In March 1995, a national health insurance policy was implemented in Taiwan. The program remburses women over 30-year of age for an annual cervico-vaginal smear(Pap smear). For those under 30-year-old, a few local governments also reimburse the cost of cervical smear by the local budget. The utilization rate of the service provided is estimated to be 13 to 15% of all women over 30-year-old.

        The effectiveness of the program should be constantly monitored as to the size of the target population screened and the quality of the cervical cytology test. An effective cervical cytology-screening program for detection of early cervical neoplasia is expected to decrease cervical cancer mortality in the years to come. Because the decrease in cervical cancer mortality has a long feed back time, it is not suitable as an on-going process of monitoring the quality and effectiveness of the program.

        Although it is important to increase the population screened, quality of cervical cytology tests is also as important. One of the least appreciated data in Taiwan is the positive rate of the test. The positive rate of cervical cytology diagnosis is a crude estimate of the efficiency of cervical smear creening program. The estimated false positive rate in Taiwan is 0.3%. Therefore, the positive rate of cervical cytology diagnosis reflects mostly the sensitivity of the test.

        First attempt at community-wide cervical cytology screening in Taiwan was conducted between 1974 to 1978 sponsored by Cancer Society, ROC(Pesus Chou, 1983). A positive rate of 1.3% was recorded. The high-grade lesions and invasive squamous cell carcinoma comprised of 0.62% of the subjects screened. The results are similar to the data from Sheffield, England, during her first cervical cytology campaign conducted in 1966 to 1968 (Table1). Even the ratio of CIS/Ca is similar. The ratio is an estimate of cervical neoplasm. In a well-screened population, the ratio is expected to be high. Considering the lack of properly trained cytotechnologists and cytopathologists at that time, I think a positive rate of 1.3% during the first such campaign in Taiwan is a resplasia, and England is not. Therefore, the data may be similar, but the sensitivity of two campaigns may not be compar able.

Table 1. Results of community-wide cervical cytology screening conducted in Taiwan.

Note: The statistics from major medical centers (the tertiary hospitals) are excluded, because they include a large number of referred known cervical cancer patients. In many cases the positive rate is in the range of 5-6% and the CIS/Ca ratio is less than 1.0.
*: Data are biopsy confirmed and extraplated to all cytology positive cases.

        The positive rate is dependent on the adequacy of cytological specimens and quality of laboratory examination. To focus on the adquacy of specimens as an important determinant for variation in positive rate, we compile three sets studies conducted by a single laboratory(Table 2). Taipei Institute of Pathology examines all the cytological specimens.

Table 2. Results of cervical smears examined by Taipei Institute of Pathology from three defferent sources.

Note: The data in the last two columns are biopsy confirmed and extrapolated.

        The first two sources of cytolgical smears are very different in its institutional characteristics. The THB are composed of community-based health stations, some of which do not have full-time physicians and are often staffed by physicians of various specialty on short-term basis. CHH is a 600-bed general community hospial. The Obstetric /Gynecology department is well staffed with strong leadership and active involvement in the cervical cytology-screening program. Since the CIS/Ca ratio is similar, we believe that they dealt with population of similar demographics. The positive rate for THB is 0.67% and that of CHH is 1.72%. An almost three-fold difference in "CIS and up" lesions is also noted. Such difference can be best explained on the basis of specimen adequacy. The last source (Taita-TIP) is a well-planned community-based research project sponsored by the Ministry of Health, ROC. In the studies, Cervex Brush is used, instead of the conventional Ayres spatula, for sampling of squamo-columnar junctions. Vaginal cell pools from posterior fornix are not collected. There was a constant feedback on the specimen adequacy to the phqsicians sampling the cytological specimens. The positive rate of 2.8% is the highest figure ever achieved in Taiwan on a community-based screening project and should become the goal to pursue in our current nation-wide screening program. The high CIS/Ca ratio is also noteworthy. It indicates that a carefully planned and executed program improve the sensitivity of detection of high-grade pre-neoplasm.

        The quality of laboratory examination is probably as important as the quality of cytological specimens obtained. Since implementation of National Health Insurance, a central information system established in the Ministry of Health enable us to examined the over all performance of the tests. The first set of data available to us shows that the overall positive rate of the test is 0.99%md the latest set, 1.19%. The are a few major laboratories which account for approximately 30% of the test performed annually have a positive rate in the range of 0.6 to 0.8%. We do not know if the problem resides in the specimen collections or laboratory examination, or combination of both. The results are apparently less than satisfactory and, in a few cases, even alarming.

       There are certain preliminary data on the nation-wide cytology laboratory performance. During the first attempt at assessing the accuracy of the positive diagnosis of cervical smears, the false positive rate is estimated to be 0.3%. A limited sampling of smears carrying"negative", diagnosis, the false negative rate is estimated to be in the order of 30 to 40%. It is important to point out that the estimate is preliminary and may be subjected to revision in the future. Because false positive rate is reasonably low, only 0.3%, the positive rate is a ref1ection of the sensitivity of laboratory screening. The low positive diagnosis rate of smears is mainly caused by high false negative rate.

       The screening sensitivity is amenable to assistance by automated image analysis. So far, FDA (Food and Drug Administration, USA) has approved two systems for cervical cytology laboratory quality control. The Papnet system has been introduced into Taiwan since 1995. Taipei Institute of Pathology has adopted the system to improve the quality of laboratory screening. We use combined screening of cytotechnologists and Papnet system. We have performed 14,506 cases. The overall positive rate is 3.75% as shown in Table 3. Also shown are age distribution and lesions identified, since it is known in Taiwan that older age groups have higher prevalence of high-grade lesion and invasive cancer. The age distribution of this set of cases is similar to other sets of data recorded. Nevertheless, an age-adjusted correction of the positive rate was performed and a similar figure of 3.56% was obtained.

Table 3. Combined screening with Papnet systemf conducted in Taipei Institute of Pathology from January 1996 to July 1997. (n=14,506)

Note: The age-adjusted positive rate is 3.56% vs. 3.75%

        It will be of great interest to compare the results with cases examined in Taipei Institute of Pathology with and without assistance of Papnet system. The Table 4 shows that the positive rate and rate of high-grade lesions are substantially higher.

Table 4.Results of the latest nation-wide data and combined screening with Papnet system.

Note: All data are cytological diagnosis.

        By combined screening with Papnet system, the positive rate of low-Fade lesions is three times and that of the high-Fade lesions is twice higher than the nation-wide average. The reasons for the better results may include: