Design Issues and Sample Size when Exposure Measurement is Inaccurate

Abstract

Measurement error often leads to biased estimates and incorrect tests in epidemiological studies. These problems can be corrected by design modifications which allow for refined statistical models, or in some situations by adjusted sample sizes to compensate a power reduction. The design options are mainly an additional replication or internal validation study. Sample size calculations for these designs are more complex, since usually there is no unique design solution to obtain a prespecified power. Thus, additionally to a power requirement, an optimal design should also fulfill the criteria of minimizing overall costs. In this review corresponding strategies and formulae are described and appraised.

Keywords

Measurement Error - Misclassification - Sample Size - Cost-efficient Design - Internal Validation - Replication

• References

• 1 Armstrong BG. The effects of measurements errors on relative risk regressions. Am J Epidemiol 1990; 132: 1176-84.
• 2 Carroll RJ, Ruppert D, Stefanski LA. Measurement error in nonlinear models. London: Chapman & Hall; 1995
• 3 Rippin G. Effekte fehlerhafter Expositions-und Confounderbestimmungen in epidemiologischen Studien (in German). Informatik, Biometrie und Epidemiologie in Medizin und Biologie 1999; 30: 72-83.
• 4 Thomas D, Stram D, Dwyer J. Exposure measurement error: influence on exposure-disease relationships and methods of correction. Ann Rev Publ Health 1993; 14: 69-93.
• 5 Bross IDJ. Misclassification in 2 2-tables. Biometrics 1954; 10: 478-86.
• 6 Brenner H. Notes of the assessment of trend in the presence of nondifferential exposure misclassification. Epidemiology 1992; 3: 420-7.
• 7 Holford RH, Stack C. Study design for epidemiologic studies with measurement error. Stat Methods Med Res 1995; 4: 339-58.
• 8 Greenland S. The effect of misclassification in the presence of covariates. Am J Epidemiol 1980; 112: 564-9.
• 9 Marshall JR, Hastrup JL. Mismeasurement and the resonance of strong confounders: uncorrelated errors. Am J Epidemiol 1996; 143: 1069-78.
• 10 Walker AM, Lanes SF. Misclassification of covariates. Stat Med 1991; 10: 1181-96.
• 11 Armstrong BG, Whittemore AS, Howe GR. Analysis of case-control data with covariate measurement error: application to diet and colon cancer. Stat Med 1989; 8: 1151-63.
• 12 Carroll RJ, Gail MH, Lubin JH. Case-control studies with errors in covariates. JASA 1993; 88: 185-99.
• 13 Chen TT. A review of methods for misclassified categorical data in epidemiology. Stat Med 1989; 8: 1095-106.
• 14 Duffy SW, Rohan TE, Day NE. Misclassification in more than one factor in a case-control study: a combination of Mantel-Haenszel and maximum likelihood approaches. Stat Med 1989; 8: 1529-36.
• 15 Greenland S. Variance estimation for epidemiologic effect estimates under misclassification. Stat Med 1988; 7: 745-57.
• 16 Greenland S. On correcting for misclassification in twin studies and other matched pair studies. Stat Med 1989; 8: 825-9.
• 17 Roeder K, Carroll RJ, Lindsay BG. A semi-parametric mixture approach to case-control studies with errors in covariables. JASA 1996; 91: 722-32.
• 18 Rosner B, Spiegelman D, Willett WC. Correction of logistic regression on relative risk estimates and confidence intervals for systematic within-person measurement error. Stat Med 1989; 8: 1051-69.
• 19 Rosner B, Spiegelman D, Willett WC. Correction of logistic regression on relative risk estimates and confidence intervals for measurement error: the case of multiple covariates measured with error. Am J Epidemiol 1990; 132: 734-45.
• 20 Greenland S. On sample size and power calculations for studies using confidence intervals. Am J Epidemiol 1988; 128: 231-6.
• 21 Spiegelman D. Cost-efficient study designs for relative risk modeling with covariate measurement error. J Stat Plan Inf 1994; 42: 187-208.
• 22 Palmgren J. Precision of double sampling estimators for comparing two probabilities. Biometrika 1987; 74: 687-94.
• 23 Greenland S. Statistical uncertainty due to misclassification: implications for validation substudies. J Clin Epidemiol 1988; 41: 1167-74.
• 24 Rippin G. A comparison of formulae for calculating cost-efficient sample sizes of case-control studies with an internal validation scheme. Biom J 2000; 42: 887-94.
• 25 Greenland S. The effect of misclassification in matched-pair case-control studies. Am J Epidemiol 1982; 116: 402-6.
• 26 Buonaccorsi JP. Double sampling for exact values in the normal discriminant model with application to binary regression. Com Stat A 1990; 19: 4569-86.
• 27 Spiegelman D, Gray R. Cost-efficient study designs for binary response data with Gaussian covariate measurement error. Bio-metrics 1991; 47: 851-69.
• 28 Reilly M. Optimal sampling strategies for two-stage studies. Am J Epidemiol 1996; 143: 92-100.
• 29 Reilly M, Pepe MS. A mean-score method for missing and auxiliary covariate data in regression models. Biometrika 1995; 82: 299-314.
• 30 Holcroft CA, Spiegelman D. Design of validation studies for estimating the odds ratio of exposure-disease relationships when exposure is misclassified. Biometrics 1999; 55: 1193-201.
• 31 Duffy SW, Maximovitch DM, Day NE. External validation, repeat determination, and precision of risk estimation in misclassified exposure data in epidemiology. J Epidemiol Community Health 1992; 46: 620-4.
• 32 Lagakos SW. Effects of mismodelling and mismeasuring explanatory variables on tests of their association with a response variable. Stat Med 1988; 7: 257-74.
• 33 Walker AM, Blettner M. Comparing imperfect measures of exposure. Am J Epidemiol 1985; 121: 783-90.