Strategies And Cost-effectiveness Analysis Of Two Breast Cancer Screening Programs In Shanghai, China

ObjectiveBreast cancer is the most common malignancy among women. Early diagnosis of breast cancer has been suggested to effectively decrease the mortality of the disease. Multiple randomized controlled trials (RCTs) conducted in western countries have individually and collectively provided strong support for the efficacy of mammography (MAM) for breast cancer screening. Based on these RCTs and observational studies, different screening guidelines have been established and proved cost-effective.Chinese women differed in breast size and density from their western counterparts. The epidemiological pattern of breast cancer is also much different. The recommended screening strategies for western women may not be suitable for Chinese women. To establish an optimal screening strategy for Chinese women, we comprehensively evaluated the cost, effect, and cost-effectiveness of two breast cancer screening programs conducted in Minhang district of Shanghai, China.Materials and Methods1. Study PopulationA biennial community-based breast cancer screening program was conducted among women aged35to74years old who lived in Qibao County, Minhang district of Shanghai, China (Qibao program) since May2008. All participants were interviewed using a structured questionnaire and had a clinical breast examination (CBE) after providing a consent form. Women with positive CBE results and women at age of45-69years old, regardless of the CBE results, were further preformed US and MAM examinations. Women with Breast Imaging Reporting and Data System (BI-RADS) category4or higher had breast biopsy. Between May23th,2008and Oct.31th,2010,12,215women without previous history of breast cancer participated in the first round of QB screening modality.During the same period, a large scale opportunistic screening (OS) was implemented among outpatient women from community healthcare centers and local secondary hospitals, women participating in a screening program on gynecological disorders and some women aged40to59years old from communities of Minhang district (Minhang program). An in-person interview was conducted using a structured questionnaire, and a CBE and/or TTM examination was performed for all participants. Women with positive result of either CBE or TTM were urged to have a further MAM and/or US examination at their own expense. Women with BI-RADS category4or higher had breast biopsy at hospitals. A total of104,809women aged35to74years old without previous history of breast cancer participated in the Minhang screening modality.A total of136,416female residents of Minhang District at ages of35to74years old who didn’t participate in any one of the breast cancer screening programs were regarded as control group.2. Data CollectionThe costs related to screening were from the Qibao Community Healthcare Center and the Maternal and Child Health Center of Minhang District. In-person interview was conducted for each participant of the two screening programs to collect demographic characteristics and possible risk factors of breast cancer by using a questionnaire. The date and results of each screening examination were recorded for each subject. The total breast cancer cases in the two screening programs were identified by the results of screening and supplemented by a record linkage with the Shanghai Cancer Registry System.The demographic characteristics of the control group were from the Vital Registry System of the Minhang district. The incident breast cancer cases were obtained by record linkage with the Shanghai Cancer Registry System. All breast cancer cancers were confirmed by medical record review. The medical record review was also applied to collect clinicopathological stages of breast cancer cases and the direct hospitalization expenses for the first treatment for each case. In-person interview was conducted for all cancer cases to collect the first outpatient cost, and direct non-medical and indirect costs during the first year after diagnosis.3. Statistical AnalysisThe screening data and results were recorded and managed using the Microsoft Excel2010. The information collected using questionnaires were key entered using Epidata3.1software. Further data cleaning and statistical analysis were performed using SAS9.2software.The characteristics of the three populations were described and compared by using x~2tests and ANOVA tests. The reporting rate was calculated to compare the incidence and prevalence of breast cancer among three populations. The screening performance of individual and combined use of clinical breast examination (CBE), ultrasonography (US) and MAM was evaluated using sensitivity (Se), specificity (Sp), Positive predictive value (PPV) and negative predictive value (NPV) based on the Qibao program, x~2test was used to compare the stage distribution of breast cancer cases in three groups. Kruskal-Wallis test was conducted to compare average cost of direct hospitalization expenses for first treatment, average first outpatient expenses, average direct non-medical cost and average indirect cost among three groups, using P<0.05as significant difference. The effectiveness of screening was evaluated by breast cancer stage improvement, which was calculated by assuming that the stage distribution is same as that in unscreened breast cancers. Missing values for the costs were simply imputed using the medians. The cost-effectiveness ratio (CER) was computed as the cost of gaining a stage improvement from a specified screening strategy when compared with the situation of no screening, which was expressed as cost per stage improved. Incremental cost-effectiveness ratio (ICER) that compared the two screening strategies was calculated by dividing the difference in total net cost and the difference in stages improved between two screening strategies.Results1. Characteristics of the study populationCompared with the control group, women participating in the screening programs were more likely to be older and had less educational level. The participants of the Qibao program were more likely to have a family history of breast cancer, had less than12years of age at menarche, more than55years of age at menopause, and more than35years old of age at first live birth.2. Breast cancer reportingA total of35,193and479breast cancer cases were identified in Qibao, Minhang and control groups, with reporting rate of215,146and144per100,000person-years, respectively. After being adjusted by age, the reporting rate changed to192,135and156per100,000person-years for Qibao, Minhang and control groups, respectively. The average age of the breast cancers detected in Qibao and Minhang programs were higher than those derived from the control group (P=0.003).3. Screening performance of the examinationsBased on the results of screening from the Qibao program, the sensitivity of CBE, US and MAM alone were63.3%,50.0%and86.7%, respectively. While MAM performed better in elder age groups than in younger ones, US had a higher sensitivity in younger age group. Combining use of the two imaging examinations increased the sensitivity for all patients (93.3%, P=0.01) and in almost all age groups. MAM had a higher sensitivity in progesterone receptor positive cancers than in those negative, while US had an inapparent higher sensitivity in hormone receptor negative cancers than in those positive.4. Stage improvement in screened casesThe percentage of early-stage breast cancer cases (stage0and I) from the Qibao, Minhang and control groups were46.9%,40.7%and38.9%, respectively. The difference did not reach significant. Comparing with the cases from the control group, the cases identified in Qibao and Minhang programs had a total of16.14and15.08stage improvements, respectively.5. Cost-effectiveness analysisThe average screening costs were208RMB Yuan for each participant and72,453RMB Yuan for each screened case in Qibao program, while they were21RMB Yuan and11,640RMB Yuan, respectively, in the Minhang program. In breast cancer cases, the average direct hospitalization expenses for the first treatment increased with the increasing TNM stages (P<0.001). It was11,024RMB Yuan for the each case detected in Qibao program, significantly higher than13,465RMB Yuan for each case screened in Minhang program and14,243RMB Yuan for each case from the control group (P<0.001). No significant difference was observed for average first outpatient expenses and average direct non-medical cost among three groups. The average indirect cost was lowest in Qibao program. The average cost was103,650RMB Yuan for each case detected in Qibao program,50,712RMB Yuan for each case screened in Minhang program and35,413yuan for each case from the control. The average cost saved was-68,237RMB Yuan for Qibao program and-15,299RMB Yuan for Minhang program, with CERs being135,291and152,179RMB Yuan per stage improved for Qibao and Minhang program, respectively. The ICER that compared the Qibao screening program with the Minhang screening program was131,086RMB Yuan per stage improved.ConclusionsThe Qibao screening program provides evidence on the improved sensitivity of combined uses of MAM and US in Chinese women. Our results suggest that the Qibao program, an organized biennial mass screening modality based on combining use of MAM and US examinations, is more effective in early detection of breast cancer than the Minhang program, an OS program. From a perspective of society, the Qibao program, albeit its much higher cost, is more cost-effective and can be applied among Chinese women in economically developed areas in China.