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 301–400 of 1,018
Carr, P. J., Rippey, J. C. R., Cooke, M. L., Bharat, C., Murray, K., Higgins, N. S., Foale, A., & Rickard, C. M. (2016) · BMJ Open · doi:10.1136/bmjopen-2015-009196 · Cited by 67
“Additionally this number would be sufficient based on guidelines suggested by Peduzzi et al(58) and Vittinghoff and McCulloch.(59) Furthermore, this would more than adequately satisfy the minimum recommendation of Steyerberg(60) for validation purpose.”
Rocchi, M., Routhier, F., Latimer‐Cheung, A. E., Ginis, K. A. M., Noreau, L., & Sweet, S. N. (2017) · Spinal Cord · doi:10.1038/sc.2016.181 · Cited by 85
A Gene Signature of Survival Prediction for Kidney Renal Cell Carcinoma by Multi-Omic Data Analysis
Hu, F., Zeng, W., & Liu, X. (2019) · International Journal of Molecular Sciences · doi:10.3390/ijms20225720 · Cited by 75
Wiebenga, J. X. M., Eikelenboom, M., Heering, H. D., van Oppen, P., & Penninx, B. W. J. H. (2020) · Australian & New Zealand Journal of Psychiatry · doi:10.1177/0004867420951256 · Cited by 75
Tiller, H., Killie, M. K., Husebekk, A., Skogen, B. R., Ni, H., Kjeldsen‐Kragh, J., & Øian, P. (2011) · Acta Obstetricia Et Gynecologica Scandinavica · doi:10.1111/j.1600-0412.2011.01269.x · Cited by 66
“The number of events per predictor variable (EPV) was therefore 6.5 and within the recommended range of EPV (16).”
BackgroundInfluentialBiliary Multifocal Chromosomal Polysomy and Cholangiocarcinoma in Primary Sclerosing Cholangitis
Eaton, J. E., Fritcher, E. G. B., Gores, G. J., Atkinson, E. J., Tabibian, J. H., Topazian, M. D., Gossard, A. A., Halling, K. C., Kipp, B. R., & Lazaridis, K. N. (2015) · The American Journal of Gastroenterology · doi:10.1038/ajg.2014.433 · Cited by 67
Why Mortality Is Increased in Health-Care-Associated Pneumonia
Rello, J., Luján, M., Gallego, M., Vallés, J., Belmonte, Y., Fontanals, D., Dı́az, E., & Lisboa, T. C. (2009) · CHEST Journal · doi:10.1378/chest.09-2175 · Cited by 66
Bonner, A., Gillespie, K. M., Campbell, K. L., Corones‐Watkins, K., Hayes, B., Harvie, B., Kelly, J. T., & Havas, K. (2018) · BMC Nephrology · doi:10.1186/s12882-018-0830-8 · Cited by 70
“Utilising the most conservative ‘rule of thumb’, 20 participants per item determined the study sample size [17, 18].”
BackgroundDarmon, M., Bourmaud, A., Reynaud, M., Rouleau, S., Meziani, F., Boivin, A., Benyamina, M., Vincent, F., Lautrette, A., Leroy, C., et al. (2018) · Intensive Care Medicine · doi:10.1007/s00134-018-5386-3 · Cited by 70
Parekh, N., Ali, K., Davies, J. G., Stevenson, J. M., BANYA, W. A. S., Nyangoma, S., Schiff, R., van der Cammen, T. J. M., Harchowal, J., & Rajkumar, C. (2019) · BMJ Quality & Safety · doi:10.1136/bmjqs-2019-009587 · Cited by 80
Pittsburgh B Compound Positron Emission Tomography in Patients With AL Cardiac Amyloidosis
Lee, S., Suh, H. Y., Park, S., Oh, S., Kwak, S., Kim, H. M., Koh, Y., Park, J., Kim, H., Cho, H., et al. (2020) · Journal of the American College of Cardiology · doi:10.1016/j.jacc.2019.11.037 · Cited by 67
Fukuda, S., Varshney, A., Fowler, B. J., Wang, S.-B., Narendran, S., Ambati, K., Yasuma, T., Magagnoli, J., Leung, H., Hirahara, S., et al. (2021) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.2022751118 · Cited by 65
“Given the low event rate (38) among NRTI users in Truven and Veterans databases, we employed Bayesian proportional hazards Cox regression models: NRTI exposure was associated with a reduced hazard of atrophic AMD in Truven (aHR = 0.”
MethodBarlati, S., Morena, D., Nibbio, G., Cacciani, P., Corsini, P., Mosca, A., Deste, G., Accardo, V., Turrina, C., Valsecchi, P., et al. (2021) · Schizophrenia Research · doi:10.1016/j.schres.2021.06.007 · Cited by 76
Acute kidney injury is associated with early cytokine changes after trauma
Bihorac, A., Baslanti, T. O., Cuenca, A. G., Hobson, C., Ang, D. N., Efron, P. A., Maier, R. V., Moore, F. A., & Moldawer, L. L. (2013) · The Journal of Trauma: Injury, Infection, and Critical Care · doi:10.1097/ta.0b013e31828586ec · Cited by 76
van Kampen, D. A., van den Berg, T., van der Woude, H.-J., Castelein, R. M., Terwee, C. B., & Willems, W. J. (2013) · Journal of Shoulder and Elbow Surgery · doi:10.1016/j.jse.2013.05.006 · Cited by 74
“In developing a prediction model, it is recommended that for each potential important predictor studied, at least 5 patients must have a certain diagnosis (in this study, 60 patients were diagnosed with traumatic anterior shoulder instability); otherwise, the risk of overestimating the model would become a problem.(26…”
BackgroundPietrzak, R. H., Whealin, J. M., Stotzer, R. L., Goldstein, M. B., & Southwick, S. M. (2011) · Journal of Psychiatric Research · doi:10.1016/j.jpsychires.2011.07.010 · Cited by 65
Osté, M. C. J., Corpeleijn, E., Navis, G. J., Velde‐Keyzer, C. A. T., Soedamah‐Muthu, S. S., van den Berg, E., Postmus, D., de Borst, M. H., Kromhout, D., & Bakker, S. J. L. (2017) · BMJ Open Diabetes Research & Care · doi:10.1136/bmjdrc-2016-000283 · Cited by 70
A Stratified Model for Psychosis Prediction in Clinical Practice
Michel, C., Ruhrmann, S., Schimmelmann, B. G., Klosterkötter, J., & Schultze-Lutter, F. (2014) · Schizophrenia Bulletin · doi:10.1093/schbul/sbu025 · Cited by 64
Liu, K. X., Ioakeim‐Ioannidou, M., Susko, M. S., Rao, A. D., Yeap, B. Y., Snijders, A. M., Ladra, M. M., Vogel, J., Zaslowe‐Dude, C., Marcus, K. J., et al. (2020) · International Journal of Radiation Oncology*Biology*Physics · doi:10.1016/j.ijrobp.2020.09.049 · Cited by 64
van Kampen, D. A., van den Berg, T., van der Woude, H.-J., Castelein, R. M., Scholtes, V. A., Terwee, C. B., & Willems, W. J. (2014) · Journal of Orthopaedic Surgery and Research · doi:10.1186/s13018-014-0070-y · Cited by 76
“For the development of a prediction model, it is recommended that, for each potential important predictor studied, at least five events (in this study, patients with a RC tear) are required to avoid the risk that overestimation might become problematic [34,35].”
MethodSakles, J. C., Javedani, P. P., Chase, E., Garst‐Orozco, J., Guillen-Rodriguez, J., & Stolz, U. (2015) · Academic Emergency Medicine · doi:10.1111/acem.12674 · Cited by 66
Nelson, G., Kiyang, L., Chuck, A., Nguyen, T., & Gramlich, L. (2016) · Current Oncology · doi:10.3747/co.23.2980 · Cited by 60
Patterns and Predictors of Recurrence after Partial Nephrectomy for Kidney Tumors
Mouracadé, P., Kara, Ö., Maurice, M. J., Dagenais, J., Malkoç, E., Nelson, R. J., & Kaouk, J. H. (2016) · The Journal of Urology · doi:10.1016/j.juro.2016.12.046 · Cited by 64
Probert-Lindström, S., Berge, J., Westrin, Å., Öjehagen, A., & Skogman Pavulans, K. (2020) · BMJ Open · doi:10.1136/bmjopen-2020-038794 · Cited by 63
Non-tuberculous mycobacterial infections of the hand
Balagué, N., Uçkay, İ., Vostrel, P., Hinrikson, H., Aaken, I. V., & Beaulieu, J. (2014) · Chirurgie de la Main · doi:10.1016/j.main.2014.12.004 · Cited by 64
Forke, C. M., Catallozzi, M., Localio, A. R., Grisso, J. A., Wiebe, D. J., & Fein, J. A. (2019) · Preventive Medicine Reports · doi:10.1016/j.pmedr.2019.100942 · Cited by 63
“We entered the latter of these five variables into a separate regression model to predict a single-variable “parent health score” that was included as a covariate in the child health model; this was done to decrease potential for overfitting the child model by having to control for an excessive number of parent variab…”
MethodDeferred Cytoreductive Nephrectomy in Patients with Newly Diagnosed Metastatic Renal Cell Carcinoma
Bhindi, B., Graham, J., Wells, J. C., Bakouny, Z., Donskov, F. N., Fraccon, A. P., Pasini, F., Lee, J., Basappa, N. S., Hansen, A. R., et al. (2020) · European Urology · doi:10.1016/j.eururo.2020.04.038 · Cited by 63
Propensity score model overfitting led to inflated variance of estimated odds ratios
Schuster, T., Lowe, W. K., & Platt, R. W. (2016) · Journal of Clinical Epidemiology · doi:10.1016/j.jclinepi.2016.05.017 · Cited by 69
5-Year imaging sequelae of concussive blast injury and relation to early clinical outcome
Mac Donald, C. L., Barber, J., Andre, J., Evans, N., Panks, C., Sun, S., Zalewski, K., Elizabeth Sanders, R., & Temkin, N. (2017) · NeuroImage: Clinical · doi:10.1016/j.nicl.2017.02.005 · Cited by 62
“The events-per-variable (EPV) ratio for the best AIC model was 7.5, which while lower than the minimum recommended EPV of 10–15 is likely still valid given the conditions of this particular analysis (Vittinghoff and McCulloch, 2007).”
BackgroundInfluentialLiang, B., Tian, Y., Chen, X., Yan, H., Yan, L., Zhang, T., Zhou, Z., Wang, L., & Dai, J. (2020) · Frontiers in Oncology · doi:10.3389/fonc.2019.01500 · Cited by 71
“For all the three models, the number of predictors was fixed to 2 to avoid overfitting (28).”
BackgroundFactors Associated with Influenza Vaccination Decisions among Patients with Mental Illness
Lorenz, R. A., Norris, M. M., Norton, L. C., & Westrick, S. C. (2013) · The International Journal of Psychiatry in Medicine · doi:10.2190/pm.46.1.a · Cited by 67
Ray, M. E., Coon, J. M., Al‐Jumaili, A. A., & Fullerton, M. (2019) · American Journal of Pharmaceutical Education · doi:10.5688/ajpe6983 · Cited by 81
“positive and negative factors associated with feelings of isolation.(20) The study was not powered to conduct a subanalysis of students by college of study.”
BackgroundChen, R. F., Stewart, W. F., Sun, J., Ng, K., & Yan, X. S. (2019) · Circulation Cardiovascular Quality and Outcomes · doi:10.1161/circoutcomes.118.005114 · Cited by 80
“More recent evidence 20 suggests that the rule of thumb is more likely to be one in at least a 5 to 9 ratio (ie, 5 to 9 outcome events for one feature in the model) to avoid unreliable modeling.”
BackgroundCai, S., Li, J., Gao, J., Pan, W., & Zhang, Y. (2022) · International Journal of Nursing Studies · doi:10.1016/j.ijnurstu.2022.104340 · Cited by 71
To IPO or Not To IPO: Risks, Uncertainty and the Decision to Go Public
Latham, S., & Braun, M. R. (2010) · British Journal of Management · doi:10.1111/j.1467-8551.2010.00707.x · Cited by 77
Bruce, J. P., Hui, A. B., Shi, W., Perez‐Ordoñez, B., Weinreb, I., Xu, W., Haibe‐Kains, B., Waggott, D. M., Boutros, P. C., O’Sullivan, B., et al. (2015) · Oncotarget · doi:10.18632/oncotarget.3005 · Cited by 65
“After fitting, the 4 miRNAs most strongly associated with DM were selected for inclusion in the risk score (RS) equation in order to maintain the number of events per variable (EVP) at ~five [39].”
Bosch, W., Heckman, M. G., Diehl, N. N., Shalev, J. A., Pungpapong, S., & Hellinger, W. C. (2011) · American Journal of Transplantation · doi:10.1111/j.1600-6143.2011.03618.x · Cited by 61
“showing the strongest single variable associations with the given endpoint were adjusted for in the multivariable analysis, allowing for no more than one variable in the model for every 10 patients who experienced the given endpoint (11).”
BackgroundMinor and Adult Domestic Sex Trafficking Risk Factors in Ohio
Chohaney, M. L. (2016) · Journal of the Society for Social Work and Research · doi:10.1086/685108 · Cited by 76
“Vittinghoof and McCulloch (2007) discovered the rule of 10 can be relaxed to five events per-variable with minimal effects on bias and efficiency.”
BackgroundHyperglycemia predicts poststroke infections in acute ischemic stroke
Zonneveld, T. P., Nederkoorn, P. J., Westendorp, W. F., Brouwer, M. C., van de Beek, D., Kruyt, N. D., For the PASS Investigators, PASS Investigators, Vermeij, J., Zock, E., et al. (2017) · Neurology · doi:10.1212/wnl.0000000000003811 · Cited by 61
Debray, T. P. A., Collins, G. S., Riley, R. D., Snell, K. I. E., Van Calster, B., Reitsma, J. B., & Moons, K. G. M. (2023) · BMJ · doi:10.1136/bmj-2022-071058 · Cited by 57
“However, some researchers have argued that an event per variable of 10 is too conservative 86 and others have suggested larger values of events per variable to avoid bias in estimated regression coefficients.”
BackgroundAlmadi, M. A., Mosli, M. H., Bohlega, M. S., Essa, M. A., AlDohan, M., Alabdallatif, T., AlSagri, T., Algahtani, F., & Mandil, A. (2015) · Saudi Journal of Gastroenterology · doi:10.4103/1319-3767.153814 · Cited by 60
“Using the rule of 10 outcome events per predictor variable,[11] and our wish to include up to 10 variables in the multivariable model, we estimated that 500 individuals would be needed to provide sufficient accuracy.”
van Walraven, C., Fergusson, D. A., Earle, C. C., Baxter, N. N., Alibhai, S. M. H., MacDonald, B., Forster, A. J., & Cagiannos, I. (2011) · Journal of Clinical Oncology · doi:10.1200/jco.2011.34.6379 · Cited by 61
Frerix, M., Stegbauer, J., Kreuter, A., & Weiner, S. M. (2014) · Arthritis Research & Therapy · doi:10.1186/ar4489 · Cited by 69
“Limitations of the study Whether the number of patients with plaque from the SSc and SLE patients was high enough to build two separate logistic regression models with a suitable number of events per variable, it was not for Cox regression and the analysis of both diseases was combined to ensure at least five patients…”
Chan, R. J., Northfield, S., Larsen, E. N., Mihala, G., Ullman, A. J., Hancock, P. A., Marsh, N. M., Gavin, N., Wyld, D., Allworth, A. M., et al. (2017) · Trials · doi:10.1186/s13063-017-2207-x · Cited by 69
“The final multivariable model breached our own self-imposed ten-events-per-variable rule, which was done with the intention to inform future studies (otherwise potentially interesting associations would not show up), and was deemed acceptable for a pilot study [35].”
Deenik, J., Tenback, D. E., Tak, E. C. P. M., Rutters, F., Hendriksen, I. J. M., & van Harten, P. N. (2019) · Schizophrenia Research · doi:10.1016/j.schres.2018.07.033 · Cited by 60
“tients recovered from MetS as a whole to make a valid corrected analysis (Vittinghoff and McCulloch, 2007).”
Uçkay, I., Kressmann, B., Malacarne, S., Toumanova, A., Jaafar, J., Lew, D., & Lipsky, B. A. (2018) · BMC Infectious Diseases · doi:10.1186/s12879-018-3253-z · Cited by 60
Factors Associated with the High Prevalence of Dementia in Older Aboriginal Australians
Radford, K. A., Lavrencic, L. M., Delbaere, K., Draper, B., Cumming, R. G., Daylight, G., Mack, H. A., Chalkley, S., Bennett, H. P., Garvey, G., et al. (2018) · Journal of Alzheimer s Disease · doi:10.3233/jad-180573 · Cited by 68
Geriatric Assessment in Older Adults with Multiple Myeloma
Wildes, T. M., Tuchman, S. A., Klepin, H. D., Mikhael, J. R., Trinkaus, K. M., Stockerl-Goldstein, K. E., Vij, R., & Colditz, G. A. (2018) · Journal of the American Geriatrics Society · doi:10.1111/jgs.15715 · Cited by 63
“Our limited sample size also did not allow for inclusion of all potential predictors in the multivariate model to avoid overfitting the model; indeed, some would argue that including three variables in our multivariate model could result in overfitting, while others have argued that as few as five events per predictor…”
BackgroundNørskov, A. K., Lange, T., Nielsen, E. E., Gluud, C. N., Winkel, P., Beyersmann, J., de Uña‐Álvarez, J., Torri, V., Billot, L., Putter, H., et al. (2020) · BMJ Evidence-Based Medicine · doi:10.1136/bmjebm-2019-111268 · Cited by 68
Cochet, H., Iriart, X., Allain-Nicolaï, A., Camaioni, C., Sridi, S., Nivet, H., Fournier, E., Dinet, M., Jalal, Z., Laurent, F., et al. (2019) · European Heart Journal - Cardiovascular Imaging · doi:10.1093/ehjci/jez068 · Cited by 65
“Assuming an expected prevalence of ventricular arrhythmia of about 17% based on prior publications,(10,14) the study was populated to enable the inclusion of three variables in multivariable models.(17) All statistical tests were two-tailed, and a P-value <0.”
MethodBrake, H. J. V. D., Walter, F. H., Rink, F. A., Essens, P. J. M. D., & van der Vegt, G. S. (2019) · Journal of Management Studies · doi:10.1111/joms.12539 · Cited by 68
“This multilevel structural equation model did not converge due to the large number of estimated parameters (see Cameron and Trivedi, 1998; Vittinghoff and McCulloch, 2007).”
BackgroundGeerts, M., Dooms, M., & Stas, L. (2021) · Sustainability · doi:10.3390/su13063148 · Cited by 75
Body Image and Gestational Weight Gain: A Prospective Study
Hill, B., Skouteris, H., McCabe, M. P., & Fuller‐Tyszkiewicz, M. (2013) · Journal of Midwifery & Women s Health · doi:10.1111/j.1542-2011.2012.00227.x · Cited by 72
Vitamin D Deficiency Predicts Decline in Kidney Allograft Function: A Prospective Cohort Study
Obi, Y., Hamano, T., Ichimaru, N., Tomida, K., Matsui, I., Fujii, N., Okumi, M., Kaimori, J.-Y., Yazawa, K., Kokado, Y., et al. (2013) · The Journal of Clinical Endocrinology & Metabolism · doi:10.1210/jc.2013-2421 · Cited by 59
“We used a propensity score (PS) of vitamin D sufficiency (25D 20 ng/mL) as a data-reduction technique by using a logistic regression model because the study size was too small to adjust for the potential confounders (Table 1) (27, 28).”
MethodSkaarup, K. G., Iversen, A. Z., Jørgensen, P. G., Olsen, F. J., Grove, G. L., Jensen, J. S., & Biering‐Sørensen, T. (2018) · European Heart Journal - Cardiovascular Imaging · doi:10.1093/ehjci/jey004 · Cited by 60
Hanazawa, T., Asayama, K., Watabe, D., Tanabe, A., Satoh, M., Inoue, R., Hara, A., Obara, T., Kikuya, M., Nomura, K., et al. (2018) · Journal of the American Heart Association · doi:10.1161/jaha.117.008509 · Cited by 59
Robustness of Clonogenic Assays as a Biomarker for Cancer Cell Radiosensitivity
Matsui, T., Nuryadi, E., Komatsu, S., Hirota, Y., Shibata, A., Oike, T., & Nakano, T. (2019) · International Journal of Molecular Sciences · doi:10.3390/ijms20174148 · Cited by 68
“It is accepted that approximately 5–10 events per variable are required to obtain reliable results from multiple regression [24,25].”
BackgroundMaheshwer, B., Wong, S. E., Polce, E. M., Paul, K. M., Forsythe, B., Bush‐Joseph, C. A., Bach, B. R., Yanke, A. B., Cole, B. J., Verma, N. N., et al. (2021) · Arthroscopy The Journal of Arthroscopic and Related Surgery · doi:10.1016/j.arthro.2021.04.058 · Cited by 58
Cardiovascular Disease Among Patients With AML and CHIP-Related Mutations
Arguelles, O. C., Schöffel, A. C., Capo‐Chichi, J., Abdel-Qadir, H. M., Schuh, A. C., Estrada, M. C., Liu, S., Gupta, V. K., Schimmer, A. D., Yee, K. W. L., et al. (2022) · JACC CardioOncology · doi:10.1016/j.jaccao.2021.11.008 · Cited by 64
“Also the ratio of predictors to outcomes was still within the acceptable range.(33) We also show that most patients with CHIPrelated mutations had concomitant non-CHIPrelated mutations.”
ResultShany, T., Wang, K., Liu, Y., Lovell, N., & Redmond, S. (2015) · Healthcare technology letters · doi:10.1049/htl.2015.0019 · Cited by 51
“For example, it has been suggested that the 1:10 rule may actually be relaxed in the case of logistic regression [64], which is noteworthy given the prevalence of this model type in SFRT research.”
Ausset, S., Auroy, Y., Verret, C., Benhamou, D., Vest, P., Cirodde, A., & Lenoir, B. (2010) · Anesthesiology · doi:10.1097/aln.0b013e3181eaacc4 · Cited by 58
Quality of first trimester risk prediction models for pre‐eclampsia: a systematic review
Brunelli, V., & Prefumo, F. (2015) · BJOG An International Journal of Obstetrics & Gynaecology · doi:10.1111/1471-0528.13334 · Cited by 64
Adverse Drug Reactions in a Population of Hospitalized Very Elderly Patients
Tangiisuran, B., Davies, G., Wright, J. E., & Rajkumar, C. (2012) · Drugs & Aging · doi:10.1007/bf03262282 · Cited by 65
“The number of variables included in the final multivariate regression followed convention whereby one variable was entered for every ten patients recruited.([30]) The selection of variables (potential risk factors for ADRs) for the univariate analysis was based on the prior studies.”
MethodWhen Societal Norms and Social Identity Collide: The Race Talk Dilemma for Racial Minority Children
Pauker, K., Apfelbaum, E. P., & Spitzer, B. (2015) · Social Psychological and Personality Science · doi:10.1177/1948550615598379 · Cited by 58
“We met requirements for logistic regression in the primary analyses presented below (see Vittinghoff & McCullock, 2007).”
MethodBarison, A., Aimo, A., Ortalda, A., Todiere, G., Grigoratos, C., Passino, C., Camici, P. G., Aquaro, G. D., & Emdin, M. (2017) · International Journal of Cardiology · doi:10.1016/j.ijcard.2017.10.043 · Cited by 65
Arenas, Á., Serrano, C. A., Quiñones, L. A., Harris, P. R., Sandoval, M., Lavanderos, M. A., Sepúlveda, R., Maquilón, S., Echeverría, A., Ríos, C., et al. (2019) · Scientific Reports · doi:10.1038/s41598-019-56399-7 · Cited by 58
Gallioli, A., Boissier, R., Territo, A., Vila-Reyes, H., Sanguedolce, F., Gaya, J. M., Regis, F., Subiela, J. D., Palou, J., & Breda, A. (2020) · Journal of Endourology · doi:10.1089/end.2019.0750 · Cited by 58
Understanding Wellbeing Among College Music Students and Amateur Musicians in Western Switzerland
Philippe, R. A., Kosirnik, C., Vuichoud, N., Williamon, A., & von Roten, F. C. (2019) · Frontiers in Psychology · doi:10.3389/fpsyg.2019.00820 · Cited by 73
“Finally, a multivariate logistic regression was performed to compare the HEM and non-HEM groups on the four QoL dimensions after controlling for sex and age; the model meets Vittinghoff and McCulloch (2007) rule of 5 – 9 outcome events per predictor variable.”
MethodBrunet‐Ratnasingham, E., Anand, S. P., Gantner, P., Dyachenko, A., Moquin‐Beaudry, G., Brassard, N., Beaudoin-Bussières, G., Pagliuzza, A., Gasser, R., Benlarbi, M., et al. (2021) · Science Advances · doi:10.1126/sciadv.abj5629 · Cited by 57
Ladouceur, M., Bruaene, A. V. D., Kauling, R. M., Budts, W., Roos‐Hesselink, J. W., Albert, S. V., Pérez, I. S., Sarubbi, B., Fusco, F., Gallego, P., et al. (2022) · European Heart Journal · doi:10.1093/eurheartj/ehac288 · Cited by 57
Gutman, J., Mwandama, D., Wiegand, R. E., Ali, D., Mathanga, D. P., & Skarbinski, J. (2013) · Journal of Infectious Diseases · doi:10.1093/infdis/jit276 · Cited by 56
Nijsse, R., Ploeger, H. W., Wagenaar, J. A., & Mughini‐Gras, L. (2016) · Parasitology Research · doi:10.1007/s00436-016-5242-8 · Cited by 61
“As the main focus of this study was on Toxocara, further results were presented for this specific helminth only.”
BackgroundInfluentialDahlman, D., Håkansson, A., Kral, A. H., Wenger, L. D., Ball, E. L., & Novak, S. P. (2016) · Substance Abuse · doi:10.1080/08897077.2016.1263592 · Cited by 65
Many places called home: the adaptive value of seasonal adjustments in range fidelity
Lafontaine, A., Drapeau, P., Fortin, D., & St‐Laurent, M. (2017) · Journal of Animal Ecology · doi:10.1111/1365-2656.12645 · Cited by 69
Association of napping with incident cardiovascular events in a prospective cohort study
Häusler, N., Haba‐Rubio, J., Heinzer, R. C., & Marques-Vidal, P. M. (2019) · Heart · doi:10.1136/heartjnl-2019-314999 · Cited by 59
Presidential Instability in Latin America: Why Institutionalized Parties Matter
Martínez, C. A. (2020) · Government and Opposition · doi:10.1017/gov.2020.18 · Cited by 64
“We must bear in mind that statistical analyses are sensitive to the low number of events per variable (EPV) ratio (Lanser 2015; Vittinghoff and McCulloch 2007).”
BackgroundLund, L. K., Vik, T., Skranes, J., Brubakk, A., & Indredavik, M. S. (2011) · Acta Paediatrica · doi:10.1111/j.1651-2227.2010.02111.x · Cited by 56
“We adjusted for potential confounders separately to avoid model over fitting (23).”
MethodIncidence and predictors of portal and splenic vein thrombosis after pure laparoscopic splenectomy
de’Angelis, N., Abdalla, S., Lizzi, V., Esposito, F., Genova, P., Roy, L., Galactéros, F., Luciani, A., & Brunetti, F. (2017) · Surgery · doi:10.1016/j.surg.2017.07.016 · Cited by 56
Muthalaly, R. G., Kwong, R. Y. K., John, R. M., van der Geest, R. J., Tao, Q., Schaeffer, B. N., Tanigawa, S., Nakamura, T., Kaneko, K., Tedrow, U. B., et al. (2018) · JACC. Cardiovascular imaging · doi:10.1016/j.jcmg.2018.07.003 · Cited by 57
“However, the event rate was still low, with only 18 arrhythmic events and 17 EHFE in total, which limits the complexity of models that can be built because of the risk for overfitting data and necessitates the use of model probability scores (21).”
BackgroundMapping Associations Between Gait Decline and Fall Risk in Mild Cognitive Impairment
Pieruccini‐Faria, F., Sarquis‐Adamson, Y., Antón‐Rodrigo, I., Noguerón‐García, A., Bray, N. W., Camicioli, R. M., Hunter, S. W., Speechley, M. R., McIlroy, B., & Montero‐Odasso, M. (2019) · Journal of the American Geriatrics Society · doi:10.1111/jgs.16265 · Cited by 68
“Stability of the regression models was assured by having a minimum of six events per fall outcome.(46) Independence of gait parameters and their membership to lower or higher levels of gait control were determined by principal components analysis (PCA) and factor analysis.”
MethodSchumacher, M., Holländer, N., Schwarzer, G., Binder, H., & Sauerbrei, W. (2012) · Handbook of Statistics in Clinical Oncology, Third Edition · doi:10.1201/b11800-30 · Cited by 56
Crash risk factors associated with injury severity of teen drivers
Duddu, V. R., Kukkapalli, V. M., & Pulugurtha, S. S. (2019) · IATSS Research · doi:10.1016/j.iatssr.2018.08.003 · Cited by 55
Zhao, H., You, J., Peng, Y., & Feng, Y. (2021) · Frontiers in Surgery · doi:10.3389/fsurg.2021.634629 · Cited by 60
“For traditional multivariate logistic regression, at least 5–10 outcome events per input variable have been recommended (19).”
BackgroundBias in Odds Ratios From Logistic Regression Methods With Sparse Data Sets
Gosho, M., Ohigashi, T., Nagashima, K., Ito, Y., & Maruo, K. (2021) · Journal of Epidemiology · doi:10.2188/jea.je20210089 · Cited by 81
Adverse pregnancy outcomes after exposure to methylphenidate or atomoxetine during pregnancy
Bro, S. P., Kjaersgaard, M. I. S., Parner, E. T., Sørensen, M. J., Olsen, J., Bech, B. H., Pedersen, L. H., Christensen, J., & Vestergaard, M. (2015) · Clinical Epidemiology · doi:10.2147/clep.s72906 · Cited by 55
“The use of ADHD medication among women of fertile age has increased almost 100-fold during the past 12 years in Denmark, but little is known about the safety of these drugs during pregnancy.(2,8,9) Animal studies have shown that high doses of MPH given during pregnancy are associated with an increased risk of malforma…”
BackgroundDavis, J. L., Worodria, W. O., Kisembo, H. N., Metcalfe, J. Z., Cattamanchi, A., Kawooya, M. G., Kyeyune, R. B., den Boon, S., Powell, K., Okello, R., et al. (2010) · PLoS ONE · doi:10.1371/journal.pone.0009859 · Cited by 63
“[10] We used conventional logistic regression to generate predicted probabilities of TB, plotted the fraction of true positives versus the fraction of false positives at various thresholds of predicted probability as a receiver operating characteristic curve, and measured model accuracy using the concordance (C) stati…”
MethodØsthus, A. A., Aarstad, A. K. H., Olofsson, J. S., & Aarstad, H. J. (2013) · JAMA Otolaryngology–Head & Neck Surgery · doi:10.1001/jamaoto.2013.1056 · Cited by 56
“unreliable results found when the rate of events per prognostic variable fall to low.(17) Dichotomized versions of all the enabled HRQOL sum scores were computed,(12) comparing the upper quartile with the lower 3 quartiles for the symptom scores and the lower quartile with the upper 3 quartiles for the functional and…”
MethodCan Social Role Theory Explain Gender Differences in Facebook Usage?
Lin, X., Li, Y., Califf, C. B., & Featherman, M. (2013) · 2013 46th Hawaii International Conference on System Sciences · doi:10.1109/hicss.2013.125 · Cited by 52
“The ratio of sample size to independent variables, including gender, is 24:1, which exceeds the rule of thumb of 10 for the sample size requirement in a regression analysis [52].”
BackgroundConfounder summary scores when comparing the effects of multiple drug exposures
Cadarette, S. M., Gagne, J. J., Solomon, D. H., Katz, J. N., & Stürmer, T. (2010) · Pharmacoepidemiology and Drug Safety · doi:10.1002/pds.1845 · Cited by 55
“As highlighted previously, one advantage of EPS over conventional multivariable regression is in reducing the number of covariate terms in the outcome model.(17,18) Thus, EPS may be a particularly relevant method when exposures are prevalent but disease is rare.”
Ekström, K., Lehtonen, J. Y. A., Hänninen, H., Kandolin, R. M., Kivistö, S., & Kupari, M. M. (2016) · Journal of the American Heart Association · doi:10.1161/jaha.115.003040 · Cited by 56
Matthews, M., Rathleff, M. S., Claus, A. P., McPoil, T. G., Nee, R., Crossley, K. M., & Vicenzino, B. (2016) · British Journal of Sports Medicine · doi:10.1136/bjsports-2016-096545 · Cited by 71
“Studies de fi ned a successful outcome using a predetermined amount of improvement in pain scores, questionnaire and/or a rating on a global rating of changes scale to stratify respondents into successful or unsuccessful outcomes.”
BackgroundMethodOmondi, F. H., Chandrarathna, S., Mujib, S., Brumme, C. J., Jin, S. W., Sudderuddin, H., Miller, R. L., Rahimi, A., Laeyendecker, O. B., Bonner, P., et al. (2019) · Journal of Virology · doi:10.1128/jvi.01832-18 · Cited by 60
“To minimize the risk of overfitting, we followed the minimum criterion of 5 degrees of freedom per included variable (62, 63) (i.”
MethodSeptal Myectomy in Hypertrophic Cardiomyopathy: National Outcomes of Concomitant Mitral Surgery
Holst, K. A., Hanson, K. T., Ommen, S. R., Nishimura, R. A., Habermann, E. B., & Schaff, H. V. (2019) · Mayo Clinic Proceedings · doi:10.1016/j.mayocp.2018.07.022 · Cited by 54
Fabrès, V., Dossin, O., Reif, C., Campos, M. A. L., Freiche, V., Maurey, C., Pilot‐Storck, F., Desquilbet, L., & Benchekroun, G. (2019) · Journal of Veterinary Internal Medicine · doi:10.1111/jvim.15421 · Cited by 55
“If more than 2 potential confounders were identified, only the 2 with the lowest P-value testing the association with the exposure of interest were selected, to have no more than 4 variables included into each multivariable model.(29) If a potential risk factor remained significantly associated with the outcome in its…”
BackgroundImportance of Households in Business Disaster Recovery
Watson, M., Xiao, Y., HELGESON, J. F., & Dillard, M. (2020) · Natural Hazards Review · doi:10.1061/(asce)nh.1527-6996.0000393 · Cited by 58
“The EPV at its lowest, however, never drops below nine, which research suggests still has a similar risk of bias to the ten EPV rule of thumb (Vittinghoff and McCulloch 2007).”
ResultSauder, J. D., & Rachlow, J. L. (2013) · Forest Ecology and Management · doi:10.1016/j.foreco.2013.11.029 · Cited by 56
“We restricted our models to 62 independent variables because of relatively small event-to-variable ratios (Peduzzi et al., 1996; Vittinghoff and McCulloch, 2007), and screened metrics for high colinearity (Pearson’s correlation coefficient P0.”
MethodThe British Trans-Atlantic Slave Trade 1650-1775
Bean, R. N. (1972) · The Journal of Economic History · doi:10.1017/s0022050700111520 · Cited by 58
Risk Factors for the Number of Sustained Injuries in Artisanal and Small-Scale Mining Operation
Ajith, M. M., Ghosh, A. K., & Jansz, J. (2020) · Safety and Health at Work · doi:10.1016/j.shaw.2020.01.001 · Cited by 68
“That is, in logistic regression analysis, the events per variable must be > 20 to avoid model instability [40,41].”
BackgroundStefanello, D., Avallone, G., Ferrari, R., Roccabianca, P., & Boracchi, P. (2011) · Journal of Veterinary Internal Medicine · doi:10.1111/j.1939-1676.2011.00822.x · Cited by 54
“The prognostic role of covariates measured on a continuous scale (eg, tumor size, age) was investigated by regression cubic splines, to allow the identification of possible nonlinear relationships.(33) For categorical variables results are shown in terms of hazard ratios (HR) and their 95% confidence intervals (CI).”
BackgroundMethodResultInfluentialBrainstem infarcts predict REM sleep behavior disorder in acute ischemic stroke
Tang, W. K., Hermann, D. M., Chen, Y. K., Liang, H. J., Liu, X. X., Chu, C. W., Ahuja, A. T., Abrigo, J., Mok, V., Ungvari, G. S., et al. (2014) · BMC Neurology · doi:10.1186/1471-2377-14-88 · Cited by 55
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.