The Study Of Social Inequality In Cervical Cancer Mortality And Screening For Efficient Screening Pattern In China

Background and objectives:Cervical cancer is the second gynecological cancer worldwide and most common in the middle-and low-income areas.Population-based cervical cancer screening as the predominent prevention method could effectively decrease the cervical cancer incidence and death rate.Although cervical cancer death rate continually decreases in China during 30 years,Although the cervical cancer death rate continually decrease in China during 30 years,the prevention and control of cervical cancer still has been facing two challenge:1)the great gap among different social groups and geographical areas are ignored because of entirely downtrend in cervical cancer death rate;2)the great discrepancy between limited public health resource in China and low efficiency in population-based screening.Unfortunately,some studies suggest that the cervical cancer incidence and death rate are increasing since 2000.So it is essential to enhance the efficiency of usage of public health resource and establish effective population-based screening mode considering the lack of medical resource in China.Some limitations are found in current studies:1)current studies could not reveal the geographical variants in cervical cancer death rate or screening coverage because of lack of systematic research in content and method,which only focused on epidemiology,secular trend and explanatory of influence factors;2)the research on social inequalities of cervical cancer death and screening is still in a blank so that it could not describe the pattern of death and screening among different socioeconomic groups;3)there is not research about the impact evaluation of the National Rural Women Cervical Cancer Screening Program.Therefore,these researches could not supply enough evidence for designing localized population-base screening mode and improve screening efficiency.This study is to provide evidence-based suggestions for health maker to design efficient and localized population-based screening mode for decreasing cervical cancer mortality according to the aspect of mortality,screening coverage and social inequalities,through:i)analyzing the cervical cancer mortality and mass screening coverage and regional disparities;ii)revealing the social inequalities magnitude and pattern in cervical cancer mortality and mass screening;iii)assessing the impact of national rural women cervical cancer screening program on coverage and magnitude of social inequalities.Materials and Methods:The dataset included in this study are abstracted form Global Burden Disease 2013(GBD2013)dataset,National Disease Surveillance Points Cause-of-Death dataset 2006-2012(CoD2006-2012),Chinese Chronic Diseases and Risk Factors Surveillance 2013,National Cervical Cancer Screening Program in Rural Areas(NACCSPRA)list 2014 and the Sixth Population Census Dataset.The study methods are as follows:1)The Analysis of Cervical Cancer Mortality and Social Inequalities in China.Two dataset were used in this study,including GBD2013 dataset and Cause-of-Death Surveillance dataset(2006-2012).The GBD2013 adopted uniform diseases definition and classification and analysis method to assess the disease burden between 1990 and 2013 in China;the CoD2006-2012 covered 161 counties/districts and the death results were adjusted by two underreporting investigation.This study abstracted cervical cancer death burden from GBD2013 and used descriptive analysis method to analyze the absolute number and crude or age-standardized rate of death and Disability Adjusted Life Years(DALY)in cervical cancer,analyze the long-term trend in 1990-2013,provincial results were mapped to show regional disparity.The direct standardization method was used to calculate age-standardized death rate for cervical cancer in 2006-2012.The Joinpoint regression model was performed to obtain annual percentage change and 95%confidential interval(95%CI)for assessing the short-term trend of mortality rate in cervical cancer from 2006 to 2012.The Possion regression-based relative index of inequality(RII)and slope index of inequality(SII)were chose and calculated to show the magnitude and pattern of social inequalities induced by geographic-level education,income and employment in cervical cancer mortality in 2006-2008 and 2010-2012,respectively.2)The Analysis of Cervical Cancer Screening Coverage and Socical Inequality in China.The data used in this study were from 2013-2014 Chinese Chronic Diseases and Risk Factors Surveillance(CDRFS2013)and The Sixth Census 2010.The CDRFS2013 was a population-based,multipurpose survey conducted every three years by the National center for Non-communicable Disease Control and Prevention.The surveillance system adopted a multi-stage stratified cluster sampling strategy and randomly selected 298 districts or counties as primary sampling units(PSU)from 31 provinces,autonomous regions and municipalities in mainland China,with stratification by population size(high/low)and mortality rate(high/low)of each province.A total of 91,816 eligible women aged 21 years or above were selected by multiple stage cluster sampling method,and information related to experience of cervical cancer screening,demographic and socioeconomic factors were collected through face-to-face interviews.Taking into account the complex sampling design,95%CI of measurement were estimated using survey data analysis methods in SAS 9.4(SAS Institute Inc.,Cary,USA).Prevalence with 95%CI of cervical screening were calculated for 31 provinces,autonomous regions and municipalities in mainland China,not including Hong Kong,Macao Special Administrative Regions,and Taiwan.These prevalence rates were mapped into a visual geographical map.Rao-Scott X2 was used to compare the screening coverage within subgroups.A series of random intercept equations with a logit link function were fitted according to three levels:individual,county and province.The fix effects of all explanatory variables were converted into OR with 95%CI to represent the magnitude of the impact of SES factors.Random intercepts,represented by variance and standard error(SE)at each level,accounted for geographical variations within counties and provinces.We also used log-binomial logistic regression to assess the RII and SII and 95%CI induced by individual-level education,household income and occupation type in cervical cancer screening.The magnitude and pattern of social inequalities were also mapped.3)The Impact Evaluation of National Cervical Cancer Screening Program in Rural Areas on the Population-based Coverage.This study design comprehensive evaluation strategies based on stratification analysis and multivariate regression analysis,propensity score method(PSM)and qualitative research.We divided the eligible subjects in CRDS2013 into two groups,including intervention group and non-intervention group.The primary sampled units(PSU)covered by NCCSPRA were considered as intervention group.We chose the Women aged 35-64 years as target group and screening rate at least once within five years as the evaluation index.The differences in cervical cancer screening rate between intervention and non-intervention group were calculated.The multilevel regression model was conducted to analysis the impact of intervention,and the interaction between intervention and living in rural areas show the magnitude of effect.The PSM was also applied for impact evaluation of intervention on cervical cancer screening coverage rate and social inequalities magnitude among targeted women.1:1 matching sample were conducted by greedy match method.Additionally,we adopted field investigation and qualitative interview with key stake holders to provide evidence for quantitative evaluation.Results:1)The Analysis of Cervical Cancer Mortality and Social Inequalities in China.In 2013,there were almost 65,943(95%UI,41,744-81,853)new diagnosed cervical cancer cases,28,010(95%UI,18,711-33,877)deaths from cervical cancer and age-standardized death rate was 3.7 per 100,000(95%UI,2.5-4.4 per 100,000),accounting for about 0.7%(95%UI,0.4-0.7%)of death.The rates of mortality from cervical cancer varied widely among provinces,with age-standardized death rates ranging from 1.5 per 100,000(95%UI,1.0-1.9 per 100,000)to 6.7 per 100,000(95%UI,5.0-8.6 per 100,000).At the national level,the age-standardized death rate decreased by 38.1%(95%UI,-60.8 to-23.0%)between 1990 and 2013.At provincial level,the decline in age-standardized rates was significant in 24 of 31 provinces ranging from-60.7%(95%UI,-77.6 to-41.7%)to 7.7%(95%UI,-57.2 to 54.2%).In 2006-2012,there is not markedly change in cervical cancer age-standardized death rate in China Mainland,however,age-standardized death rate in eastern increased by 2.3%(95%CI,0.1-4.7%)annually in rural areas in Eastern China and declined by 6.3%(95%CI,-8.9 to-3.7%)in rural areas in Central China.In 2010-2012,the age-adjusted mortality in women with lower education level was 1.28(95%CI,1.08-1.51)times than those with higher education level,and increase by 0.39 per 100,000(95%CI,0.16-0.61 per 100,000).The magnitude of social inequalities of cervical cancer mortality in urban areas was more than those in rural areas.Compared with inequalities in 2006-2008,the magnitude of social inequalities decreased significantly in western China,however,increased in Central and Eastern China.2)The Analysis of Cervical Cancer Screening Coverage and Social Inequality in China.In 2013,about 21.4%(95%CI,19.6-23.1%)participants reported having ever screened for cervical cancer,with a higher coverage among urban women than their rural counterparts(25.2%vs.18.0%,P<0.0001).Significant geographical variations of screening coverage were identified at provincial region ranging from 48.4%(95%CI,39.4-57.5%)to 9.8%(95%CI,0.1-22.6%).About 10 provinces had coverage rage more than 25%,however,7 provinces less than 15%.After controlling for covariates,individual education(P<0.0001);medical insurance(P<0.0001),occupation type(P<0.0001),household income(P<0.0001)and geographical education status(P<0.05),geographical per capita income level(P<0.01)had significantly association with uptake of cervical cancer screening among women aged 21 years or older.Education-induced RII in cervical cancer screening coverage was 2.96(95%CI,2.82-3.11),and corresponding age-adjusted SII was 26.4%(95%CI,25.2-27.6%);income-induced RII was 1.94(95%CI,1.84-2.04),and corresponding age-adjusted 18.0%(95%CI,16.6-19.3%);occupation-induced RII was 1.93(95%CI,1.84-2.03)and corresponding SII was 15.6%(95%CI,14.4-16.8%).The magnitude of inequalities in urban areas was more than rural areas.The magnitude and pattern of social inequalities in cervical cancer screening coverage were obviously different among provinces.3)The Impact Evaluation of National Cervical Cancer Screening Program in Rural Areas on the Population-based Coverage.The coverage rate of intervention group in rural areas was higher than that of non-intervention group in urban areas(24.0%,95%CI,21.8-26.3%vs.15.6%,95%CI,13.4-17.8%),however,the difference was not significant between groups in urban areas(28.5%,95%CI,26.1-31.0%vs.26.3%,95%CI,24.1-28.4%).The multilevel logistic regression analysis showed that the intervention had significantly positive effect with coverage rate OR=1.64(95%CI,1.28-2.10,P<0.0001),and also the interaction between intervention and rural areas OR=1.12(95%CI,1.00-1.25,P=0.0423).The similar results were seen in propensity score method.In whole 1:1 matching samples,the screening coverage in intervention group was 7.6%(95%CI,6.9-8.3%)more than that in non-intervention group,and OR=1.63(95%CI,1.56-1.71);in urban 1:1 matching samples,the screening coverage in intervention group was 6.1%(95%CI,4.9-7.3%)more than that in non-intervention group,and OR=1.44(95%CI,1.34-1.54);in rural 1:1 matching samples,the screening coverage in intervention group was 8.2%(95%CI,7.3-9.1%)more than that in non-intervention group,and OR=1.78(95%CI,1.67-1.90).The magnitude of social inequalities induced by education,income and occupation were 2.57(95%CI,2.38-2.78),1.45(95%CI,1.34-1.58),1.42(95%CI,1.31-1.54)in intervention group and lower than those in non-intervention group 3.32(95%CI,3.00-3.67),2.12(95%CI,1.91-2.35),and 2.07(95%CI,1.88-2.29),respectively.Conclusions:1)The age-standardized cervical cancer mortality in China was relatively lower than the global average mortality;however,the great geographical variations in cervical cancer death and secular trend were identified in this study.The pattern of distribution and tendency in cervical cancer mortality varies across different regions,in which "high and low death burden" and "upward and downward trend" are concurrent.This suggests that distinguished national prevention and control strategy of cervical cancer should be taken for special areas according death rate and change pattern.2)The national coverage of population-based cervical cancer screening in China remained low and varies by regional socioeconomic status.Moreover,the geographical patterns in coverage are reversed with the geographical patterns in mortality,showing that enlarging the population-based coverage rate in regions with high cervical cancer mortality is currently priority in our prevention and control strategy.In addition,individual and geographical socioeconomic factors are significantly associated with screening uptake.So it is of importance to consider the affordability and accessibility of screening service in increasing coverage in focused regions and groups.3)Significant social inequalities are identified in cervical cancer mortality and screening coverage.Moreover,there is obviously "inequality inverse" phenomenon in cervical cancer mortality and screening coverage,indicating that women in lower socioeconomic status have higher cervical cancer mortality but lower usage of cervical cancer screening.Great regional disparity in magnitude and pattern of social inequalities are also identified at provincial level.So,it is necessary to monitor the change of inequalities for health policy maker to reduce socioeconomic inequalities for future scale-up of cervical cancer prevention and improve screening efficiency.4)The National Rural Women Cervical Cancer Screening Program significantly increases the screening coverage and reduces the social inequalities among women living in urban and rural areas at the same time.However,the magnitude of increased coverage is lower than expected coverage.The evaluation indicates that the geographical scope of program is scaled up but the screening efficiency is probably negatively impacted by repeated screening.5)These findings highlight the importance of organized cervical screening program delivering to socioeconomically deprived women,which is line with current government strategies.In the future,localized cervical cancer screening mode should be considered according to mortality,screening coverage and magnitude and pattern of social inequalities.Diversification recruitment strategy should be used to increase the screening coverage among women with lower SES.Screening medical record system should be established to reduce the amount of repeated screening.Further investigations and intercommunicating monitoring systems should be strengthened to provide timely evidence for impact evaluation and adjusting public health strategies.