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 401–500 of 1,018
Periodontal disease in patients with ankylosing spondylitis
Pischon, N., Pischon, T., Gülmez, E., Kröger, J., Purucker, P., Kleber, B. M., Landau, H., Jost-Brinkmann, P.-G., Schlattmann, P., Zernicke, J., et al. (2009) · Annals of the Rheumatic Diseases · doi:10.1136/ard.2008.097212 · Cited by 54
Conservative treatment of mallet finger: A systematic review
Valdes, K. A., Naughton, N., & Algar, L. A. (2015) · Journal of Hand Therapy · doi:10.1016/j.jht.2015.03.001 · Cited by 59
Canine Perivascular Wall Tumors
Avallone, G., Boracchi, P., Stefanello, D., Ferrari, R., Rebughini, A., & Roccabianca, P. (2013) · Veterinary Pathology · doi:10.1177/0300985813503565 · Cited by 54
“The adequate ratio was suggested to be 10(18,31) or, more recently, 5.(34) When the incidence of the events does not allow the investigation of the joint prognostic role of several variables, as occurred for this caseload, a suggested strategy is to reduce the number of variables in the regression model by data reduct…”
BackgroundMethodLott, J. P., Wititsuwannakul, J., Lee, J., Ariyan, S., Narayan, D., Kluger, H. H., & Lazova, R. Z. (2014) · Journal of the American Academy of Dermatology · doi:10.1016/j.jaad.2014.08.026 · Cited by 54
Too many covariates and too few cases? – a comparative study
Chen, Q., Nian, H., Zhu, Y., Talbot, H. K., Griffin, M. R., & Harrell, F. E. (2016) · Statistics in Medicine · doi:10.1002/sim.7021 · Cited by 55
Forslund, E. B., Jørgensen, V., Franzén, E., Opheim, A. I., Seiger, �., Ståhle, A., Hultling, C., Stanghelle, J. K., Roaldsen, K. S., & Wahman, K. (2017) · Journal of Rehabilitation Medicine · doi:10.2340/16501977-2177 · Cited by 54
Milk Metabolomics Data Reveal the Energy Balance of Individual Dairy Cows in Early Lactation
Xu, W., Vervoort, J. J. M., Saccenti, E., van Hoeij, R. J., Kemp, B., & van Knegsel, A. T. M. (2018) · Scientific Reports · doi:10.1038/s41598-018-34190-4 · Cited by 66
At odds with Europe: explaining populist radical right voting in Central and Eastern Europe
Santana, A., Zagórski, P., & Rama, J. (2020) · East European Politics · doi:10.1080/21599165.2020.1737523 · Cited by 73
Alhakak, A. S., Biering-Sørensen, S. R., Møgelvang, R., Modin, D., Jensen, G. B., Schnohr, P., Iversen, A. Z., Svendsen, J. H., Jespersen, T., Gislason, G., et al. (2020) · European Heart Journal - Cardiovascular Imaging · doi:10.1093/ehjci/jeaa287 · Cited by 56
Rosanò, A., Martinelli, E., Fusina, F., Albani, F., Caserta, R., Morandi, A., Dell’Agnolo, P., Dicembrini, A., Mansouri, L., Marchini, A., et al. (2020) · Critical Care Medicine · doi:10.1097/ccm.0000000000004752 · Cited by 57
Noguchi, Y., Asayama, K., Staessen, J. A., Inaba, M., Ohkubo, T., Hosaka, M., Satoh, M., Kamide, K., Awata, T., Katayama, S., et al. (2013) · Journal of Hypertension · doi:10.1097/hjh.0b013e328361732c · Cited by 53
“In general, Cox models should be used with at least 10 events per variable, and is acceptable with 5–9 events per variable [46].”
Toward A Simple Diagnostic Index for Acute Uncomplicated Urinary Tract Infections
Knottnerus, B. J., Geerlings, S. E., van Charante, E. P. M., & ter Riet, G. (2013) · The Annals of Family Medicine · doi:10.1370/afm.1513 · Cited by 61
“0 18 (9) 12 6 67 1 40 (20) 28 12 70 2 37 (19) 23 14 62 3 37 (19) 22 15 59 4 24 (12) 14 10 58 ≥5 40 (20) 21 19 53 False urge to urinate No 54 (28) 30 24 57 A little 82 (42) 53 29 65 Quite much 43 (22) 27 16 63 Very much 17 (9) 10 7 59 More frequent micturition than usual No 12 (6) 3 9 25 A little 62 (32) 31 31 50 Quite…”
Koyanagi, A., Stickley, A., & Haro, J. M. (2015) · Schizophrenia Research · doi:10.1016/j.schres.2015.07.041 · Cited by 56
“Since the number of individuals who attempted suicide in the previous year was small, the number of control variables for Model 3 was restricted only when suicide attempt was the outcome to avoid small-sample bias (Vittinghoff and McCulloch, 2007).”
BackgroundDiet Quality and Mortality Risk in Metabolically Obese Normal-Weight Adults
Park, Y., Fung, T. T., Steck, S. E., Zhang, J., Hazlett, L. J., Han, K., Lee, S., & Merchant, A. T. (2016) · Mayo Clinic Proceedings · doi:10.1016/j.mayocp.2016.06.022 · Cited by 57
Sandini, M., Ruscic, K. J., Ferrone, C. R., Warshaw, A. L., Qadan, M., Eikermann, M., Lillemoe, K. D., & Castillo, C. F. F.-D. (2018) · Annals of Surgical Oncology · doi:10.1245/s10434-018-6827-5 · Cited by 57
Belser, C. T., Shillingford, M. A., Daire, A. P., Prescod, D. J., & Dagley, M. A. (2018) · The Professional Counselor · doi:10.15241/ctb.8.3.262 · Cited by 70
“However, Field (2009) and Vittinghoff and McCulloch (2006) recommended having a minimum of five cases per outcome per predictor, which the sample achieved for all independent variables.”
BackgroundJohnson, J. K., Lapin, B. R., Green, K., & Stilphen, M. (2020) · Physical Therapy · doi:10.1093/ptj/pzaa181 · Cited by 58
Bartko, P. E., Heinze, G., Graf, S., Clavel, M.-A., Khorsand, A., Bergler-Klein, J., Burwash, I. G., Dumesnil, J. G., Sénéchal, M., Baumgartner, H., et al. (2013) · Circulation: Cardiovascular Imaging · doi:10.1161/circimaging.112.980201 · Cited by 52
“For these models, we used bias-corrected multivariable Cox regression to account for the limited number of events per variable.(26,27) The incremental value of these models was evaluated using likelihood ratio tests.”
MethodNicolazzo, C., Busetto, G. M., Gradilone, A., Sperduti, I., Giudice, F. D., Loreni, F., Cortesi, E., BERARDINIS, E. D., Gazzaniga, P., & Raimondi, C. (2019) · The Oncologist · doi:10.1634/theoncologist.2018-0784 · Cited by 52
“Points of attention in this multivariable analysis are the long-term follow-up and the adequate number of outcome events [10].”
BackgroundCandel, B. G. J., Duijzer, R., Gaakeer, M. I., ter Avest, E., Sir, Ö., Lameijer, H., Hessels, R. A. P. A., Reijnen, R., van Zwet, E. W., de Jonge, E., et al. (2022) · Emergency Medicine Journal · doi:10.1136/emermed-2020-210628 · Cited by 55
Machine Learning–Based Prediction Models for Delirium: A Systematic Review and Meta-Analysis
Xie, Q., Wang, X., Pei, J., Wu, Y., Guo, Q., Su, Y., Yan, H., Nan, R., Chen, H., & Dou, X. (2022) · Journal of the American Medical Directors Association · doi:10.1016/j.jamda.2022.06.020 · Cited by 58
Myaskovsky, L., Posluszny, D. M., Schulz, R., DiMartini, A. F., Switzer, G. E., Dabbs, A. D., McNulty, M. L., Kormos, R. L., Toyoda, Y., & Dew, M. A. (2012) · American Journal of Transplantation · doi:10.1111/j.1600-6143.2012.04243.x · Cited by 56
“Second, we had to limit the number of predictors included because we did not have a sufficient ratio of events to predictor variables to examine all possible predictors (60,61).”
BackgroundDoes occiput posterior position in the second stage of labour increase the operative delivery rate?
Carseldine, W. J., Phipps, H., Zawada, S. F., Campbell, N. T., Ludlow, J., Krishnan, S., & de Vries, B. S. (2013) · Australian and New Zealand Journal of Obstetrics and Gynaecology · doi:10.1111/ajo.12041 · Cited by 52
Chemotherapy use and patient treatment preferences in advanced colorectal cancer
Zafar, S. Y., Malin, J. L., Grambow, S. C., Abbott, D. H., Kolimaga, J. T., Zullig, L. L., Weeks, J. C., Ayanian, J. Z., Kahn, K. L., Ganz, P. A., et al. (2012) · Cancer · doi:10.1002/cncr.27815 · Cited by 52
“If necessary to prevent over fitting, the least significant variables were removed to attain model degrees of freedom to an effective sample size ratio of 7.(13) Multiple imputation was used to address item nonresponse for survey-based variables and was performed centrally by the CanCORS Statistical Coordinating Cente…”
MethodHunt, J. E., Talling, P. J., Clare, M. A., Jarvis, I., & Wynn, R. B. (2014) · Geochemistry Geophysics Geosystems · doi:10.1002/2014gc005232 · Cited by 55
Leroy, S., Romanello, C., Galetto-Lacour, A., Bouissou, F., Fernàndez-López, A., Smolkin, V., Gurgoz, M. K., Bressan, S., Karavanaki, K., Tuerlinckx, D., et al. (2011) · The Journal of Pediatrics · doi:10.1016/j.jpeds.2011.03.008 · Cited by 56
Seventeen years of adult congenital heart surgery: a single centre experience☆
Putman, L. M., van Gameren, M., Meijboom, F. J., de Jong, P. L., Roos‐Hesselink, J. W., Witsenburg, M., Takkenberg, J. J. M., & Bogers, A. J. J. C. (2009) · European Journal of Cardio-Thoracic Surgery · doi:10.1016/j.ejcts.2009.01.046 · Cited by 54
“For multivariate analysis backward conditional stepping and 5—9 events per variable, as proposed by Vittinghoff and McCulloch were used [19].”
MethodPredictors of Response to Cognitive-Behavioral Therapy for Body Dysmorphic Disorder
Greenberg, J. L., Phillips, K. A., Steketee, G. S., Hoeppner, S. S., & Wilhelm, S. (2019) · Behavior Therapy · doi:10.1016/j.beth.2018.12.008 · Cited by 58
“We then combined the significant univariate predictors identified into multiple logistic regression models with no more than 2 predictors to stay within the recommended lower limit of 5–9 events per variable that is still likely to produce adequate model performance (Vittinghoff and McCulloch, 2007).”
MethodSources of variation in weaned age among wild chimpanzees in Gombe National Park, Tanzania
Lonsdorf, E. V., Stanton, M. A., Pusey, A. E., & Murray, C. M. (2019) · American Journal of Physical Anthropology · doi:10.1002/ajpa.23986 · Cited by 58
Challenges Raised by Mediation Analysis in a High-Dimension Setting
Blum, M. G. B., Valeri, L., François, O., Cadiou, S., Siroux, V., Lepeule, J., & Slama, R. (2020) · Environmental Health Perspectives · doi:10.1289/ehp6240 · Cited by 58
“…logistic regression) estimators, the exposure–outcome regression model including all potential mediators (Equation 4) provides estimation of regression parameters with a prohibitively large variance as p increases and reaches a fraction of n (Sur and Candès 2019; Vittinghoff and McCulloch 2007).”
BackgroundChanges in experienced value of everyday occupations after nature-based vocational rehabilitation
Pálsdóttir, A. M., Grahn, P., & Persson, D. (2013) · Scandinavian Journal of Occupational Therapy · doi:10.3109/11038128.2013.832794 · Cited by 104
“All variables of significance for the aim of the study were included in the models in order to find the best model (54,55).”
Fracture Risk Assessment: State of the Art, Methodologically Unsound, or Poorly Reported?
Collins, G. S., & Michaëlsson, K. (2012) · Current Osteoporosis Reports · doi:10.1007/s11914-012-0108-1 · Cited by 51
Statistical primer: checking model assumptions with regression diagnostics†
Hickey, G. L., Kontopantelis, E., Takkenberg, J. J. M., & Beyersdorf, F. (2018) · Interactive Cardiovascular and Thoracic Surgery · doi:10.1093/icvts/ivy207 · Cited by 75
Invernizzi, M., Corti, C., Lopez, G., Michelotti, A., Despini, L., Gambini, D., Lorenzini, D., Guerini‐Rocco, E., Maggi, S., Noale, M., et al. (2018) · BMC Cancer · doi:10.1186/s12885-018-4851-2 · Cited by 55
“In the Cox multivariable model employed, the rule of 10 events per factors was relaxed, as previously described [39].”
Chi, N., Hu, H., Wang, C.-Y., Chan, L., Peng, C., Novak, V., & Hu, C. (2018) · Stroke · doi:10.1161/strokeaha.118.022481 · Cited by 54
Myocardial Scar and Mortality in Chronic Aortic Regurgitation
Malahfji, M., Senapati, A., Tayal, B., Nguyen, D. T., Graviss, E. A., Nagueh, S. F., Reardon, M. J., Quiñones, M. A., Zoghbi, W. A., & Shah, D. J. (2020) · Journal of the American Heart Association · doi:10.1161/jaha.120.018731 · Cited by 54
“The AVR group had a higher RVol (53 [38–80] versus 32 [26–42]; P <0.001), higher RF (43% [35–51] versus 35% [31–40]; P <0.001), and greater left ventricular remodeling.”
InfluentialWilliams, M. W., & Williams, M. N. (2012) · Ethics & Behavior · doi:10.1080/10508422.2011.653291 · Cited by 58
“A minimum of 5 events per variable was indicated as appropriate for logistic regression in the simulation study of Vittinghoff and McCulloch (2007). Bootstrapping bias-corrected confidence intervals across 1000 (re)samples was therefore used to confirm the robustness of relationships; the risk-seeking (95% bias-correc…”
MethodAssessment and Analysis of Workplace Violence in a Greek Tertiary Hospital
Mantzouranis, G., Fafliora, E., Bampalis, V. G., & Christopoulou, I. (2014) · Archives of Environmental & Occupational Health · doi:10.1080/19338244.2013.879564 · Cited by 51
Clinical relevance of ErbB-2/HER2 nuclear expression in breast cancer
Schillaci, R., Guzmán, P., Cayrol, F., Béguelin, W., Flaqué, M. C. D., Proietti, C. J., Pineda, V., Palazzi, J., Frahm, I., Charreau, E. H., et al. (2012) · BMC Cancer · doi:10.1186/1471-2407-12-74 · Cited by 51
“Adjustment for significant confounders was done to avoid increased bias and variability, unreliable confidence interval coverage, and problems with the model associated to the small size of our sample [16].”
BackgroundStiell, I. G., Clement, C. M., Aaron, S. D., Rowe, B. H., Perry, J. J., Brison, R. J., Calder, L. A., Lang, E., Borgundvaag, B., Forster, A. J., et al. (2014) · Canadian Medical Association Journal · doi:10.1503/cmaj.130968 · Cited by 59
“nal validation with bootstrapping showed that the estimate and classification performance for risk scores were stable.(36) We chose not to adjust the analysis, despite the fact that some patients were enrolled more than once.”
Methodde Vries, N. K. S., van der Veere, C. N., Reijneveld, S. A., & Bos, A. F. (2013) · PLoS ONE · doi:10.1371/journal.pone.0064654 · Cited by 50
“[19].”
Differences in international opportunity identification between native and immigrant entrepreneurs
Vinogradov, E., & Jørgensen, E. (2016) · Journal of International Entrepreneurship · doi:10.1007/s10843-016-0197-5 · Cited by 52
“When EPV is so much higher than 16 (Vittinghoff and McCulloch 2006), the differences between regressions with 116 or 864 cases are practically unobservable.”
BackgroundNovel digital signatures of tissue phenotypes for predicting distant metastasis in colorectal cancer
Sirinukunwattana, K., Snead, D. R. J., Epstein, D. B. A., Aftab, Z., Mujeeb, I., Tsang, Y. W., Cree, I. A., & Rajpoot, N. (2018) · Scientific Reports · doi:10.1038/s41598-018-31799-3 · Cited by 57
Prevalence of suicidal behaviour among nursing and midwifery college students in Ghana
Quarshie, E. N., Cheataa-Plange, H. V., Annor, F., Asare‐Doku, W., & Lartey, J. K. S. (2019) · Nursing Open · doi:10.1002/nop2.271 · Cited by 58
“Thus, to avoid an unstable logistic regression model due to the sparse data problem, we applied the relaxed rule of at least five events per variable (Vittinghoff & McCulloch, 2007) to select the socio‐demographic variables included as correlates in the final logistic regression model.”
MethodGilmore, S. J., Hahne, A. J., Davidson, M., & McClelland, J. A. (2019) · BMC Musculoskeletal Disorders · doi:10.1186/s12891-019-2806-7 · Cited by 54
Merckx, J., Cooke, S., Tal, T. E., Laxer, R. M., Bitnun, A., Morris, S. K., Yeh, E. A., Yea, C., Gill, P., Papenburg, J., et al. (2022) · Canadian Medical Association Journal · doi:10.1503/cmaj.210873 · Cited by 51
Computed Tomography and Echocardiography in Patients With Acute Pulmonary Embolism
George, E., Kumamaru, K. K., Ghosh, N., Quesada, C. G., Wake, N., Bedayat, A., Dunne, R. M., Saboo, S. S., Khandelwal, A., Hunsaker, A. R., et al. (2013) · Journal of Thoracic Imaging · doi:10.1097/rti.0000000000000048 · Cited by 51
Association Between Serum Cholesterol and Noncardiovascular Mortality in Older Age
Newson, R. S., Felix, J. F., Heeringa, J., Hofman, A., Witteman, J. C. M., & Tiemeier, H. W. (2011) · Journal of the American Geriatrics Society · doi:10.1111/j.1532-5415.2011.03593.x · Cited by 53
Causal effect of low-income areas on shared dockless e-scooter use
Frías-Martínez, V., Sloate, E., Manglunia, H., & Wu, J. (2021) · Transportation Research Part D Transport and Environment · doi:10.1016/j.trd.2021.103038 · Cited by 50
Risk factors associated with endometriosis among infertile Iranian women
Moini, A., Malekzadeh, F., Amirchaghmaghi, E., Kashfi, F., Akhoond, M. R., Saei, M., & Mirbolok, M. H. (2013) · Archives of Medical Science · doi:10.5114/aoms.2013.35420 · Cited by 58
Coagulase-negative staphylococcus sepsis in preterm infants and long-term neurodevelopmental outcome
Alshaikh, B. N., Yee, W., Lodha, A. K., Henderson, E., Yusuf, K., & SAUVE, R. S. (2013) · Journal of Perinatology · doi:10.1038/jp.2013.155 · Cited by 57
“94 In this cohort only 60 infants had major NDI, limiting the study to 5 variables in addition to the exposure to CoNS sepsis under the rule of 10 events for each variable in the multivariable logistic regression analysis.[110, 111] Of all the covariates found in our selection approach,”
Remission rate of implant-related infections following revision surgery after fractures
Al‐Mayahi, M., Betz, M., Müller, D. A., STERN, R. E., Tahintzi, P., Bernard, L., Hoffmeyer, P. J., Suvà, D., & Uçkay, İ. (2013) · International Orthopaedics · doi:10.1007/s00264-013-2092-1 · Cited by 50
“We included five to ten outcome events per predictor variable [15].”
BackgroundWhiting, S. W., Potenza, M. N., Park, C. L., McKee, S. A., Mazure, C. M., & Hoff, R. A. (2016) · Journal of Behavioral Addictions · doi:10.1556/2006.5.2016.027 · Cited by 57
Network Analysis to Risk Stratify Patients With Exercise Intolerance
Oldham, W. M., Oliveira, R. K. F., Wang, R., Opotowsky, A. R., Rubins, D. M., Hainer, J. M., Wertheim, B. M., Alba, G. A., Choudhary, G., Tornyos, A., et al. (2018) · Circulation Research · doi:10.1161/circresaha.117.312482 · Cited by 49
Fitzpatrick, K. E., Kurinczuk, J. J., Bhattacharya, S., & QUIGLEY, M. A. (2019) · PLoS Medicine · doi:10.1371/journal.pmed.1002913 · Cited by 50
“The above adjustments were only performed when there was a minimum of 5–9 outcome events per coefficient in the model [25,26].”
BackgroundCare complexity individual factors associated with adverse events and in-hospital mortality
Adamuz, J., Juvé‐Udina, M., González‐Samartino, M., Jiménez‐Martínez, E., Tàpia‐Pérez, M., López‐Jiménez, M., Romero‐García, M., & Delgado‐Hito, P. (2020) · PLoS ONE · doi:10.1371/journal.pone.0236370 · Cited by 53
“The number of variables included in the multivariate analysis was restricted following the rule that there should be at least ten events per variable [25].”
MethodRuud, E. A., Stavem, K., Geitung, J. T., Borthne, A., Søyseth, V., & Ashraf, H. (2020) · European Radiology · doi:10.1007/s00330-020-07449-6 · Cited by 49
“In some cases, the pigtail catheter was left in the pleural cavity when the patient was returned [24] to the referring department.”
BackgroundMethodGait Retraining and Incidence of Medial Tibial Stress Syndrome in Army Recruits
Sharma, J., Weston, M., Batterham, A. M., & Spears, I. R. (2014) · Medicine & Science in Sports & Exercise · doi:10.1249/mss.0000000000000290 · Cited by 62
“Due to a low ratio of events to predictor variables, confidence intervals for the hazard ratio were obtained using a bias-corrected and accelerated bootstrap resampling method (5000 resamples with replacement) (38).”
BackgroundMethodInfluentialGriesdale, D. E. G., Sekhon, M. S., Menon, D. K., Lavinio, A., Donnelly, J., Robba, C., Sekhon, I. S., Taylor, A., Henderson, W. R., Turgeon, A. F., et al. (2014) · Neurocritical Care · doi:10.1007/s12028-014-0096-3 · Cited by 49
Perez, C., Huttner, A., Assal, M., Bernard, L., Lew, D., Hoffmeyer, P. J., & Uçkay, İ. (2010) · Journal of Antimicrobial Chemotherapy · doi:10.1093/jac/dkq043 · Cited by 48
Hay, A. D., Sterne, J. A. C., Hood, K., Little, P. S., Delaney, B. C., Hollingworth, W., Wootton, M., Howe, R. A., MacGowan, A. P., Lawton, M., et al. (2016) · The Annals of Family Medicine · doi:10.1370/afm.1954 · Cited by 59
Carter, A. R., McAvoy, M. P., Siegel, J. S., Hong, X., Astafiev, S. V., Rengachary, J., Zinn, K., Metcalf, N. V., Shulman, G. L., & Corbetta, M. (2016) · Cortex · doi:10.1016/j.cortex.2016.12.009 · Cited by 52
Imöhl, M., Fitzner, C., Perniciaro, S., & van der Linden, M. (2017) · PLoS ONE · doi:10.1371/journal.pone.0180757 · Cited by 49
Hevesi, M., Sanders, T. L., Pareek, A., Milbrandt, T. A., Levy, B. A., Stuart, M. J., Saris, D. B. F., & Krych, A. J. (2018) · Cartilage · doi:10.1177/1947603518786545 · Cited by 49
“Previous literature has suggested that approximately 5 to 10 events are required per variable modeled for a reasonably well-powered analysis.(33,34) Additionally, because of the nature of the epidemiological project from which our data were obtained, the accuracy of diagnosis is dependent on accurate record keeping, a…”
BackgroundLee‐Jenkins, S. S. Y., & Robinson, S. A. (2018) · Environmental Toxicology and Chemistry · doi:10.1002/etc.4284 · Cited by 59
“Wemet the sample size recommendations of >10 observations/independent variable (Vittinghoff and Mcculloch 2007) for all treatments except 10mg/L thiamethoxam, and thus the results are presented and interpreted without that treatment.”
Backgroundvan Kranendonk, K. R., Treurniet, K. M., Boers, A. M. M., Berkhemer, O. A., van den Berg, L. A., Chalos, V., Lingsma, H. F., van Zwam, W. H., van der Lugt, A., van Oostenbrugge, R. J., et al. (2019) · Stroke · doi:10.1161/strokeaha.118.024255 · Cited by 53
“Last, although we selected variables with care, it should be considered that some associations might have occurred by chance due to the large number of variables that were included in the analysis.(38) In conclusion, scores indicating severe strokes, such as poor collaterals and high NIHSS score, are associated with H…”
MethodBundles of Lean Automation practices and principles and their impact on operational performance
Tortorella, G. L., Saurin, T. A., Filho, M. G., Samson, D., & Kumar, M. (2021) · International Journal of Production Economics · doi:10.1016/j.ijpe.2021.108106 · Cited by 68
Predictors of relationship dissolution in lesbian, gay, and heterosexual adoptive parents.
Goldberg, A. E., & Garcia, R. (2015) · Journal of Family Psychology · doi:10.1037/fam0000095 · Cited by 53
“Because of a relatively small number of couples dissolving their unions (i.e., splitting up/divorcing) in our final sample (n 15), we minimized the number of predictors included in our final model (Vittinghoff & McCulloch, 2007) by trimming variables that did not approach statistical significance.”
MethodAustin, P. C. (2008) · Journal of Clinical Epidemiology · doi:10.1016/j.jclinepi.2007.11.014 · Cited by 51
Eid, H. O., & Abu‐Zidan, F. M. (2014) · World Journal of Surgery · doi:10.1007/s00268-014-2745-2 · Cited by 48
“In fact, however, others have recently argued that ‘‘the rule of ten’’ should be relaxed [26].”
BackgroundJalil, M. H. A., Hammour, K. A., Alsous, M. M., Awad, W. B., Hadadden, R., Bakri, F. G., & Fram, K. M. (2017) · Scientific Reports · doi:10.1038/s41598-017-12431-2 · Cited by 48
Values, Attitudes Toward Interpersonal Violence, and Interpersonal Violent Behavior
Seddig, D., & Davidov, E. (2018) · Frontiers in Psychology · doi:10.3389/fpsyg.2018.00604 · Cited by 81
“…and attitudes) and nearly 50 events (i.e., about 25 events per predictor) non-linear estimates are considered trustworthy (Peduzzi et al., 1996; Vittinghoff and McCulloch, 2007).14 14Analyzing these data as (Poisson or negative binomial) count variables is not applicable because very few…”
BackgroundFein, J. A., Shimoni, A., Labopin, M., Shem‐Tov, N., Yerushalmi, R., Magen, H., Furie, N., Kopel, E., Danylesko, I., Nagler, A., et al. (2018) · Leukemia · doi:10.1038/s41375-018-0185-y · Cited by 51
Elements of Palliative Care in the Last 6 Months of Life: Frequency, Predictors, and Timing
Ernecoff, N. C., Wessell, K. L., Hanson, L. C., Dusetzina, S. B., Shea, C. M., Weinberger, M., & Bennett, A. V. (2019) · Journal of General Internal Medicine · doi:10.1007/s11606-019-05349-0 · Cited by 50
Calderón‐Parra, J., Mills-Sánchez, P., Moreno‐Torres, V., Bravo, S. T., Romero‐Sánchez, I., Moreno, B. B., Calvo‐Salvador, M., Portero‐Azorín, F., Fernández, S. G.-M., Múñez, E., et al. (2022) · Mycoses · doi:10.1111/myc.13434 · Cited by 53
Impact of Galectin-3 Circulating Levels on Frailty in Elderly Patients with Systolic Heart Failure
Komici, K., Gnemmi, I., Bencivenga, L., Vitale, D. F., Rengo, G., Stefano, A. D., & Eleuteri, E. (2020) · Journal of Clinical Medicine · doi:10.3390/jcm9072229 · Cited by 47
“Taking into account the study sample size, the rule of “at least five observed events for each tested variable” [28] and parsimonious criteria, the following 10 factors were selected as possible confounders to test the possible association of Gal-3 levels with the probability of frailty presence in the logistic model;…”
MethodTan, C., Shojaei, E., Wiener, J., Shah, M., Koivu, S. L., & Silverman, M. (2020) · JAMA Network Open · doi:10.1001/jamanetworkopen.2020.12974 · Cited by 54
Goyal, R., Chen, H., Abughosh, S., Holmes, H. M., Candrilli, S. D., & Johnson, M. L. (2023) · Cancer · doi:10.1002/cncr.34675 · Cited by 52
“For the second line, a reduced model was built to avoid extreme estimates and CIs, a common occurrence in multivariable analyses involving a very small number of events per variable (typically if less than 10).(31,32) All analyses were performed using SAS statistical software, version 9.”
MethodTwelve-month psychosis-predictive value of the ultra-high risk criteria in children and adolescents
Armando, M., Pontillo, M., Crescenzo, F. D., Mazzone, L., Monducci, E., Cascio, N. L., Santonastaso, O., Pucciarini, M. L., Vicari, S., SCHIMMELMANN, B. G., et al. (2015) · Schizophrenia Research · doi:10.1016/j.schres.2015.10.033 · Cited by 58
“For descriptive purposes, notwithstanding the recommended 5:1 relation of number of events to number of predictors (Vittinghoff and McCulloch, 2007), potential predictors showing at least a trend-level result of p b 0.”
BackgroundMethodInfluentialKeyser, T. L., & Zarnoch, S. J. (2014) · Forest Ecology and Management · doi:10.1016/j.foreco.2014.01.045 · Cited by 47
“Adequate sample size was defined as those species where the number of observations was greater than five times the number of independent predictor variables (Vittinghoff and McCulloch, 2006).”
Circulating Inflammatory Mediators as Potential Prognostic Markers of Human Colorectal Cancer
Caro, G. D., Carvello, M., Pesce, S., Erreni, M., Marchesi, F., Todoric, J., Sacchi, M. M., Montorsi, M., Allavena, P., & Spinelli, A. (2016) · PLoS ONE · doi:10.1371/journal.pone.0148186 · Cited by 53
“To avoid overfitting of the COX model [21, 22] because of the relatively low amount of events (n = 9) in our cohort, each inflammatory mediator was separately entered in a multivariate model including TNM stage or one other inflammatory mediator (data not shown).”
BackgroundMethodvan den Bunt, G., Liakopoulos, A., Mevius, D. J., Geurts, Y., Fluit, A. C., Bonten, M. J. M., Mughini‐Gras, L., & van Pelt, W. (2016) · Journal of Antimicrobial Chemotherapy · doi:10.1093/jac/dkw443 · Cited by 61
Voorveld, H. A. M., Smit, E. G., Neijens, P. C., & Bronner, A. E. (2016) · Journal of Advertising Research · doi:10.2501/jar-2016-044 · Cited by 63
“Logistic regression analyses require a minimum of five to ten events per predictor (Vittinghoff and McCulloch, 2007).”
Paterson, K. L., Kasza, J., Hunter, D. J., Hinman, R. S., Menz, H. B., Peat, G., & Bennell, K. L. (2016) · Osteoarthritis and Cartilage · doi:10.1016/j.joca.2016.12.003 · Cited by 52
Duwe, G. (2018) · Journal of Experimental Criminology · doi:10.1007/s11292-018-9325-7 · Cited by 80
Determinants of Dropout From Correctional Offender Treatment
Brünner, F., Neumann, I., Yoon, D., Rettenberger, M., Stück, E., & Briken, P. (2019) · Frontiers in Psychiatry · doi:10.3389/fpsyt.2019.00142 · Cited by 58
“In a second data-driven approach, the model with the best fit was identified via stepwise backward elimination per likelihoodratio-test (48).”
MethodRisk of reoperative valve surgery for endocarditis associated with drug use
Mori, M., Mahmood, S. U. B., Schranz, A. J., Sultan, I. S., Axtell, A. L., Sarsour, N. Y., Hiesinger, W., Boskovski, M. T., Hirji, S. A., Kaneko, T., et al. (2019) · Journal of Thoracic and Cardiovascular Surgery · doi:10.1016/j.jtcvs.2019.06.055 · Cited by 49
Sun, Y., Wu, G., Cheng, K. S., Chen, A., Neoh, K. H., Chen, S., Tang, Z., Lee, P. F., Dai, M., & Han, R. P. S. (2019) · EBioMedicine · doi:10.1016/j.ebiom.2019.07.044 · Cited by 48
“model in order to ensure the validity of themodel by satisfying theminimum of 5 events per variable rule of thumb [29].”
MethodFoody, J. M., Turpin, R. S., Tidwell, B. A., Lawrence, D. F. E., & Schulman, K. L. (2017) · Europe PMC (PubMed Central) · Cited by 46
Kinoshita, H., Saito, J., Takekawa, D., Ohyama, T., Kushikata, T., & Hirota, K. (2021) · PLoS ONE · doi:10.1371/journal.pone.0254654 · Cited by 49
Tavare, A. N., Alfuraih, A. M., Hensor, E. M. A., Astrinakis, E., Gupta, H., & Robinson, P. J. (2018) · Radiology · doi:10.1148/radiol.2018180950 · Cited by 47
“With 13 independent variables, a minimum of 175 participants (65 with malignant lesions) were required to provide at least ive events per variable, recommended in rules of thumb for logistic regression (16).”
MethodAssaye, A. M., Wiechula, R., Schultz, T., & Feo, R. (2021) · International Journal of Nursing Practice · doi:10.1111/ijn.13031 · Cited by 52
Zhu, Y.-J., Chang, X., Zhou, R., Chen, Y., Ma, H., Xiao, Z., Qu, X., Liu, Y., Liu, L., Li, Y., et al. (2022) · Lung Cancer · doi:10.1016/j.lungcan.2022.03.006 · Cited by 52
Yang, H., Sun, B., Ma, W., Fan, L., Xu, K., Jia, Y., Xu, J., Wang, Z., & Yao, F. (2022) · EBioMedicine · doi:10.1016/j.ebiom.2022.104265 · Cited by 52
Postoperative delirium following transcatheter aortic valve implantation: a historical cohort study
Tse, L., Bowering, J., Schwarz, S. K. W., Moore, R. L., Burns, K. D., & Barr, A. M. (2014) · Canadian Journal of Anesthesia/Journal canadien d anesthésie · doi:10.1007/s12630-014-0254-2 · Cited by 46
“Renal insufficiency, n (%) 28 (38) 18 (42) 0.”
Wärnberg, M. G., Roy, N., Khajanchi, M. U., Kumar, V., Dharap, S. B., Felländer‐Tsai, L., Petzold, M., Bhoi, S. K., Saha, M. L., & von Schreeb, J. (2014) · PLoS ONE · doi:10.1371/journal.pone.0105606 · Cited by 46
“We used ten events per free parameter to calculate sample size [27,28].”
MethodViciano, J., D’Anastasio, R., & Capasso, L. (2014) · Archives of Oral Biology · doi:10.1016/j.archoralbio.2014.09.003 · Cited by 54
Jones, M. E., Campbell, G., Patel, D., Brunner, E., Shatapathy, C. C., Murray-Thomas, T., van Staa, T. P., & Motsko, S. (2013) · Cardiovascular Psychiatry and Neurology · doi:10.1155/2013/647476 · Cited by 46
“To ensure that the resultingmodels were not overparameterised, the number of variables to be included in the model was selected using the rule of thumb of five outcome events per parameter [7, 8].”
MethodFan, H., Ji, M., Huang, J., Yue, P., Yang, X., Wang, C., & Ying, W. (2019) · International Journal of Nursing Studies · doi:10.1016/j.ijnurstu.2018.10.008 · Cited by 47
Rane, A. A., Tyser, A. R., Presson, A. P., Zhang, C., & Kazmers, N. H. (2019) · The Journal Of Hand Surgery · doi:10.1016/j.jhsa.2019.03.015 · Cited by 46
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.