整群抽样调查数据分析中应正确计算抽样误差

为了澄清整群抽样调查数据分析中正确计算抽样误差的必要性，以在某市15岁及以上人群中开展的一次两阶段整群抽样调查为例，分别采用适用于单纯随机抽样数据的方法和考虑了复杂抽样设计的方法对数据进行分析。结果显示，忽略对复杂抽样设计的考虑，不恰当的采用适用于单纯随机抽样数据的方法进行数据分析，不仅有可能大大低估样本统计量的抽样误差，在进行假设检验时，甚至会得到错误的结果，故正确分析和报告整群抽样调查数据的抽样误差是非常必要的。     

复杂抽样调查总体特征量及其方差的估计

目的 介绍一种基于设计的复杂抽样调查数据总体特征量及其方差的估计方法.方法 采用抽样权重估计有关的总体特征量,用泰勒级数线性法估计其方差,并给出具体的分析实例.结果 加权估计量是总体特征量的近似无偏估计,所构建的95%CI可用于有关估计量准确度的评价及假设检验.结论 基于设计的复杂抽样调查数据分析方法说明了调查设计的特征,在大规模调查情况下都能得到可靠的结论.     

死因监测整群抽样与不等概率抽样设计方案的比较

为比较不同整群抽样设计方法 的抽样误差及设计效应,评价不等概率抽样在死因监测中的应用效果.以陕西省107个县(市、区)作为抽样框架,采用等概率整群抽样和不等概率整群抽样等设计方案抽取样本,用复杂抽样方法 计算不同方案样本的抽样误差和设计效应.不同的抽样方案得到不同的抽样误差估计,分层整群抽样的标准误小于完全随机整群抽样;不等概率抽样(πPS抽样)的设计效率虽略逊于等概率的完全随机整群抽样,但扩大了监测范围.结论 :对于抽样框架明确的整群抽样调查数据,在统计分析时不应脱离预先设定的抽样设计方案和设计参数.死因监测采用不等概率抽样设计,能增加样本的权重,提高死亡率的地区代表性.     

权重引入对大型入户调查慢性病患病数据统计分析结果的影响

目的探讨在分析大型入户调查数据时引入权重的必要性。方法以2015年四川省分级诊疗需方调查15岁及以上居民的慢性病患病数据为例,分别比较引入和不引入权重时人口结构、参数点估计值以及统计推断结果的差异。结果相对于四川省第六次人口普查总体,引入权重前四川省分级诊疗需方调查样本人口结构偏老龄化,而权重引入后的人口结构基本和四川省第六次人口普查数据保持一致;未加权前15岁及以上居民慢性病患病率高于2013年四川省卫生服务调查结果,加权之后慢性病患病率显著下降,城市地区下降更为明显;与不引入权重的两水平logistic回归模型相比中,引入权重后模型绝大多数回归系数的P值有所增大。按照0.05的检验水准,婚姻状况(已婚)、基本医疗保险由不拒绝无效假设变为拒绝无效假设,而学历(高中/技校/中专)则由拒绝无效假设变为不拒绝无效假设。结论大型家庭入户调查需考虑复杂抽样的影响,在分析过程中忽视不等概率抽样、层次结构以及调查样本的人口学特征,可能带来错误的估计和结论。     

常州市社区人群糖尿病流行特征抽样调查分析

目的 了解常州市社区人群糖尿病流行特征,为评价该地区糖尿病的流行现状并采取相应的预防措施提供依据.方法 采用整群抽样方法,对常州市区25岁以上8 578名常住居民进行流行病学抽样调查和相关指标的检查.结果 糖尿病总患病率为5.60%.其中男性患病率为5.12%,女性患病率为5.95%,男女糖尿病的患病率均随年龄增长而升高.不同文化程度、经济收入、体重指数、血脂水平糖尿病的患病率有统计学意义.结论 根据调查结果,寻找糖尿病的相关影响因素,对糖尿病的重点人群采取综合干预措施,有效预防糖尿病.     

考虑多阶段抽样设计的误差估计

多阶段随机抽样是公共卫生开展人群抽样调查的常用设计。多阶段抽样设计下获得的样本具有复杂样本的特征,存在群效应或数据不独立,若不考虑抽样设计,通常会低估抽样误差或增加统计推断Ⅰ类错误的风险。由于复杂样本误差估计形式较复杂,目前常用统计软件均默认采用极群方差估计策略来简化样本结构,即假设样本来自于一阶段整群抽样,忽略除第一阶段抽样外的所有抽样设计,从而实现对误差的近似估计。然而,在初级抽样单元入样比较高时,后继抽样阶段对误差的贡献不可忽略,极群方差估计策略可能导致无效的误差估计。本文旨在介绍考虑多阶段抽样设计下的误差估计方法,并通过对现实数据进行多阶段模拟抽样,探讨在不同抽样设计下,极群方差估计策略和考虑多阶段抽样设计下的误差估计差异。模拟结果显示,随初级抽样单元入样比的增加,极群方差估计策略估计的误差出现不同程度的偏倚,且随入样比增加偏倚加重;而考虑多阶段抽样设计下的误差估计则较准确反映误差水平,可得到准确的统计推断结果。     

唐山汉族人群高血压流行病学特征及相关因素分析

目的 了解唐山地区居民高血压患病情况及其影响因素.方法 采用整群抽样调查的方法,选取唐山市区8个社区,对20岁以上居民4 651人进行高血压患病情况调查,并对高血压患病的影响因素进行分析.结果 高血压患病率为27.31％.男性患病率(24.70％)较女性患病率(29.73％)低.单因素分析结果为,性别、年龄、糖尿病、冠心病、家族史、吸烟、饮酒和腹部肥胖为高血压的危险因素.结论 唐山汉族人群高血压患病率相对较高,可能与生活环境和基因有关.     

九江市社区人群糖尿病流行特征抽样调查分析

目的了解九江市社区人群糖尿病流行特征,为评价本地区糖尿病的流行现状并采取相应的预防措施提供依据。方法采用整群抽样方法,对九江市区25岁以上8 578名常住居民进行流行病学抽样调查和相关指标的检查。结果糖尿病总患病率为5.60%。其中男性患病率为5.12%,女性患病率为5.95%,男女糖尿病的患病率均随年龄增长而升高。不同文化程度、经济收入、体质指数、血脂水平糖尿病的患病率间差异有统计学意义。结论根据调查结果,寻找糖尿病的相关影响因素,对糖尿病的重点人群采取综合干预措施,有效预防糖尿病。     

产前保健服务利用指标的设计效应初探

目的 计算整群抽样调查所得产前保健服务利用指标的设计效应,为今后妇幼卫生领域的相关调查设计提供参考. 方法 利用联合国儿童基金会“母子健康综合项目”2011年9月～10月对中国西部6省/自治区12县2岁以下儿童母亲调查得到的数据,根据调查的抽样方法,计算各产前保健服务利用指标的设计效应.结果 被调查地区的各产前保健服务利用指标的设计效应在2.8 ～4.9之间.反映产前检查内容的内容合格率和至少接受4项检查比例两个指标的设计效应(分别为4.64、4.90)大于其他指标. 结论 整群抽样时指标的设计效应较大,不同产前保健服务利用指标、不同县同一指标的设计效应不同.在设计整群抽样调查时,需要根据设计效应,调整样本量,以保证样本统计量的精确度.     

河北省3632名城乡社区老年人慢性病及老年痴呆的现况调查

目的 了解河北省内城乡社区老年人群老年痴呆的流行现况.方法 采用多级分层整群抽样调查,利用简易智能状态检查表(MMSE)、日常生活能力量表(ADL)及相应的老年痴呆诊断标准,数据用SPSS 13.0软件进行统计学分析.结果 河北省城乡≥60岁老年人慢性病的总患病率为64.84%(2355/3632),男女性患病率和总患病率均随年龄增高而升高.老年痴呆总患病率为7.24%(263/3632),其中阿尔茨海默病(AD)为4.87%(177/3632),位于高血压(32.35%)、糖尿病(11.37%)、慢性阻塞性肺疾病(9.25%)、冠心病(8.84%)和脑卒中(7.16%)之后的第六位.AD 的患病率有随年龄增加而升高及随文化程度增高而降低的趋势.结论 河北省社区老年人包括AD在内的慢性病总患病率较高,AD已成为老年人主要慢性病,患病率随年龄增高而升高.     

Diabetes in Bolivia.

OBJECTIVE: To measure the prevalence of diabetes mellitus (DM), hypertension, obesity, and related risk factors in major cities in Bolivia. METHODS: A population-based survey was conducted in four Bolivian cities: La Paz, El Alto, Santa Cruz, and Cochabamba. The total sample size was chosen to be 2,948 persons. The overall response rate was 86%, with the rate varying somewhat among the four cities. DM was diagnosed through an oral glucose tolerance test (OGTT) 2 hours after an overload of 75 grams of glucose, using World Health Organization criteria. RESULTS: The overall prevalence of DM in the four urban areas combined was 7.2% (95% confidence interval (CI): 6.2%-8.3%) and of impaired glucose tolerance (IGT) was 7.8%. A total of 73.1% (95% CI: 65.0%-81.0%) of those previously diagnosed with DM and 73.7% (95% CI: 61.0%-86.4%) of newly diagnosed cases were overweight, according to measurements of body mass index. Hypertension was found in 36.5% (95% CI: 27.6%-45.5%) of known diabetics and in 36.6% (95% CI: 23.0%-50.1%) of newly diagnosed cases, compared to only 15.9% (95% CI: 14.3%-17.5%) among people without DM. The disease was most common among older persons and those with little education. CONCLUSIONS: Diabetes is a genuine public health problem in Bolivia. Further, the high prevalence of IGT that was found suggests that diabetes prevalence will increase in the near future in the country unless prevention strategies are implemented.     

A National Study of Obesity Prevalence and Trends by Type of Rural County

CONTEXT: Obesity is epidemic in the United States, but information on this trend by type of rural locale is limited. PURPOSE: To estimate the prevalence of and recent trends in obesity among US adults residing in rural locations. METHODS: Analysis of data from the Behavioral Risk Factor Surveillance System (BRFSS) for the years 1994-1996 (n = 342,055) and 2000-2001 (n = 385,384). The main outcome measure was obesity (body mass index [BMI] > or = 30), as determined by calculating BMI from respondents' self-reported height and weight. RESULTS: In 2000-2001, the prevalence of obesity was 23.0% (95% confidence interval [CI] 22.6%-23.4%) for rural adults and 20.5% (95% CI 20.2%-20.7%) for their urban counterparts, representing increases of 4.8% (95% CI 4.2%-5.3%) and 5.5% (95% CI 5.1%-5.9%), respectively, since 1994-1996. The highest obesity prevalence occurred in rural counties in Louisiana, Mississippi, and Texas; obesity prevalence increased for rural residents in all states but Florida over the study period. African Americans had the highest obesity prevalence of any group, up to 31.4% (95% CI 29.1%-33.6) in rural counties adjacent to urban counties. The largest difference in obesity prevalence between those with a college education compared with those without a high school diploma occurred in urban areas (18.4% [95% CI 17.9%-18.9%] vs 23.5% [95% CI 22.5%-24.5%], respectively); the smallest difference occurred in small, remote rural counties (20.3% [95% CI 18.7%-21.9%] versus 22.3% [95% CI 20.7%-24.0%], respectively). CONCLUSIONS: The prevalence of obesity is higher in rural counties than in urban counties; obesity affects some residents of rural counties disproportionately.     

Smoking, exposure to secondhand smoke, and smoking restrictions in Tijuana, Mexico.

OBJECTIVE: To estimate the prevalence of tobacco use, exposure to secondhand smoke, and smoking restrictions in the home and workplace among residents of Tijuana, one of Mexico's largest cities. METHODS: This cross-sectional household survey was conducted in Tijuana, Baja California, Mexico, during 2003 and 2004. A population-based sample of 400 Tijuana adult residents responded to a tobacco survey, and 397 of the surveys were analyzed. RESULTS: About 22.9% (95% confidence interval (CI): 18.7%-27.1%) of Tijuana adults reported current smoking, and 53.9% (95% CI: 48.8%-58.9%) reported chronic exposure to secondhand smoke. Approximately 44.4% (95% CI: 37.9%-50.9%) of Tijuana adults had a nonsmoking policy in their workplace, while 65.8% (95% CI: 61.0%-70.6%) of Tijuana households were smoke-free. CONCLUSIONS: The results underline the need for increased tobacco control efforts, particularly stricter enforcement of existing passive smoking regulations, in order to expand protection from secondhand smoke from private settings to public ones and to curb the tobacco epidemic in Tijuana and elsewhere in Mexico.     

Twenty-year trends in the prevalence of disability in China

Objective

To evaluate changes in the age-adjusted prevalence of disability in transitional China from 1987 to 2006.

Methods

Data from nationally representative surveys conducted in 1987 and 2006 were used to calculate age-adjusted disability prevalence rates by applying appropriate sample weights and directly adjusting to the age distribution of the 1990 Chinese population. Trends were assessed in terms of average annual percentage change.

Findings

The estimated number of disabled people in China in 1987 and 2006 was 52.7 and 84.6 million, respectively, corresponding to a weighted prevalence of 4.9% and 6.5%. The age-adjusted prevalence of disability decreased by an average of 0.5% per year (average annual percentage change, AAPC: −0.5%; 95% confidence interval, CI: −0.7 to −0.4) during 1987–2006. However, it increased by an average of 0.3% (AAPC: 0.3%; 95% CI: 0.1 to 0.5) per year in males and by an average of 1.0% (AAPC: 1.0%; 95% CI: 0.8 to 1.2) per year among rural residents, whereas among females it showed an average annual decrease of 1.5% (AAPC: −1.5%; 95% CI: −1.7 to −1.3) and among urban residents, an average annual decrease of 3.9% (AAPC: −3.9%; 95% CI: −4.3 to −3.5). Despite significant declining trends for hearing and speech, intellectual and visual disabilities, the annual age-adjusted prevalence of physical and mental disabilities increased by an average of 11.2% (AAPC: 11.2%; 95% CI: 10.5 to 11.9) and 13.3% (AAPC: 13.3%; 95% CI: 10.7 to 16.2), respectively.

Conclusion

In China, the age-adjusted prevalence of disability has declined since 1987, with inconsistencies dependent on the type of disability. These findings call for continuing and specific efforts to prevent disabilities in China.     

河南省2020年1～69岁人群丙型肝炎流行状况调查

目的了解2020年河南省1～69岁人群丙型肝炎(丙肝)流行状况和特征。方法估计样本量为5 827人。采用多阶段抽样, 于2020年8-12月在河南省抽取8个县(区), 每个县(区)抽取2个调查点, 对1～69岁的人群进行问卷调查, 并采集静脉血液标本进行HCV抗体、核酸和基因型检测。结果调查5 165人, 其中男性占44.76%(2 312/5 165), 女性占55.24%(2 853/5 165)。1～69岁的人群抗-HCV、HCV RNA调整阳性率分别为0.69%(95%CI:0.68%～0.70%)和0.20%(95%CI:0.19%～0.21%)。男性抗-HCV和HCV RNA调整阳性率分别为0.48%(95%CI:0.46%～0.50%)和0.09%(95%CI:0.08%～0.10%);女性为0.86%(95%CI:0.85%～0.87%)和0.30%(95%CI:0.28%～0.32%)。抗-HCV和HCV RNA阳性率均呈随年龄增长而增加的趋势。城市人群抗-HCV和HCV RNA调整阳性率分别为0.87%(95%CI:0.86%～0.88%)和0.28%(95%CI:...     

湖南益阳农村老年糖尿病前期人群血脂异常及其影响因素分析

目的 了解湖南益阳农村地区老年糖尿病前期人群血脂异常的患病率水平和影响因素。方法 2015年4月～7月期间,通过问卷调查和实验室检测收集老年糖尿病前期人群一般情况和实验室结果。描述分析老年糖尿病前期人群血脂异常的患病率;用二分类非条件Logistic回归分析老年糖尿病前期人群血脂异常的相关影响因素。结果 益阳农村地区老年糖尿病前期人群血脂异常患病率为28.3%（95% CI:24.2%～32.5%）,其中,男性血脂异常的患病率为25.6%（95% CI:19.1%～32.0%）,女性血脂异常的患病率为30.3%（95% CI:24.6%～36.0%）。经二分类非条件Logistic回归分析结果显示,高龄、高身体质量指数、腰臀比偏高、高血压、吸烟、体力活动不足以及膳食不合理是老年糖尿病前期人群血脂异常发生的危险因素。结论 益阳农村地区老年糖尿病前期人群血脂异常的患病率较高,应做好健康监测和教育工作,以控制血脂异常发生的危险因素,从而降低该人群血脂异常的发生率。     

中国内地居民2000-2009年被动吸烟率Meta分析

目的 运用Meta分析综合评价中国内地居民2000-2009年被动吸烟率,并分析其在性别、城/乡及不同经济发展水平地区、场所等因素间的差别.方法 检索万方数据库、维普信息资源系统、中国期刊全文数据库、中国生物医学文献数据库及PubMed数据库,收集所有关于居民吸烟的调查研究,再从中筛选出有关被动吸烟的研究报告.各资料间进行异质性检验,以确定采用固定模型或随机模型进行合并分析,采用秩相关检验法进行发表偏倚的评估.结果 共入选相关文献19篇,累计不吸烟人数为195 349人,被动吸烟人数为70 781人,总被动吸烟率为47.04%(95%CI:38.88%～55.27%).将被动吸烟率按照性别、城/乡、研究年份、研究地区和被动吸烟场所进行分层分析,男、女性被动吸烟率分别为44.80%(95%CI:34.07%～55.79%)和49.09%(95%C:39.62%～58.59%),P＜0.05;城市、农村地区合并的被动吸烟率分别为46.10%(95%CI:28.88%～63.82%)和47.55%(95%CI:17.85%～78.25%),P＜0.05;研究年份在2000-2004年被动吸烟率合并为47.59%(95%CI:38.31%～56.95%),2005-2009年为46.90%(95%CI:33.19%～60.87%),P＜0.05;东、西部地区居民被动吸烟率分别为41.38%(28.88%～54.47%)和74.38%(95%CI:59.08%～87.10%),P＜0.05;家庭、工作场所、公共场所被动吸烟率分别为73.03%(95%CI: 60.41%～84.00%)、14.72%(95%CI:8.83%～21.82%)和25.90%(95%CI:5.65%～54.24%),P＜0.05.结论 合并的被动吸烟率女性高于男性,农村地区高于城市,研究年份为2005-2009年的被动吸烟率低于2000-2004年,西部地区居民被动吸烟率高于东部地区,家庭内被动吸烟率高于工作场所和公共场所.

Abstract：

Objective To analyze the prevalence of passive smoking among inland residents in China from 2000 to 2009 and to analyze the differences between sex, urban/rural geographic distribution, different levels of economic development etc.. Methods Electronic search strategy was carried out, using WanFang database, China Journal Full-text database, VIP database, CBM and PubMed database to collect data on smoking, and passive smoking status, among residents in China.Fixed effects model or random effects model was employed according to statistical tests for homogeneity. Publication bias was assessed by rank correlation test. All statistical analysis was conducted with R 2.8.0. Results Nineteen studies were selected with a total of 195 349 non-smokers and 70 781 passive smokers involved. The overall prevalence of passive smoking was 47.04%(95%CI: 38.88%-55.27%). The prevalence of passive smoking was stratified by factors as sex, urban/rural, year and areas of the study, and areas where passive smoking was studied. The pooled prevalence rates of passive smoking were as follows: 44.80% (95%CI: 34.07%-55.79%) and 49.09%(95%CI:39.62%-58.59%) ,P＜0.05 for male and female;46.10%(95%CI:28.88%-63.82%),47.55%(95% CI: 17.85%-78.25% ), P＜0.05 for urban and rural, respectively. The pooled prevalence rates of passive smoking were 47.59% (95% CI: 38.31%-56.95% ) in the study year of 2000-2004 and 46.90% (95%CI: 33.19%-60.87% ) in 2005-2009 (P＜0.05). The pooled prevalence rates of passive smoking for eastern and western areas were 41.38%(28.88%-54.47%) and 74.38%(95%CI: 59.08% -87.10% ) (P＜0.05), and 73.03% (95%CI: 60.41% - 84.00% ), 14.72% (95%CI: 8.83%-21.82% )and 25.90% (95% CI: 5.65% - 54.24% ) for family, workplace and public place, respectively (P＜0.05). Conclusion The pooled prevalence of passive smoking was higher in females than males, in rural than in urban and in the western area than in the eastern areas. The prevalence of passive smoking in the study year of 2005-2009 was lower than of 2000-2004. The pooled passive smoking rate in the family was higher than in the workplace or in public.     

贵州省城乡高血压病流行现状

目的 了解贵州省城乡高血压病流行现状,探索城乡间高血压病患病的异同,为制定适宜干预措施提供科学依据.方法 采用按规模大小成比例的概率抽样法（PPS）的方法,在贵州省12个县（市、区）对18岁以上人群开展调查,收集贵州省城乡居民高血压患病相关信息,数据分析使用SPSS 19.0统计完成.结果 人群平均收缩压和平均舒张压分别为（125.76±21.64） mmHg （95％CI：126.20～125.32）和（78.43±12.45）mmHg （95％CI：78.68～78.17）;高血压患病率为27.8％,标化率为29.5％,城市地区患病率为28.8％,农村地区患病率为27.4％,城乡患病率无显著性差异.结论 贵州省成人高血压患病率较高,农村高血压防治工作应得到重视,应针对性的开展防治工作.     

Hypertension: Development of a prediction model to adjust self-reported hypertension prevalence at the community level

ABSTRACT: BACKGROUND: Accurate estimates of hypertension prevalence are critical for assessment of population health and for planning and implementing prevention and health care programs. While self-reported data is often more economically feasible and readily available compared to clinically measured HBP, these reports may underestimate clinical prevalence to varying degrees. Understanding the accuracy of self-reported data and developing prediction models that correct for underreporting of hypertension in self-reported data can be critical tools in the development of more accurate population level estimates, and in planning population-based interventions to reduce the risk of, or more effectively treat, hypertension. This study examines the accuracy of self-reported survey data in describing prevalence of clinically measured hypertension in two racially and ethnically diverse urban samples, and evaluates a mechanism to correct self-reported data in order to more accurately reflect clinical hypertension prevalence. METHODS: We analyze data from the Detroit Healthy Environments Partnership (HEP) Survey conducted in 2002 and the National Health and Nutrition Examination (NHANES) 2001--2002 restricted to urban areas and participants 25 years and older. We re-calibrate measures of agreement within the HEP sample drawing upon parameter estimates derived from the NHANES urban sample, and assess the quality of the adjustment proposed within the HEP sample RESULTS: Both self-reported and clinically assessed prevalence of hypertension were higher in the HEP sample (29.7 and 40.1, respectively) compared to the NHANES urban sample (25.7 and 33.8, respectively). In both urban samples, self-reported and clinically assessed prevalence is higher than that reported in the full NHANES sample in the same year (22.9 and 30.4, respectively). Sensitivity, specificity and accuracy between clinical and self-reported hypertension prevalence were 'moderate to good' within the HEP sample and 'good to excellent' within the NHANES sample. Agreement between clinical and self-reported hypertension prevalence was 'moderate to good' within the HEP sample (kappa =0.65; 95 % CI = 0.63-0.67), and 'good to excellent' within the NHANES sample (kappa =0.75; 95%CI = 0.73-0.80). Application of a 'correction' rule based on prediction models for clinical hypertension using the national sample (NHANES) allowed us to re-calibrate sensitivity and specificity estimates for the HEP sample. The adjusted estimates of hypertension in the HEP sample based on two different correction models, 38.1% and 40.5%, were much closer to the observed hypertension prevalence of 40.1 %. CONCLUSIONS: Application of a simple prediction model derived from national NHANES data to self-reported data from the HEP (Detroit based) sample resulted in estimates that more closely approximated clinically measured hypertension prevalence in this urban community. Similar correction models may be useful in obtaining more accurate estimates of hypertension prevalence in other studies that rely on self-reported hypertension.     

Trends in prevalence of complete tooth loss among Australians, 1979--2002

Edentulism is a key indicator of the oral health status of populations and is associated with reduced quality of life. OBJECTIVE: To describe temporal trends in the prevalence of edentulism in the Australian adult population. METHODS: Data were obtained from four national surveys of persons aged 15 years and over conducted in 1979, 1987/88, 1994 and 2002. Prevalence estimates and standard errors were calculated for each survey for males and females and each State/Territory. Birth cohort analysis was undertaken to track changes in prevalence across successive surveys. Data from the 1987/88, 1994 and 2002 surveys were age-standardised to the 1979 resident population estimates for each State and Territory and crude and adjusted prevalence estimates were compared. RESULTS: The crude prevalence of edentulism declined from 20.5% (95% CI 20.1-20.7) in 1979 to 8.0% (95% CI 7.2-8.8) in 2002. The 2002 age-standardised estimate of 7.1% (95% CI 6.5-7.7) was only marginally lower. There were substantial reductions in prevalence of edentulism in all ages, sexes and State/Territories of Australia during the 23-year period. There were no significant changes in edentulism prevalence in three birth cohorts born since 1915. CONCLUSION: Edentulism prevalence at least halved between 1979 and 2002 among all ages, sexes and States/Territories of Australia. There was no significant change in prevalence of edentulism in cohorts born since 1915, refuting the notion that the risk of edentulism increases due to ageing, suggesting instead today's elderly had historically high rates of extraction prior to the 1950s that have not been experienced by subsequent cohorts.