An overview of the statistical methods reported by studies using the Canadian community health survey

Say farewell to bland regression reporting: Three forest plot variations for visualizing linear models

Regression ranks among the most popular statistical analysis methods across many research areas, including psychology. Typically, regression coefficients are displayed in tables. While this mode of presentation is information-dense, extensive tables can be cumbersome to read and difficult to interpret. Here, we introduce three novel visualizations for reporting regression results. Our methods allow researchers to arrange large numbers of regression models in a single plot. Using regression results from real-world as well as simulated data, we demonstrate the transformations which are necessary to produce the required data structure and how to subsequently plot the results. The proposed methods provide visually appealing ways to report regression results efficiently and intuitively. Potential applications range from visual screening in the model selection stage to formal reporting in research papers. The procedure is fully reproducible using the provided code and can be executed via free-of-charge, open-source software routines in R.

Statement on methods in sport injury research from the 1st METHODS MATTER Meeting, Copenhagen, 2019

High quality sports injury research can facilitate sports injury prevention and treatment. There is scope to improve how our field applies best practice methods—methods matter (greatly!). The 1st METHODS MATTER Meeting, held in January 2019 in Copenhagen, Denmark, was the forum for an international group of researchers with expertise in research methods to discuss sports injury methods. We discussed important epidemiological and statistical topics within the field of sports injury research. With this opinion document, we provide the main take-home messages that emerged from the meeting.

Statement on Methods in Sport Injury Research From the First METHODS MATTER Meeting, Copenhagen, 2019

High-quality sports injury research can facilitate sports injury prevention and treatment. There is scope to improve how our field applies best-practice methods-methods matter (greatly!). The first METHODS MATTER meeting, held in January 2019 in Copenhagen, Denmark, was the forum for an international group of researchers with expertise in research methods to discuss sports injury methods. We discussed important epidemiological and statistical topics within the field of sports injury research. With this opinion document, we provide the main take-home messages that emerged from the meeting. Meeting participants agreed that the definition of sport injury depends on the research question and context. It was considered essential to be explicit about the goal of the research effort and to use frameworks to illustrate the assumptions that underpin measurement and the analytical strategy. Complex systems were discussed to illustrate how potential risk factors can interact in a nonlinear way. This approach is often a useful alternative to identifying single risk factors. Investigating changes in exposure status over time is important when analyzing sport injury etiology, and analyzing recurrent injury, subsequent injury, or injury exacerbation remains challenging. The choice of statistical model should consider the research question, injury measure (eg, prevalence, incidence), type and granularity of injury data (categorical or continuous), and study design. Multidisciplinary collaboration will be a cornerstone for future high-quality sport injury research. Working outside professional silos in a diverse, multidisciplinary team benefits the research process, from the formulation of research questions and designs to the statistical analyses and dissemination of study results in implementation contexts. This article has been copublished in the British Journal of Sports Medicine and the Journal of Orthopaedic & Sports Physical Therapy. J Orthop Sports Phys Ther 2020;50(5):226-233. doi:10.2519/jospt.2020.9876.

Conducting gender-based analysis of existing databases when self-reported gender data are unavailable: the GENDER Index in a working population

OBJECTIVES

Growing attention has been given to considering sex and gender in health research. However, this remains a challenge in the context of retrospective studies where self-reported gender measures are often unavailable. This study aimed to create and validate a composite gender index using data from the Canadian Community Health Survey (CCHS).

METHODS

According to scientific literature and expert opinion, the GENDER Index was built using several variables available in the CCHS and deemed to be gender-related (e.g., occupation, receiving child support, number of working hours). Among workers aged 18-50 years who had no missing data for our variables of interest (n = 29,470 participants), propensity scores were derived from a logistic regression model that included gender-related variables as covariates and where biological sex served as the dependent variable. Construct validity of propensity scores (GENDER Index scores) were then examined.

RESULTS

When looking at the distribution of the GENDER Index scores in males and females, they appeared related but partly independent. Differences in the proportion of females appeared between groups categorized according to the GENDER Index scores tertiles (p < 0.0001). Construct validity was also examined through associations between the GENDER Index scores and gender-related variables identified a priori such as choosing/avoiding certain foods because of weight concerns (p < 0.0001), caring for children as the most important thing contributing to stress (p = 0.0309), and ability to handle unexpected/difficult problems (p = 0.0375).

CONCLUSION

The GENDER Index could be useful to enhance the capacity of researchers using CCHS data to conduct gender-based analysis among populations of workers.

Risk Factors for Bicycling Injuries in Children and Adolescents: A Systematic Review

CONTEXT

Child and adolescent bicycling is beneficial, but injuries occur and can be severe and costly.

OBJECTIVE

To systematically review the individual and environmental factors associated with bicycling injury risk in children and adolescents.

DATA SOURCES

Fourteen electronic databases were searched.

STUDY SELECTION

Two authors independently assessed potentially relevant articles for eligibility. The inclusion criteria were as follows: bicyclists younger than 20 years old; examined individual and environmental characteristics of bicycling crashes; compared injured and uninjured bicyclists or bicyclists with different types or severity of injury; study designs with a predetermined comparison group; and published in English from January 1990 to May 2015. The exclusion criteria were outcomes related to helmet use, helmet legislation, or mountain biking, and comparisons of census-based injury rates.

DATA EXTRACTION

Data on study design, setting, population, injury definitions, injury risk factors, and results were extracted. Risk of bias was assessed by using the Newcastle-Ottawa Scales.

RESULTS

Fourteen articles were included. Lower socioeconomic status, riding on the road, riding in rural compared with urban areas, and riding on the sidewalk were associated with bicycling injury. Bicycling safety education did not protect children against future injury. Injuries related to a motor vehicle collision were more severe than other bicycling injuries.

LIMITATIONS

Study heterogeneity prevented meta-analyses. Study quality was affected by inadequate definitions of study groups and self-reported data.

CONCLUSIONS

Lower socioeconomic status and riding location were associated with bicycling injury and severity increased with motor vehicle collisions. The bicycling environment is a promising avenue for prevention.

Visualization of the IMIA Yearbook of Medical Informatics Publications over the Last 25 Years

BACKGROUND

The last 25 years have been a period of innovation in the area of medical informatics. The International Medical Informatics Association (IMIA) has published, every year for the last quarter century, the Yearbook of Medical Informatics, collating selected papers from various journals in an attempt to provide a summary of the academic medical informatics literature. The objective of this paper is to visualize the evolution of the medical informatics field over the last 25 years according to the frequency of word occurrences in the papers published in the IMIA Yearbook of Medical Informatics.

METHODS

A literature review was conducted examining the IMIA Yearbook of Medical Informatics between 1992 and 2015. These references were collated into a reference manager application to examine the literature using keyword searches, word clouds, and topic clustering. The data was considered in its entirety, as well as segregated into 3 time periods to examine the evolution of main trends over time. Several methods were used, including word clouds, cluster maps, and custom developed web-based information dashboards.

RESULTS

The literature search resulted in a total of 1210 references published in the Yearbook, of which 213 references were excluded, resulting in 997 references for visualization. Overall, we found that publications were more technical and methods-oriented between 1992 and 1999; more clinically and patient-oriented between 2000 and 2009; and noted the emergence of "big data", decision support, and global health in the past decade between 2010 and 2015. Dashboards were additionally created to show individual reference data, as well as, aggregated information.

CONCLUSION

Medical informatics is a vast and expanding area with new methods and technologies being researched, implemented, and evaluated. Determining visualization approaches that enhance our understanding of literature is an active area of research, and like medical informatics, is constantly evolving as new software and algorithms are developed. This paper examined several approaches for visualizing the medical informatics literature to show historical trends, associations, and aggregated summarized information to illustrate the state and changes in the IMIA Yearbook publications over the last quarter century.

Quality Reporting of Multivariable Regression Models in Observational Studies: Review of a Representative Sample of Articles Published in Biomedical Journals

Controlling for confounders is a crucial step in analytical observational studies, and multivariable models are widely used as statistical adjustment techniques. However, the validation of the assumptions of the multivariable regression models (MRMs) should be made clear in scientific reporting. The objective of this study is to review the quality of statistical reporting of the most commonly used MRMs (logistic, linear, and Cox regression) that were applied in analytical observational studies published between 2003 and 2014 by journals indexed in MEDLINE.Review of a representative sample of articles indexed in MEDLINE (n = 428) with observational design and use of MRMs (logistic, linear, and Cox regression). We assessed the quality of reporting about: model assumptions and goodness-of-fit, interactions, sensitivity analysis, crude and adjusted effect estimate, and specification of more than 1 adjusted model.The tests of underlying assumptions or goodness-of-fit of the MRMs used were described in 26.2% (95% CI: 22.0-30.3) of the articles and 18.5% (95% CI: 14.8-22.1) reported the interaction analysis. Reporting of all items assessed was higher in articles published in journals with a higher impact factor.A low percentage of articles indexed in MEDLINE that used multivariable techniques provided information demonstrating rigorous application of the model selected as an adjustment method. Given the importance of these methods to the final results and conclusions of observational studies, greater rigor is required in reporting the use of MRMs in the scientific literature.

Research design and statistical methods in Pakistan Journal of Medical Sciences (PJMS)

OBJECTIVE

This article compares the study design and statistical methods used in 2005, 2010 and 2015 of Pakistan Journal of Medical Sciences (PJMS).

METHODS

Only original articles of PJMS were considered for the analysis. The articles were carefully reviewed for statistical methods and designs, and then recorded accordingly. The frequency of each statistical method and research design was estimated and compared with previous years.

RESULTS

A total of 429 articles were evaluated (n=74 in 2005, n=179 in 2010, n=176 in 2015) in which 171 (40%) were cross-sectional and 116 (27%) were prospective study designs. A verity of statistical methods were found in the analysis. The most frequent methods include: descriptive statistics (n=315, 73.4%), chi-square/Fisher's exact tests (n=205, 47.8%) and student t-test (n=186, 43.4%). There was a significant increase in the use of statistical methods over time period: t-test, chi-square/Fisher's exact test, logistic regression, epidemiological statistics, and non-parametric tests.

CONCLUSION

This study shows that a diverse variety of statistical methods have been used in the research articles of PJMS and frequency improved from 2005 to 2015. However, descriptive statistics was the most frequent method of statistical analysis in the published articles while cross-sectional study design was common study design.

[Use of multiple regression models in observational studies (1970-2013) and requirements of the STROBE guidelines in Spanish scientific journals]

BACKGROUND

In medicine and biomedical research, statistical techniques like logistic, linear, Cox and Poisson regression are widely known. The main objective is to describe the evolution of multivariate techniques used in observational studies indexed in PubMed (1970-2013), and to check the requirements of the STROBE guidelines in the author guidelines in Spanish journals indexed in PubMed.

METHODS

A targeted PubMed search was performed to identify papers that used logistic linear Cox and Poisson models. Furthermore, a review was also made of the author guidelines of journals published in Spain and indexed in PubMed and Web of Science.

RESULTS

Only 6.1% of the indexed manuscripts included a term related to multivariate analysis, increasing from 0.14% in 1980 to 12.3% in 2013. In 2013, 6.7, 2.5, 3.5, and 0.31% of the manuscripts contained terms related to logistic, linear, Cox and Poisson regression, respectively. On the other hand, 12.8% of journals author guidelines explicitly recommend to follow the STROBE guidelines, and 35.9% recommend the CONSORT guideline.

CONCLUSIONS

A low percentage of Spanish scientific journals indexed in PubMed include the STROBE statement requirement in the author guidelines. Multivariate regression models in published observational studies such as logistic regression, linear, Cox and Poisson are increasingly used both at international level, as well as in journals published in Spanish.

A systematic review of the statistical methods in prospective cohort studies investigating the effect of medications on cognition in older people

BACKGROUND

There is increasing awareness that medications can contribute to cognitive decline. Prospective cohort studies are rich sources of clinical data. However, investigating the contribution of medications to cognitive decline is challenging because both medication exposure and cognitive impairment can be associated with attrition of study participants, and medication exposure status may change over time. The objective of this review was to investigate the statistical methods in prospective cohort studies assessing the effect of medications on cognition in older people.

METHODS

A systematic literature search was conducted to identify prospective cohort studies of at least 12 months duration that investigated the effect of common medications or medication classes (anticholinergics, antihistamines, hypnotics, sedatives, opioids, statins, estrogens, testosterone, antipsychotics, anticonvulsants, antidepressants, anxiolytics, antiparkinson agents and bronchodilators) on cognition in people aged 65 years and older. Data extraction was performed independently by two investigators. A descriptive analysis of the statistical methods was performed.

RESULTS

A total of 44 articles were included in the review. The most common statistical methods were logistic regression (24.6% of all reported methods), Cox proportional hazards regression (22.8%), linear mixed-effects models (21.1%) and multiple linear regression (14.0%). The use of advanced techniques, most notably linear mixed-effects models, increased over time. Only 6 articles (13.6%) reported methods for addressing missing data.

CONCLUSIONS

A variety of statistical methods have been used for investigating the effect of medications on cognition in older people. While advanced techniques that are appropriate for the analysis of longitudinal data, most notably linear mixed-effects models, have increasingly been employed in recent years, there is an opportunity to implement alternative techniques in future studies that could address key research questions.

Important statistical points to improve and promote the methodology of the articles on medical sciences, particularly nephrology and kidney; a review article

Background: Quality of articles’ methodology is one of the important factors which is considered by researchers.

Objectives: This study was conducted to determine statistical guidelines on promotion of methodology’s quality in the articles concerning medical sciences, particularly nephrology, to assist authors and reviewers.

Materials and Methods: This study is a systematic review. Initially, the keywords "Epidemiologic Methods/analysis" [Mesh] OR "Epidemiologic Methods/epidemiology" and "reporting" were selected in Medline database. Then, reliable databases were searched for relevant publications. Being relevant, containing viewpoints, and recommending statistical guidelines as well as approval of at least two of the three examiners of articles were determined as the inclusion criteria into the study.

Results: Two hundred relevant articles were retrieved. Thirty-two articles met the inclusion criteria. By the examined articles, 30 applied points have determinative role for improving and promoting quality of articles methodology. Of the important points, introducing and describing target community and statistical population, mentioning article title, introducing independent and dependent variables as well as confounders, reporting sample size for subgroups and the whole study, summarizing the data according to their statistical distribution (reporting mean and standard deviation for data with normal distribution), reporting the type of rate (incidence, survival), ratio (odds, hazard) or risk (absolute, relative, difference) with 95% CI and the used software could be mentioned.

Conclusion: The most important factors contributing greatly to the quality of articles’ methodology on nephrology were reported in the present study. Applying these factors by articles authors and reviewers could lead to improve articles’ and journals’ quality. In addition, use of the findings of the present study in articles’ materials and methods could avoid research errors.

Validity of administrative data in recording sepsis: a systematic review

Introduction

Administrative health data have been used to study sepsis in large population-based studies. The validity of these study findings depends largely on the quality of the administrative data source and the validity of the case definition used. We systematically reviewed the literature to assess the validity of case definitions of sepsis used with administrative data.

Methods

Embase and MEDLINE were searched for published articles with International Classification of Diseases (ICD) coded data used to define sepsis. Abstracts and full-text articles were reviewed in duplicate. Data were abstracted from all eligible full-text articles, including ICD-9- and/or ICD-10-based case definitions, sensitivity (Sn), specificity (Sp), positive predictive value (PPV) and negative predictive value (NPV).

Results

Of 2,317 individual studies identified, 12 full-text articles met all eligibility criteria. A total of 38 sepsis case definitions were tested, which included over 130 different ICD codes. The most common ICD-9 codes were 038.x, 790.7 and 995.92, and the most common ICD-10 codes were A40.x and A41.x. The PPV was reported in ten studies and ranged from 5.6% to 100%, with a median of 50%. Other tests of diagnostic accuracy were reported only in some studies. Sn ranged from 5.9% to 82.3%; Sp ranged from 78.3% to 100%; and NPV ranged from 62.1% to 99.7%.

Conclusions

The validity of administrative data in recording sepsis varied substantially across individual studies and ICD definitions. Our work may serve as a reference point for consensus towards an improved and harmonized ICD-coded definition of sepsis.

Electronic supplementary material

The online version of this article (doi:10.1186/s13054-015-0847-3) contains supplementary material, which is available to authorized users.