Relaxing the rule of ten events per variable in logistic and Cox regression.

Department of Epidemiology and Biostatistics, University of California-San Francisco, 185 Berry Street, San Francisco, CA 94107, USA.
American Journal of Epidemiology (Impact Factor: 4.98). 04/2007; 165(6):710-8. DOI: 10.1093/aje/kwk052
Source: PubMed

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.

1 Follower
  • [Show abstract] [Hide abstract]
    ABSTRACT: Patient satisfaction is an important metric of health-care quality. Accidental awareness under general anaesthesia (AAGA) is a serious complication of anaesthesia care which may go unrecognised in the immediate perioperative period but leads to long-term psychological harm for affected patients. The SNAP-1 study aimed to measure patient satisfaction with anaesthesia care and the incidence of AAGA, reported on direct questioning within 24 h of surgery, in a large multicentre cohort. A secondary aim of SNAP-1 was to test the effectiveness of a new network of Quality Audit and Research Coordinators in NHS anaesthetic departments, to achieve widespread study participation and high patient recruitment rates. This manuscript describes the study methodology. SNAP-1 was a prospective observational cohort study. The study protocol was approved by the National Research Ethics Service. All UK NHS hospitals with anaesthetic departments were invited to participate. Adult patients undergoing any type of non-obstetric surgery were recruited in participating hospitals on 13th and 14th May 2014. Demographic data were collected by anaesthetists providing perioperative care. Patients were then approached within 24 h of surgery to complete two questionnaires-the Bauer patient satisfaction questionnaire (to measure patient reported outcome) and the modified Brice questionnaire (to detect possible accidental awareness). Completion of postoperative questionnaires was taken as evidence of implied consent. Results were recorded on a standard patient case report form, and local investigators entered anonymised data into an electronic database for later analysis by the core research team. Preliminary analyses indicate that over 15,000 patients were recruited across the UK, making SNAP-1 the largest NIHR portfolio-adopted study in anaesthesia to date. Both descriptive and analytic epidemiological analyses will be used to answer specific questions about the patient perception of anaesthesia care overall and in surgical sub-specialties and to determine the incidence of AAGA. The SNAP-1 study recruited a large number of UK hospitals and thousands of perioperative patients using newly established networks in the UK anaesthetic profession. The results will provide benchmarking information to aid interpretation of patient satisfaction data and also determine the incidence of AAGA reported on a single postoperative visit.
    12/2015; 4(1). DOI:10.1186/s13741-015-0011-2
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Objectives: This study intends to evaluate whether the belief that condoms are 100% effective in protecting against HIV infection is associated with sexual risk behaviours among youth. Methods: A cross-sectional study was performed in representative samples of high-school students in the Philippines, El Salvador and Peru. Participants completed a self-administered questionnaire. Students were asked about the risk of HIV transmission if one has sex using condoms. They were also asked to indicate whether they had ever had sexual relations and whether they used a condom in their first sexual relation. The sample was composed of 8994 students, aged 13–18. Results: One out of seven adolescents believed condoms are 100% effective (safe-sex believers). Those adolescents were 82% more likely to have had sex than those without such belief, after adjusting for confounders (OR=1.82; 95% CI 1.51 to 2.21). On the contrary, no association was found between risk perception and condom use. Subgroup and sensitivity analyses produced similar results. Conclusions: This is, to the best of our knowledge, the first study conducted specifically to evaluate this phenomenon and that has used the same questionnaire and the same data collection protocol in three different developing countries from Asia, Central and South America. These results reasonably suggest that there could be an association between safe sex beliefs and sexual initiation. Longitudinal studies are needed to better understand this possible association as it could influence how to better promote sexual health.
    BMJ Open 04/2015; 5(4). DOI:10.1136/bmjopen-2015-007826 · 2.06 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: No predictive models for long-term mortality in critically ill patients with acute kidney injury (AKI) exist. We aimed to develop and validate two predictive models for one-year mortality in patients with AKI based on data 1) on intensive care unit (ICU) admission and 2) on the third day (D3) in the ICU. This substudy of the FINNAKI study comprised 774 patients with early AKI (diagnosed within 24 hours of ICU admission). We selected predictors a priori based on previous studies, clinical judgment, and differences between one-year survivors and non-survivors in patients with AKI. We validated the models internally with bootstrapping. Of 774 patients, 308 [39.8%, 95% confident interval (CI) 36.3-43.3] died during one year. Predictors of one-year mortality on admission were: advanced age, diminished premorbid functional performance, co-morbidities, emergency admission, and resuscitation or hypotension preceding ICU admission. The area under the receiver operating characteristic curve (AUC) (95% CI) for the admission model was 0.76 (0.72-0.79) and the mean bootstrap-adjusted AUC 0.75 (0.74-0.75). Advanced age, need for mechanical ventilation on D3, number of co-morbidities, higher modified SAPS II score, the highest bilirubin value by D3, and the lowest base excess value on D3 remained predictors of one-year mortality on D3. The AUC (95% CI) for the D3 model was 0.80 (0.753-0.846) and by bootstrapping 0.79 (0.77-0.80). The prognostic performance of the admission data-based model was acceptable, but not good. The D3 model for one-year mortality performed fairly well in patients with early AKI.
    Critical care (London, England) 03/2015; 19(1):125. DOI:10.1186/s13054-015-0848-2