Standardizing criteria for logistic regression models

Prediction equations of forced oscillation technique: the insidious role of collinearity

Many studies have reported reference data for forced oscillation technique (FOT) in healthy children. The prediction equation of FOT parameters were derived from a multivariable regression model examining the effect of age, gender, weight and height on each parameter. As many of these variables are likely to be correlated, collinearity might have affected the accuracy of the model, potentially resulting in misleading, erroneous or difficult to interpret conclusions.

The aim of this work was: To review all FOT publications in children since 2005 to analyze whether collinearity was considered in the construction of the published prediction equations. Then to compare these prediction equations with our own study. And to analyse, in our study, how collinearity between the explanatory variables might affect the predicted equations if it was not considered in the model. The results showed that none of the ten reviewed studies had stated whether collinearity was checked for. Half of the reports had also included in their equations variables which are physiologically correlated, such as age, weight and height. The predicted resistance varied by up to 28% amongst these studies. And in our study, multicollinearity was identified between the explanatory variables initially considered for the regression model (age, weight and height). Ignoring it would have resulted in inaccuracies in the coefficients of the equation, their signs (positive or negative), their 95% confidence intervals, their significance level and the model goodness of fit. In Conclusion with inaccurately constructed and improperly reported models, understanding the results and reproducing the models for future research might be compromised.

Quality of reporting of multivariable logistic regression models in Chinese clinical medical journals

Multivariable logistic regression (MLR) has been increasingly used in Chinese clinical medical research during the past few years. However, few evaluations of the quality of the reporting strategies in these studies are available.To evaluate the reporting quality and model accuracy of MLR used in published work, and related advice for authors, readers, reviewers, and editors.A total of 316 articles published in 5 leading Chinese clinical medical journals with high impact factor from January 2010 to July 2015 were selected for evaluation. Articles were evaluated according 12 established criteria for proper use and reporting of MLR models.Among the articles, the highest quality score was 9, the lowest 1, and the median 5 (4-5). A total of 85.1% of the articles scored below 6. No significant differences were found among these journals with respect to quality score (χ = 6.706, P = .15). More than 50% of the articles met the following 5 criteria: complete identification of the statistical software application that was used (97.2%), calculation of the odds ratio and its confidence interval (86.4%), description of sufficient events (>10) per variable, selection of variables, and fitting procedure (78.2%, 69.3%, and 58.5%, respectively). Less than 35% of the articles reported the coding of variables (18.7%). The remaining 5 criteria were not satisfied by a sufficient number of articles: goodness-of-fit (10.1%), interactions (3.8%), checking for outliers (3.2%), collinearity (1.9%), and participation of statisticians and epidemiologists (0.3%). The criterion of conformity with linear gradients was applicable to 186 articles; however, only 7 (3.8%) mentioned or tested it.The reporting quality and model accuracy of MLR in selected articles were not satisfactory. In fact, severe deficiencies were noted. Only 1 article scored 9. We recommend authors, readers, reviewers, and editors to consider MLR models more carefully and cooperate more closely with statisticians and epidemiologists. Journals should develop statistical reporting guidelines concerning MLR.

Evaluation of logistic regression reporting in current obstetrics and gynecology literature

OBJECTIVE

To evaluate the quality of logistic regression reporting in the obstetrics and gynecology literature.

METHODS

All original papers published in 2005 and 2006 in four leading obstetrics and gynecology journals were manually searched for the use of logistic regression. One hundred four articles that used logistic regression were randomly selected (13 from each journal and each year) and evaluated according to previously established criteria for reporting logistic regression analyses. Rates of compliance with these criteria were calculated separately for each journal and weighted according to the number of articles using logistic regression in each of the journals in the same period to obtain an overall estimate.

RESULTS

Logistic regression was used in 34.2% of all original research articles (724 of 2,234) in the four journals for the study period. Statistical significance of estimates was reported in 96% of examined articles. Criteria of variable selection for the logistic regression model were reported in 76% of articles, and coding of variables was described in 83%. Overfitting (models with too many variables for the number of outcome events) occurred in 57% of studies. The majority of examined articles insufficiently reported information for the remaining criteria-testing for interactions (18%), conformity to a linear gradient of continuous variables (9%), goodness of fit (3.6%), assessment of multi-collinearity (0.46%), and validation of the model (0%).

CONCLUSION

Logistic regression has become a standard statistical method in obstetrics and gynecology literature. Although some standards are mostly fulfilled, there is still considerable room for improvement.

LEVEL OF EVIDENCE

III.

Regresión logística en la literatura psiquiátrica: evaluación de los artículos publicados entre 2002 y 2005 en una prominente revista

INTRODUCCIÓN: Muchos artículos de revistas especializadas en psiquiatría usan la regresión logística para identificar el efecto de variables independientes sobre la probabilidad de ocurrencia de un evento. Evaluaciones de artículos publicados en otras especialidades médicas han mostrado que el reporte de la regresión logística es incompleto en muchos casos, lo cual implica dificultades para su interpretación. OBJETIVO: Evaluar la calidad, en cuanto al análisis de regresión logística y a al cumplimiento de criterios de validez interna de los artículos publicados en una revista de alto factor de impacto en psiquiatría. METODOLOGÍA: Dos revisores independientes seleccionaron (búsqueda manual) los artículos publicados en Archives of General Psychiatry (2002-2005) que usaran regresión logística y los evaluaron con criterios de validez interna del Journal of American Medical Association (JAMA) y un instrumento diseñado para valorar la calidad de los modelos logísticos (CML). RESULTADOS: De 121 artículos evaluados, 85 (70,2%) cumplían criterios de validez interna de JAMA. En relación con el instrumento CML, el criterio más reportado fue la codificación de las variables independientes (90,9%), seguido por el reporte del proceso de selección de variables independientes incluidas al inicio del estudio (87,6%) y por la inclusión del RR u OR del modelo con su respectivo IC (82,6%). El criterio menos reportado fue la evaluación de la bondad del ajuste (9,1%), seguido por el reporte del proceso de ajuste del modelo (24,8%). CONCLUSIONES: Aunque la mayoría de los artículos cumplen con altos estándares editoriales, la forma en que se reporta la regresión logística debe ser mejorada.

A review of two journals found that articles using multivariable logistic regression frequently did not report commonly recommended assumptions

BACKGROUND AND OBJECTIVE

To examine if commonly recommended assumptions for multivariable logistic regression are addressed in two major epidemiological journals.

METHODS

Ninety-nine articles from the Journal of Clinical Epidemiology and the American Journal of Epidemiology were surveyed for 10 criteria: six dealing with computation and four with reporting multivariable logistic regression results.

RESULTS

Three of the 10 criteria were addressed in 50% or more of the articles. Statistical significance testing or confidence intervals were reported in all articles. Methods for selecting independent variables were described in 82%, and specific procedures used to generate the models were discussed in 65%. Fewer than 50% of the articles indicated if interactions were tested or met the recommended events per independent variable ratio of 10:1. Fewer than 20% of the articles described conformity to a linear gradient, examined collinearity, reported information on validation procedures, goodness-of-fit, discrimination statistics, or provided complete information on variable coding. There was no significant difference (P>.05) in the proportion of articles meeting the criteria across the two journals.

CONCLUSION

Articles reviewed frequently did not report commonly recommended assumptions for using multivariable logistic regression.

Simplified therapy with zidovudine, lamivudine, and abacavir for very nonadherent, treatment-failing patients

OBJECTIVE

To assess the effectiveness of a simplified therapy for very nonadherent patients who had previously failed with HAART.

METHOD

We performed a prospective open-label study of antiretroviral-experienced patients. Dosing schedule comprised (co-formulated) zidovudine, lamivudine, and abacavir bid. Eligible patients had to have plasma HIV RNA >5000 copies/mL, previous therapy, and very poor adherence to the medication regimen.

RESULTS

Eighty-five patients were included (mean viral load, 4.4 log/mL; mean CD4, 240 cells/mL; IDUs, 78%; methadone maintenance program, 42%; AIDS, 28%). Number of previous therapies: one, 53%; two, 28%; three or more, 19%. In the intent-to-treat analysis at 1 year, 38 patients (44.7%) achieved viral load below 500 copies/mL. Adherence greater than 90% of prescribed drugs was reported in 49% of patients, adverse events were reported in 17.6%, mortality in 6%, and lost to follow-up in 26%. The factors associated with virologic failure were nonadherence (odds ratio [OR], 4.4; 95% CI 1.5-12.3), baseline CD4 cell count <200 cells/mL (OR, 3.4; 95% CI 1.3-8.9; p =.01), and more than one previous treatment (OR, 2.7; 95% CI 1.1-6.9).

CONCLUSION

Regarding previously very nonadherent patients, this simplified combination therapy containing three NRTIs obtained satisfactory results in ART-experienced patients. However, more aggressive interventions to enhance adherence are needed to improve results.

An appraisal of multivariable logistic models in the pulmonary and critical care literature

OBJECTIVE

Multivariable modeling techniques are appearing in today's medical literature with increasing frequency. Improper reporting of these statistical models can potentially make the results of a study inaccurate, misleading, or difficult to interpret. We performed a manual literature search of five international pulmonary and critical care journals to determine the accuracy in the reporting of logistic regression modeling strategies.

DESIGN

We examined all of the published manuscripts for 12 potential limitations in the reporting of important statistical methodologies over a 6-month period from July 1, 2000, until December 31, 2000.

RESULTS

Of the 81 articles that included multivariable logistic regression analyses, only 65% (53 analyses) properly reported the coding classification of pertinent independent variables that were included in the final model. An odds ratio and confidence interval were reported for the independent variables included in the final model for 79% (64 analyses) and 74% (60 analyses), respectively. Only 12% (10 articles) referenced whether interaction terms or effect modifications were examined, 1% (1 article) reported testing for collinearity, and only 16% (13 articles) included a goodness-of-fit analysis of the logistic model. The type of statistical package was reported in 69% (56 articles). Finally, approximately 39% of the articles (22 of 57) may have overfit the logistic regression model, leading to potentially unreliable regression coefficients and odds ratios.

CONCLUSIONS

Our results indicate that the reporting of multivariable logistic regression analyses in the pulmonary and critical care literature is often incomplete, therefore making it difficult for the reader to accurately interpret the manuscript. We recommend the implementation of adequate guidelines that will lead to overall improvements in the reporting and possibly to the conducting of multivariable analyses in the pulmonary medicine and critical care medicine literature.

Logistic regression in the medical literature: standards for use and reporting, with particular attention to one medical domain

Logistic regression (LR) is a widely used multivariable method for modeling dichotomous outcomes. This article examines use and reporting of LR in the medical literature by comprehensively assessing its use in a selected area of medical study. Medline, followed by bibliography searches, identified 15 peer-reviewed English-language articles with original data, employing LR, published between 1985 and 1999, pertaining to patient interest in genetic testing for cancer susceptibility. Articles were examined for each of 10 criteria for proper use and reporting of LR models. Substantial shortcomings were found in both use of LR and reporting of results. For many studies, the ratio of the number of outcome events to predictor variables (events per variable) was sufficiently small to call into question the accuracy of the regression model. Additionally, no studies reported validation analysis, regression diagnostics, or goodness-of-fit measures. It is recommended that authors, reviewers, and editors pay greater attention to guidelines concerning the use and reporting of LR models.