1 Division of Biostatistics and Epidemiology, Department of Molecular and Translational Medicine, Paul L. Foster School of Medicine, Texas Tech University Health Sciences Center El Paso, El Paso Texas
Find articles by Alok Kumar Dwivedi2 Division of Biostatistics and Epidemiology, Department of Environmental Health, University of Cincinnati, Cincinnati Ohio
Find articles by Rakesh Shukla1 Division of Biostatistics and Epidemiology, Department of Molecular and Translational Medicine, Paul L. Foster School of Medicine, Texas Tech University Health Sciences Center El Paso, El Paso Texas
2 Division of Biostatistics and Epidemiology, Department of Environmental Health, University of Cincinnati, Cincinnati Ohio
Alok Kumar Dwivedi, Email: ude.cshutt@ideviwd.kola . Corresponding author.* Correspondence
Alok Kumar Dwivedi PhD, Division of Biostatistics and Epidemiology, Department of Molecular and Translational Medicine, Paul L. Foster School of Medicine, Texas Tech University Health Sciences Center El Paso, 5001 El Paso Drive, El Paso, TX 79905.
Email: ude.cshutt@ideviwd.kola ,
Statistical analysis according to design features and objectives is essential to ensure the validity and reliability of the study findings and conclusions in biomedical research. Heterogeneity in reporting study design elements and conducting statistical analyses is often observed for the same study design and study objective in medical literatures. Sometimes, researchers face a lot of predicaments using appropriate statistical approaches highlighted by methodologists for a specific study design either due to lack of accessibility or understanding of statistical methods or unavailability of checklists related to design and analysis in a concise format. The purpose of this review is to provide the checklist of statistical analysis and methods in biomedical research (SAMBR) to applied researchers.
We initially identified the important steps of reporting design features that may influence the choice of statistical analysis in biomedical research and essential steps of data analysis of common studies. We subsequently searched for statistical approaches employed for each study design/study objective available in publications and other resources. Compilation of these steps produced SAMBR guidance document, which includes three parts. Applied researchers can use part (A) and part (B) of SAMBR to describe or evaluate research design features and quality of statistical analysis, respectively, in reviewing studies or designing protocols. Part (C) of SAMBR can be used to perform essential and preferred evidence‐based data analysis specific to study design and objective.
We believe that the statistical methods checklists may improve reporting of research design, standardize methodological practices, and promote consistent application of statistical approaches, thus improving the quality of research studies. The checklists do not enforce the use of suggested statistical methods but rather highlight and encourage to conduct the best statistical practices. There is a need to develop an interactive web‐based application of the checklists for users for its wide applications.
Keywords: checklists, evidence‐based statistical practice, statistical analysis, statistical methodsThe overall quality and utility of biomedical research in generating proper evidence depend, in part, on the appropriate execution of research design, statistical methods, and interpretation of results and their quality reporting as well. Recently, a systematic review study identified nonadherence to standards of methodological approaches required by the Agency for Healthcare Research and Quality for research based on the National Inpatient Sample database even in high‐quality publications. 1 It is found that the appropriate use of methods and their standardized reporting helps in improving the quality of studies. 2 However, inconsistencies exist in methodological practices for similar study designs with the same objective/hypothesis. As a result, the quality of methodological standards in biomedical studies is often incredulous.
Guidelines and recommendations exist for assessing the quality of a study, or appropriate reporting and interpretation of results (www.equator‐network.org). Similarly, numerous statistical guidelines were developed for biomedical researchers to minimize misconduct of statistical approaches and improve the quality of biomedical studies. 3 , 4 , 5 , 6 However, these statistical guidelines mainly focus on improving the reporting of statistical methods used in studies. Unfortunately, guidance support is nonexistent for assessing best statistical practices of different types of studies as per the design features. Due to the lack of methodological standards checklist, misuse and abuse of statistical approaches in biomedical research have been noticed for a long time. 7 , 8
In recent years, novel statistical methods, computational program codes to analyze the complex problems, and statistical software for the ease of application of statistical methods and reporting have grown substantially. Numerous studies proposed alternative efficient and accurate approaches for specific study designs or distributional conditions and provided up‐to‐date statistical methods by comparing their performance on real data and extensive simulation studies. 9 , 10 However, the use of state‐of‐the‐art appropriate statistical methods for design and analysis of research studies is minimal in practice due to a lack of guidance for applied statisticians and applied researchers as recognized in the strengthening analytical thinking for observational studies (STRATOS). 11 For example, predictive intervals are computed and reported rarely in published meta‐analysis, 12 , 13 risk ratio models are rarely being used for the analysis of cross‐sectional or interventional studies even in high impact clinical journals, 14 , 15 , 16 inappropriate use and presentation of statistical modeling depending on the objective of the model building is common in published works, 17 inappropriate uses of graphs in animal studies and inappropriate interpretations of the results have also been noticed in biomedical studies as well. 18 Such examples and many more like these demonstrate that the use of appropriate statistical methods, accurate interpretations of results, and their reporting are not according to evidence‐based statistical methods and analysis. Thus, there is a need to develop checklists for evaluating the quality of statistical practices and a guidance document for promoting evidence‐based statistical analysis.
In the era of reproducible research, to increase the reproducibility, validity, and integrity of the research findings, we suggest following evidence‐based statistical practices in publications by use of appropriate statistical methods and their reporting relevant for specific objectives. Specifically, we (a) summarize the reporting elements of design features in studies to determine appropriate statistical analysis, (b) develop essential steps to be conducted in data analysis of common studies for promoting best statistical practices, and (c) provide evidence‐based essential and preferred choices of statistical methods for data analysis in different studies. Overall, the intention of the review is to provide checklists of statistical analysis and methods in biomedical research (SAMBR) according to specific objectives in different studies.
Initially, we identified the purpose and objectives of commonly employed study designs such as clinical trials, observational studies, and laboratory studies in biomedical research through various resources that may influence the choice of statistical analysis in studies. We also identified the essential steps to be followed in common studies to evaluate adherence to the best statistical practice in biomedical research. State‐of‐the‐art available statistical methods were identified for analyses (both unadjusted and adjusted and sensitivity) and reporting from high‐quality publications of biostatistics/epidemiology journals and other resources. The identified statistical methods were classified and linked with study designs and study objectives. When a clear choice did not exist, the decision was based on the qualitative evaluation of the statistical methods by comparing with other competing approaches in terms of statistical properties, assumptions, interpretation, and recommendations suggested by the researchers. The essential and preferred statistical procedures and appropriate references for employing each statistical method were provided under each study design and objective. Altogether, these procedures set the checklists for evidence‐based statistical analysis and their reporting for a specific study design in view of study purpose and objectives. Figure 1 shows the components of SAMBR and provides navigation to appropriate SAMBR checklist table as per study design and objective. Figure 2 summarizes the essential steps of data analysis according to common study designs/objectives.
Flow chart for selecting the appropriate checklist table specific to study design and objective
