Whole genome sequencing for epidemiological studies of tuberculosis: a systematic review of reporting practices and factors associated with reporting quality of STROME-ID

Background: Whole-genome sequencing (WGS) has the potential to improve the understanding of tuberculosis (TB) epidemiology. However, standardized reporting is necessary to enhance the reproducibility and interpretation of WGS results in genomic epidemiologystudies of TB to better inform public health decision-making. In 2014, guidelines called STROME-ID were published to provide recommendations for reporting in genomic epidemiology studies. Reporting practices before and after its publication were compared, and the correlation between STROME-ID reporting quality and study-level characteristics were also explored.Methods: This study is registered on PROSPERO (CRD42017064395). MEDLINE, Embase Classic and Embase were searched on May 3, 2017 (updated April 23, 2019). 976 titles and abstracts were screened, with 114 full-texts eligible for inclusion. The proportion of STROME-ID criteria reported was tabulated for each article, and differences in means were compared before and after STROME-ID’s publication date using a t-test. A 6-month lag period after STROME-ID was included to account for articles in-press; sensitivity analyses were also performed. Quasi-Poisson and tobit regression were used to assess whether h-index (HI), journal impact factor (IF), sample size (SS), and geographic region of the senior author’s primary affiliation were correlated with the count and proportion of STROME-ID criteria met. Results: The proportion of applicable criteria met in included articles ranged from 16.3-75.0% (mean 49.9%, ± 11.88%), with no difference between mean proportions of criteria comparing before and after guideline publication. HI was not included in the adjusted regression analysis. Only SS was significantly associated with a greater proportion of STROME-ID criteria met. Conclusion: Reporting quality in genomic epidemiology studies of tuberculosis is variable, despite publication of STROME-ID guidelines. Future studies should investigate factors affecting adherence to these guidelines to improve the value and utility of evidence. Journal endorsement may be needed to support this