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
Items 901–1,000 of 1,018
Levine, A., Chanchlani, N., Hussey, S., Ziv‐Baran, T., Escher, J. C., Dias, J. A., Veres, G., Koletzko, S., Turner, D., Kolho, K.-L., et al. (2019) · Journal of Crohn s and Colitis · doi:10.1093/ecco-jcc/jjz111 · Cited by 29
Tsujimoto, K., Mizuno, K., Kobayashi, Y., Tanuma, A., & Liu, M. (2020) · European Journal of Physical and Rehabilitation Medicine · doi:10.23736/s1973-9087.19.05595-3 · Cited by 27
Sytsma, V. A., Connealy, N. T., & Piza, E. L. (2020) · Crime & Delinquency · doi:10.1177/0011128720910961 · Cited by 32
Bai, D. L., Liu, T.-W., Chou, H., & Hsu, Y. (2020) · PLoS ONE · doi:10.1371/journal.pone.0241276 · Cited by 30
“When calculating the sample size for a multiple logistic regression model, the rule of ten events per variable is widely used [42], and recently, the rule of 20 or 50 events per variable has been increasingly suggested [43].”
MethodCalciphylaxis in end-stage kidney disease: outcome data from the United Kingdom Calciphylaxis Study
Chinnadurai, R., Huckle, A. L., Hegarty, J., Kalra, P. A., & Sinha, S. (2021) · Journal of Nephrology · doi:10.1007/s40620-020-00908-9 · Cited by 27
Phillips, K. A., Greenberg, J. L., Hoeppner, S. S., Weingarden, H. M., O’Keefe, S. M., Keshaviah, A. P., Schoenfeld, D., & Wilhelm, S. (2021) · Journal of Affective Disorders · doi:10.1016/j.jad.2021.03.011 · Cited by 32
Aerde, N. V., Meersseman, P., Debaveye, Y. A., Wilmer, A. P., Gunst, J., Casaer, M. P., Wauters, J., Wouters, P. J., Gosselink, R., den Berghe, G. V., et al. (2021) · Thorax · doi:10.1136/thoraxjnl-2020-216720 · Cited by 32
Reaffirming the relationship between routine activities and violent victimization in prison
McNeeley, S. (2022) · Journal of Criminal Justice · doi:10.1016/j.jcrimjus.2022.101883 · Cited by 27
Loneliness and mortality: The moderating effect of positive affect
Beller, J. (2022) · Applied Psychology Health and Well-Being · doi:10.1111/aphw.12354 · Cited by 37
Artificially Intelligent Nanoarray Detects Various Cancers by Liquid Biopsy of Volatile Markers
Amor, R. E., Zinger, A., Broza, Y. Y., Schroeder, A., & Haick, H. (2022) · Advanced Healthcare Materials · doi:10.1002/adhm.202200356 · Cited by 28
Mori, M., Palumbo, D., Muffatti, F., Partelli, S., Mushtaq, J., Andreasi, V., Prato, F., Ubeira-Gabellini, M. G., Palazzo, G., Falconi, M., et al. (2022) · European Radiology · doi:10.1007/s00330-022-09351-9 · Cited by 27
Lindau, S., Nadermann, M., Ackermann, H., Bingold, T. M., Stephan, C., Kempf, V. A. J., Herzberger, P., Beiras-Fernández, A., Zacharowski, K. D., & Meybohm, P. (2015) · Journal of Intensive Care · doi:10.1186/s40560-015-0097-0 · Cited by 27
“The number of cases and these five predictor variables in Cox models stick to the rule of thumb, using around ten outcome events per predictor variable [12, 13].”
MethodSurgical predictors of acute postoperative pain after hip arthroscopy
Tan, C. O., Chong, Y. M., Tran, P., Weinberg, L., & Howard, W. (2015) · BMC Anesthesiology · doi:10.1186/s12871-015-0077-x · Cited by 27
“The total number of variables included in the regression model did not exceed the conservative ratio of 1 variable per 10 samples [10].”
BackgroundYoung women with family history of breast cancer and their risk factors for benign breast disease
Berkey, C. S., Tamimi, R. M., ROSNER, B. A., Frazier, A. L., & Colditz, G. A. (2011) · Cancer · doi:10.1002/cncr.26519 · Cited by 28
“Logistic model estimates obtained from data including <10 outcome events (BBD cases) should be interpreted with great caution, as there may be a 10% bias away from the null in the estimated effect of any continuous risk factor.(31) But the general consistency of our estimated alcohol effects, across family history sub…”
MethodMann, F., Fisher, H. L., & Johnson, S. (2014) · Journal of Mental Health · doi:10.3109/09638237.2014.910641 · Cited by 31
“A recommended ‘‘rule of thumb’’ is to have at least 10 times as many observations as predictor variables, in order to reduce the Type II error risk (discussed in Vittinghoff & McCulloch, 2007).”
BackgroundLee, E. Y., Neuman, M. I., Lee, N. J., Johnson, V. M., Zurakowski, D., Tracy, D. A., & Boiselle, P. M. (2012) · American Journal of Roentgenology · doi:10.2214/ajr.11.8005 · Cited by 27
“10 from univariate analysis treated as candidate variables in the model [23].”
MethodBatte, A., Katahoire, A., Chimoyi, A., Ajambo, S., Tibingana, B., & Banura, C. (2015) · BMC Public Health · doi:10.1186/s12889-015-1420-3 · Cited by 27
“05 were entered in the logistic regression model using the rule of 10 events per variable [30].”
MethodSchumm, W. R. (2015) · Comprehensive Psychology · doi:10.2466/17.cp.4.24 · Cited by 32
“…4, Article 24Comprehensive Psychology when they had between 141 and 153 cases, creating a ratio of cases to independent variables far below the recommended level of 10 ( Schumm, et al ., 1980 , p. 253), even below the level of fi ve to nine deemed as acceptable by Vittinghoff and McCulloch (2007 ).”
Backgroundde Man, H., Mughini‐Gras, L., Schimmer, B., Friesema, I. H. M., Husman, A. D. R., & van Pelt, W. (2015) · Epidemiology and Infection · doi:10.1017/s0950268815002654 · Cited by 35
Tibial component rotation in total knee arthroplasty
Feczkó, P. Z., Pijls, B. G. C. W., van Steijn, M. J., van Rhijn, L. W., Arts, J. J. C., & Emans, P. J. (2016) · BMC Musculoskeletal Disorders · doi:10.1186/s12891-016-0940-z · Cited by 27
“Various studies have suggested that for each variable studied in multiple regression analysis at least 10 observations are required [38–40] although a recent study showed that this number could be lower in certain circumstances [41].”
BackgroundLi, G., Huang, W., Zhang, L., Tian, Z., Zheng, W., Wang, T., Zhang, T., & Zhang, W. (2018) · Diabetes/Metabolism Research and Reviews · doi:10.1002/dmrr.3003 · Cited by 29
Bushnaq, S., Qeadan, F., Thacker, T., Abbas, M., & Carlson, A. P. (2018) · Interventional Neurology · doi:10.1159/000487960 · Cited by 27
Meeks, J. R., Bambhroliya, A. B., Meyer, E. G., Slaughter, K. B., Fraher, C. J., Sharrief, A. Z., Bowry, R., Ahmed, W. O., Tyson, J. E., Miller, C. C., et al. (2019) · International Journal of Stroke · doi:10.1177/1747493019827763 · Cited by 28
“This renders an adequate sample size for multivariable model building utilizing the convention of 5–20 events per included variable.(18,19) Our final logistic regression model comprises of five important variables, whereas the modified Poisson model has been adjusted for six variables.”
MethodRemiro‐Azócar, A., Heath, A., & Baio, G. (2021) · UCL Discovery (University College London) · Cited by 26
Ginzburg, K., Kutz, I., Koifman, B., Roth, A., Kriwisky, M., David, D., & Bleich, A. (2015) · Annals of Behavioral Medicine · doi:10.1007/s12160-015-9744-x · Cited by 28
“As was shown by Vittinghoff and McCulloch [49], Cox models can be used with a ratio of five outcome events per predictor variable.”
MethodReber, A., Moldovan, A., Dunkel, N., Emonet, S., Rohner, P., Tahintzi, P., Hoffmeyer, P. J., Harbarth, S., & Uçkay, İ. (2012) · Journal of Infection · doi:10.1016/j.jinf.2011.12.023 · Cited by 26
Silvestri, L. A., Clark, M. C., & Moonie, S. A. (2012) · Journal of Nursing Education and Practice · doi:10.5430/jnep.v3n6p21 · Cited by 40
“Vittinghoff and McCulloch (2006) noted a minimum of 10 cases per predictor variable is adequate for logistic regression analysis ([51]).”
BackgroundHigh-frequency oscillation in adults: A utilization review*
Adhikari, N. K. J., Bashir, A., Lamontagne, F., Mehta, S., Ferguson, N. D., Zhou, Q. X., Hand, L. E., Czarnecka, K., Cook, D. J., Granton, J. T., et al. (2011) · Critical Care Medicine · doi:10.1097/ccm.0b013e318226675e · Cited by 26
Zytoon, M. A.-M., & Basahel, A. (2017) · International Journal of Environmental Research and Public Health · doi:10.3390/ijerph14030229 · Cited by 33
“If multivariate logistic analysis was used, regression coefficients would be biased in both positive and negative directions [29,31,32] (also OR would be extremely high or low).”
MethodVarotti, G., Dodi, F., Terulla, A., Santori, G., Mariottini, G., Bertocchi, M., Marchese, A. E., & Fontana, I. (2017) · Transplant Infectious Disease · doi:10.1111/tid.12757 · Cited by 26
Pedersen, J., Flampouri, S., Bryant, C., Liang, X., Mendenhall, N. P., Li, Z., Liu, M., & Muren, L. P. (2019) · Radiotherapy and Oncology · doi:10.1016/j.radonc.2019.09.017 · Cited by 26
Powers, B. D., Rothermel, L. D., Fleming, J. B., Strosberg, J., & Anaya, D. A. (2019) · Journal of Gastrointestinal Surgery · doi:10.1007/s11605-019-04433-4 · Cited by 26
Uptake of polygenic risk information among women at increased risk of breast cancer
Yanes, T., Meiser, B., Kaur, R., Scheepers‐Joynt, M., McInerny, S., Taylor, S., Barlow-Stewart, K. K., Antill, Y. C., Salmon, L., Smyth, C., et al. (2019) · Clinical Genetics · doi:10.1111/cge.13687 · Cited by 27
“However, we included all eligible variables in our regression model based on the theoretical framework used, and on significance level cutoffs as their inclusion is justified due to statistically significant associations.(42) Finally, it should be noted that research is still ongoing to identify SNPs associated with b…”
MethodAlBadri, A., Tahhan, A. S., Sabbak, N., Alkhoder, A. A., Liu, C., Ko, Y., Vaccarino, V., Martini, A., Sidoti, A., Goodwin, C., et al. (2020) · Journal of the American Heart Association · doi:10.1161/jaha.119.015515 · Cited by 26
Kesar, N., Winkelmann, R., Oppermann, J., Ghanaati, S., Martin, D., Neumayer, T., Balster, S., Rödel, C. M., Rödel, F., von der Grün, J., et al. (2020) · Oral Oncology · doi:10.1016/j.oraloncology.2020.104931 · Cited by 26
“To avoid overfitting [34], we strived for a minimum of 10 events per variable (EPV) in each multivariate analysis: Age as a multi-influenced variable was excluded a priori from any Cox model.”
MethodEstep, C. (2017) · The Accounting Review · doi:10.2308/tar-2017-0303 · Cited by 25
O’Sullivan, O., Barker-Davies, R. M., Thompson, K., Bahadur, S., Gough, M. R. C., Lewis, S., Martin, M., Segalini, A., Wallace, G., Phillip, R., et al. (2021) · BMJ Military Health · doi:10.1136/bmjmilitary-2021-001856 · Cited by 26
Strømme, M. F., Mellesdal, L., Bartz-Johannessen, C. A., Kroken, R. A., Krogenes, M. L., Mehlum, L. I., & Johnsen, E. (2021) · Schizophrenia Research · doi:10.1016/j.schres.2021.07.009 · Cited by 27
“As a rule of thumb, the number of outcome events per predictor variable should be 10 or more in Cox regression models (Vittinghoff and McCulloch, 2007).”
Zou, S., Wang, Z., Bhura, M., & Tang, K. (2020) · Public Health · doi:10.1016/j.puhe.2022.01.021 · Cited by 29
Zhai, Y., Ryou, M. K., & Thompson, C. C. (2022) · Gastrointestinal Endoscopy · doi:10.1016/j.gie.2022.04.014 · Cited by 26
Teixeira, J. F., Silva, C., & Sá, F. M. E. (2023) · Journal of Transport Geography · doi:10.1016/j.jtrangeo.2023.103651 · Cited by 31
Mortality surrogates in combined pulmonary fibrosis and emphysema
Zhao, A., Gudmundsson, E., Moğulkoç, N., van Moorsel, C. H. M., Corte, T. J., Vasudev, P., Romei, C., Chapman, R., Wallis, T., Denneny, E., et al. (2023) · European Respiratory Journal · doi:10.1183/13993003.00127-2023 · Cited by 32
“Cox models were used with a minimum of 8 outcome events per predictor covariate [10].”
MethodRosin, G., de Boniface, J., Karthik, G., Frisell, J., Bergh, J. C. S., & Hartman, J. (2014) · British Journal of Cancer · doi:10.1038/bjc.2014.398 · Cited by 29
“This is especially evident when the association is plausible and hypothesised a priori (Vittinghoff and McCulloch, 2007).”
BackgroundOlsson, M. O., Brådvik, L., Öjehagen, A., & Håkansson, A. (2016) · Drug and Alcohol Dependence · doi:10.1016/j.drugalcdep.2016.03.009 · Cited by 31
Keane, D. F., Gardiner, C., Lindley, E. J., Lines, S., Woodrow, G., & Wright, M. (2016) · Nutrients · doi:10.3390/nu8110702 · Cited by 29
“The inclusion of 7 covariates with only 36 events would generally indicate the potential for over-fitting, although computer modelling has suggested that models with greater than 5 events per covariate are not particularly susceptible to problems [29].”
Gruenbaum, S. E., Toscani, L., Fomberstein, K., Ruskin, K. J., Dai, F., Qeva, E., Rosa, G., Meng, L., & Bilotta, F. (2017) · Anesthesia & Analgesia · doi:10.1213/ane.0000000000001946 · Cited by 26
The impact of contralateral carotid artery stenosis on outcomes after carotid endarterectomy
Pothof, A. B., Soden, P. A., Fokkema, M. T. M., Zettervall, S. L., Deery, S. E., Bodewes, T. C. F., de Borst, G. J., & Schermerhorn, M. L. (2017) · Journal of Vascular Surgery · doi:10.1016/j.jvs.2017.04.032 · Cited by 26
“We performed multivariable logistic regression to test the independent association between CCO and our primary and secondary outcomes.(15,16) In addition, we tested the association between the degree of CCS and our primary and secondary outcomes, using <50% CCS as the referent.”
Dupont, G., Flory, L., Morel, J., Lukaszewicz, A., Patoir, A., Presles, É., Monneret, G., & Molliex, S. B. (2018) · PLoS ONE · doi:10.1371/journal.pone.0205237 · Cited by 28
Zelis, N., Buijs, J., de Leeuw, P. W., van Kuijk, S. M. J., & Stassen, P. M. (2020) · European Journal of Internal Medicine · doi:10.1016/j.ejim.2020.02.021 · Cited by 28
Norder, G., Roelen, C. A. M., Bültmann, U., & van der Klink, J. J. (2015) · Scandinavian Journal of Work Environment & Health · doi:10.5271/sjweh.3501 · Cited by 30
“The statistical power of a Cox regression model including 3 independent variables and 27 events (ie, SA episodes due to mood disorders) is debatable, although Vittinghoff and McCulloch (12) showed that CI coverage, type I error, and relative bias with 5–9 events per variable were comparable to those with 10–16 events…”
MethodHaugnes, H. S., Laurell, A., Stierner, U. K., Bremnes, R. M., Dahl, O., Cavallin-Ståhl, E. K., & Cohn-Cedermark, G. E. (2011) · Acta Oncologica · doi:10.3109/0284186x.2011.641507 · Cited by 26
Lifestyle and incident dementia: A COSMIC individual participant data meta‐analysis
Asbroeck, S. V., Köhler, S., van Boxtel, M. P. J., Lipnicki, D. M., Crawford, J. D., Costa, E., Lima‐Costa, M. F., Blay, S. L., Shifu, X., Wang, T., et al. (2024) · Alzheimer s & Dementia · doi:10.1002/alz.13846 · Cited by 30
“53 The proportional hazard assumption was assessed by testing of Schoen-feld residuals.”
MethodTwelve practical recommendations for developing and applying clinical predictive models
Feng, G., Xu, H., Wan, S., Wang, H., Chen, X., Magari, R. T., Han, Y. K., Wei, Y., & Gu, H. (2024) · The Innovation Medicine · doi:10.59717/j.xinn-med.2024.100105 · Cited by 32
Variability in transfusion practice and effectiveness of strategies to improve it.
Shander, A., Puzio, T., & Javidroozi, M. (2012) · Journal of Cardiothoracic and Vascular Anesthesia · doi:10.1053/j.jvca.2012.04.003 · Cited by 23
Bennett, J. (2013) · Ecography · doi:10.1111/j.1600-0587.2013.00393.x · Cited by 28
“However, recent studies assert the importance of signal strength in predictor variables over prevalence in the performance of binomial models (Vittinghoff and McCulloch 2007, Jim é nezValverde et al. 2009).”
BackgroundDialysate Cytokine Levels do not Predict Encapsulating Peritoneal Sclerosis
Goodlad, C., Tam, F. W. K., Ahmad, S., Bhangal, G., North, B. V., & Brown, E. A. (2014) · Peritoneal Dialysis International · doi:10.3747/pdi.2012.00305 · Cited by 27
Despard, M. (2016) · Human Services Organizations Management Leadership & Governance · doi:10.1080/23303131.2016.1140101 · Cited by 37
“The number of observations per predictor variable was 31, which exceeded the recommendation of 10 observations per variable for logistic regression (Hosmer & Lemeshow, 2000; Peng, Lee, & Ingersoll, 2002; Vittinghoff & McCulloch, 2007).”
MethodSonoelastography of Musculoskeletal Soft Tissue Masses
Pass, B., Johnson, M., Hensor, E. M. A., Gupta, H., & Robinson, P. J. (2016) · Journal of Ultrasound in Medicine · doi:10.7863/ultra.15.11065 · Cited by 25
Dash Cam videos on YouTube™ offer insights into factors related to moose-vehicle collisions
Rea, R. V., Johnson, C. J., Aitken, D. A., Child, K. N., & Hesse, G. (2018) · Accident Analysis & Prevention · doi:10.1016/j.aap.2018.02.020 · Cited by 25
Resistin is a prognostic factor for death in type 2 diabetes
Kapłon‐Cieślicka, A., Tymińska, A., Rosiak, M., Ozierański, K., Peller, M., Eyileten, C., Kondracka, A., Pordzik, J., Mirowska-Guzel, D. M., Opolski, G., et al. (2018) · Diabetes/Metabolism Research and Reviews · doi:10.1002/dmrr.3098 · Cited by 25
“An adequate event per predictor variable (EPV) value was maintained.(11) For all tests, p < 0.”
MethodCui, S., Luo, Y., Tseng, H.-H., Haken, R. K. T., & Naqa, I. E. (2018) · IEEE Transactions on Radiation and Plasma Medical Sciences · doi:10.1109/trpms.2018.2884134 · Cited by 26
“As a result, an elaborate cross-validation scheme was implemented, where the performance of models is tested outside of the training data for relatively fair evaluation of generalization [30].”
Brown, E., Tohira, H., Bailey, P. M., Fatovich, D. M., Pereira, G. F., & Finn, J. (2019) · Emergency Medicine Australasia · doi:10.1111/1742-6723.13244 · Cited by 25
“In-hospital mortality 38 (24) 63 (29) 69 (35) 0.”
A risk score for predicting respiratory complications after thoracic surgery
Pipanmekaporn, T., Bunchungmongkol, N., Punjasawadwong, Y., Lapisatepun, W., Tantraworasin, A., & Saokaew, S. (2019) · Asian Cardiovascular and Thoracic Annals · doi:10.1177/0218492319835994 · Cited by 28
Byambasukh, O., Osté, M. C. J., Neto, A. W. G., van den Berg, E., Navis, G. J., Bakker, S. J. L., & Corpeleijn, E. (2020) · Journal of Clinical Medicine · doi:10.3390/jcm9020415 · Cited by 25
“This is now fulfilled for all analyses [34,35].”
Kneale, D., Thomas, J., & French, R. (2020) · The Journals of Gerontology Series B · doi:10.1093/geronb/gbaa071 · Cited by 33
Magel, J. S., Kim, J., Fritz, J. M., & Freburger, J. K. (2020) · Physical Therapy · doi:10.1093/ptj/pzaa100 · Cited by 27
Blood Transfusion and Adverse Graft-related Events in Kidney Transplant Patients
Massicotte‐Azarniouch, D., Sood, M. M., Fergusson, D. A., Chassé, M., Tinmouth, A. T., & Knoll, G. A. (2021) · Kidney International Reports · doi:10.1016/j.ekir.2021.01.015 · Cited by 27
Impact of visceral fat area in patients with chronic kidney disease
Manabe, S., Kataoka, H., Mochizuki, T., Iwadoh, K., Ushio, Y., Kawachi, K., Watanabe, K., Watanabe, S., Akihisa, T., Makabe, S., et al. (2021) · Clinical and Experimental Nephrology · doi:10.1007/s10157-021-02029-4 · Cited by 30
Evaluating how lethal management affects poaching of Mexican wolves
Louchouarn, N. X., Santiago‐Ávila, F. J., Parsons, D. R., & Treves, A. (2021) · Royal Society Open Science · doi:10.1098/rsos.200330 · Cited by 27
Jiang, M., Li, C.-L., Chen, R., Tang, S.-C., Lv, W., Luo, X., Chuan, Z.-R., Jin, C.-Y., Liao, J., Cui, X., et al. (2021) · European Journal of Radiology · doi:10.1016/j.ejrad.2021.109781 · Cited by 26
Muro, A., Feliu‐Soler, A., & Castellà, J. (2021) · Women & Health · doi:10.1080/03630242.2021.1954133 · Cited by 31
Mason, A. E., Kasl, P., Hartogensis, W., Natale, J. L., Dilchert, S., Dasgupta, S., Purawat, S., Chowdhary, A., Anglo, C., Veasna, D., et al. (2022) · Vaccines · doi:10.3390/vaccines10020264 · Cited by 27
“Due to the proportion of observations with right-censored values of the outcome variable, we adopted a semi-parametric approach using Cox regression models [21], using the RBD antibody result as the dependent variable.”
MethodCan Perceived Support for Entrepreneurship Keep Great Faculty in the Face of Spinouts?
Nicolaou, N., & Souitaris, V. (2015) · Journal of Product Innovation Management · doi:10.1111/jpim.12274 · Cited by 30
Associations between a highly invasive species and native macrophytes differ across spatial scales
Thomaz, S. M., & Michelan, T. S. (2011) · Biological Invasions · doi:10.1007/s10530-011-0008-9 · Cited by 26
“According to Vittinghoff and McCulloch (2006), the probability of type I error and the relative bias that is found in 10–16 events are almost the same as found in 5–9 events, and the coverage and bias will fall within acceptable levels.”
BackgroundKhanna, R. R., Vittinghoff, E., Maselli, J. H., & Auerbach, A. D. (2011) · Journal of General Internal Medicine · doi:10.1007/s11606-011-1871-x · Cited by 28
Self-citationMark, D. G., Morehouse, J. W., Hung, Y., Kene, M. V., Elms, A. R., Liu, V. X., Ballard, D. W., & Vinson, D. R. (2014) · Critical Care · doi:10.1186/s13054-014-0496-y · Cited by 26
“All final regression models have a minimum of five event outcomes per independent variable with at least thirty total events [46].”
BackgroundCorrelates of nutritional behavior in individuals with multiple sclerosis
Plow, M. A., Finlayson, M. L., & Cho, C. (2012) · Disability and health journal · doi:10.1016/j.dhjo.2012.05.007 · Cited by 28
Factors affecting mortality of hospitalized chest trauma patients in United Arab Emirates
Aleassa, E. M., Al-Marashda, M. J., El-Sherif, A., Eid, H. O., & Abu‐Zidan, F. M. (2013) · Journal of Cardiothoracic Surgery · doi:10.1186/1749-8090-8-57 · Cited by 25
“Nevertheless, other methodologists [22] have recently argued that “the rule of ten” should be relaxed.”
BackgroundBaldelli, I., Cardoni, G., Franchelli, S., Fregatti, P., Friedman, D., Pesce, M., Ponte, E., Santori, G., & Santi, P. (2016) · Plastic & Reconstructive Surgery · doi:10.1097/prs.0000000000002180 · Cited by 25
Estimating Effects with Rare Outcomes and High Dimensional Covariates: Knowledge is Power
Balzer, L. B., Ahern, J., Galea, S., & van der Laan, M. J. (2016) · Epidemiologic Methods · doi:10.1515/em-2014-0020 · Cited by 31
Rodrigues, J. N., Zhang, W., Scammell, B. E., Chakrabarti, I., Russell, P. G., Fullilove, S., Davidson, D., & Davis, T. R. C. (2016) · Journal of Hand Surgery (European Volume) · doi:10.1177/1753193416660045 · Cited by 29
“However, more recent examination of this rule has suggested that five to nine outcome events per predictor variable may be acceptable (Vittinghoff and McCulloch, 2007), in which case 60–108 poor functional outcomes would be needed.”
BackgroundLim, S., Carabini, L. M., Kim, R. B., Khanna, R., Dahdaleh, N. S., & Smith, Z. A. (2016) · The Spine Journal · doi:10.1016/j.spinee.2016.09.018 · Cited by 27
Portable mTBI Assessment Using Temporal and Frequency Analysis of Speech
Daudet, L., Yadav, N., Perez, M. C., Poellabauer, C., Schneider, S. L., & Huebner, A. R. (2016) · IEEE Journal of Biomedical and Health Informatics · doi:10.1109/jbhi.2016.2633509 · Cited by 25
Collin, Y., Paré, A., Belblidia, A., Letourneau, R., Plasse, M., Dagenais, M., Turcotte, S., Martel, G., Roy, A., Lapointe, R. W., et al. (2018) · International Journal of Surgery · doi:10.1016/j.ijsu.2018.11.029 · Cited by 26
Prediction Model for Nodal Disease Among Patients With Non–Small Cell Lung Cancer
Verdial, F. C., Madtes, D. K., Hwang, B., Mulligan, M. S., Odem‐Davis, K., Waworuntu, R. L., Wood, D. E., & Farjah, F. (2019) · The Annals of Thoracic Surgery · doi:10.1016/j.athoracsur.2018.12.041 · Cited by 25
“Based on an empirically-supported heuristic for powering regression analyses[23], we constructed a logistic-regression model to estimate the probability of nodal disease and limited it to six degrees-of-freedom to mitigate overfitting.”
MethodFriedl, N., Berger, T., Krieger, T., Caspar, F., & Holtforth, M. G. (2019) · Psychotherapy Research · doi:10.1080/10503307.2019.1664782 · Cited by 29
“A commonly used rule of thumb states that at least 10 observations per predictor are necessary to not exceed an acceptable level of bias (Peduzzi, Concato, Feinstein, & Holford, 1995; Vittinghoff & McCulloch, 2007).”
Mumbwe, M. C., McIsaac, D., Jarman, A., & Bould, M. D. (2019) · Anesthesia & Analgesia · doi:10.1213/ane.0000000000004464 · Cited by 26
Garnier, F., Daubin, D., Larcher, R., Bargnoux, A., Platon, L., Brunot, V., Aarab, Y., Besnard, N., Dupuy, A., Jung, B., et al. (2020) · Critical Care Medicine · doi:10.1097/ccm.0000000000004218 · Cited by 26
Two-hour glucose predicts the development of hypertension over 5 years: the AusDiab study
Boyko, E. J., Barr, E., Zimmet, P., & Shaw, J. E. (2007) · Journal of Human Hypertension · doi:10.1038/sj.jhh.1002316 · Cited by 27
Prevalence and Predictors of Osteoporosis Risk in Orthopaedic Patients
Rozental, T. D., Shah, J. A., Chacko, A. T., & Zurakowski, D. (2009) · Clinical Orthopaedics and Related Research · doi:10.1007/s11999-009-1162-6 · Cited by 26
“This allowed us to test 15 variables, which was appropriate given the number of patients with abnormally low QUS values based on our T-score definition [34].”
How much running is too much? Identifying high-risk running sessions in a 5200-person cohort study
Schuster Brandt Frandsen, J., Hulme, A., Parner, E. T., Møller, M., Lindman, I., Abrahamson, J., Sjørup Simonsen, N., Sandell Jacobsen, J., Ramskov, D., Skejø, S., et al. (2025) · British Journal of Sports Medicine · doi:10.1136/bjsports-2024-109380 · Cited by 25
Surawattanasakul, V., Siviroj, P., & Kiratipaisarl, W. (2024) · PLoS ONE · doi:10.1371/journal.pone.0312839 · Cited by 38
“…potential for a high number of determinant variables compared to the sample size, which could be unsuitable for regression analyses, with potentially fewer than 10 burnout cases per variable with 24 degrees of freedom [27–29], the investigators chose the confounder summary score concept [30, 31].”
BackgroundAge, climate and economic disparities drive the current state of global dam safety
Moreno-Rodenas, A., Mantilla-Jones, J. D., & Valero, D. (2025) · Nature Water · doi:10.1038/s44221-025-00402-1 · Cited by 31
“All models where fitted with at least 30–60 events per fitted variable, complying with the recommended >10–20 ratio for survival analysis 91,92 .”
Zhang, Y., Banks, C. J., & Heaven, S. (2012) · Cited by 24
Maternal and neonatal outcomes in obese women who lose weight during pregnancy
Bauer, C. M. C., Bernhard, K. A., Greer, D. M., & Merrill, D. (2016) · Journal of Perinatology · doi:10.1038/jp.2015.202 · Cited by 27
Nair, S. C., Welsing, P. M. J., Choi, I. Y. K., Roth, J., Holzinger, D., Bijlsma, J. W. J., van Laar, J. M., Gerlag, D. M., Lafeber, F. P. J. G., & Tak, P. P. (2016) · PLoS ONE · doi:10.1371/journal.pone.0152362 · Cited by 24
Margini, C., Murgia, G., Stirnimann, G., Gottardi, A. D., Semmo, N., Casu, S., Bosch, J., Dufour, J., & Berzigotti, A. (2018) · Hepatology Communications · doi:10.1002/hep4.1201 · Cited by 26
“Alanine aminotransferase (U/L) 63 (7-296) 44 (15-204) 66 (7-296) 0.”
MethodInfluentialMonroy‐Gonzalez, A. G., Tio, R. A., de Groot, J. C., Boersma, H. H., Prakken, N. H. J., de Jongste, M. J. L., Alexánderson, E., & Slart, R. H. J. A. (2018) · Journal of Nuclear Cardiology · doi:10.1007/s12350-018-1448-8 · Cited by 24
Postoperative oscillatory brain activity as an add-on prognostic marker in diffuse glioma
Belgers, V., Numan, T., Kulik, S. D., Hillebrand, A., Hamer, P. C. D. W., Geurts, J. J. G., Reijneveld, J. C., Wesseling, P., Klein, M., Derks, J., et al. (2020) · Journal of Neuro-Oncology · doi:10.1007/s11060-019-03386-7 · Cited by 26
Schall, D. L. (2020) · Cogent Economics & Finance · doi:10.1080/23322039.2020.1777813 · Cited by 25
“…estimations with samples larger than 200 observations (Schaefer, 1983) and a minimum of 10 outcome events per predictor variable (or even less; see Vittinghoff & McCulloch, 2007) should generally not suffer significantly from small sample bias or encounter other problems (such as unreliable…”
Background
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.