Did your real-world study follow STROBE guidelines?

https://ift.tt/2PQB7te

Is your observational research study following best practices? Is your methodology transparent? To help answer these questions, the Enhancing the QUAlity and Transparency Of health Research (EQUATOR) Network created the STROBE guidelines. The STROBE guidelines–an acronym for The Strengthening the Reporting of Observational Studies in Epidemiology–aim to improve the transparency of the methods behind observational research studies. The full guideline checklist is here, and I provide an overview an commentary below.

The Introductionsection should clearly explain the current literature, the gap in the existing literature, and the objective of the study.

The Methods section needs to describe how the study was conducted. Specific key elements include:

• Study design: Is this retrospective of prospective? Cohort, case control or cross sectional?
• Setting: Which country? Which care setting? How were people recruited? Over what time period were outcomes observed?
• Participants: Describe inclusion and exclusion criteria. If the study is a matching study (e.g., propensity score matching) describe how the matching was done.
• Variables: Key variable categories include the outcome, intervention/exposures, and other independent variables (i.e., predictors, confounders, effect modifiers)
• Data source(s): Describe the data source for each variable used in the study. Often times the setting and data source may overlap.
• Sample size: If this was a prospective study, how was the sample size arrived at? Was there a power calculation? If this was retrospective, how big is the population from which the sample was drawn from?
• Statistical methods: Describe how the impact of the exposure on the outcome was modeled. Provide justification for the approach. Describe any subgroups analyzed and justification for investigating said subgroups. Describe how missing data (if relevant) will be addressed statistically. Discuss any sampling strategy or re-weighting of the sample. Include a description of any sensitivity analyses.

The Results section should also be presented in a clear manner.

• Participants. A flow diagram or table should be used either in the body text or in an appendix to show how the study inclusion and exclusion criteria affected the sample size.
• Descriptive data. Researchers should provide basic demographic and relevant health information. For economic studies, baseline utilization and cost information should be shown. For cohort studies, typical follow-up time should be shown, particularly for unbalanced panels (i.e., variable follow-up time). In this section or in the outcomes section, the number of people per exposure group should be shown.
• Outcomes (unadjusted). This section will compare outcomes between different exposure groups. This would be differences in outcomes before propensity score matching, or before regression adjustment, or before implementing an instrumental variables strategy. The goal is to show whether differences in the main results are due to raw differences or if the statistical analysis used changes the results. For time to event studies, reporting the raw number of events.
• Main results. Show the main results. The structure of this section should parallel the structure of your statistical methods section.
• Other analyses: Typically this section would have any subgroup analysis, sensitivity analysis, or falsification tests/robustness checks.

While the methods and results sections aim to clearly describe what was done, the Discussion section of the paper aims places the results in context and identify key limitations. There are generally four sections.

• Key results: The first paragraph of the discussion summarizes the results.
• Interpretation: This section describes how this result fits within existing literature. If the results are similar to those in other studies, that is helpful. If they are different, explain propose some hypotheses why they could be different
• Generalizability: Authors should describe the contexts to which the study results are most versus least generalizable to give context to results.
• Limitations. This section identify key study limitations. If the limitation affects generalizability, that should be stated clearly. If the limitation is likely to bias the results, it is helpful to provide a direction of the bias or try bound the magnitude of such bias. If other methods could be superior, the authors should clarify why these other methods were not selected for this study.

Other information should also be shared such as study funding, author disclosures, and acknowledgements. Where more detailed methods or results are needed, researchers should use supplementary appendices where available.

The STROBE guidelines are not perfect and are focused more on epidemiological than health economic studies. Nevertheless, these guidelines provide a helpful outline for how researchers should report the results of real-world, observational data studies in peer-reviewed journals.

from Verdure https://bit.ly/3aibYiA