that the Pap test is the equivalent of a cervical carcinoma insurance policy that will pay a substantial dollar benefit if a woman who received any Pap testing, even a woman who underwent screening only periodically, develops cervical carcinoma or dies of cervical carcinoma.
Available evidence clearly indicates that the Pap test has significant limits for effectiveness. For example, to my knowledge all studies published to date have demonstrated that as the frequency of Pap screening decreases from annual testing to less frequent testing, the incidence rates of cervical carcinoma diagnosis and death increase. [5][6][7] Nevertheless, even doubling or tripling the relative risk of developing cervical carcinoma in less frequently screened women still may be discounted as only "small" increases in "absolute risk" and may be believed to be consistent with an "acceptable" (to whom?) and "low underlying probability of disease." Epidemiologic model evaluations, which nearly always are taken from a group rather than an individual patient perspective, have argued that "efficient" and "cost-effective" screening programs for cervical carcinoma can add up to only 32.4 days of average life expectancy, compared with a gain in life expectancy of 46 days with the total elimination of cervical carcinoma. 8 This type of model evaluation suggests that the last 30% gain in life