The reporting of studies using routinely collected health data was often insufficient

Lars G Hemkens,Eric I Benchimol,Sinéad M Langan,Matthias Briel,Benjamin Kasenda,Jean-Marie Januel,Emily Herrett,Erik Von Elm

doi:10.1016/j.jclinepi.2016.06.005

Abstract

ObjectivesTo assess reporting quality of studies using routinely collected health data (RCD) to inform the REporting of studies Conducted using Observational Routinely collected health Data (RECORD) guideline development. Study Design and SettingPubMed search for observational studies using RCD on any epidemiologic or clinical topic. Sample of studies published in 2012. Evaluation of five items based on the STrengthening the Reporting of OBservational studies in Epidemiology (STROBE) guideline and eight newly developed items for RCD studies. ResultsOf 124 included studies, 39 (31.5%) clearly described its design in title or abstract. Complete information to frame a focused research question, that is, on the population, intervention/exposure, and outcome, was provided for 51 studies (41.1%). In 44 studies where definitions of codes or classification algorithms would be necessary to operationalize such a research question, only nine (20.5%) reported all items adequately. In 81 studies describing multivariable analyses, 54 (66.7%) reported all variables used for modeling and 34 (42.0%) reported basic details required for replication. Database linkage was reported adequately in 12 of 41 studies (29.3%). Statements about data sharing/availability were rare (5/124; 4%). ConclusionMost RCD studies are insufficiently reported. Specific reporting guidelines and more awareness and education on their use are urgently needed.

Highlights

Collected health data (RCD) are defined as data collected for purposes other than research [1,2]
We selected items addressing these reporting domains in STROBE, and we developed a set of new items deemed important for reporting of research using routinely collected health data (RCD)
Inadequate or incomplete reporting has been shown in observational studies on general medical interventions that were published before introduction of STROBE [6] and in more recent evaluations that addressed specific research areas, including cancer [7], hand surgery [8], dermatology [9], plastic surgery [10], or magnetic resonance imaging [11]

Summary

Introduction

Collected health data (RCD) are defined as data collected for purposes other than research [1,2]. Process, and quickly access large amounts of such data led to increasing collection and usage for health research. Using such novel data sources involves unique challenges for research reporting, for example, the description of database characteristics or record linkage methodology [3]. Poor reporting wastes efforts and resources [4]. Guidelines such as the STrengthening the Reporting of OBservational studies in Epidemiology (STROBE) statement have been developed and endorsed by many journals to improve reporting of biomedical research [5]. Inadequate or incomplete reporting has been shown in observational studies on general medical interventions that were published before introduction of STROBE [6] and in more recent evaluations that addressed specific research areas, including cancer [7], hand surgery [8], dermatology [9], plastic surgery [10], or magnetic resonance imaging [11]

Objectives

Methods

Results