【作者】 张楠；

【导师】 孙强；

【作者基本信息】 山东大学 ， 社会医学与卫生事业管理， 2022， 博士

【摘要】 研究目的上消化道癌是我国高发的恶性肿瘤。筛查和早诊早治是当前最有效的癌症防治手段之一,“内镜+指示性活检”是上消化道癌筛查的“金标准”和首选技术,但我国上消化道癌内镜筛查起步较晚,如何制定适合不同农村地区的筛查策略仍缺少可靠证据。本研究旨在以山东省农村地区为研究现场,在三重目标(Triple Aim)框架下,综合分析各上消化道癌内镜筛查方案的效果、成本及经济性和人群筛查体验,为制定“成本效果好、社会可负担、群众愿接受”的上消化道癌内镜筛查策略提供决策依据,促进上消化道癌筛查效率和效益的提升。研究方法基于三重目标框架,采用模型长期预测和横断面分析相结合的研究设计,对内镜筛查策略进行综合分析与考量:引入卫生经济学评价和偏好分析理念确定10个测量指标;通过构建上消化道癌联合筛查Markov决策模型模拟不同筛查方案下人群健康维度、成本及经济性维度指标的长期趋势;利用离散选择实验(Discrete Choice Experiment,DCE)量化分析人群内镜筛查偏好,探讨筛查服务体验情况。以山东省东部的济南市、潍坊市,中部的济宁市、泰安市和西部的聊城市、滨州市为样本市,分两阶段选取相应的县级行政区收集数据。第一阶段,以平阴县、临朐县、曲阜市、肥城市、东昌府区和无棣县为样本地区,收集构建Markov模型所需数据,共整理筛查数据9万余条,调查医院5家、受检者171人、上消化道病变患者1117人;第二阶段,在临朐县、肥城市和东昌府区各整群抽取2-5个村,以40-70岁、既往未被诊断为癌症的村民为调查对象,完成DCE问卷调查986人。主要研究结果1.2015-2020年,样本地区共完成内镜初筛86183人,检出阳性病例1823例,病变检出率、早诊率、治疗率分别为2.12%、82.17%和91.11%,检出情况存在地区、年龄和性别差异;应随访病例11700例,实际完成8773例,随访率74.98%,检出阳性368例,检出率约为初筛的2倍,随访率与早诊率呈同向变化趋势。2.本研究构建的上消化道癌内镜筛查Markov模型稳定、可信。从人群健康维度看,内镜筛查能有效降低上消化道癌发病率和死亡率,提高人群期望寿命、生命年(Life Year,LY)和质量调整生命年(Quality-adjusted Life Year,QALY),但各方案对指标的改善各有优势。筛查间隔为1年、2年和5年时,筛查起始年龄越早,效果越好;10年、15年时,起始年龄和是否随访对效果的影响较大,效果最好的均是40岁开始筛查并随访的方案,最差分别是以55岁、45岁为起始年龄且不随访的方案;终生筛查1次时,起始年龄越晚效果越好。成本和经济性维度,通过增量成本-效果/效用分析(Incremental Cost-effectiveness/Cost-utility Ratio,ICER/ICUR)各筛选出8个优势方案。与不筛查相比,各优势方案的ICER、ICUR分别在40000元/LY和35000元/QALY以下;与相邻方案相比,大多数在30000元/LY和20000元/QALY以下。当支付意愿值为30000元/LY(QALY)时,40岁开始筛查、5年筛查间隔并随访的方案能获得最大净收益;随支付意愿值的升高,间隔越短、起始年龄越早方案的净收益逐步增加,当达到100000元/LY和70000元/QALY及以上时,40岁开始筛查、每年1次的方案能获得最大净收益。3.以偏好为代理指标探讨筛查体验:农村人群偏好较短筛查间隔、定期随访、死亡风险降低大、无痛和免费的筛查,筛查技术是影响偏好的最重要属性;满足偏好能改善筛查体验和提高参与率,但人群存在偏好异质性。人群对无痛内镜的支付意愿最高(686.01元),若不提供随访服务,理论上应补偿受检者57.02元。不同属性水平对参与率的影响不同,提高自付费用,参与率下降,采用无痛内镜,参与率将大幅提高;在自付费用300元、500元时,也可通过调整和优化其他属性使参与率达80%以上。结论与建议三重目标框架为癌症筛查策略的优化提供了新的思路,但效果好的方案其经济性不一定好,成本低、经济性好的方案,在效果和体验上的表现可能欠佳。根据“成本效果好、社会可负担、群众愿接受”的优化目标和“应因地制宜,可因人而异”的理念,在将无痛内镜和定期随访作为必备要素的基础上,经济水平好、筛查资源充足的地区应更加重视人群健康获益和筛查体验,可在40岁开始筛查并采用1-2年的筛查间隔;经济水平一般的地区,应适当考虑筛查的可负担性,可在40或45岁开始筛查,首选5年,次选10年的筛查间隔,此时能兼顾三个维度;经济水平较差的可继续推行终生1次筛查,起始年龄应提高至50岁或55岁,但人群健康获益和体验将受到较大影响。在地区内部,可针对不同人群分类设置不同筛查间隔和10-30%的费用自付比例。根据上述结论,提出如下政策建议:1)强化政府责任,确保癌症筛查措施落实落地;2)筛查策略的制定应坚持“成本效果好、社会可负担、群众愿接受”的目标,既要重视筛查的效果和经济性,还应满足人群偏好,并根据资源可用性做出权衡;3)探索建立癌症筛查多方筹资和成本分担机制;4)多措施提升癌症筛查能力,推进筛查的规范化、同质化;5)开展相关科学研究,为筛查策略的及时调整和持续优化提供量化证据。创新点与不足创新点:1)基于三重目标框架,提出了“成本效果好、社会可负担、群众愿接受”的优化目标和“应因地制宜,可因人而异”的优化理念,为筛查等公共卫生干预措施的优化提供了新的视角;2)构建了上消化道癌联合筛查Markov决策模型,避免了仅评价单一癌症时人为放大成本或降低效果的问题;3)在国内首次采用DCE测量人群上消化道癌内镜筛查偏好,克服了传统态度调查不能实现属性间权衡和量化分析的问题。不足:Markov模型的部分参数来源于文献资料和二手数据,模型的本土化程度不够;部分数据通过横断面调查获取,可能存在回忆偏倚和抽样偏倚;通过DCE得到的陈述性偏好与实际的行为与体验是否一致仍需进一步研究证实。更多还原

【Abstract】 ObjectivesUpper gastrointestinal cancer(UGC)is one of the high-incidence cancers in China.Screening is one of the most realistic and effective measures for cancer prevention and control.Endoscopy followed by indicative biopsy is the gold standard for screening and diagnosis of UGC,which has been written into guidelines and expert consensus as the preferred screening method.However,with a short time and less practice of endoscopic screening,China has limited realiable evidence on how to formulate screening strategies suitable for different rural regions.The study aimed to comprehensively analyze the effectiveness,cost and economics and experience of UGC endoscopic screening under Triple Aim framework,to provide decisionmaking evidence for the optimizing of UGC endoscopic screening strategy within the goals of"Good cost-effectiveness,Affordable to the society and Acceptable to the mass",and to promote the efficiency and benefit of UGC endoscopic screening.MethodsBased on the Triple Aim framework,this study adopted a combination design of model long-term prediction and cross-sectional analysis.A total of 10 measurement indicators were determined after introducing the concept of health economic evaluation and preference analysis.A specific Markov decision model for UGC joint endoscopic screening was constructed to analyze the long-term trend of population health dimension and cost dimension indicators.Discrete Choice Experiment(DCE)was used to measure residents’ preference of endoscopic screening,and then the screening experience were explored by the obtained preference.Thereafter,the study achieved a comprehensive analysis of UGC joint endoscopic screening strategy.Jinan City and Weifang City in the East,Jining City and Tai’an City in the Middle,Liaocheng City and Binzhou City in the West of Shandong Province were selected as the sample areas in which certain county-level administrative regions were extracted to carry out a twostage data collection and questionnaire survey.In the first stage,Pingyin,Linqu,Qufu,Feicheng,Dongchangfu and Wudi were selected as sample counties,and 90000 screening cases were analyzed,5 medical institutions,171 screening participants and 1117 patients with upper gastrointestinal diseases were investigated.In the second stage,2-5 villages were randomly chosen from Linqu,Feicheng and Dongchangfu.Villagers aged 40-70 who had never been diagnosed with cancer were randomly investigated.A total of 986 participants was complished the DCE survey.Results1.A total of 86183 villagers was screened by endoscopy in sample areas,and 1823 positive cases were detected in the year 2015 to 2020.The overall detection rate,early diagnosis rate and treatment rate were 2.12%,82.17%and 91.11%,respectively,but there were differences between regions,age groups and sex groups.A total of 8773 in 11700 cases were followed-up,with a follow-up rate of 74.98%.368 positive cases were detected in the follow-up period,with a detection rate of 2 times higher than that of primary screening.An obvious same direction change trend between the follow-up rate and the early diagnosis rate was observed,that is,the early diagnosis rate was also higher in the areas with high follow-up rate.2.The Markov prediction results were in line with the actual given that the differences were within the interpretable range,so the Marokv model in the study was stable and reliable.Within the population health dimension,screening could effectively reduce the UGC cumulative morbidity(range,3.63-5.42%)and mortality(range,2.68-4.04%),and improve life expectancy(LE),life year(LY)and quality-adjusted life year(QALY)of the population.When the screening intervals were 1,2 or 5 years,the earlier the screening starting age,the better the effectiveness.When the intervals were 10 or 15 years,the starting age of 40 with follow-up could obtain the best health benefit,and the starting age of 55 or 45 without follow-up were the worst ones.In the case of screening once in a lifetime,a later screening starting age would be better choice for more effectiveness,conversely.In the cost and economic dimension,8 dominant options were found out by the incremental cost-effectiveness ratio(ICER)and the incremental cost-utility ratio(ICUR),respectively.Compared with none screening,the ICER and ICUR of each dominant option were less than￥40000/LY and ￥35000/QALY,respectively.Compared with adjacent options,most of ICER and ICUR were below ￥30000/LY and ￥20000/QALY,with the highest value of ￥85799.40/LY and ￥66764.06/QALY,respectively.With a ￥30000/LY(QALY)willingness to pay threshold,the option that screening every 5 years,starting at age 40 and with follow-up could obtain the maximum net benefit.As the threshold increasing,the net benefit of options with shorter screening interval and earlier starting age would gradually elevate accordingly.When the threshold reached above ￥100000/LY or ￥70000/QALY,the option that screening once a year from age 40 could produce the maximum net benefit.3.DCE results were used as proxy indicators to reflect the experience and participation rate.Overall,the respondents preferred the screening program with a high-risk reduction of UGC-related mortality,shorter interval,follow-up,less pain,and lower out-of-pocket cost.The screening technique attribute had the most important impact on residents(accounted for 43.61%),followed by out-of-pocket costs and screening interval.Satisfying preferences can improve the screening experience,but there was preference heterogeneity in different populations.When converting endoscopy to painless endoscopy,the respondents were willing to pay ￥686.01.Theoretically,if regular follow-up was not provided,the respondents should be compensated ￥57.02.In the scenario prediction,the participation rate would decrease accordingly as out-of-pocket costs increase,and it would increase by 89.94%if painless endoscopy was used.Even if the out-of-pocket costs were ￥300 and ￥500,participation rates could reach to 92.72%and 83.99%by adjusting and optimizing other attributes.Conclusions and ImplicationsThe Triple Aim provides a novel framework for the evaluation and optimization of cancer screening strategies.But the performance of different screening options varied in three dimensions.A screening option with good effectiveness might not be economical,and a lowcost and economical option might not perform well in effectiveness and experience.Based on the Triple Aim,the formulation of UGC endoscopic screening strategies should firstly clarify the weights and priorities of the three dimensions and followed the principle of "adapting to local conditions" and "being able to vary from person to person".After taking painless endoscopy and regular follow-up as essential components,areas with good economic level and sufficient screening resources should pay more attention to population health and experience and could perform screening at age 40 with a 1 year or 2 years screening interval.Areas with medium economic levels should properly considered the affordability of screening,which can adopt a starting age of 40 or 45 and a screening interval of 5 or 10 years.Areas with poor economic level and limited resources can perform cancer screening for target population once in their lifetime,but the starting age should be raised to 50 or 55.But the health benefits and experience will be negatively affected under these circumstances.In addition,varied screening intervals can be set for different age groups within certain regions.Pluralistic screening financing mechanisms shoule be explored and a classified 10-30%co-pay rate could be set for people with different incomes,which will help for the sustainable development of cancer screening.According to the above conclusions,the following policy recommendations were proposed:1)To strengthen government responsibility to ensure that cancer screening measures are implemented.2)The determination of UGC endoscopic screening strategies should adhere to the goals of "Good cost-effectiveness,Affordable to the society and Acceptable to the mass",It is necessary not only to pay attention to the effectiveness and economics of the screening,but also to improve the experience and participation rate of the target population and to make tradeoffs according to the availability of resources.3)Reasonable multi-party funding and costsharing mechanism of cancer screening should be explored.4)Various measures should be taken to improve cancer screening capabilities in various regions,and to promote the standardization and homogeneity of screening.5)Supporting scientific studies to provide quantitative evidences for the timely adjustment and continuous optimization of screening strategies.Innovations and LimitationsInnovations:1)This study firstly introduced the Triple Aim into the evaluation and optimization of UGC screening,and then put forward the optimization goals of "Good costeffectiveness,Affordable to the society and Acceptable to the mass" and the improvement ideas of "adapting to local conditions,being able to vary from person to person",which provided a new perspective for the formulation and evaluation of public health interventions.2)A specific Markov model for UGC joint screening was constructed,which avoided the issues of enlarging cost or reducing effectiveness when prevous studies only evaluated a single cancer.3)This was the first DCE study to measure the preference of residents’ endoscopic screening in China,which overcame the problems that traditional attitude survey can not achieve trade-offs and quantitative analysis of attributes.Limitations:A part of Markov model parameters came from literature and second-hand data which lower the model’s localization level and might affect the accuracy of the prediction results.Recalling bias and sampling bias might be exsit in the questionnaire survey.Whether the stated preference obtained through DCE was consistent with the actual behavior and screening experience still need to be confirmed by further research.更多还原