Management of Patient-Reported Outcome (PRO) Alerts in Clinical Trials: A Cross Sectional Survey

Abstract

Background: Assessment of patient-reported outcomes (PROs) provides valuable information to inform patient-centered care, but may also reveal 'PRO alerts': psychological distress or physical symptoms that may require an immediate response. Ad-hoc management of PRO alerts in clinical trials may result in suboptimal patient care or potentially bias trial results. To gain greater understanding of current practice in PRO alert management we conducted a national survey of personnel involved in clinical trials with a PRO endpoint.

Methods and findings: We conducted a national cross-sectional survey of 767 UK-based research nurses, data managers/coordinators, trial managers and chief/principal investigators involved in clinical trials using PROs. Respondents were self-selected volunteers from a non-randomised sample of eligible individuals recruited via 55 UK Clinical Research Collaboration Registered Clinical Trials Units and 19 Comprehensive Local Research Networks. Questions centred on the proportion of trial personnel encountering alerts, how staff responded to PRO alerts and whether current guidance was deemed sufficient to support research personnel. We undertook descriptive analyses of the quantitative data and directed thematic analysis of free-text comments. 20% of research nurses did not view completed PRO questionnaires and were not in a position to discover alerts, 39-50% of the remaining respondent group participants reported encountering PRO alerts. Of these, 83% of research nurses and 54% of data managers/trial coordinators reported taking action to assist the trial participant, but less than half were able to record the intervention in the trial documentation. Research personnel reported current PRO alert guidance/training was insufficient.

Conclusions: Research personnel are intermittently exposed to PRO alerts. Some intervene to help trial participants, but are not able to record this intervention in the trial documentation, risking co-intervention bias. Other staff do not check PRO information during the trial, meaning alerts may remain undiscovered, or do not respond to alerts if they are inadvertently encountered; both of which may impact on patient safety. Guidance is needed to support PRO alert management that protects the interests of trial participants whilst avoiding potential bias.