Evaluation of reporting quality for observational studies using routinely collected health data in pharmacovigilance

Objectives: To appraise the reporting quality of studies which concerned linezolid-related thrombocytopenia referring to REporting of studies Conducted using Observational Routinely collected health Data (RECORD) statement. Methods: Medline, Embase, Cochrane library and clinicaltrial.gov were searched for observational studies concerning linezolid-related thrombocytopenia using routinely collected health data from 2000 to 2017. Two reviewers screened potential eligible articles and extracted data independently. Finally, reporting quality assessment was performed by two senior researchers using RECORD statement. Results: Of 25 included studies, 11 (44.0%) mentioned the type of data in the title and/or abstract. In 38 items derived from RECORD statement, the median number of items reported in the included studies was 22 (inter-quartile range 18–27). Inadequate reporting issues were discovered in the following aspects: validation studies of the codes or algorithms, study size estimation, quantitative variables, subgroup statistical methods, missing data, follow-up/matching or sampling strategy, sensitivity analysis and cleaning methods, funding and role of funders and accessibility of protocol, raw data. Conclusion: This study provides the evidence that the reporting quality of post-marketing safety evaluation studies conducted using routinely collected health data was often insufficient. Future stakeholders are encouraged to endorse the RECORD guidelines in pharmacovigilance.