1、第一步第一步选择课题,陈述假设选择课题,陈述假设第二步第二步制订研究方案制订研究方案第三步第三步收集资料收集资料第四步第四步整理和分析资料整理和分析资料第五步第五步解释结果解释结果第一步第一步选择课题,陈述假设选择课题,陈述假设第二步第二步制订研究方案制订研究方案第三步第三步收集资料收集资料第四步第四步整理和分析资料整理和分析资料第五步第五步解释结果解释结果 Classification BMI(kg/m2)Risk of Co-morbidities UnderweightBMI=25 kg/m2)and obesity(BMI=30 kg/m2)for both men and women
2、.Source:Garrow JS&Webber J,1985;WHO Report,1997 High correlation with laboratory-based measures of adiposity for population study.Source:Yalloefer et al,1997 Inexpensive and practical for most clinical settings BMI DOES NOT ACCOUNT 1)The wide variation in body fat distribution;2)The nature of obesit
3、y across different individuals and populations;3)The joint relation of body composition and body size to health outcomes Source:Michaels KB,et al,1998,WHO Report,1997 Body fat distribution is a more powerful predictor for cardiovascular risk factor,disease,and mortality than BMIAdditional measuremen
4、ts might usefully supplement BMI in identifying individuals at increased risk due to abdominal fat accumulation(腰(腰围围)Waist circumference(WC)is a convenient and simple measure of abdominal adipose tissue that correlates closely with BMI and total body fat,and has an independent association over BMI
5、with cardiovascular disease(CVD)related risksWaist circumferences in white men of 90 and 100 cm and women of 83 and 93 cm represent the same risk of having cardiovascular disease(CVD)as respective BMIs of 25 and 30 kg/m2.However,WC cutoff points for other race-ethnic groups remain unknown.(Zhu et al
6、,2002 Am J Clin Nutr)Current WHO recommended criteria for WCCurrent WHO recommended criteria for WC Risk of obesity-associated metabolic complications Increased Substantially increased (Action level I)(Action level II)Men(n=904)/=94 cm(37 inches)/=102 cm(40 inches)Women(n=1014)/=80 cm(32 inches)/=88
7、 cm(35 inches)Cited from Obesity preventing and managing the global epidemic:Report of A WHO Consultation on Obesity.Men Women Age(yrs)51.0 (14.1)50.8 (14.0)Weight(kg)75.9 (13.7)65.9 (14.9)Height(cm)170.7 (7.0)158.2 (6.5)BMI(kg/m2)26.0 (4.1)26.3 (5.5)WC(cm)93.3 (11.9)82.0 (12.3)Cited from M.Lean et
8、al.,BMJ 1995 To determine sex-/race ethnicity-specific WC cutoff points by investigating the relationships of BMI and WC with CVD risk factors usinga USA population representative data.Source:National Center for Health Statistic,CDC,2005Source:WHO Report,1997Diabetes Gallbladder disease Hypertension
9、DyslipidaemiaInsulin resistenceBreathlessnessSleep apneaCoronary heart diseaseOsteoarthritis(knees)Hyperuricemia GoutCancer(breast,endometrial,colon)Reproductive hormone abnormalitiesPolycystic ovary syn.Impaired fertilityLow back painIncreased anesthetic Diseases%All diseases N (2)Cancer 22.8%557,2
10、71 (4)Chronic lower respiratory disease 5.1%124,816 (5)Unintentional injuries 4.4%106,742 (7)Influenza and pneumonia 2.7%65,681 (8)Alzheimers Disease 2.4%58,866 (9)Kidney disease 1.7%40,974 (10)Infection 1.4%33,865 (11)Suicide 1.3%(12)Liver disease 1.1%(14)Assault homicide 0.7%(15)All other causes 1
11、7.4%Source:Modified from NCHS CDC,and death of cause()第一步第一步选择课题,陈述假设选择课题,陈述假设第二步第二步制订研究方案制订研究方案第三步第三步收集资料收集资料第四步第四步整理和分析资料整理和分析资料第五步第五步解释结果解释结果 它必须遵循随机化(它必须遵循随机化(randomization)的原则。)的原则。随机抽样是研究的样本由总体中抽取,每个单位都有同等机会被抽中。随机抽样是研究的样本由总体中抽取,每个单位都有同等机会被抽中。1 1)(simple random samplingsimple random sampling)最基
12、本方法,对象编号,)最基本方法,对象编号,在用随机数字表或抽签、摸球等进行抽样。此法用于数目不大的情况下。在用随机数字表或抽签、摸球等进行抽样。此法用于数目不大的情况下。2 2)(systematic samplingsystematic sampling),此法是按照一定顺序,机械地每),此法是按照一定顺序,机械地每隔一定数量的单位抽取一个单位,又称间隔抽样或机械抽样。隔一定数量的单位抽取一个单位,又称间隔抽样或机械抽样。3 3)(stratified sampling)(stratified sampling)是先将研究对象按主要特征(如性别、是先将研究对象按主要特征(如性别、年龄、职业、
13、教育程度等)分为几层,然后再在各层中进行随机抽样。年龄、职业、教育程度等)分为几层,然后再在各层中进行随机抽样。4 4)(cluster samplingcluster sampling),抽到的不是个体而是由个体所组成的),抽到的不是个体而是由个体所组成的集体(即群体)。如村,班级,居民小组等。便于组织,节约人力物力,集体(即群体)。如村,班级,居民小组等。便于组织,节约人力物力,因而多用于大规模调查。缺点是抽样误差大,分析工作量也较大。因而多用于大规模调查。缺点是抽样误差大,分析工作量也较大。5 5)(multistage samplingmultistage sampling)又称多阶段
14、抽样。这是大型调查时)又称多阶段抽样。这是大型调查时常用的一种抽样方法。例如某学校常用的一种抽样方法。例如某学校1000010000名大学生,名大学生,1616个系。每个系的学个系。每个系的学生数相差不多。如要调查生数相差不多。如要调查800800名学生体重,身高和体脂百分比含量,抽样名学生体重,身高和体脂百分比含量,抽样方法如下:先自方法如下:先自1616个系随机抽取个系随机抽取2 2个系,在从这个系,在从这2 2个系随机各抽取个系随机各抽取400400名学名学生,这就是两级抽样。生,这就是两级抽样。(cross-section study)指某个时间断面(时点或很短的时间内)进行的调查研究
15、,指某个时间断面(时点或很短的时间内)进行的调查研究,其目的是了解某个时点的现状,所以也称为现况研究。在医其目的是了解某个时点的现状,所以也称为现况研究。在医学中通常用于描述某种疾病或健康状况如肥胖和心血管疾病学中通常用于描述某种疾病或健康状况如肥胖和心血管疾病等的分布特征、探讨疾病或健康状况与某些因素的联系如心等的分布特征、探讨疾病或健康状况与某些因素的联系如心血管疾病与人体脂肪分布的关系、并评价防治措施的效果等。血管疾病与人体脂肪分布的关系、并评价防治措施的效果等。是了解某病患病率或健康状况,于一定时间内对一定范是了解某病患病率或健康状况,于一定时间内对一定范围的人群中每一成员所作的调查或
16、检查。围的人群中每一成员所作的调查或检查。是通过调查某一人群中有代表性的部分(统计学上是通过调查某一人群中有代表性的部分(统计学上称为样本)。根据抽取样本所调查出的结果估计出该人群某称为样本)。根据抽取样本所调查出的结果估计出该人群某病的患病率,或某些特征的情况。抽样调查较为常用。它的病的患病率,或某些特征的情况。抽样调查较为常用。它的重点在于所选择的样本是否具有代表性。重点在于所选择的样本是否具有代表性。是将一个范围明确的人群按是否是将一个范围明确的人群按是否于某可疑因素及其暴露水于某可疑因素及其暴露水平分为不同的亚组,追踪其各自的平分为不同的亚组,追踪其各自的。(prospective c
17、ohort study)是指研究的确定)是指研究的确定与分组是根据研究开始时研究对象的暴露状况而定。与分组是根据研究开始时研究对象的暴露状况而定。(historical cohort study)是指研究对象的确)是指研究对象的确定与分组是根据研究开始时研究者已掌握的有关研究对象在定与分组是根据研究开始时研究者已掌握的有关研究对象在过去某个时点暴露状况的历史材料做出的。过去某个时点暴露状况的历史材料做出的。(ambispective cohort study)也称混合型队列)也称混合型队列研究,即在历史性队列研究之后,继续前瞻性观察一段时间。研究,即在历史性队列研究之后,继续前瞻性观察一段时间
18、。的确定和暴露人群的选择。的确定和暴露人群的选择。人群是否有足够的数量,人群是否比较稳定以及随访是否方便人群是否有足够的数量,人群是否比较稳定以及随访是否方便等条件等。在确定暴露人群同时要考虑对照人群的选择。对照等条件等。在确定暴露人群同时要考虑对照人群的选择。对照的选择也要注意与暴露队列的均衡性和可比性,即除研究的暴的选择也要注意与暴露队列的均衡性和可比性,即除研究的暴露因素外,其它非研究因素,如年龄、性别、职业、教育、社露因素外,其它非研究因素,如年龄、性别、职业、教育、社会经济状况等都应尽可能地与暴露因素队列相似。会经济状况等都应尽可能地与暴露因素队列相似。的定义和选择是一关键的步骤。的
19、定义和选择是一关键的步骤。在病因研究中,最灵敏的终结指标是发病率,反映疾病发病在病因研究中,最灵敏的终结指标是发病率,反映疾病发病风险与暴露的关系。但须有检测系统(疾病发病登记系统),风险与暴露的关系。但须有检测系统(疾病发病登记系统),以确定个体的发病时点。其次的选择就是死亡率。一般来说,以确定个体的发病时点。其次的选择就是死亡率。一般来说,死亡登记系统在许多地区较为完善,通过死亡登记系统容易检死亡登记系统在许多地区较为完善,通过死亡登记系统容易检测终结事件的发生。但与发病率相比,除受暴露危险因素影响测终结事件的发生。但与发病率相比,除受暴露危险因素影响外,死亡率还受治疗条件和社会环境的混杂
20、因素的影响。外,死亡率还受治疗条件和社会环境的混杂因素的影响。目标人群中选择符合研究个体作为病例,以未患该病或不具该条件的个体目标人群中选择符合研究个体作为病例,以未患该病或不具该条件的个体作为对照,调查他们既往有关因素的暴露情况,估计患病优势比。作为对照,调查他们既往有关因素的暴露情况,估计患病优势比。(population-based case-control study)和以)和以(hospital-based case-control study)。前者的目)。前者的目标人群为一自然人群,研究对象是其中符合纳入标准的某病确诊的病例及标人群为一自然人群,研究对象是其中符合纳入标准的某病确
21、诊的病例及可作为对照正常人。后者的目标人群可以理解成为一个特殊人群,即普通可作为对照正常人。后者的目标人群可以理解成为一个特殊人群,即普通人群中凡有病即会来研究所在的医院就诊的人们,故其对照可以是非研究人群中凡有病即会来研究所在的医院就诊的人们,故其对照可以是非研究疾病的其他疾病的病人。疾病的其他疾病的病人。(unmatched case-control study)和和(matched case-control study)。一般以性别和年龄匹配为常见。如性别。一般以性别和年龄匹配为常见。如性别为匹配因素,一个病例为女性,则该病例的对照也必须为女性。其中匹为匹配因素,一个病例为女性,则该病例
22、的对照也必须为女性。其中匹配一个对照者称配一个对照者称1:1配对;匹配多个对照者称配对;匹配多个对照者称1:R配比,如配比,如1:2,1:3匹匹配等,但一般不应超过配等,但一般不应超过1:4匹配。匹配。Subjects&Methods The Third National Health and Nutrition Examination Survey The Third National Health and Nutrition Examination Survey (NHANES-III)(NHANES-III)Design Cross-sectional Setting 89 locati
23、ons nationwide Baseline survey time 1988 1994 Sample size 18,110 Race-ethnicity non-Hispanic black(black)Mexican American(MA)non-Hispanic white(white)第一步第一步选择课题,陈述假设选择课题,陈述假设第二步第二步制订研究方案制订研究方案第三步第三步收集资料收集资料第四步第四步整理和分析资料整理和分析资料第五步第五步解释结果解释结果 第一步第一步选择课题,陈述假设选择课题,陈述假设第二步第二步制订研究方案制订研究方案第三步第三步收集资料收集资料第四步
24、第四步整理和分析资料整理和分析资料第五步第五步解释结果解释结果Typing error,Error checking(numerical,categorical,Dates),Typing error,Error checking(numerical,categorical,Dates),Handling Missing Data,OutlierHandling Missing Data,Outlier Subjects&Methods The Third National Health and Nutrition Examination Survey The Third National H
25、ealth and Nutrition Examination Survey (NHANES-III)(NHANES-III)Design Cross-sectional Setting 89 locations nationwide Baseline survey time 1988 1994 Sample size 18,110 Race-ethnicity non-Hispanic black(black)Mexican American(MA)non-Hispanic white(white)Age of subjects(yrs)20-90 Pregnant women 195 Fo
26、od/beverages 6-h prior to venipuncture,demographic,SES,or dietary information missing 6946 Subjects used in analyses 10,969 men:5313 (black 1337,MA 1564,white 2412)women:5656 (black 1577,MA 1427,white 2652)Weighted size 120 million Outcome 1:Subjects with one or more of the following risk factorsOut
27、come 1:Subjects with one or more of the following risk factorsv Dyslipidemia LDL 160 mg/dl,HDL35 mg/dl(125 mg/dl or Current medication for diabetesv High Blood Pressure Systolic 140 mmHg,Diastolic 90 mmHg or Current medication for hypertensionOutcome 2:Subjects with 1 or more,2 or more,or 3 or more
28、Outcome 2:Subjects with 1 or more,2 or more,or 3 or more of the following risk factors:of the following risk factors:(1)Abdominal Obesity waist 102 cm(88 cm for women)(2)Triglyceride 150 mg/dL(3)HDL Cholesterol 40 mg/dL(110 mg/dL (5)Blood Pressure Systolic 130 or Diastolic 85 mmHg Developing race-et
29、hnicity specific BMI and WC Odds Ratio EquationsDeveloping race-ethnicity specific BMI and WC Odds Ratio Equations1).Dependent Variable:individuals with CVD risk factors vs.individuals with no risk factors 2).Predictor Variables:BMI or WC 3).Covariates:Age,interaction of age with BMI or WC,smoking a
30、nd drinking status,economic status,education level,physical activity,and menopausal statusThe analyses were repeated excluding subjects who had a CVD ordiabetes history(I.e.,history of type 2 diabetes,hypertension,heartattach,congestive heart failure,or stroke)but did not have any at the time of sur
31、vey.Logistic Regression ModelDeveloping odds ratio equations from regression modelsDeveloping odds ratio equations from regression modelswhere 1 and 2 are coefficient parameters of BMI and WC derived from logistic regression models X ref is the reference point (set at BMI of 25th percentile in sex-e
32、thnicity-specific population)Odds Ratio and BMI024681012141618202224262830323436384042Fig 1.Body Mass Index(kg/m2)Odds RatioORs menOdds Ratio and WC02468101214667074788286909498102106110114118Fig 2.Waist Circumference(cm)Odds RatioORs women第一步第一步选择课题,陈述假设选择课题,陈述假设第二步第二步制订研究方案制订研究方案第三步第三步收集资料收集资料第四步第
33、四步整理和分析资料整理和分析资料第五步第五步解释结果解释结果40.7176.483.426.792.3Age(yr)Height(cm)Weight(kg)BMI(kg/m2)WC(cm)41.4163.177.028.992.9Prevalence of CVD risk factors(Outcome 1)Men Black MA WhiteWomen Black MA White High glucose High blood pressure Dyslipidemia Subjects with one or more above items 36.7170.177.926.993.8
34、44.9176.583.526.796.638.5157.168.827.990.646.9162.568.826.187.9LDL(mg/dL)HDL(mg/dL)Blood Pressure(mmHg)systolic diastolicGlucose(mg/dL)123.952.0126.478.096.6120.845.5121.775.698.1128.244.8124.876.797.3122.456.9121.673.996.2118.152.0117.071.096.2125.855.8120.772.192.45.414.428.038.85.624.335.748.66.5
35、28.135.150.16.614.438.147.33.724.137.948.65.128.523.943.7 WC LDL-cho HDL-cho Systolic Diastolic Glucose Men Black WC(cm)-0.290*-0.329 0.308*0.268*0.206*BMI(kg/m2)0.916 0.241 -0.325 0.186 0.197 0.163 MA WC -0.168*-0.310 0.314*0.381*0.220*BMI 0.907 0.108 -0.319 0.208 0.352 0.177White WC -0.112*-0.282
36、0.241*0.268 0.233*BMI 0.897 0.067 -0.270 0.138 0.258 0.182 WomenBlack WC -0.211*-0.284 0.304*0.253*0.256*BMI 0.907 0.150 -0.277 0.217 0.207 0.183 MA WC -0.195*-0.235*0.339*0.341*0.263*BMI 0.883 0.102 -0.263 0.207 0.307 0.214White WC -0.263*-0.280 0.348*0.312 0.279*BMI 0.883 0.188 -0.276 0.221 0.296
37、0.204 Odds ratio and BMI in men02468101214161820161820222426283032343638404244Body mass index(kg/m2)Odds ratio for CVD risk factorsBlackHispanicWhiteOdds ratio and WC in men02468101214161820768084889296100104108112116120124128132136140Waist circumference(cm)Odds ratio for CVD risk factorsBlackHispan
38、icWhiteOdds ratio and BMI in women02468101214161820161820222426283032343638404244Body mass index(kg/m2)Odds ratio for CVD risk factorsBlackHispanicWhiteOdds ratio and WC in women0246810121416182062667074788286909498102106110114118122126Waist circumference(cm)Odds ratio for CVD risk factorsBlackHispa
39、nicWhite 70.486.498.8111.2123.5 72.488.8101.4113.9126.7 74.691.3104.1116.9129.872.488.7101.2113.7126.3 (inch)72 (28)89 (35)101 (40)114 (44)127 (50)70.083.593.9104.4114.8 69.683.393.9104.4115.1 69.283.494.2105.1116.0 69.683.193.6104.0114.4 (inch)70 (27)83 (33)94 (37)104 (41)115 (45)BMI cut points(kg/
40、m2)WC cut points corresponding to BMI cut points(cm)BlackMexican AmericanWhiteAverageRoundedMenOne or more12586.488.991.388.9893098.8102.4104.3101.8102Two or more12586.388.691.388.7893098.2100.9104.0101.0101Three or more12585.788.891.188.6893096.4101.9103.5100.6101WomenOne or more2583.683.783.483.68
41、43094.396.294.494.995Two or more2583.283.083.483.2833092.592.894.493.293Three or more2582.782.582.782.6833090.492.091.591.391 Source:Shankuan Zhu et al.,AJCN 2005SummarySummary1)Our results suggest that a WC of 89 cm for men and 83 cm for women in all three ethnicities which corresponding to BMI of
42、25,may represent an action level for limiting future weight gain,while WC of 101 cm for men and 94 cm for women which corresponding to BMI of 30,may suggest the need for risk reduction and weight loss in three major ethnic groups2)Waist circumference is a better predictor than BMI,and could detect 1
43、0%more people with CVD relatedrisks than BMI in three ethnicities both men and women.第一步第一步选择课题,陈述假设选择课题,陈述假设第二步第二步制订研究方案制订研究方案第三步第三步收集资料收集资料第四步第四步整理和分析资料整理和分析资料第五步第五步解释结果解释结果 BackgroundWaist circumference(WC)is a convenient and simple measure of abdominal adiposity and can partially compensate for
44、the shortcoming of BMI by bringing regional fat into consideration In addition,studies have shown that BMI and WC have independentassociations with CVD indicating the potential utility of utilizing WC in addition to BMI in clinical practice.A combined measure of BMI and WC(BMI-WC)may be preferable A
45、 combined measure of BMI and WC(BMI-WC)may be preferable to BMI or WC alone for the identification of CVD risks.to BMI or WC alone for the identification of CVD risks.1)Combined measure using current cutoffs for BMI and WC2)Combined measure derived from statistical models联系方式:联系方式:浙江大学医学院浙江大学医学院 研究楼研究楼C321,C319电话:电话:0571-8820-8520,0571-8820-8590,0571-8820-8591E-mail: