ABSTRACT
OBJECTIVE: The case-cohort design combines the advantages of a prospective cohort study and the efficiency of a case-control design. Usually a Cox proportional-hazards model is used for the analyses. However, adaptation of the model is necessary because of the sampling. We compared three methods that were proposed in the literature, which differ in weighting of study subjects: Prentice's, Barlow's, and Self and Prentice's method. STUDY DESIGN AND SETTING: In a cohort of 17,357 women we studied the relationship between body mass index and cardiovascular disease (n=821) with varying subcohort sizes (sampling fraction=0.005, 0.01, 0.05, 0.10, 0.15). RESULTS: Even with a sampling fraction of 0.01, all three methods showed identical estimates and standard errors (SE). With sampling fractions >or=0.10, results of the case-cohort analyses were similar to the full-cohort analyses. With simulations, the three methods provided different results if the full cohort is small (<1,250 subjects, subcohort=10%, 8% failures) or if the subcohort size was smaller than 15% (full cohort of 1,000 observations, 8% failures). The difference between the methods did not change with the number of failures or with different effect sizes. CONCLUSION: In the above-mentioned situations, the effect estimates and SE of Prentice's method most resembled the estimates of the full-cohort estimates.
Subject(s)
Body Mass Index , Cardiovascular Diseases/epidemiology , Data Interpretation, Statistical , Aged , Analysis of Variance , Case-Control Studies , Cohort Studies , Computer Simulation , Epidemiologic Methods , Female , Humans , Middle Aged , Netherlands/epidemiology , Proportional Hazards Models , Prospective Studies , Risk Assessment/methodsABSTRACT
BACKGROUND: The recent development of tissue array technology has potentiated large-scale retrospective cohort studies using archival formalin-fixed, paraffin-embedded tissues. This study evaluates the potential for using archival head and neck cancer tissue in such arrays. METHOD: Tissue array blocks were made from 184 head and neck cancer specimens. Three core tissue biopsies (0.6 mm x 3-4 mm) were taken from individual "donor" paraffin-embedded tumor blocks and arrayed into a new "recipient" paraffin block. Immunohistochemical (IHC) analyses were performed using antibodies recognizing cyclin-D1, Rb, and EGFR. IHC was scored on a 6-point scale for extent and a 3-point scale for intensity. We compared the staining of tissue array disks with staining of full tissue sections. RESULTS: Seventy-four percent (475 of 640) of samples placed into tissue arrays were confirmed to represent tumor tissue. The remaining samples were lost during processing or contained too few tumor cells. Only 6% of cases were completely lost, whereas 55%, 28%, and 11% of cases were judged on 3, 2, or 1 disk, respectively. Cohen's kappa coefficient was 0.66 for cyclin-D1, 0.40 for EGFR, and 0.41 for Rb. CONCLUSIONS: Tissue array technology is a rapid and efficient method for retrospective analysis of protein expression and is a promising tool for validation of prognostic markers in large series of head and neck squamous cell carcinomas. The agreement in scoring of the full section and the tissue arrays is reasonable. Discordance is probably due to intraobserver variation and lack of robustness of the scoring inherent of the proteins studied.