Harnessing electronic healthcare data for wound care research: Wound registry analytic guidelines for less-biased...

Publications based on large healthcare databases that contain data pertaining to wound-related outcomes are starting to appear more frequently. However, concern exists in regard to study design adequacy, the methodology used to minimize misclassifications, bias, and confounding, and lack of full reporting. The STROBE guidelines were published to encourage fuller reporting of observational studies and have now been extended using the RECORD statement to better document routinely collected healthcare data. In this paper, elements of the RECORD statement have been used to create guidelines for study design, cohort matching, reporting criteria, and analysis frameworks in regard to analyses of populations involving comparative effectiveness research. It is recommended that researchers present full data analysis with minimal inclusion and exclusion criteria and preplanned subgroups analyses rather than attempt to emulate randomized controlled trials, as patterns of product administration are likely to be vastly different to those using controlled trials; moreover, missing data are very common. Suggestions for creating better matched cohorts, classification of wound- and patient-related variables, and a rationale for reporting at a minimum a particular set of benchmarks to better characterize wound care populations is also presented. Adherence to these guidelines would improve the credibility of studies and make comparisons between studies much easier. Finally, an adaptation of the Cochrane risk of bias tool is presented in connection with the proposed guidelines for systematic reviewers to assess these kinds of retrospective studies.