Log-binomial models: exploring failed convergence

Emerging Themes in Epidemiology

Gordon Hilton Fick²  Misha Eliasziw³  Tyler Williamson¹ 

[1] Departments of Family Medicine and Public Health Sciences, Queen’s University, Kingston, ON, Canada;Department of Community Health Sciences, University of Calgary, Calgary, AB, Canada;Department of Public Health and Community Medicine, Tufts University, Boston, MA, USA

关键词: Logistic regression alternatives;    Maximum likelihood estimates;    Likelihood estimation;    Log relative risk;    Method of maximum likelihood;    Relative risk;    Failed convergence;    Non-convergence;    Log-binomial;   

Others  :  810385
DOI  :  10.1186/1742-7622-10-14

 received in 2013-04-23, accepted in 2013-12-05,  发布年份 2013

【 摘 要 】

Background

Relative risk is a summary metric that is commonly used in epidemiological investigations. Increasingly, epidemiologists are using log-binomial models to study the impact of a set of predictor variables on a single binary outcome, as they naturally offer relative risks. However, standard statistical software may report failed convergence when attempting to fit log-binomial models in certain settings. The methods that have been proposed in the literature for dealing with failed convergence use approximate solutions to avoid the issue. This research looks directly at the log-likelihood function for the simplest log-binomial model where failed convergence has been observed, a model with a single linear predictor with three levels. The possible causes of failed convergence are explored and potential solutions are presented for some cases.

Results

Among the principal causes is a failure of the fitting algorithm to converge despite the log-likelihood function having a single finite maximum. Despite these limitations, log-binomial models are a viable option for epidemiologists wishing to describe the relationship between a set of predictors and a binary outcome where relative risk is the desired summary measure.

Conclusions

Epidemiologists are encouraged to continue to use log-binomial models and advocate for improvements to the fitting algorithms to promote the widespread use of log-binomial models.

【 授权许可】

   
2013 Williamson et al.; licensee BioMed Central Ltd.

【 图 表 】

Figure 1.

Figure 2.

Figure 3.

【 参考文献 】

• [1]Davies H, Crombie I, Tavakoli M: When can odds ratios mislead? Br Med J 1998, 316(7136):989.
• [2]Sedgwick P: Case-control studies: measures of risk. BMJ 2013, 346:f1185.
• [3]van Belle G: Statistical Rules of Thumb. Wiley-Interscience: New York; 2002.
• [4]Silvestrini M, Altamura C, Cerqua R, Pedone C, Balucani C, Luzzi S, Bartolini M, Provinciali L, Vernieri F: Early activation of intracranial collateral vessels influences the outcome of spontaneous internal carotid artery dissection. Stroke 2011, 42:139-143.
• [5]Spiegelman D, Hertzmark E: Easy SAS calculations for risk or prevalence ratios and differences. Am J Epidemiol 2005, 162(3):199-200.
• [6]Knol MJ, Le Cessie S, Algra A, Vandenbroucke JP, Groenwold RH: Overestimation of risk ratios by odds ratios in trials and cohort studies: alternatives to logistic regression. Can Med Assoc J 2012, 184(8):895-899.
• [7]Fang J: Using SAS Procedures FREQ, GENMOD, LOGISTIC, and PHREG to Estimate Adjusted Relative Risks – A Case Study. In SAS Global Forum 2011, 4–11 April 2011. Las Vegas: SAS Institute Inc.; 2011:345-2011.
• [8]Wacholder S: Binomial regression in GLIM: estimating risk ratios and risk differences. Am J Epidemiol 1986, 123:174-184.
• [9]Lee J: Odds ratio or relative risk for cross-sectional data? Int J Epidemiol 1994, 23:201-203.
• [10]Schouten EG, Dekker JM, Kok FJ, Le Cessie S, van Houwelingen HC, Pool J, Vanderbroucke JP: Risk ratio and rate ratio estimation in case-cohort designs: hypertension and cardiovascular mortality. Stat Med 1993, 12(18):1733-1745.
• [11]Zhang J, Yu KF: What’s the relative risk? A method of correcting the odds ratio in cohort studies of common outcomes. J Am Med Assoc 1998, 280(19):1690-1691.
• [12]Deddens JA, Petersen MR: Re: “Estimating the relative risk in cohort studies and clinical trials of common outcomes”. Am J Epidemiol 2004, 159(2):213-214.
• [13]Zou G: A modified poisson regression approach to prospective studies with binary data. Am J Epidemiol 2004, 159(7):702-706.
• [14]Lumley T, Ma S, Kronmal R: Relative Risk, Regression in Medical Research: Models, Contrasts, Estimators, and Algorithms. U Washington Working Papers: BE Press; 2006. [http://biostats.bepress.com/uwbiostat/paper293 webcite]
• [15]R Development Core Team: R: A Language and Environment for Statistical Computing, Reference Index Version 2.12.1. Vienna: R Foundation for Statistical Computing; 2011. [http://www.R-project.org webcite]
• [16]StataCorp: Stata Statistical Software: Release 11. College Station, TX: StataCorp LP; 2009.
• [17]SAS: SAS/STAT(R) 9.2 user’s guide. 2011. [http://support.sas.com/documentation/cdl/en/statug/63962/PDF/default/statug.pdf webcite]
• [18]IBM Corp: IBM SPSS Statistics for Windows, Version 19.0. Armonk: IBM Corp.; 2010.
• [19]Kalbfleisch J: Probability and Statistical Inference - Vol. 2: Statistical Inference, 2nd edition. New York: Springer; 1985.
• [20]Lange K: Optimization. New York: Springer; 2004.
• [21]Berlin Heidelberg J: Exhaustive search, combinatorial optimization and enumeration: exploring the potential of raw computing power. In SOFSEM 2000: Theory and Practice of Informatics, volume 1963. Edited by Hlavác V, Jeffery KG, Wiedermann J. Berlin, Heidelberg: Springer; 2000:18-35.
• [22]Trakhtenbrot BA: A survey of russian approaches to perebor (Brute-force searches) algorithms. IEEE Ann Hist Comput 1984, 6(4):384-400.
• [23]Kirkpatrick S, Gelatt CD, Vecchi MP: Optimization by simulated annealing. Science 1983, 220(4598):671-680.
• [24]StataCorp: Stata Base Reference Manual: Release 11. College Station: StataCorp; 2009.
• [25]Huber P: The behavior of maximum likelihood estimates under non-standard conditions. Proc Fifth Berkeley Symp Math Stat Probability 1967, 1:221-223.
• [26]Kirkpatrick L, Feeney B: A Simple Guide to SPSS for Version 17.0. Belmont: Cengage Learning; 2010.