Relaxing the Rule of Ten Events per Variable in Logistic and Cox Regression
The DOI matches its Crossref record; 9 of the 9 references it cites were found in a registry or an index.
Abstract
The rule of thumb that logistic and Cox models should be used with a minimum of 10 outcome events per predictor variable (EPV), based on two simulation studies, may be too conservative. The authors conducted a large simulation study of other influences on confidence interval coverage, type I error, relative bias, and other model performance measures. They found a range of circumstances in which coverage and bias were within acceptable levels despite less than 10 EPV, as well as other factors that were as influential as or more influential than EPV. They conclude that this rule can be relaxed, in particular for sensitivity analyses undertaken to demonstrate adequate control of confounding.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“Following the generally accepted rule of thumb that 5–10 events are required per covariate in logit or Cox regression models (Peduzzi et al., 1995; Peduzzi et al., 1996; Vittinghoff & McCulloch, 2007), we assessed whether or not each paper could have supported fully saturated regression models with product terms between all of the psychosocial problems considered.”
From Syndemics of psychosocial problems and HIV risk: A systematic review of empirical tests of the disease interaction concept (2015) · cites this paper as Vittinghoff & McCulloch, 2007
In the citing paper
“Because so few students used psychotherapy, we limited the number of explanatory variables to the number of observed events divided by five (14).”
From Mental health treatment seeking among veteran and civilian community college students (2017) · cites this paper as [14]
In the citing paper
“This sample size was chosen taking into account approximate sample size estimation criteria [38,39]. We worked with the criteria that 10 subjects would be needed per parameter included in a statistical model used to adjust for confounding.”
From Relationships, love and sexuality: what the Filipino teens think and feel (2009) · cites this paper as [38]
In the citing paper
“As Vittinghoff and McCulloch 27 recommended, the 10 outcome events per predictor variable rule was relaxed.”
From Mental health services for parents who lost a child to cancer: if we build them, Will they come? (2015) · cites this paper as [27]
In the citing paper
“The requirements regarding events per variable were fulfilled owing to a considerable number of injuries in the present dataset.34”
From How much running is too much? Identifying high-risk running sessions in a 5200-person cohort study (2025) · cites this paper as [34]
Citations
Cites
A simulation study of the number of events per variable in logistic regression analysis
Peduzzi, P., Concato, J., Kemper, E., Holford, T. R., & Feinstein, A. R. (1996) · Journal of Clinical Epidemiology · doi:10.1016/s0895-4356(96)00236-3 · Cited by 9,655 · cited 1×
“(2, 3) more closely, we also examined models with all binary predictors.”
BackgroundModeling and variable selection in epidemiologic analysis
Greenland, S. (1989) · American Journal of Public Health · doi:10.2105/ajph.79.3.340 · Cited by 2,208
Peduzzi, P., Concato, J., Feinstein, A. R., & Holford, T. R. (1995) · Journal of Clinical Epidemiology · doi:10.1016/0895-4356(95)00048-8 · Cited by 2,154 · cited 2×
“(2, 3) more closely, we also examined models with all binary predictors.”
BackgroundResultThe Abuse of Power: The Pervasive Fallacy of Power Calculations for Data Analysis
Hoenig, J. M., & Heisey, D. M. (2001) · The American Statistician · doi:10.1198/000313001300339897 · Cited by 2,046
Sample-Size Formula for the Proportional-Hazards Regression Model
Schoenfeld, D. A. (1983) · Biometrics · doi:10.2307/2531021 · Cited by 1,096
Concato, J., Peduzzi, P., Holford, T. R., & Feinstein, A. R. (1995) · Journal of Clinical Epidemiology · doi:10.1016/0895-4356(95)00510-2 · Cited by 797 · cited 1×
“The rule of thumb that logistic and Cox models should be used with a minimum of 10 events per predictor variable (EPV) is based on two simulation studies (1–3).”
MethodSection 4 Long-term survival results in medically and surgically randomized patients
Detre, K. M., Peduzzi, P. N., Scott, S. M., & Davies, B. (1985) · Progress in Cardiovascular Diseases · doi:10.1016/0033-0620(85)90019-2 · Cited by 24
Section 2 Design and baseline characteristics
Peduzzi, P. N., Detre, K. M., & Gage, A. A. (1985) · Progress in Cardiovascular Diseases · doi:10.1016/0033-0620(85)90017-9 · Cited by 15
Harrell, F., Jr. (1996) · doi:10.1002/(sici)1097-0258(19960229)15:4<361::aid-sim168>3.0.co;2-4
Cited by
Moons, K. G. M., Altman, D. G., Reitsma, J. B., Ioannidis, J. P. A., Macaskill, P., Steyerberg, E. W., Vickers, A. J., Ransohoff, D. F., & Collins, G. S. (2015) · Annals of Internal Medicine · doi:10.7326/m14-0698 · Cited by 5,147
Harrell (2015) · Springer Series in Statistics · doi:10.1007/978-3-319-19425-7 · Cited by 6,878
Mapping genomic loci implicates genes and synaptic biology in schizophrenia
Trubetskoy, V., Pardiñas, A. F., Qi, T., Panagiotaropoulou, G., Awasthi, S., Bigdeli, T. B., Bryois, J., Chen, C.-Y., Dennison, C. A., Hall, L. S., et al. (2022) · Nature · doi:10.1038/s41586-022-04434-5 · Cited by 2,964
Calculating the sample size required for developing a clinical prediction model
Riley, R. D., Ensor, J., Snell, K. I. E., Harrell, F. E., Martin, G. P., Reitsma, J. B., Moons, K. G. M., Collins, G., & van Smeden, M. (2020) · BMJ · doi:10.1136/bmj.m441 · Cited by 2,917
Moons, K. G. M., de Groot, J. A. H., Bouwmeester, W., Vergouwe, Y., Mallett, S., Altman, D. G., Reitsma, J. B., & Collins, G. S. (2014) · PLoS Medicine · doi:10.1371/journal.pmed.1001744 · Cited by 2,018
“The number of events-pervariable (EPV) is commonly used to calculate the sample size, where attaining a sample size with an EPV of ten or more is frequently recommended to avoid overfitting [97–101].”
MethodMoons, K. G. M., Wolff, R. F., Riley, R. D., Whiting, P. F., Westwood, M., Collins, G. S., Reitsma, J. B., Kleijnen, J., & Mallett, S. (2019) · Annals of Internal Medicine · doi:10.7326/m18-1377 · Cited by 1,669
Riley, R. D., Snell, K. I., Ensor, J., Burke, D. L., Harrell Jr, F. E., Moons, K. G., & Collins, G. S. (2019) · Statistics in Medicine · doi:10.1002/sim.7992 · Cited by 1,409
“When Nagelkerke's R2, ln Lnull, and n are available, the R(2)CS_app can be calculated by rearranging Equation (14) to give”
BackgroundMethodPrognosis and prognostic research: what, why, and how?
Moons, K. G. M., Royston, P., Vergouwe, Y., Grobbee, D. E., & Altman, D. G. (2009) · BMJ · doi:10.1136/bmj.b375 · Cited by 1,337
“although a recent study showed that this number could be lower in certain circumstances.”
BackgroundBeaugerie, L., Brousse, N., Bouvier, A. M., Colombel, J. F., Lémann, M., Cosnes, J., Hébuterne, X., Cortot, A., Bouhnik, Y., Gendre, J. P., et al. (2009) · The Lancet · doi:10.1016/s0140-6736(09)61302-7 · Cited by 1,115
The number of subjects per variable required in linear regression analyses
Austin, P. C., & Steyerberg, E. W. (2015) · Journal of Clinical Epidemiology · doi:10.1016/j.jclinepi.2014.12.014 · Cited by 1,059
“In a more recent study, Vittinghoff and McCulloch [10] found that this requirement of 10 EPV could be relaxed in the context of confounder adjustment, to requiring five to nine EPV.”
MethodBrantsch, K. D., Meisner, C., Schönfisch, B., Trilling, B., Wehner-Caroli, J., Röcken, M., & Breuninger, H. (2008) · The Lancet Oncology · doi:10.1016/s1470-2045(08)70178-5 · Cited by 1,054
Reporting Recommendations for Tumor Marker Prognostic Studies (REMARK): Explanation and Elaboration
Altman, D. G., McShane, L. M., Sauerbrei, W., & Taube, S. E. (2012) · PLoS Medicine · doi:10.1371/journal.pmed.1001216 · Cited by 1,013
“In this situation sample size need not be as large as the 10-25 events per variable rule would recommend [90] and other sample size calculation methods that appropriately account for correlation of the marker with the other variables are available [82,91].”
BackgroundVariable selection strategies and its importance in clinical prediction modelling.
Chowdhury, M. Z. I., & Turin, T. C. (2020) · Family Medicine and Community Health · doi:10.1136/fmch-2019-000262 · Cited by 1,011
“Other rules also exist, such as the ‘one in twenty rule’, 10 ‘one in fifty rule’ 11 or ‘five to nine events per variable rule’, 12 depending on the research question(s).”
Sparse data bias: a problem hiding in plain sight
Greenland, S., Mansournia, M. A., & Altman, D. G. (2016) · BMJ · doi:10.1136/bmj.i1981 · Cited by 898
Sample size for binary logistic prediction models: Beyond events per variable criteria
van Smeden, M., Moons, K. G., de Groot, J. A., Collins, G. S., Altman, D. G., Eijkemans, M. J., & Reitsma, J. B. (2019) · Statistical Methods in Medical Research · doi:10.1177/0962280218784726 · Cited by 791
Afilalo, J., Eisenberg, M. J., Morin, J.-F., Bergman, H., Monette, J., Noiseux, N., Perrault, L. P., Alexander, K. P., Langlois, Y., Dendukuri, N., et al. (2010) · Journal of the American College of Cardiology · doi:10.1016/j.jacc.2010.06.039 · Cited by 746
Comparison of Propensity Score Methods and Covariate Adjustment
Elze, M. C., Gregson, J., Baber, U., Williamson, E., Sartori, S., Mehran, R., Nichols, M., Stone, G. W., & Pocock, S. J. (2017) · Journal of the American College of Cardiology · doi:10.1016/j.jacc.2016.10.060 · Cited by 713
“However, more recent opinions favor relaxing this rule of thumb([3]).”
BackgroundStrangfeld, A., Schäfer, M., Gianfrancesco, M. A., Lawson-Tovey, S., Liew, J. W., Ljung, L., Mateus, E. F., Richez, C., Santos, M. J., Schmajuk, G., et al. (2021) · Annals of the Rheumatic Diseases · doi:10.1136/annrheumdis-2020-219498 · Cited by 667
Guidelines for reporting mortality and morbidity after cardiac valve interventions
Akins, C. W., Miller, D. C., Turina, M. I., Kouchoukos, N. T., Blackstone, E. H., Grunkemeier, G. L., Takkenberg, J. J. M., David, T. E., Butchart, E. G., Adams, D. H., et al. (2008) · The Journal of Thoracic and Cardiovascular Surgery · doi:10.1016/j.jtcvs.2007.12.002 · Cited by 665
Common pitfalls in statistical analysis: Logistic regression
Ranganathan, P., Pramesh, C. S., & Aggarwal, R. (2017) · Perspectives in Clinical Research · doi:10.4103/picr.picr_87_17 · Cited by 652
“However, the validity of this thumb rule has been questioned.[4]”
BackgroundNo rationale for 1 variable per 10 events criterion for binary logistic regression analysis
van Smeden, M., de Groot, J. A. H., Moons, K. G. M., Collins, G. S., Altman, D. G., Eijkemans, M. J. C., & Reitsma, J. B. (2016) · BMC Medical Research Methodology · doi:10.1186/s12874-016-0267-3 · Cited by 595
“In the first approach we used likelihood non-convergence as a criterion for removing simulation data sets, as was done in previous studies [12, 13].”
InfluentialReporting and Methods in Clinical Prediction Research: A Systematic Review
Bouwmeester, W., Zuithoff, N. P. A., Mallett, S., Geerlings, M. I., Vergouwe, Y., Steyerberg, E. W., Altman, D. G., & Moons, K. G. M. (2012) · PLoS Medicine · doi:10.1371/journal.pmed.1001221 · Cited by 600
“A frequently mentioned rule of thumb is ‘‘10 events needed per candidate predictor’’ [12,25,26,30,54].”
BackgroundAdequate sample size for developing prediction models is not simply related to events per variable
Ogundimu, E. O., Altman, D. G., & Collins, G. S. (2016) · Journal of Clinical Epidemiology · doi:10.1016/j.jclinepi.2016.02.031 · Cited by 546
“Some have suggested that the 10 EPV rules can be relaxed [5], and others that no single EPV rule of thumb can guarantee accurate estimates of regression coefficients [6].”
Rauh, V., Arunajadai, S., Horton, M., Perera, F., Hoepner, L., Barr, D. B., & Whyatt, R. (2011) · Environmental Health Perspectives · doi:10.1289/ehp.1003160 · Cited by 524
Five myths about variable selection
Heinze, G., & Dunkler, D. (2017) · Transplant International · doi:10.1111/tri.12895 · Cited by 528
“Current recommendations suggest that a minimum of 5–15 EPV should be available, depending on context [15,18].”
BackgroundHow to Read Articles That Use Machine Learning: Users’ Guides to the Medical Literature
Liu, Y., Chen, P.-H. C., Krause, J., & Peng, L. (2019) · JAMA · doi:10.1001/jama.2019.16489 · Cited by 516
Developing prediction models for clinical use using logistic regression: an overview
Shipe, M. E., Deppen, S. A., Farjah, F., & Grogan, E. L. (2019) · Journal of Thoracic Disease · doi:10.21037/jtd.2019.01.25 · Cited by 512
“However, this method has largely been replaced as models with higher and lower EVP are all susceptible to bias (17).”
MethodNguyen, T.-L., Collins, G. S., Spence, J., Daurès, J.-P., Devereaux, P. J., Landais, P., & Le Manach, Y. (2017) · BMC Medical Research Methodology · doi:10.1186/s12874-017-0338-0 · Cited by 512
“regression adjustments after matching [10, 11], as described in the following subsection.”
MethodGuidelines for Reporting Mortality and Morbidity After Cardiac Valve Interventions
Akins, C. W., Miller, D. C., Turina, M. I., Kouchoukos, N. T., Blackstone, E. H., Grunkemeier, G. L., Takkenberg, J. J. M., David, T. E., Butchart, E. G., Adams, D. H., et al. (2008) · The Annals of Thoracic Surgery · doi:10.1016/j.athoracsur.2007.12.082 · Cited by 481
Merlo, M., Pyxaras, S. A., Pinamonti, B., Barbati, G., Di Lenarda, A., & Sinagra, G. (2011) · Journal of the American College of Cardiology · doi:10.1016/j.jacc.2010.11.030 · Cited by 430
Pavlou, M., Ambler, G., Seaman, S., De Iorio, M., & Omar, R. Z. (2016) · Statistics in Medicine · doi:10.1002/sim.6782 · Cited by 418
“This cut-off has been challenged as not being based on convincing scientific reasoning [3], and in fact, it is based on empirical evidence for obtaining valid estimates of standard errors for regression coefficients, rather than reliable risk prediction [9, 10].”
Ersbøll, M., Valeur, N., Mogensen, U. M., Andersen, M. J., Møller, J. E., Velazquez, E. J., Hassager, C., Søgaard, P., & Køber, L. (2013) · Journal of the American College of Cardiology · doi:10.1016/j.jacc.2013.02.061 · Cited by 408
“The multivariate statistical models suffer from overfitting due to the small number of events, which warrant caution for the interpretation, but this strategy can be acceptable when the main focus is on controlling for confounders rather than building prediction models (29).”
BackgroundT1 mapping and survival in systemic light-chain amyloidosis
Banypersad, S. M., Fontana, M., Maestrini, V., Sado, D. M., Captur, G., Petrie, A., Piechnik, S. K., Whelan, C. J., Herrey, A. S., Gillmore, J. D., et al. (2015) · European Heart Journal · doi:10.1093/eurheartj/ehu444 · Cited by 393
“analysis can still be performed in these situations.(26) Whether ECV and pre-contrast T1—which are not entirely concordant—provide different pathological insights is at this stage unknown.”
Montero-Odasso, M. M., Sarquis-Adamson, Y., Speechley, M., Borrie, M. J., Hachinski, V. C., Wells, J., Riccio, P. M., Schapira, M., Sejdic, E., Camicioli, R. M., et al. (2017) · JAMA Neurology · doi:10.1001/jamaneurol.2017.0643 · Cited by 381
“4) is considered robust for models with binary end point outcomes.(69) Our annual conversion rate toADwas 7%,which is in linewith clinicbased studies, including amnestic MCI, but higher than populationstudies; thus,our results aregeneralizable toclinicbasedsettingsonly.”
ResultClinical Presentation of Patients with Ebola Virus Disease in Conakry, Guinea
Bah, E. I., Lamah, M.-C., Fletcher, T., Jacob, S. T., Brett-Major, D. M., Sall, A. A., Shindo, N., Fischer, W. A., Lamontagne, F., Saliou, S. M., et al. (2015) · New England Journal of Medicine · doi:10.1056/nejmoa1411249 · Cited by 386
Postcolonoscopy colorectal cancers are preventable: a population-based study
le Clercq, C. M. C., Bouwens, M. W. E., Rondagh, E. J. A., Bakker, C. M., Keulen, E. T. P., de Ridder, R. J., Winkens, B., Masclee, A. A. M., & Sanduleanu, S. (2014) · Gut · doi:10.1136/gutjnl-2013-304880 · Cited by 379
Reporting recommendations for tumor marker prognostic studies (REMARK): explanation and elaboration
Altman, D. G., McShane, L. M., Sauerbrei, W., & Taube, S. E. (2012) · BMC Medicine · doi:10.1186/1741-7015-10-51 · Cited by 372
“In this situation, sample size need not be as large as the 10 to 25 events per variable rule would recommend [86] and other sample size calculation methods that appropriately account for correlation of the marker with the other variables are available [78,87].”
BackgroundJakobsen, J. C., Katakam, K. K., Schou, A., Hellmuth, S. G., Stallknecht, S. E., Leth-Møller, K., Iversen, M., Banke, M. B., Petersen, I. J., Klingenberg, S. L., et al. (2017) · BMC Psychiatry · doi:10.1186/s12888-016-1173-2 · Cited by 366
Device-Related Thrombosis After Percutaneous Left Atrial Appendage Occlusion for Atrial Fibrillation
Fauchier, L., Cinaud, A., Brigadeau, F., Lepillier, A., Pierre, B., Abbey, S., Fatemi, M., Franceschi, F., Guedeney, P., Jacon, P., et al. (2018) · Journal of the American College of Cardiology · doi:10.1016/j.jacc.2018.01.076 · Cited by 352
“Finally, the number of events was relatively low, considering the number of variables in the multivariable analysis, but this may be appropriate for adequate control of confounding in such a study (29).”
BackgroundPsychedelics and mental health: a population study
Krebs, T. S., & Johansen, P.-Ø. (2013) · PLoS ONE · doi:10.1371/journal.pone.0063972 · Cited by 347
“It is typically recommended to have at least 10 events per predictor variable for multivariate logistic regression, although recent simulation studies suggest as few as 5 events per predictor variable is sufficient [27].”
BackgroundThe relationship between liquidity risk and credit risk in banks
Imbierowicz, B., & Rauch, C. (2014) · Journal of Banking & Finance · doi:10.1016/j.jbankfin.2013.11.030 · Cited by 338
Early Corticosteroids in Severe Influenza A/H1N1 Pneumonia and Acute Respiratory Distress Syndrome
Brun-Buisson, C., Richard, J.-C. M., Mercat, A., Thiébaut, A. C. M., & Brochard, L. (2011) · American Journal of Respiratory and Critical Care Medicine · doi:10.1164/rccm.201101-0135oc · Cited by 338
“10), allowing a maximum of eight variables for entry in the final model (18); because corticosteroid therapy was prescribed at varying times after initiation of MV (Figure E1), it was entered in the model as a time-dependent covariate.”
BackgroundStortecky, S., Schoenenberger, A. W., Moser, A., Kalesan, B., Jüni, P., Carrel, T., Bischoff, S., Schoenenberger, C.-M., Stuck, A. E., Windecker, S., et al. (2012) · JACC: Cardiovascular Interventions · doi:10.1016/j.jcin.2012.02.012 · Cited by 332
School resource officers and the criminalization of student behavior
Theriot, M. T. (2009) · Journal of Criminal Justice · doi:10.1016/j.jcrimjus.2009.04.008 · Cited by 327
“…in multivariate analysis (Knofczynski & Mundfrom, 2008), this study used the popular rule of thumb that one independent variable per ten sample members is appropriate (Harrel, Lee, Matchar, & Reichert, 1985; Peduzzi, Concato, Kemper, Holford, & Feinstein, 1996; Vittinghoff & McCulloch, 2007).”
MethodCourvoisier, D. S., Combescure, C., Agoritsas, T., Gayet-Ageron, A., & Perneger, T. V. (2011) · Journal of Clinical Epidemiology · doi:10.1016/j.jclinepi.2010.11.012 · Cited by 320
“proposed by Vittinghoff and McCulloch [7].”
ResultInfluentialRodrigues, M. A., Samarasekera, N., Lerpiniere, C., Humphreys, C., McCarron, M. O., White, P. M., Nicoll, J. A. R., Sudlow, C. L. M., Cordonnier, C., Wardlaw, J. M., et al. (2018) · The Lancet Neurology · doi:10.1016/s1474-4422(18)30006-1 · Cited by 309
Psychedelics not linked to mental health problems or suicidal behavior: A population study
Johansen, P.-Ø., & Krebs, T. S. (2015) · Journal of Psychopharmacology · doi:10.1177/0269881114568039 · Cited by 302
Ziadi, M. C., deKemp, R. A., Williams, K., Guo, A., Renaud, J. M., Chow, B. J. W., Klein, R., Ruddy, T. D., Aung, M., Garrard, L., et al. (2012) · Journal of Nuclear Cardiology · doi:10.1007/s12350-011-9506-5 · Cited by 301
“Although the sample size of 120 may be small, it is powered to detect the effect of SSS and MFR on 3-vessel CAD with only these three other variables considered in the adjustment.(26) To show the incremental value of MFR over SSS, the adjusted model with SSS ? MFR was compared to the model with SSS only using a likeli…”
MethodSchoenenberger, A. W., Stortecky, S., Neumann, S., Moser, A., Juni, P., Carrel, T., Huber, C., Gandon, M., Bischoff, S., Schoenenberger, C.-M., et al. (2013) · European Heart Journal · doi:10.1093/eurheartj/ehs304 · Cited by 299
“However, the number of 22 patients with functional decline was sufficient to avoid over-fitting in the applied univariable and bivariable models.(32) The finding that despite limited sample size the frailty index and all its subcomponents were strong predictors further supports the conclusion of .”
ResultNarayanamurthy, G., & Tortorella, G. (2021) · International Journal of Production Economics · doi:10.1016/j.ijpe.2021.108075 · Cited by 295
“In regression analysis, which is the procedure conducted in our study, many researchers (e.g. Concato et al., 1995; Peduzzi et al., 1995; Vittinghoff and McCulloch, 2007) say that there should be at least”
Guidelines for reporting mortality and morbidity after cardiac valve interventions
Akins, C. W., Miller, D. C., Turina, M. I., Kouchoukos, N. T., Blackstone, E. H., Grunkemeier, G. L., Takkenberg, J. J. M., David, T. E., Butchart, E. G., Adams, D. H., et al. (2008) · European Journal of Cardio-Thoracic Surgery · doi:10.1016/j.ejcts.2007.12.055 · Cited by 289
“events as the minimum per risk factor considered in the model has been advocated for both logistic regression(14) and Cox regression,(15) although this minimum could be lowered a bit.(16) In cases of few events per risk factor, resampling techniques can be used to test model validity.”
BackgroundStatistical Foundations for Model-Based Adjustments
Greenland, S., & Pearce, N. (2015) · Annual Review of Public Health · doi:10.1146/annurev-publhealth-031914-122559 · Cited by 274
van Blitterswijk, M., DeJesus-Hernandez, M., Niemantsverdriet, E., Murray, M. E., Heckman, M. G., Diehl, N. N., Brown, P. H., Baker, M. C., Finch, N. A., Bauer, P. O., et al. (2013) · The Lancet Neurology · doi:10.1016/s1474-4422(13)70210-2 · Cited by 274
Lin, X., Featherman, M., Brooks, S. L., & Hajli, N. (2019) · Information Systems Frontiers · doi:10.1007/s10796-018-9831-1 · Cited by 273
“The ratio of sample size to independent variables, including gender, is 45:1, which exceeds the rule of thumb of 10 for the sample size requirement in a regression analysis (Vittinghoff and McCulloch 2007).”
BackgroundPrognostic Value of 64-Slice Cardiac Computed Tomography
Chow, B. J. W., Wells, G. A., Chen, L., Yam, Y., Galiwango, P., Abraham, A., Sheth, T., Dennie, C., Beanlands, R. S., & Ruddy, T. D. (2010) · Journal of the American College of Cardiology · doi:10.1016/j.jacc.2009.10.039 · Cited by 270
Marsh, N., Webster, J., Larsen, E., Cooke, M., Mihala, G., & Rickard, C. M. (2018) · Journal of Hospital Medicine · doi:10.12788/jhm.2867 · Cited by 267
“Location: hand cubital fossa lower forearm wrist upper forearm 582 (37) 418 (27) 264 (17) 190 (12) 106 (7) 85 (15) 61 (15) 38 (14) 21 (11) 22 (21) 84 (14) 75 (20) 44 (17) 35 (18) 27 (25) 59 (10) 41 (10) 23 (9) 21 (11) 8 (8)”
InfluentialCatastrophic health expenditure and impoverishment in Turkey
Yardim, M. S., Cilingiroglu, N., & Yardim, N. (2010) · Health Policy · doi:10.1016/j.healthpol.2009.08.006 · Cited by 266
Cowley, L. E., Farewell, D. M., Maguire, S., & Kemp, A. M. (2019) · Diagnostic and Prognostic Research · doi:10.1186/s41512-019-0060-y · Cited by 263
The Effects of Corpus Use on Second Language Vocabulary Learning: A Multilevel Meta-analysis
Lee, H., Warschauer, M., & Lee, J. H. (2019) · Applied Linguistics · doi:10.1093/applin/amy012 · Cited by 256
Flowers, H. L., Silver, F. L., Fang, J., Rochon, E., & Martino, R. (2013) · Journal of Communication Disorders · doi:10.1016/j.jcomdis.2013.04.001 · Cited by 249
Clinical outcome and pathological features associated with NRAS mutation in cutaneous melanoma
Devitt, B., Liu, W., Salemi, R., Wolfe, R., Kelly, J., Tzen, C., Dobrovic, A., & McArthur, G. (2011) · Pigment Cell & Melanoma Research · doi:10.1111/j.1755-148x.2011.00873.x · Cited by 246
“In general, 10 events are required for each HR (Vittinghoff and Mcculloch, 2007), a rule which is observed in the simpler multivariate analysis (Breslow thickness and mutational status) but violated in the larger analysis.”
MethodMorbidity and Survival Probability in Burn Patients in Modern Burn Care*
Jeschke, M. G., Pinto, R., Kraft, R., Nathens, A. B., Finnerty, C. C., Gamelli, R. L., Gibran, N. S., Klein, M. B., Arnoldo, B. D., Tompkins, R. G., et al. (2015) · Critical Care Medicine · doi:10.1097/ccm.0000000000000790 · Cited by 243
Sparks, J. A., Wallace, Z. S., Seet, A. M., Gianfrancesco, M. A., Izadi, Z., Hyrich, K. L., Strangfeld, A., Gossec, L., Carmona, L., Mateus, E. F., et al. (2021) · Annals of the Rheumatic Diseases · doi:10.1136/annrheumdis-2021-220418 · Cited by 238
A modern maximum-likelihood theory for high-dimensional logistic regression
Sur, P., & Candès, E. J. (2019) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.1810420116 · Cited by 232
“A line of simulation-based results exists to guide practitioners about the sample size required to avoid finite-sample problems (36, 37).”
BackgroundAnderson, L. A., Peters, C. L., Park, B. B., Stoddard, G. J., Erickson, J. A., & Crim, J. R. (2009) · The Journal of Bone and Joint Surgery-American Volume · doi:10.2106/jbjs.g.01198 · Cited by 233
Tsai, A. C., & Burns, B. F. O. (2015) · Social Science & Medicine · doi:10.1016/j.socscimed.2015.06.024 · Cited by 227
“…accepted rule of thumb that 5– 10 events are required per covariate in logit or Cox regression models (Peduzzi et al., 1995; Peduzzi et al., 1996; Vittinghoff & McCulloch, 2007), we assessed whether or not each paper could have supported fully saturated regression models with product terms…”
MethodStatistical primer: multivariable regression considerations and pitfalls†
Grant, S. W., Hickey, G. L., & Head, S. J. (2019) · European Journal of Cardio-Thoracic Surgery · doi:10.1093/ejcts/ezy403 · Cited by 225
Immunoglobulin signature predicts risk of post-acute COVID-19 syndrome
Cervia, C., Zurbuchen, Y., Taeschler, P., Ballouz, T., Menges, D., Hasler, S., Adamo, S., Raeber, M. E., Bächli, E., Rudiger, A., et al. (2022) · Nature Communications · doi:10.1038/s41467-021-27797-1 · Cited by 226
Slama, R., Morgenstern, V., Cyrys, J., Zutavern, A., Herbarth, O., Wichmann, H.-E., Heinrich, J., & the LISA Study Group (2007) · Environmental Health Perspectives · doi:10.1289/ehp.10047 · Cited by 222
“…regression model and their variance (“overfit-ting”) may be a concern if there are < 10 events per variable (Harrell 2001; Peduzzi et al. 1996), whereas other researchers consider that bias remains infrequent with as few as 5–9 events per variable in the model (Vittinghoff and McCulloch 2007).”
BackgroundBurnout and Risk of Coronary Heart Disease: A Prospective Study of 8838 Employees
Toker, S., Melamed, S., Berliner, S., Zeltser, D., & Shapira, I. (2012) · Psychosomatic Medicine · doi:10.1097/psy.0b013e31826c3174 · Cited by 222
“To avoid an excessive number of covariates in the analysis, we estimated four multivariate Cox proportional hazard regression models in which the number of predictors was restricted (34).”
MethodMultivariable Modeling Strategies
Harrell, F. E. (2015) · Springer Series in Statistics · doi:10.1007/978-3-319-19425-7_4 · Cited by 221
Mughini Gras, L., Smid, J. H., Wagenaar, J. A., de Boer, A. G., Havelaar, A. H., Friesema, I. H. M., French, N. P., Busani, L., & van Pelt, W. (2012) · PLoS ONE · doi:10.1371/journal.pone.0042599 · Cited by 221
“To support the accuracy of inferences of regression models with ,5 cases per variable, bias-corrected bootstrap confidence intervals were also calculated (1000 replications) and compared with the standard ones [38].”
Hickey, G. L., Dunning, J., Seifert, B., Sodeck, G., Carr, M. J., Burger, H. U., & Beyersdorf, F. (2015) · European Journal of Cardio-Thoracic Surgery · doi:10.1093/ejcts/ezv168 · Cited by 221
“Although the rule of thumb has been the requirement of 10 EPV, this is not a strict necessity and in general is application dependent [52, 53].”
BackgroundDouketis, J. D., Gu, C. S., Schulman, S., Ghirarduzzi, A., Pengo, V., & Prandoni, P. (2007) · Annals of Internal Medicine · doi:10.7326/0003-4819-147-11-200712040-00007 · Cited by 220
“Patients, n 2052 1628 424 Mean age (range), y 62 (16–96) 62 (16–96) 61 (18–91) Women, n (%) 1082 (53) 893 (55) 189 (45) Mean duration of previous anticoagulant therapy (range), mo 6 (3–39) 6 (3–39) 6 (6–6) Duration of previous anticoagulant therapy, n (%) 3 mo 540 (26) 540 (33) NA 3 to 6 mo 1241 (60) 817 (50) 424 (60)…”
BackgroundLoneliness, social isolation, their synergistic interaction, and mortality.
Beller, J., & Wagner, A. (2018) · Health Psychology · doi:10.1037/hea0000605 · Cited by 217
Predictors of disease relapse in IgG4-related disease following rituximab
Wallace, Z. S., Mattoo, H., Mahajan, V. S., Kulikova, M., Lu, L., Deshpande, V., Choi, H. K., Pillai, S., & Stone, J. H. (2016) · Rheumatology · doi:10.1093/rheumatology/kev438 · Cited by 216
Nelson, G., Kiyang, L. N., Crumley, E. T., Chuck, A., Nguyen, T., Faris, P., Wasylak, T., Basualdo‐Hammond, C., McKay, S., Ljungqvist, O., et al. (2016) · World Journal of Surgery · doi:10.1007/s00268-016-3472-7 · Cited by 213
“Using the rule of at least ten observations per predictor variable in multiple regression models [17, 18], this sample size enabled use of regression models that included all potential confounding factors, for LOS, complications and readmissions as outcome variables.”
MethodKindler, S., Samietz, S., Houshmand, M., Grabe, H. J., Bernhardt, O., Biffar, R., Kocher, T., Meyer, G., Völzke, H., Metelmann, H.-R., et al. (2012) · The Journal of Pain · doi:10.1016/j.jpain.2012.09.004 · Cited by 212
Peffault de Latour, R., Porcher, R., Dalle, J.-H., Aljurf, M., Korthof, E. T., Svahn, J., Willemze, R., Barrenetxea, C., Mialou, V., Soulier, J., et al. (2013) · Blood · doi:10.1182/blood-2013-01-479733 · Cited by 213
“12 mo 22% (17-27) 44% (35-52) 14% (10-19) 24% (18-31) 60 mo 27% (22-33) 47% (38-56) 19% (14-26) 27% (20-35)”
InfluentialWhen Optimism Hurts: Inflated Predictions in Psychiatric Neuroimaging
Whelan, R., & Garavan, H. (2014) · Biological Psychiatry · doi:10.1016/j.biopsych.2013.05.014 · Cited by 209
“cases per predictor is a common (18), although not a universal (19), recommendation.”
Darmon, M., Schortgen, F., Vargas, F., Liazydi, A., Schlemmer, B., Brun-Buisson, C., & Brochard, L. (2011) · Intensive Care Medicine · doi:10.1007/s00134-010-2050-y · Cited by 208
“Given the number of events (22 patients with persistent AKI), three variables were entered into the model (1 variable included in the model for 7 events) [28].”
MethodFong, K., Truong, V., Foote, C. J., Petrisor, B., Williams, D., Ristevski, B., Sprague, S., & Bhandari, M. (2013) · BMC Musculoskeletal Disorders · doi:10.1186/1471-2474-14-103 · Cited by 200
Thavendiranathan, P., Zhang, L., Zafar, A., Drobni, Z. D., Mahmood, S. S., Cabral, M., Awadalla, M., Nohria, A., Zlotoff, D. A., Thuny, F., et al. (2021) · Journal of the American College of Cardiology · doi:10.1016/j.jacc.2021.01.050 · Cited by 194
Geissbühler, M., Hincapié, C. A., Aghlmandi, S., Zwahlen, M., Jüni, P., & da Costa, B. R. (2021) · BMC Medical Research Methodology · doi:10.1186/s12874-021-01310-0 · Cited by 195
Valpione, S., Mundra, P. A., Galvani, E., Campana, L. G., Lorigan, P., De Rosa, F., Gupta, A., Weightman, J., Mills, S., Dhomen, N., et al. (2021) · Nature Communications · doi:10.1038/s41467-021-24343-x · Cited by 194
Steenstra, I. A., Munhall, C., Irvin, E., Oranye, N., Passmore, S., Van Eerd, D., Mahood, Q., & Hogg-Johnson, S. (2017) · Journal of Occupational Rehabilitation · doi:10.1007/s10926-016-9666-x · Cited by 193
“The high quality study examined 7 constructs in a population with 98 events, which indicates sufficient power for this study [40].”
BackgroundPhysical Performance and Frailty in Chronic Kidney Disease
Reese, P. P., Cappola, A. R., Shults, J., Townsend, R. R., Gadegbeku, C. A., Anderson, C., Baker, J. F., Carlow, D., Sulik, M. J., Lo, J. C., et al. (2013) · American Journal of Nephrology · doi:10.1159/000355568 · Cited by 187
Outcomes of Children With Severe Bronchopulmonary Dysplasia Who Were Ventilator Dependent at Home
Cristea, A. I., Carroll, A. E., Davis, S. D., Swigonski, N. L., & Ackerman, V. L. (2013) · Pediatrics · doi:10.1542/peds.2012-2990 · Cited by 187
“72 Private and Medicaid 14 (17) 2 (11) Private 18 (22) 2 (11) Initial hospital stay Length (mo) 10 (8–12) 10.”
Extracorporeal membrane oxygenation in postcardiotomy patients: Factors influencing outcome
Kumar, T. K. S., Zurakowski, D., Dalton, H., Talwar, S., Allard-Picou, A., Duebener, L. F., Sinha, P., & Moulick, A. (2010) · The Journal of Thoracic and Cardiovascular Surgery · doi:10.1016/j.jtcvs.2010.02.034 · Cited by 180
Goodwill Impairment: The Effects of Public Enforcement and Monitoring by Institutional Investors
Glaum, M., Landsman, W. R., & Wyrwa, S. (2018) · The Accounting Review · doi:10.2308/accr-52006 · Cited by 172
Huelin, P., Solà, E., Elia, C., Solé, C., Risso, A., Moreira, R., Carol, M., Fabrellas, N., Bassegoda, O., Juanola, A., et al. (2019) · Hepatology · doi:10.1002/hep.30592 · Cited by 178
“Serum BAs (mol/L)‡ 51 (17-98) 47 (23-96) 33 (8-86) 74 (23-133)”
MethodInfluentialLiang, B., Yan, H., Tian, Y., Chen, X., Yan, L., Zhang, T., Zhou, Z., Wang, L., & Dai, J. (2019) · Frontiers in Oncology · doi:10.3389/fonc.2019.00269 · Cited by 178
“Recent studies show that the one tenth rule is generally too conservative, and can be relaxed for the logistic and Cox regression (18).”
BackgroundAttitudes Toward e-Mental Health Services in a Community Sample of Adults: Online Survey
March, S., Day, J., Ritchie, G., Rowe, A., Gough, J., Hall, T., Yuen, C. Y. J., Donovan, C. L., & Ireland, M. (2018) · Journal of Medical Internet Research · doi:10.2196/jmir.9109 · Cited by 158
“As some dichotomous response categories were endorsed by relatively few respondents, to account for potential problems arising from too many predictors in the final combined models (eg, biased estimates), we used nonparametric bootstrapping techniques to estimate bias-corrected confidence intervals (CIs), in line with…”
Wynants, L., Bouwmeester, W., Moons, K. G. M., Moerbeek, M., Timmerman, D., Van Huffel, S., Van Calster, B., & Vergouwe, Y. (2015) · Journal of Clinical Epidemiology · doi:10.1016/j.jclinepi.2015.02.002 · Cited by 171
“Some researchers have proposed upward [4,5] and downward [6] adjustments for the EPV guideline, stating that the required EPV is influenced by the size of the predictor effects, the correlations among predictors, the prevalence of dichotomous predictors, and the predictor selection strategy.”
BackgroundSamarasekera, N., Fonville, A., Lerpiniere, C., Farrall, A. J., Wardlaw, J. M., White, P. M., Smith, C., Al-Shahi Salman, R., Addison, A., Ahmad, K., et al. (2015) · Stroke · doi:10.1161/strokeaha.114.007953 · Cited by 174
“We prespecified these variables on the basis of their clinical relevance, known or hypothesized influences on outcome,(8,66) and the number of outcome events per variable because 5 to 9 outcome events per variable is valid, especially if variables are selected a priori and the associations found are plausible in the c…”
HIV status in discordant couples in sub-Saharan Africa: a systematic review and meta-analysis
Eyawo, O., de Walque, D., Ford, N., Gakii, G., Lester, R. T., & Mills, E. J. (2010) · The Lancet Infectious Diseases · doi:10.1016/s1473-3099(10)70189-4 · Cited by 174
Sikkema, M., Looman, C. W. N., Steyerberg, E. W., Kerkhof, M., Kastelein, F., van Dekken, H., van Vuuren, A. J., Bode, W. A., van der Valk, H., Ouwendijk, R. J. T., et al. (2011) · American Journal of Gastroenterology · doi:10.1038/ajg.2011.153 · Cited by 173
“A generalized linear model was made to calculate risk ratios (RRs) with corresponding 95p confidence intervals (CIs) (”
MethodDeng, B., Zhou, B., Zhang, S., Zhu, Y., Han, L., Geng, Y., Jin, Z., Liu, H., Wang, D., Zhao, Y., et al. (2013) · PLoS ONE · doi:10.1371/journal.pone.0080802 · Cited by 166
“Multivariate logistic regression with stepwise backward elimination was used to investigate associations between non-severity and severity in the confirmed cases, and Cox regression was used to model outcomes in the severe cases, allowing a maximum of eight variables for entry in the final model [17].”
MethodSong, P., Wang, H., Theodoratou, E., Chan, K. Y., & Rudan, I. (2018) · Journal of Global Health · doi:10.7189/jogh.08.010804 · Cited by 164
Masci, P. G., Schuurman, R., Andrea, B., Ripoli, A., Coceani, M., Chiappino, S., Todiere, G., Srebot, V., Passino, C., Aquaro, G. D., et al. (2013) · Circulation: Cardiovascular Imaging · doi:10.1161/circimaging.113.000438 · Cited by 164
“Spironolactone, n (%) 20 (34) 9 (41) 11 (31) 0.”
Metadata
- Authors
- Vittinghoff, E., & McCulloch, C. E.
- Affiliations
- University of California, San Francisco (US)
- Venue
- American Journal of Epidemiology
- Publisher
- Oxford University Press (OUP)
- Published
- 2007-01-12
- Volume
- 165
- Issue
- 6
- Pages
- 710-718
- ISSN
- 0002-9262 (ISSN-L)
- 1476-6256
- OpenAlex
- W2130373985
- Cited by
- 4,359
- References
- 9
- Citations by year
- 2026: 642 · 2025: 394 · 2024: 313 · 2023: 315 · 2022: 306 · 2021: 351 · 2020: 332 · 2019: 281 · 2018: 282 · 2017: 205 · 2016: 194 · 2015: 171 · 2014: 146 · 2013: 132 · 2012: 94
- FWCI
- 20.12
- Type
- journal article
- Language
- en
- Open access
- closed
- Locations
- American Journal of Epidemiology (publishedVersion)
- PubMed (publishedVersion)
- Topics
- Physical Sciences › Mathematics › Statistics and Probability › Statistical Methods in Epidemiology
- Physical Sciences › Mathematics › Statistics and Probability › Statistical Methods and Inference
- Physical Sciences › Mathematics › Statistics and Probability › Advanced Causal Inference Techniques
- Keywords
- sample size, logistic regression, Cox proportional hazards model, confidence interval coverage, type I error, relative bias, simulation study, sensitivity analysis, confounding adjustment, small sample size
- MeSH terms
- Data Interpretation, Statistical, Factor Analysis, Statistical, Humans, Predictive Value of Tests, Sensitivity and Specificity, Bias, Confounding Factors, Epidemiologic, Prevalence, Multivariate Analysis, Confidence Intervals, Logistic Models, Proportional Hazards Models, Case-Control Studies, Sample Size
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.