Estimation Prevalence Of The Cervical Cancer And The Precancerous Lesions Of Jiangsu Province

Objective: In order to estimate the prevalence of the cervical cancer and the precancerous lesions in Jiangsu Province including southern Jiangsu, central Jiangsu, northern Jiangsu through the population-based screening, and estimate their age-adjusted prevalence and the 95% confidence intervals when the age compositon is different from the standard population structure.Research subjects and methods: Based on the study on cervical cancer screening specifications and follow-up of the women married in Jiangsu Province, we have the six project sites completedly in 2007 of a total eligible sample of 77187 cases. Number of patients in this study consists of three parts: the patients of pathological diagnosed in the pap smear positive, the patients of not diagnosed in the pap smear positive and the patients of the false-negative in the pap smear negative. Estimated the number of the patients of not diagnosed in the pap smear positive by bootstrap re-sampling technique, and estimated the number of the patients of the false-negative in the pap smear negative by the prior information of pap smear sensitivity and specificity from the Meta-analysis. When the age structure of these regions are different from the standard population structure, we need estimate the age-adjusted prevalence and their 95% confidence intervals by the normal approximation method, the exact method and the Gamma distribution method.Results: The prevalence of the cervical cancer and the precancerous lesions in Jiangsu Province of 35-65 years married women in 2007 is 0.81% (95% CI: 0.73%, 0.92%), which southern Jiangsu is 0.56% (95%CI: 0.48%, 0.66%), central Jiangsu is 0.78% (95%CI: 0.70%, 0.88%), northern Jiangsu is 0.88% (95%CI: 0.79%, 1.00%). The difference among the age groups is statistical significant in Jiangsu and the three regions (p<0.001, p<0.001, p=0.020, p=0.001). The population under the age of 50 has the higher prevalence than the population over the age of 50 in southern Jiangsu and central Jiangsu. In northern Jiangsu, the people with 45-49 years and 60-65 years has the first highest prevalence, the second is age of 40-44, and the third is age of 50-59. Northern Jiangsu has the highest prevalence, the second is centre Jiangsu, the third is southern Jiangsu, the difference is statistical significant (p=0.001). After the age-adjusted, the prevalence is 0.77% in Jiangsu (95%CI of the normal approximation method: 0.71%, 0.83%; 95%CI of the exact method: 0.71%, 0.83%; 95%CI of the Gamma distribution method: 0.71%, 0.83%), which southern Jiangsu is 0.51% (95%CI of the normal approximation method: 0.39%, 0.63%;95%CI of the exact method: 0.40%, 0.64%;95%CI of the Gamma distribution method: 0.40%, 0.65%),central Jiangsu is 0.79% (95%CI of the normal approximation method: 0.69%, 0.89%;95%CI of the exact method: 0.69%, 0.90%;95%CI of the Gamma distribution method: 0.69%, 0.90%),northern Jiangsu is 0.89% (95%CI of the normal approximation method: 0.79%, 0.99%;95%CI of the exact method: 0.79%, 1.00%;95%CI of the Gamma distribution method: 0.79%, 1.00%). Age-adjusted prevalence remains the highest in northern Jiangsu, central Jiangsu in the second, southern Jiangsu has the lowest prevalence of the three regions, and the difference is statistical significant (p <0.001). The three ways for estimation the 95% confidence interval is consistent.Conclusion: In this study, diagnosed by the pathological diagnosis (gold standard) only some persons of pap smear positive when there is lost, but we estimate the prevalence of the cervical cancer and the precancerous lesions in Jiangsu Province is well lower than the high-incidence areas and slightly higher than the low-incidence areas through Bayesian theory and the bootstrap re-sampling technique. The prevalence is different among the three regions. Although it is low for the prevalence of some chronic disease, the rate and its 95% confidence intervals are perfectly consistent from the large population-based study by the normal approximation method, the exact method and the Gamma distribution method. So we may apply to the normal approximation for estimating the age-adjusted prevalence in practice.