Even though Papanicolaou and Traut demonstrated, as long ago as 1941, the potential for cervical cytology (the Pap test) to detect pre-malignant change or early invasive cancer in an apparently clinically normal cervix; cervical cancer (CC) is the second most common female cancer in the world and accounts for 15% of all cancers. In the developing world, where 80% of CC cases are diagnosed, it is the most common. The total number of deaths worldwide due to this type of cancer is 235,000 annually, and it is estimated that if by the year 2,000 an effective CC prevention program is not introduced, this figure will rise to 276,000 deaths per year.

Given that in some countries programs that use the Pap test have failed, especially where funds for health care are limited, alternative strategies have been suggested for preventing CC deaths. These include the theoretical possibility of anti-human papillomavirus (HPV) vaccine, as there is no vaccine yet available, because current evidence indicates that HPV infections are the main cause of CC, particularly the sexually transmitted types (16, 18, 31, 33, 35, and 45), and direct examination of the cervix. Nevertheless, Pap test screening detection programs still remain the most feasible and cost-effective tools available for reducing CC deaths.

Cervical cancer screening was introduced, without previous evaluation, by means of randomized controlled trials. Evidence of the benefits of screening is shown by the following: (a) comparisons of mortality rates before and after the introduction of screening programs; (b) comparison of the incidence of cancer or mortality rates in regions or countries with different levels of screening; (c) computer simulation; and (d) epidemiologic studies.

The effectiveness of a CC screening strategy is judged by the reduction in the incidence of invasive disease and the number of deaths. Screening programs have been in operation in parts of Europe and North America for over 35 years, and it is in the Nordic countries that an organized screening campaign has shown the best results in reducing death rates. Comparisons have been made of mortality trends in Denmark, Finland, Iceland, Norway, and Sweden, countries with similar educational and socio-economic structures, which share common behavioral and lifestyle patterns and where health care systems are similar. In Iceland, where the widest age range was covered, mortality was reduced by 80%; in Finland, the figure was 50%, and in Sweden, 34%. Where coverage was lower, mortality was reduced by 25%; Norway, with a screening rate of only 5% (until the 1970s), had at 10% the smallest reduction in CC deaths. World Health Organization (WHO) data (world standard-adjusted) on rates of CC mortality for the years 1950–1954 to 1985–1989 show a significant decline in the overall age-adjusted rates in all countries in Western Europe, the U.S., Canada, Australia, and New Zealand. Rates have also declined in Japan, Hong Kong, and Singapore, while they have remained fairly steady in Eastern Europe, Mexico, Venezuela, Colombia, and Chile. Gustafsson et al. determined the incidence rates of invasive CC before cytological screening was introduced. The simulation model found essential similarities in age-incidence curves in different parts of the world, suggesting that the development of CC is similar, irrespective of etiological and ethnic differences.

Mexico has had a national cervical cancer-screening program (NCCSP) since 1974. In spite of this, during the period from 1980–1985 the number of officially reported deaths from this disease was close to 62,000. To obtain mortality trends, data were analyzed from death certificates issued in Mexico during the period from 1980–1985 in which the cause of death was given as CC; regional variations were obtained for the same period by using a Poisson regression model. Twenty-four states in Mexico showed an increased mortality risk when compared with Mexico City; seven states showed a steady or downward trend. Although the NCCSP in Mexico has had a sufficient infrastructure and the resources to conduct 3,516,000 Pap tests each year for an estimated population in 1996 of over 16.5 million women between 25 and 65 years of age, the official Mexican protocol (until 1996) recommends annual Pap tests for women after they have become sexually active, and does not establish an upper or lower age limit. In addition, although the official diagnostic nomenclature used in gynecologic cytology is based on dysplasias, some health institutions in Mexico continue to use the now obsolete Papanicolaou nomenclature. Until 1996, women with a colposcopic diagnosis of HPV, and even including those with mild dysplasia, were sent to a colposcopy clinic for treatment with cryosurgery, electrosurgery or lasers, according to Mexico’s official technical protocol; this may well have led to high costs with few benefits. In addition to these problems, the Mexican screening program lacks the epidemiologic surveillance mechanisms that could ensure follow-up and treatment of the abnormalities detected.

The authors present an NCCSP evaluation model based on the idea that it is composed of three basic elements: women at risk, healthcare providers, and health service use. The descriptive studies are combined with population-based case control and cross-sectional studies. The data analysis was carried out for combination, variance analysis, partial correlation coefficients, and no conditional logistic regression. All of the women, from several studies, were invited to participate; they received information concerning the objectives, benefits, and the confidentiality of the information. If they agreed, they were asked to read and sign the consent form.

The National Institute of Public Health of Mexico carried out an evaluation, until 1997, of the flaws found in the following steps of the screening procedure: Pap smear sampling quality; cytological diagnosis validity; the effectiveness of Mexico’s screening program; compliance of women according to their CC risk level; and determinants of failure to participate. A detailed analysis by means of an epidemiologic study, as described previously, was carried out for each step. The principal discovery was that the lack of NCCSP effectiveness in Mexico is due mainly to factors associated with quality and coverage.

Pap smear quality is deficient; a randomized sampling of 6,011 smears culled from 31,378 smears diagnosed as negative in a regional sample of the screening program in Mexico City were re-evaluated in terms of specimen quality. Of the samples evaluated at the Mexico City General Hospital, 64% did not show the presence of endocervical cells, mucus, and squamous metaplasia.

Samples were also taken at 16 Ministries of Health and at the Mexican Institute of Social Security cervical cancer-screening centers throughout Mexico. Each center received a randomized batch of 90 cytology specimens with a positive prevalence [from cervical intraepithelial neoplasm (CIN) II to invasive cancer] of between 1.5 and 36%. To simulate routine working conditions, diagnoses from a given observation unit were made by different cytotechnologists. The reading centers presented overall sensitivity indices of between 46 and 90%. According to the gold standard, of the 1,440 smears taken at the 16 diagnostic centers, 315 were positive, 78 specimens were misdiagnosed as negative, 25 of which had moderate to severe dysplasia, 16 cancer in situ, and 37 invasive CC. The mean cross-center sensitivity was 65%.