Relative risks and confidence intervals were easily computed indirectly from multivariable logistic regression.

OBJECTIVE To assess alternative statistical methods for estimating relative risks and their confidence intervals from multivariable binary regression when outcomes are common. STUDY DESIGN AND SETTING We performed simulations on two hypothetical groups of patients in a single-center study, either randomized or cohort, and reanalyzed a published observational study. Outcomes of interest were the bias of relative risk estimates, coverage of 95% confidence intervals, and the Akaike information criterion. RESULTS According to simulations, a commonly used method of computing confidence intervals for relative risk substantially overstates statistical significance in typical applications when outcomes are common. Generalized linear models other than logistic regression sometimes failed to converge, or produced estimated risks that exceeded 1.0. Conditional or marginal standardization using logistic regression and bootstrap resampling estimated risks within the [0,1] bounds and relative risks with appropriate confidence intervals. CONCLUSION Especially when outcomes are common, relative risks and confidence intervals are easily computed indirectly from multivariable logistic regression. Log-linear regression models, by contrast, are problematic when outcomes are common.

[1]  S. T. Buckland,et al.  An Introduction to the Bootstrap. , 1994 .

[2]  S. Chokron,et al.  Race, sex, and physicians' referrals for cardiac catheterization. , 1999, The New England journal of medicine.

[3]  J Carpenter,et al.  Bootstrap confidence intervals: when, which, what? A practical guide for medical statisticians. , 2000, Statistics in medicine.

[4]  Xiaonan Xue,et al.  Estimating the relative risk in cohort studies and clinical trials of common outcomes. , 2003, American journal of epidemiology.

[5]  D L Sackett,et al.  An assessment of clinically useful measures of the consequences of treatment. , 1988, The New England journal of medicine.

[6]  K A Schulman,et al.  The effect of race and sex on physicians' recommendations for cardiac catheterization. , 1999, The New England journal of medicine.

[7]  D. Altman,et al.  The odds ratio , 2000, BMJ : British Medical Journal.

[8]  F. Davidoff.  Race, sex, and physicians' referrals for cardiac catheterization. , 1999, The New England journal of medicine.

[9]  H. Tómasson.  Risk scores from logistic regression: unbiased estimates of relative and attributable risk. , 1995, Statistics in medicine.

[10]  L E Daly,et al.  Confidence limits made easy: interval estimation using a substitution method. , 1998, American journal of epidemiology.

[11]  E. Campbell,et al.  Preparedness of internal medicine and family practice residents for treating common conditions. , 2002, JAMA.

[12]  Frank Davidoff,et al.  Predicting Clinical States in Individual Patients , 1996, Annals of Internal Medicine.

[13]  Sander Greenland,et al.  Model-based estimation of relative risks and other epidemiologic measures in studies of common outcomes and in case-control studies. , 2004, American journal of epidemiology.

[14]  S Wacholder,et al.  Binomial regression in GLIM: estimating risk ratios and risk differences. , 1986, American journal of epidemiology.

[15]  Leonard E. Brahman.  Confidence Intervals Assess Both Clinical Significance and Statistical Significance , 1991 .

[16]  X. Xue,et al.  Correcting the odds ratio in cohort studies of common outcomes. , 1999, JAMA.

[17]  Anthony C. Davison,et al.  Bootstrap Methods and Their Application , 1998 .

[18]  J. Zhang,et al.  What's the relative risk? A method of correcting the odds ratio in cohort studies of common outcomes. , 1998, JAMA.

[19]  Sander Greenland,et al.  Modern Epidemiology 3rd edition , 1986 .

[20]  G. Zou,et al.  A modified poisson regression approach to prospective studies with binary data. , 2004, American journal of epidemiology.

[21]  Douglas G Altman,et al.  Odds ratios should be avoided when events are common , 1998, BMJ.

[22]  Kevin J Dombkowski,et al.  Follow-up testing among children with elevated screening blood lead levels. , 2005, JAMA.

[23]  W D Flanders,et al.  Large sample confidence intervals for regression standardized risks, risk ratios, and risk differences. , 1987, Journal of chronic diseases.

[24]  S. Lipsitz,et al.  Quasi-likelihood estimation for relative risk regression models. , 2005, Biostatistics.

[25]  S Greenland,et al.  Interpreting Model Coefficients When the True Model Form Is Unknown , 1993, Epidemiology.

[26]  Edward L. Korn,et al.  Analysis of Health Surveys , 1999 .

[27]  M. Bracken,et al.  Clinically useful measures of effect in binary analyses of randomized trials. , 1994, Journal of clinical epidemiology.

[28]  Compliance among pharmacies in California with a prescription-drug discount program for Medicare beneficiaries. , 2002, The New England journal of medicine.

[29]  Peter Cummings,et al.  Association of helmet use with death in motorcycle crashes: a matched-pair cohort study. , 2002, American journal of epidemiology.

[30]  D. Sackett,et al.  The number needed to treat: a clinically useful measure of treatment effect , 1995, BMJ.

[31]  Donna Spiegelman,et al.  Easy SAS calculations for risk or prevalence ratios and differences. , 2005, American journal of epidemiology.

[32]  J. LeLorier,et al.  Expressing the magnitude of adverse effects in case-control studies: “the number of patients needed to be treated for one additional patient to be harmed” , 2000, BMJ : British Medical Journal.

[33]  R. Simon,et al.  Confidence intervals for reporting results of clinical trials. , 1986, Annals of internal medicine.

[34]  J. Deeks.  When can odds ratios mislead? , 1998 .

[35]  Douglas G. Altman,et al.  Practical statistics for medical research , 1990 .

[36]  D. Margolis,et al.  Risk factors associated with the failure of a venous leg ulcer to heal. , 1999, Archives of dermatology.

[37]  J M Robins,et al.  Estimation of a common effect parameter from sparse follow-up data. , 1985, Biometrics.

[38]  J. Hardin,et al.  Generalized linear models: extensions to the binomial family , 2000 .

[39]  Sander Greenland,et al.  Modelling Risk Ratios from Matched Cohort Data: An Estimating Equation Approach , 1994 .

[40]  J. Lindsey,et al.  Choosing among generalized linear models applied to medical data. , 1998, Statistics in medicine.

[41]  A. Robbins,et al.  What's the relative risk? A method to directly estimate risk ratios in cohort studies of common outcomes. , 2002, Annals of epidemiology.