The Reporting Of MEdication use in Observational studies (ROMEO) Statement.

Abstract

To the Editor, The value of comprehensive and standardised reporting in clinical research is widely recognised. Standardisation facilitates reliability, comparability and reproducibility. The urgent need for guidance in relation to the reporting of medication use in observational studies has been recognised [1]. We conducted a systematic review of prospective cohort studies of health and ageing that identified substantial variability in the reporting of medication use [2]. The objective of this work was to develop consensus criteria for the reporting of medication use in observational studies. A multi-disciplinary expert consensus group was convened inMelbourne, Australia, in October 2014. The consensus group included 15 national experts in clinical pharmacology, geriatrics, internal medicine, epidemiology, pharmacoepidemiology and clinical pharmacy (see Appendix). Participants were provided with a synthesis of published literature highlighting the variability in reporting of medication use in observational studies. A modified form of the nominal group technique (NGT) was utilised [3]. This is a recognised method for developing consensus criteria. Structured small group discussion and feedback cycles were conducted to prioritise and refine key reporting criteria. The final criteria were circulated to and agreed upon by all participants. This study was approved by theMonashUniversity HumanResearch Ethics Committee. All participants provided written informed consent. The ‘Reporting Of MEdication use in Observational studies’ (ROMEO) criteria comprise 16-items across six domains (Table 1). The ROMEO criteria were developed to assist authors comprehensively report medication use. The consensus group envisages that the ROMEO criteria will be best used in conjunction with other reporting criteria, such as the STROBE Statement [4]. It has been recognised that there are possible additional benefits when the STROBE Statement is implemented in conjunction with discipline-specific criteria [5–7]. Like the STROBE Statement, the ROMEO criteria were not developed as a quality assessment tool. The ROMEO criteria primarily relate to reporting medication use in cross-sectional studies and repeated assessments of medication use in prospective cohort studies. However, it should have broad applicability to a range of observational study designs utilised in primary care, hospitals and institutional care. We acknowledge that not all the ROMEO criteria are relevant to other study designs. This is because pharmacoepidemiology encompasses a diverse range of study designs (e.g. prospective cohort studies, retrospective cohort studies, self-controlled case series, case-control studies), and each methodology has its own reporting considerations. In particular, there are issues specific to the analysis and reporting of medication use with administrative claims data that are not addressed in the ROMEO criteria. We anticipate that ‘Reporting of studies Conducted using Observational Routinely collected Data’ (RECORD) reporting guidelines should fulfil this purpose [8]. * Susan G. Poole Susan.Poole@monash.edu