Design and Analysis Techniques for CaseControl Studies Instructor: 李奕慧 yihwei@mail.tcu.edu.tw 1 Lecture Overview 1. 2. 3. 4. 5. Case-Control Study Example: ”Risk factors associated with lung cancer in Hong Kong” OR for multiple exposure levels Confounding factors Methods of Controlling (adjusting for) confounders 2 Epidemiologic Study Design Analytical studies Intervention studies Clinical trials Observational studies Cohort studies Case-control studies 3 Case-control study Exposed Cases Non-exposed Study Population Exposed Controls Non-exposed 4 Selection of cases Establish a strict diagnostic criteria for the disease: Examples: Type 1 diabetes in children: severe symptoms, very high BG, marked glycosuria, and ketonuria. Type 2 diabetes: few if any symptoms, Slightly elevated BG, diagnosis “complicated”. 5 Selection of cases Population-based cases: include all subjects or a random sample of all subjects with the disease at a single point or during a given period of time in the defined population: Danish childhood diabetes register Hospital-based cases: All patients in a hospital department at a given time 6 Selection of Controls Principles of Control Selection: Study base: Controls can be used to characterise the distribution of exposure Comparable-accuracy Equal reliability in the information obtained from cases and controls no systematic misclassification Overcome confounding Elimination of confounding through control selection matching or stratified sampling 7 Selection of Controls General population controls: registries, households, telephone sampling costly and time consuming recall bias eventually high non-response rate Hospitalised controls: Patients at the same hospital as the cases Easy to identify Less recall bias Higher response rate 8 Ascertainment of Disease and exposure status External sources: Death certificates, disease registries, Hospital and physicians records etc. Internal sources: Questionnaires and interviews, information from a surrogate (spouses or mother of children), biological sampling( e.g. antibody) 9 Bias in Case-Control studies Selection bias Non-response Detection bias cases and controls are identified not independently of the exposure Observation bias Recall Bias: Cases are more likely to remember exposure than controls 10 Strengths in Case-control Quick, inexpensive Well-suited to the evaluation of diseases with long latency period Rare diseases Examine multiple etiologic factors for a single disease 11 Limitations in Case-control Case-control study Not rare exposure Incidence rates cannot be estimated unless the study is population based Selection Bias and recall bias 12 Risk factors associated with lung cancer in Hong Kong Lung Cancer 40 (2003) 131-140 13 Chi-Square Tests Value Pearson ChiSquare Asymp. Sig. (2-sided) df 0.257a 1 Exact Sig. (1sided) Exact Sig. (2-sided) .613 Risk Estimate 95% Confidence Interval Value Odds Ratio for Marital (other / married) .880 Lower .535 Upper 1.446 14 Multiple Exposure Levels Exposure level Cases Controls OR A1 B1 OR1 B2 B3 OR2 Low A2 A3 Not exposed C D Reference High Medium OR3 15 Multiple Exposure Levels A significant (P<0.05) increasing trend in the OR was found between nonsmokers, ex- and current smokers; and increasing amount of smoking among the ever smokers. Lung cancer.sav Lung Cancer 40 (2003) 131/140 16 smoking * case Crosstabulation smoking Total nonsmoker Count % within case exsmoker Count % within case current smoker Count % within case Count % within case case case control 52 96 24.5% 45.3% 68 87 32.1% 41.0% 92 29 43.4% 13.7% 212 212 100.0% 100.0% Total 148 34.9% 155 36.6% 121 28.5% 424 100.0% Chi-Square Tests Pearson Chi-Square Value 48.212a df Asymp. Sig. (2-sided) 2 .000 Likelihood Ratio 50.088 2 .000 Linear-by-Linear Association N of Valid Cases 42.734 1 .000 424 抽煙與罹患肺癌有 關, Case中抽煙者佔較 高的比例(43% vs 13.7%) a. 0 cells (.0%) have expected count less than 5. The minimum expected count is 60.50. 17 Data > Select cases > 18 只選smoking=2, 3的資料進行分析 Risk Estimate 95% Confidence Interval Value Odds Ratio for 1.443 smoking (exsmoker / nonsmoker) Lower Upper .908 2.293 1.249 .942 1.656 For cohort case = control .865 .721 1.039 N of Valid Cases 303 For cohort case = case Exsmoker 罹患肺癌 是nonsmoker的1.4 倍, 95%CI (0.9, 2.3) Exsmoker與 nonsmoker罹癌機率 沒有顯著差異 19 Confounding factors (干擾因素) Confounder: Variable is associated with both the disease and the exposure variable. 20 Method for control for confounders 1. 2. 3. Study design: restriction/ matching/ randomization Statistical adjustment: Standardization; e.g. age standardized (where age is a confounder) Stratified by confounder (Mantel-Haenszel test) Incorporate the confounder into a regression analysis as a covariate. (logistic regression approach) 21 Restriction Example 研究主旨:二手煙(ETS, exposure)與罹患肺癌 (disease)的關係 confounder: 研究對象本身是否抽煙 為了避免干擾只分析ETS對nonsmoker的影響 22 Stratified Analysis 23 將性別當作分層(stratum)的因子 smoking * case * sex Crosstabulation Count sex male smoking ex- and current smoker nonsmoker female Total smoking ex- and current smoker nonsmoker Total case case control 160 116 Total 276 52 212 13 96 212 6 148 424 19 106 119 113 119 219 238 Lung cancer2.sav 24 Sex-Specific OR for smoking Risk Estimate sex male female 95% Confidence Interval Value Lower Upper 2.55 1.68 3.85 Odds Ratio for smoking (ex- and current smoker / nonsmoker) N of Valid Cases Odds Ratio for smoking (ex- and current smoker / nonsmoker) 424 2.31 N of Valid Cases 238 0.85 6.30 可以將男士的OR與女士的OR合併嗎? 怎麼併? Lung cancer2.sav 25 Thank you! 26