Who participates in aphasia research? An analysis of the REhabilitation and recovery of peopLE with Aphasia after StrokE (RELEASE) data set

Background: People with aphasia present with different of demographic, language, stroke, and impairment profiles. A Cochrane review found that two-thirds of included trials had comparison data on participants at baseline (Brady, Kelly, Godwin, Enderby, & Campbell,2016). Participant selection may have a considerable impact on the effective- ness of therapeutic interventions (Hallowell,2008) while participant description is vital to interpretate the clinical relevance to patient subgroup (external validity). Inadequate description of aphasia research participants has been highlighted (Brady et al.,2016; Brookshire,1983; Roberts, Code, & McNeil,2003) as have data collection recommenda- tions and reporting in prospective aphasia research studies. Aim: We examined the reporting of participant descriptors in aphasia research studies from the REhabilitation and recovery of peopLE with Aphasia after StrokE (RELEASE) data set. Methods & Procedures: Two independent researchers accessed data from completed aphasia studies included in the NIHR funded (HS & DR–14/04/22) RELEASE database. We gathered and extracted data using published reports, electronic data sets, and (where possible) unpublished information directly from contributing collaborators (LW, KvB, BB, MB). We examined information on the following participant characteristics– age, sex, time since stroke, lesion and aetiology, aphasia severity, years of education, handedness, cognition, mood, vision, hearing, stroke severity, and co-morbidities. We also considered language of data collection, multi-lingualisim, living context, socio- economic status, dysarthria, and apraxia. We reported frequencies or medians for categorical or continuous factors, respectively, across studies (where such information was available) enabling comparison of the reporting standards recommendations. Outcomes & Results: Data were available from 174 aphasia studies (5932 individual participant data (IPD) conducted across 28 countries and published between 1973 and 2017. Data were derived from 58 case series or cohorts, 43 randomised controlled trials, 68 non-randomised controlled trials, and 5 registries. Severity of stroke in 18/174 = 10% (IPD = 1515) studies and concomitant medical history were only selectively reported (Table 1). We also observed some inconsistencies in the formatting of data reported; for example, in some studies, aphasia severity was reported within severity bandings (mild– moderate–severe) rather than an objective numeric value reflecting performance on a specific language assessment. Conclusions: Heterogeneity continues to be evident in the reporting of demographic and clinical data in aphasia studies. Consistent reporting of key recommended variables in standard formats is required in order to adequately adjust for covariates in analyses, for external validity purposes, for comparisons between studies, and to facilitate future meta-analyses.