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 801–900 of 1,018
Vandepitte, S., Putman, K. D., Noortgate, N. J. V. D., Verhaeghe, S. T. L., Mormont, E., Wilder, L. V., Smedt, D. D., & Annemans, L. (2018) · Dementia and Geriatric Cognitive Disorders · doi:10.1159/000494023 · Cited by 37
Yan, C., Richard, M. A., Gibson, C. J., He, J., Bosworth, A., Crossman, D. K., Singh, P., Hageman, L., Kalra, R., Armenian, S. H., et al. (2024) · Journal of Clinical Oncology · doi:10.1200/jco.23.02547 · Cited by 36
New-onset supraventricular arrhythmia during septic shock: prevalence, risk factors and prognosis
Seemann, A., Boissier, F., Razazi, K., Carteaux, G., de Prost, N., Brun‐Buisson, C., & Dessap, A. M. (2015) · Annals of Intensive Care · doi:10.1186/s13613-015-0069-5 · Cited by 30
“In view of the 30 events observed, we did not strictly follow the classical rule of a minimum of 10 outcome events per predictor variable in the multivariable analysis [29], but our analysis fulfilled the minimum of 5 events per predictor variable proposed by recent simulation studies [30].”
MethodHandwriting in healthy people aged 65 years and over
van Drempt, N., McCluskey, A., & Lannin, N. A. (2011) · Australian Occupational Therapy Journal · doi:10.1111/j.1440-1630.2011.00923.x · Cited by 30
“Factors thought less likely to be predictors of legibility, such as level of education, pen pressure, error corrections and mode of handwriting, were not included in the equation in line with recommendations by Vittinghoff and McCulloch (2007).”
ResultMoe, A., Hellzèn, O., Ekker, K., & Enmarker, I. (2012) · Aging & Mental Health · doi:10.1080/13607863.2012.717257 · Cited by 40
“The low beta-values for men ( n = 41) in relation to SOC and mental health and STS and physical health may be explained by a limited sample, but the sample number was within the critical limit (Vittinghoff & McCulloch, 2007 ).”
BackgroundMcIlroy, D. R., Farkas, D., Matto, M., & Lee, H. T. (2015) · Critical Care Medicine · doi:10.1097/ccm.0000000000000927 · Cited by 33
Presidential survival in South America: Rethinking the role of democracy
Martínez, C. A. (2015) · International Political Science Review · doi:10.1177/0192512115604904 · Cited by 36
Roelen, C. A. M., Thorsen, S. V., Heymans, M. W., Twisk, J. W. R., Bültmann, U., & Bjørner, J. B. (2016) · Disability and Rehabilitation · doi:10.1080/09638288.2016.1247471 · Cited by 30
“Another asset is that the prediction model was developed in a random sample with sufficient statistical power in terms of the number of LTSA events to provide robust risk estimates.[36] The split-sample design with validation of LTSA prediction models in employees other than those used to develop the prediction model,…”
MethodDeterminants of each domain of the Short Physical Performance Battery in COPD
Bernabéu-Mora, R., Giménez-Giménez, L. M., Montilla‐Herrador, J., García-Guillamón, G., García-Vidal, J. A., & Medina‐Mirapeix, F. (2017) · International Journal of COPD · doi:10.2147/copd.s138402 · Cited by 33
“However, this does not pose a major problem regarding the validity of the presently tested logistic models using balance as a dependent variable, because the general rule of 10 events per predictor variable can be relaxed when that number is close to 10.(45)”
BackgroundBuhmann, C., Flügel, T., Bihler, M., Gerloff, C. P., Niessen, A. C., Hidding, U., Nienstedt, J. C., & Pflug, C. (2018) · Parkinsonism & Related Disorders · doi:10.1016/j.parkreldis.2018.10.031 · Cited by 35
Berner, C., Erlacher, L., Fenzl, K. H., & Dorner, T. E. (2019) · International Journal of Rheumatology · doi:10.1155/2019/4709645 · Cited by 81
“For all other variables, the rule of thumb that logistic models should be used with a minimum of 10 events per predictor variable applies [43].”
MethodMendoza-Álvarez, A., Guillén‐Guío, B., Baez‐Ortega, A., Hernandez-Perez, C., Lakhwani-Lakhwani, S., Maeso, M.-D.-C., Lorenzo-Salazar, J. M., Morales, M., & Flores, C. (2019) · Frontiers in Genetics · doi:10.3389/fgene.2019.00439 · Cited by 32
“A formal survival analysis adjusting for clinical and demographic confounders was not pursued because the minimal number of events that are needed per variable (Vittinghoff and McCulloch, 2007; Riley et al., 2018) was out of reach because of the limited sample size.”
MethodMcDonald, S. D., Yu, Z., Blyderveen, S. V., Schmidt, L. A., Sword, W. A., Vanstone, M. G., Biringer, A. B., McDonald, H., & Beyene, J. (2020) · PLoS ONE · doi:10.1371/journal.pone.0233774 · Cited by 34
“…rule of “10 events per vari-able” in the regression model, [37] although the more lenient five events per variable in the regression model is sometimes considered to allow additional factors to be included in the regression model, [38] resulting in the inclusion of nine predictor variables.”
Outcomes after a Grammont-style reverse total shoulder arthroplasty?
Tashjian, R. Z., Hillyard, B., Childress, V., Kawakami, J., Presson, A. P., Zhang, C., & Chalmers, P. N. (2020) · Journal of Shoulder and Elbow Surgery · doi:10.1016/j.jse.2020.04.027 · Cited by 30
“adjusted for age, sex, BMI, Charlson comorbidty index, implant company, humeral version, length of follow-up, and postoperative diagnosisdexcept for reoperation, which was limited by the number of events (we adjusted for age, sex, Charlson comorbidty index, and length of follow-up).(45) We reported results from linear…”
MethodWettergren, L., Ljungman, L., Obol, C. M., Eriksson, L. E., & Lampic, C. (2020) · BMC Cancer · doi:10.1186/s12885-020-07175-8 · Cited by 36
Aorto-Iliac Artery Calcification Prior to Kidney Transplantation
Benjamens, S., Rijkse, E., Velde‐Keyzer, C. A. T., Berger, S. P., Moers, C., de Borst, M. H., Yakar, D., Slart, R. H. J. A., Dor, F. J. M. F., Minnee, R. C., et al. (2020) · Journal of Clinical Medicine · doi:10.3390/jcm9092893 · Cited by 31
“For the number of included variables in the multivariable analyses, we followed the recommendation by Vittinghoff and McCulloch of 5–9 events per predictor variable [35].”
MethodThe mortality burden of non-trauma fracture for adults with cerebral palsy
Whitney, D. G., Bell, S., Hurvitz, E. A., Peterson, M. D., Caird, M. S., & Jepsen, K. J. (2020) · Bone Reports · doi:10.1016/j.bonr.2020.100725 · Cited by 32
“This range provides coverage and bias within acceptable levels similar to the rule of thumb of 10 or more events per number of covariates for Cox models (Vittinghoff and McCulloch, 2007).”
MethodPetersen, A. S., Lund, N. L. T., Jensen, R., & Barloese, M. C. J. (2020) · Cephalalgia · doi:10.1177/0333102420970455 · Cited by 30
“In a logistic regression analysis, triptan response was negatively associated with age (29) and not associated with male sex or eCH, but the number of variables tested was very high (33).”
BackgroundVergaro, G., Aimo, A., Campora, A., Castiglione, V., Prontera, C., Masotti, S., Musetti, V., Chianca, M., Valleggi, A., Spini, V., et al. (2021) · Amyloid: Journal of Protein Folding Disorders · doi:10.1080/13506129.2021.1966624 · Cited by 25
Mazzoni, A., Brooke, H. L., Berntsen, S., Nordin, K., & Demmelmaier, I. (2021) · BMC Cancer · doi:10.1186/s12885-021-08996-x · Cited by 37
“at least five participants per predictor variable in the smallest group) [36].”
BackgroundLavrencic, L. M., Delbaere, K., Broe, G. A., Daylight, G., Draper, B., Cumming, R. G., Garvey, G., Allan, W., Hill, T. Y., Lasschuit, D., et al. (2022) · Neurology · doi:10.1212/wnl.0000000000013295 · Cited by 35
“To reduce bias, multivariable analyses were run with the Firth penalized maximum likelihood regression for imbalanced or rare events data (firthlogit) in Stata version 16 (<5 events per variable).(36)”
MethodMedetalibeyoğlu, A., Şenkal, N., Köse, M., Çatma, Y., Çaparali, E. B., Erelel, M., Öncül, O., Bahat, G., & Tükek, T. (2020) · The journal of nutrition health & aging · doi:10.1007/s12603-020-1499-9 · Cited by 30
Amor, R. E., Levy, J., Broza, Y. Y., Vangravs, R., Rapoport, S., Zhang, M., Wu, W., Leja, M., Behar, J. A., & Haick, H. (2023) · ACS Sensors · doi:10.1021/acssensors.2c02422 · Cited by 31
Hagel, B. E., Romanow, N. T. R., Morgunov, N. J., Embree, T., Couperthwaite, A. B., Voaklander, D. C., & Rowe, B. H. (2014) · Accident Analysis & Prevention · doi:10.1016/j.aap.2013.12.014 · Cited by 36
Spaderna, H., Weidner, G., Koch, K., Kaczmarek, I., Wagner, F., & Smits, J. M. (2011) · The Journal of Heart and Lung Transplantation · doi:10.1016/j.healun.2011.07.018 · Cited by 30
Thromboprophylaxis patterns and determinants in critically ill patients: a multicenter audit
Lauzier, F., Muscedere, J., Deland, É., Kutsogiannis, D. J., Jacka, M. J., Heels‐Ansdell, D., Crowther, M. A., Cartin‐Ceba, R., Cox, M. J., Zytaruk, N. L., et al. (2014) · Critical Care · doi:10.1186/cc13844 · Cited by 37
“We used multilevel modeling, which allows concurrent analysis of center and patient factors (fixed baseline characteristics and variable patient-days) as determinants of administration, and avoided overfitting [46].”
MethodIncidencia y factores asociados al abandono del tratamiento antituberculoso
Cáceres, F. D. M., & Orozco, L. C. (2007) · Biomédica · doi:10.7705/biomedica.v27i4.170 · Cited by 29
Neuropsychological correlates of Pisa syndrome in patients with Parkinson's disease
Vitale, C., Falco, F., Trojano, L., Erro, R., Moccia, M., Allocca, R., Agosti, V., Santangelo, F., Barone, P., & Santangelo, G. (2015) · Acta Neurologica Scandinavica · doi:10.1111/ane.12514 · Cited by 32
Mallhi, T. H., Khan, A. H., Sarriff, A., Adnan, A. S., Khan, Y. H., & Jummaat, F. B. (2016) · Postgraduate Medical Journal · doi:10.1136/postgradmedj-2015-133582 · Cited by 30
Risk factors for infections due to carbapenem-resistantKlebsiella pneumoniaeafter open heart surgery
Salsano, A., Giacobbe, D. R., Sportelli, E., Olivieri, G. M., Brega, C., Biase, C. D., Coppo, E., Marchese, A. E., Bono, V. D., Viscoli, C., et al. (2016) · Interactive Cardiovascular and Thoracic Surgery · doi:10.1093/icvts/ivw228 · Cited by 30
Tymińska, A., Kapłon‐Cieślicka, A., Ozierański, K., Peller, M., Balsam, P., Marchel, M., Crespo-Leiro, M. G., Maggioni, A. P., Jankowska, E. A., Dróżdż, J., et al. (2017) · The American Journal of Cardiology · doi:10.1016/j.amjcard.2017.03.035 · Cited by 35
Lee, T. H., Lee, C., Ng, C. Y., Chang, M., Chang, S.-W., Fan, P., Chung, W., Tian, Y., Chen, Y., & Chang, C. (2018) · PLoS ONE · doi:10.1371/journal.pone.0203642 · Cited by 33
Microinvasive breast carcinoma: An analysis from ten Senonetwork Italia breast centres
Costarelli, L., Cianchetti, E., Corsi, F., Friedman, D., Ghilli, M., Lacaria, M. T., Menghini, L., Murgo, R., Ponti, A., Rinaldi, S., et al. (2018) · European Journal of Surgical Oncology · doi:10.1016/j.ejso.2018.09.024 · Cited by 30
Prison and Violent Political Extremism in the United States
Lafree, G. D., Jiang, B., & Porter, L. C. (2019) · Journal of Quantitative Criminology · doi:10.1007/s10940-019-09412-1 · Cited by 35
Role of leadership in the digitalisation of manufacturing organisations
Tortorella, G. L., Prashar, A., Antony, J., Cawley, A. F. M., Vassolo, R. S., & Sony, M. (2023) · Journal of Manufacturing Technology Management · doi:10.1108/jmtm-09-2022-0312 · Cited by 44
Hernández-Rocha, C., Walshe, M., Birch, S., Sabic, K., Korie, U. M., Chasteau, C., Miladinova, V., Sabol, W. B., Mengesha, E., Hanna, M., et al. (2023) · Journal of Crohn s and Colitis · doi:10.1093/ecco-jcc/jjad186 · Cited by 29
Environmental Determinants of Bicycling Injuries in Alberta, Canada
Romanow, N. T. R., Couperthwaite, A. B., McCormack, G. R., Nettel‐Aguirre, A., Rowe, B. H., & Hagel, B. E. (2012) · Journal of Environmental and Public Health · doi:10.1155/2012/487681 · Cited by 37
“Again, we relaxed the 10% “rule” [23], and potential confounders were added to the model following a 5–9 event per variable guideline [24, 25, 28].”
MethodPredictors for postoperative nausea and vomiting after xenon-based anaesthesia
Schaefer, M. S., Apfel, C. C., Sachs, H.-J., Stuttmann, R., Bein, B., Tonner, P. H., Hein, M., Neukirchen, M., Reyle‐Hahn, M., & Kienbaum, P. (2015) · British Journal of Anaesthesia · doi:10.1093/bja/aev115 · Cited by 28
Reichert, J. A., Hlavinka, P. F., & Stolzfus, J. C. (2014) · Pharmacotherapy The Journal of Human Pharmacology and Drug Therapy · doi:10.1002/phar.1464 · Cited by 29
“In addition, a direct logistic regression model was created to determine the effect of pharmacologic VTE prophylaxis on hemorrhage, with 28 events sufficient to support four model covariates.(22) The model adjusted for platelet count, maximum INR, and antiplatelet medication.”
MethodMcDonald, S. D., Park, C. K., Timm, V., Schmidt, L. A., Neupane, B., & Beyene, J. (2013) · Journal of Obstetrics and Gynaecology Canada · doi:10.1016/s1701-2163(15)30757-x · Cited by 40
Natalini, G., Tuzzo, D., Rosano, A., Testa, M., Grazioli, M., Pennestrì, V., Amodeo, G., Berruto, F., Fiorillo, M., Peratoner, A., et al. (2016) · Annals of Intensive Care · doi:10.1186/s13613-016-0158-0 · Cited by 30
“” We anticipated that the enrollment of 100 patients would give about 40 events, allowing us to evaluate up to eight explanatory variables in a logistic model with “complete PEEP-absorber” as the outcome variable [19].”
Interactions of glyphosate use with farm characteristics and cropping patterns in Central Europe
Wiese, A., Schulte, M., Theuvsen, L., & Steinmann, H. (2017) · Pest Management Science · doi:10.1002/ps.4542 · Cited by 35
Geostatistical distribution modelling of two invasive crayfish across dendritic stream networks
Filipe, A. F., Quaglietta, L., Ferreira, M., Magalhães, M. F., & Beja, P. R. (2017) · Biological Invasions · doi:10.1007/s10530-017-1492-3 · Cited by 35
“This may be a consequence, at least partly, of the relatively small number of presences detected in our study (16), which can cause problems in logistic models (Vittinghoff and McCulloch 2007).”
BackgroundMorris, D., Skalina, T., Singh, T. P., Moxon, J. V., & Golledge, J. (2018) · Journal of the American Heart Association · doi:10.1161/jaha.118.009943 · Cited by 31
“Inclusion criteria for the study were the following: (1) lower extremity CT angiogram conducted as part of routine clinical assessment; (2) evidence of PAD on the CT angiogram defined as ≥50% stenosis in at least 1 lower limb artery; and (3) at least 1 follow-up assessment as an inpatient or outpatient.”
BackgroundHeld, U., Burgstaller, J. M., Wertli, M. M., Pichierri, G., Winklhofer, S. F.-X., Brunner, F., Porchet, F., Farshad, M., & Steurer, J. (2018) · PLoS ONE · doi:10.1371/journal.pone.0207126 · Cited by 30
“According to Vittinghoff and McCulloch [34], this rule can be relaxed to 5 to 9 EPVs allowing for up to 22 prognostic indicators if 9 EPVs are taken as threshold.”
BackgroundMüller, M., Curatolo, M., Limacher, A., Neziri, A. Y., Treichel, F., Battaglia, M., Arendt‐Nielsen, L., & Jüni, P. (2019) · European Journal of Pain · doi:10.1002/ejp.1356 · Cited by 33
“A sample size of 140 with 42 patients developing CLBP after 6 months will allow the inclusion of approximately nine variables in a multivariable model (Vittinghoff & McCulloch, 2007).”
BackgroundChiti, L. E., Martano, M., Ferrari, R., Boracchi, P., Giordano, A., Grieco, V., Buracco, P., Iussich, S., Giudice, C., Miniscalco, B., et al. (2019) · Veterinary and Comparative Oncology · doi:10.1111/vco.12534 · Cited by 29
Jensen, M. T., Søgaard, P., Gustafsson, I., Bech, J. W., Hansen, T. F., Almdal, T., Theilade, S., Biering‐Sørensen, T., Jørgensen, P. G., Galatius, S., et al. (2019) · Diabetologia · doi:10.1007/s00125-019-05009-2 · Cited by 31
Ten-Year Experience With the Anterior Approach to Total Hip Arthroplasty at a Tertiary Care Center
Gofton, W. T., Ibrahim, M. M., Kreviazuk, C., Kim, P. R., Feibel, R. J., & Beaulé, P. E. (2019) · The Journal of Arthroplasty · doi:10.1016/j.arth.2019.12.025 · Cited by 29
Sullivan, A. L., Kulkarni, T., & Chhuon, V. (2020) · Exceptional Children · doi:10.1177/0014402920905548 · Cited by 34
“For this study, subgroup sizes met recommendations for logistic regression (e.g., Vittinghoff & McCulloch, 2007).”
MethodTirkkonen, J., Skrifvars, M. B., Parr, M. J., Tamminen, T., & Åneman, A. (2020) · Resuscitation · doi:10.1016/j.resuscitation.2020.02.022 · Cited by 29
Ramspek, C. L., Verberne, W. R., van Buren, M., Dekker, F. W., Bos, W. J. W., & van Diepen, M. (2020) · Clinical Kidney Journal · doi:10.1093/ckj/sfaa021 · Cited by 32
“The goal of this selection method was to prevent overfitting while retaining enough variables for an accurate prediction [18, 19].”
MethodAnderson, R. E., Keihani, S., Moses, R. A., Nocera, A. P., Selph, J. P., Becerra, C. M. C., Baradaran, N., Glavin, K., Broghammer, J. A., Arya, C. S., et al. (2020) · The Journal of Urology · doi:10.1097/ju.0000000000001075 · Cited by 29
Townsend, R., Manji, A., Allotey, J., Heazell, A. E. P., Jorgensen, L., Magee, L. A., Mol, B. W., Snell, K. I. E., Riley, R. D., Sandall, J., et al. (2020) · BJOG An International Journal of Obstetrics & Gynaecology · doi:10.1111/1471-0528.16487 · Cited by 32
“(14) However, the optimal sample size is actually context specific and may be higher or lower than 10,(15,16) taking into account outcome prevalence, the magnitude of predictor effects and the expected fit of the model (R2).”
BackgroundMental health outcomes in communities exposed to Armed Conflict Experiences
Trujillo, S., Cardona, L. S. G., López, J. D., Acosta, A., & Trujillo, N. (2021) · BMC Psychology · doi:10.1186/s40359-021-00626-2 · Cited by 36
Pelt, K. L. V., Caccese, J. B., Eckner, J. T., Putukian, M., Brooks, M. A., Cameron, K. L., Houston, M. N., Posner, M. A., Jackson, J. C., McGinty, G. T., et al. (2021) · Journal of sport and health science/Journal of Sport and Health Science · doi:10.1016/j.jshs.2021.01.004 · Cited by 32
Matsumoto, Y., Shivappa, N., Sugioka, Y., Tada, M., Okano, T., Mamoto, K., Inui, K., Habu, D., Hébert, J. R., & Koike, T. (2021) · Arthritis Research & Therapy · doi:10.1186/s13075-021-02478-y · Cited by 35
“the event rate was greater than 5 to guarantee a confidence interval [26], the number of items entered was set to 7.”
BackgroundJose, A., Shah, S. A., Anwar, N., Jones, C. R., Sherman, K. E., & Elwing, J. (2021) · Liver Transplantation · doi:10.1002/lt.26091 · Cited by 31
“These variables included age of donor and recipient, pretransplant MELD score, sex of recipient, race/ ethnicity and body mass index of recipient, cold ischemia time, pretransplant serum albumin and serum sodium, history of recipient malignancy, presence of split- liver donation, OPTN region where transplantation occu…”
BackgroundRegression and Machine Learning Methods to Predict Discrete Outcomes in Accounting Research
Krupa, J., & Minutti‐Meza, M. (2022) · Journal of Financial Reporting · doi:10.2308/jfr-2021-010 · Cited by 33
Manara, J. R., Salmon, L. J., Kilani, F. M., de Camino, G. Z., Monk, C. I., Sundaraj, K., Pinczewski, L. A., & Roe, J. P. (2022) · The American Journal of Sports Medicine · doi:10.1177/03635465221125467 · Cited by 33
“A multivariate Cox model of predictors of CACL survival was not performed, as there were insufficient numbers of positive events of CACL rupture occurring in those who did not RTS (n = 0 in females and 2 in males), which is susceptible to bias with positive events numbering \5.(53) The significant factors on univariat…”
MethodPhan, K., Sandoval, Y.-H. G., Gorgui, J., El‐Messidi, A., Gagnon, R. J., Abenhaim, H. A., Rahme, E., & Daskalopoulou, S. S. (2023) · BJOG An International Journal of Obstetrics & Gynaecology · doi:10.1111/1471-0528.17430 · Cited by 33
Haugaard, S. B., Andersen, O., Hansen, T. W., Eugen‐Olsen, J., Linneberg, A. R., Madsbad, S., Olsen, M. H., Jørgensen, T., Borch‐Johnsen, K., & Jeppesen, J. (2011) · Diabetic Medicine · doi:10.1111/j.1464-5491.2011.03513.x · Cited by 30
“Based on recent statistical research [16], we planned to evaluate a minimum of five events per predictor variable to work with stable statistical models.”
MethodEffectiveness of child protective services interventions as indicated by rates of recidivism
Solomon, D., & Asberg, K. K. (2012) · Children and Youth Services Review · doi:10.1016/j.childyouth.2012.08.014 · Cited by 38
Papageorgiou, S., & Cho, Y. (2013) · Language Testing · doi:10.1177/0265532213499750 · Cited by 29
“The ordinal logistic regression is a statistical approach to derive cut scores to separate subjects into a number of ordered categories, but it requires a large sample size (Vittinghoff & McCulloch, 2007).”
Methodvan Deijck, R. H. P. D., Hasselaar, J. G. J., Verhagen, S. C. A. H. H. V. M., Vissers, K. C. P., & Koopmans, R. T. C. M. (2016) · Journal of Pain and Symptom Management · doi:10.1016/j.jpainsymman.2015.12.327 · Cited by 30
Teong, A. C. A., Diong, A. X., Omar, S. Z., & Tan, P. C. (2017) · Scientific Reports · doi:10.1038/s41598-017-12410-7 · Cited by 33
Dengue fever mortality score: A novel decision rule to predict death from dengue fever
Huang, C., Hsu, C.-C., Guo, H., Su, S., & Lin, H.-J. (2017) · Journal of Infection · doi:10.1016/j.jinf.2017.09.014 · Cited by 29
Subramaniam, M., Abdin, E., Shahwan, S., Satghare, P., Vaingankar, J. A., Sendren, J. R., Picco, L., Chua, B. Y., Ng, B. T., Chong, S. A., et al. (2017) · General Hospital Psychiatry · doi:10.1016/j.genhosppsych.2017.11.009 · Cited by 34
Koskela, M., Jahnukainen, T., Endén, K., Arikoski, P., Kataja, J., Nuutinen, M., & Ylinen, E. (2019) · Pediatric Nephrology · doi:10.1007/s00467-019-04238-2 · Cited by 33
Clark, V. A. (2015) · Journal of Experimental Criminology · doi:10.1007/s11292-014-9216-5 · Cited by 28
Vijverberg, R., Ferdinand, R. F., Beekman, A. T. F., & van Meijel, B. (2018) · BMC Psychiatry · doi:10.1186/s12888-018-1874-9 · Cited by 28
“By using this step-by-step approach, we created multivariable regression models that did not violate the statistical rule of 10 events per 1 variable and ensured the validity of the analysis [31, 32].”
MethodSoussi, S., Tarvasmäki, T., Kimmoun, A., Ahmadiankalati, M., Azibani, F., dos Santos, C. C., Duarte, K., Gayat, É., Jentzer, J. C., Harjola, V., et al. (2024) · EClinicalMedicine · doi:10.1016/j.eclinm.2024.103013 · Cited by 31
Chaudhuri, A. A., Binkley, M. S., Rigdon, J., Carter, J., Aggarwal, S., Dudley, S. A., Qian, Y., Kumar, K. A., Hara, W. Y., Gensheimer, M. F., et al. (2016) · Radiotherapy and Oncology · doi:10.1016/j.radonc.2016.05.007 · Cited by 29
“Results of our 3 separate multivariate analyses, each of which involved 3–5 variables [35], showed that only elevated mean non-target lung SUV (p < 0.”
BackgroundEnterococci in orthopaedic infections: Who is at risk getting infected?
Uçkay, I., Pires, D., Agostinho, A., Guanziroli, N., Öztürk, M., Bartolone, P., Tscholl, P. M., Betz, M., & Pittet, D. (2017) · Journal of Infection · doi:10.1016/j.jinf.2017.06.008 · Cited by 28
Prognostic value of multiple cytokine analysis in colorectal cancer: a systematic review
Gunawardene, A. R., Dennett, E., & Larsen, P. (2018) · Journal of Gastrointestinal Oncology · doi:10.21037/jgo.2018.07.11 · Cited by 31
“Recently, calls have been made to relax this rule in certain situations although this is mainly applicable to tests of sensitivity rather than prognostic studies (43).”
BackgroundEckard, T. G., Padua, D. A., Mauntel, T. C., Frank, B. S., Pietrosimone, B., Begalle, R. L., Goto, S., Clark, M., & Kucera, K. L. (2018) · Physical Therapy in Sport · doi:10.1016/j.ptsp.2018.10.009 · Cited by 36
Chandrasekhar, D., Rothfeder, R., Xiao, Y., & Finn, D. (2018) · Journal of Environmental Planning and Management · doi:10.1080/09640568.2018.1492907 · Cited by 31
“Vittinghoff and McCulloch (2007) suggest that the ratio of events (the smaller of the two DV frequency values) to IVs in the regression equation should be no lower than 5:1.”
BackgroundThe efficacy of neutron radiation therapy in treating salivary gland malignancies
Timoshchuk, M.-A., Dekker, P., Hippe, D. S., Parvathaneni, U., Liao, J. J., Laramore, G. E., & Dillon, J. K. (2018) · Oral Oncology · doi:10.1016/j.oraloncology.2018.11.006 · Cited by 28
Hreinsson, J. P., Sigurdardóttir, R., Lund, S. H., & Bjornsson, E. S. (2018) · Scandinavian Journal of Gastroenterology · doi:10.1080/00365521.2018.1532019 · Cited by 31
“Therefore, a minimum of 233–420 patients had to be included in the train data to have 5–9 event outcomes per variable [16].”
BackgroundNausheen, F., Niknafs, N., MacLean, D. J., Olvera, D. J., Wolfe, A. C., Pennington, T. W., & Davis, D. P. (2019) · Scandinavian Journal of Trauma Resuscitation and Emergency Medicine · doi:10.1186/s13049-019-0614-6 · Cited by 30
Incidence and risk factors for gangrene in patients with systemic sclerosis from the EUSTAR cohort
Mihai, C., Distler, O., Gheorghiu, A. M., Constantin, P. I., Dobrota, R. D., Jordan, S., Smith, V. M., HACHULLA, E., Henes, J., Siegert, E., et al. (2019) · Lara D. Veeken · doi:10.1093/rheumatology/kez558 · Cited by 28
McNeeley, S., & Donley, C. (2020) · Criminal Justice and Behavior · doi:10.1177/0093854820959394 · Cited by 35
“Scholars argue there should be a minimum of 5 to 10 events for each independent variable that is included in a logistic regression model (Penduzzi et al., 1996; Vittinghoff & McCulloch, 2007).”
BackgroundHaining, K., Karagiorgou, O., Gajwani, R., Groß, J., Gumley, A. I., Lawrie, S. M., Schwannauer, M., Schultze‐Lutter, F., & Uhlhaas, P. J. (2020) · Early Intervention in Psychiatry · doi:10.1111/eip.13075 · Cited by 29
“This outcome variable was prioritised as it did not violate the events per variable rule of 5:1 suggested by Vittinghoff and McCulloch (2007).”
MethodDetection of Single Cancer Cells in Blood with Artificially Intelligent Nanoarray
Amor, R. E., Broza, Y. Y., & Haick, H. (2021) · ACS Nano · doi:10.1021/acsnano.1c01741 · Cited by 29
Hoerler, R., van Dijk, J., Patt, A., & Duce, A. D. (2021) · Transportation Research Part D Transport and Environment · doi:10.1016/j.trd.2021.102861 · Cited by 30
Elmghaamez, I. K., Attah‐Boakye, R., Adams, K., & Agyemang, J. (2021) · Thunderbird International Business Review · doi:10.1002/tie.22244 · Cited by 30
Hooper, D. K., Varnell, C. D., Rich, K. L., Carle, A. C., Huber, J. F., Mostajabi, F., Dahale, D. S., Pai, A. L. H., Goebel, J., & Modi, A. C. (2021) · American Journal of Kidney Diseases · doi:10.1053/j.ajkd.2021.06.021 · Cited by 32
Castillo-Oti, J. M., Galván-Manso, A. I., Callejas-Herrero, M. R., Vara-González, L., Salas-Herrera, F., & Muñoz, P. (2021) · Nutrients · doi:10.3390/nu14010084 · Cited by 30
“The criterion followed was that of including one independent variable for every 10 cases of the dependent variable in the final model [29].”
BackgroundComartin, E. B., Burgess‐Proctor, A., Kubiak, S. P., & Kernsmith, P. D. (2017) · Violence and Victims · doi:10.1891/0886-6708.33.1.53 · Cited by 33
“We were unable to test full logistic regression models because of the study’s sample size (Vittinghoff & McCulloch, 2007).”
MethodKoopman, J. E., Zweedijk, B. E., Hundepool, C. A., Duraku, L. S., Smit, J. M., Wouters, R. M., Selles, R. W., van der Avoort, D. J. J. C., Blomme, R. A. M., de Boer, H. L., et al. (2022) · The Journal Of Hand Surgery · doi:10.1016/j.jhsa.2022.04.017 · Cited by 28
Xu, L., Cai, L., Zhu, Z., & Chen, G. (2023) · BMC Endocrine Disorders · doi:10.1186/s12902-023-01368-5 · Cited by 32
“However, Cox regression models should be used with a minimum of 10 outcome events per predictor variable (EPV) [28].”
MethodHall, M. K., Taylor, R. A., Luty, S., Allen, I. E., & Moore, C. (2016) · The American Journal of Emergency Medicine · doi:10.1016/j.ajem.2016.02.059 · Cited by 35
Widinghoff, C., Berge, J., Wallinius, M., Billstedt, E., Hofvander, B., & Håkansson, A. (2018) · Journal of Gambling Studies · doi:10.1007/s10899-018-9785-8 · Cited by 36
de Haan-Du, J., Plas, M., Absalom, A. R., van Leeuwen, B. L., & de Bock, G. H. (2020) · Journal of Surgical Oncology · doi:10.1002/jso.25836 · Cited by 32
“A rule of at least five events per predictor variable in the multivariable analysis was applied.(27) Odds ratios (ORs) and their corresponding 95% confidence intervals (95% CIs) were estimated and are reported.”
MethodFactors predicting patient-reported functional outcome scores after humeral shaft fractures
Shields, E. J. W., Sundem, L., Childs, S. M., Maceroli, M. A., Humphrey, C. A., Ketz, J. P., & Gorczyca, J. T. (2015) · Injury · doi:10.1016/j.injury.2015.01.027 · Cited by 28
“We limited our analysis to 12 variables: [17] age, follow-up length, surgical versus non-operative treatment, body mass index (BMI), presence of associated fractures, fracture location (proximal, middle and distal third), radial nerve palsy, smoking status, CCI score, insurance type, classification as high-energy mech…”
Prognostic factors, symptom evolution, and quality of life of posttraumatic trigeminal neuropathy
der Cruyssen, F. V., Peeters, F., Laat, A. D., Jacobs, R., Politis, C., & Renton, T. F. (2021) · Pain · doi:10.1097/j.pain.0000000000002408 · Cited by 29
“However, based on simulation studies, a minimum of 10 events per variable are required, rendering a sample size of 230 patients.(31,46) Missing data were handled by listwise exclusion to build the logistic regression model after verifying the randomness and frequency of missing values.”
Zu, W., Huang, X., Xu, T., Du, L., Wang, Y., Wang, L., & Nie, W. (2023) · PLoS ONE · doi:10.1371/journal.pone.0287308 · Cited by 40
Mathes, T., Pape-Köhler, C., Moerders, L., Lux, E. A., & Neugebauer, E. A. M. (2017) · Pain Medicine · doi:10.1093/pm/pnx242 · Cited by 27
Predictors and Outcomes of Recurrent Venous Thromboembolism in Elderly Patients
Lauber, S., Limacher, A., Tritschler, T., Stalder, O., Méan, M., Righini, M., Aschwanden, M., Beer, J. H., Frauchiger, B., Osterwalder, J., et al. (2018) · The American Journal of Medicine · doi:10.1016/j.amjmed.2017.12.015 · Cited by 29
“14 predictor variable.(31) In addition, we included only patient baseline characteristics into the model in a sensitivity analysis, omitting all treatment variables.”
MethodPrevalence, course and psychosis-predictive value of negative symptoms in 22q11.2 deletion syndrome
Schneider, M., Armando, M., Schultze‐Lutter, F., Pontillo, M., Vicari, S., Debbané, M., & Eliez, S. (2018) · Schizophrenia Research · doi:10.1016/j.schres.2018.10.014 · Cited by 32
“This ratio was reported to commonly ensure sufficient confidence interval coverage for 1 and the related type I error rate of the test of H0, little bias in the estimate of 1, and thereby sufficient power (Vittinghoff and McCulloch, 2007).”
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.