Rapid evidence identification in reviews of post-COVID syndrome: inclusion of preprint sources and timeliness of review completion

Date & Time
Wednesday, September 6, 2023, 12:30 PM - 2:00 PM
Location Name
Session Type
Information retrieval
Misso K1, Green R1, Garg M1, Kisomi M1
1Maverex Ltd., UK

Background: The COVID-19 pandemic resulted in substantial morbidity and mortality on a global scale. Subsequent to initial infection, the impact of post-acute sequelae SARS-CoV-2 infection (PASC) has become an increasing health concern. Prevalence of PASC has been estimated at 48% of COVID-19 survivors, experiencing at least one unresolved symptom.1 The pandemic kickstarted rapid changes in evidence production and communication, with availability of preprints via nonprofit online servers allowing rapid dissemination while the disease and evidence evolved. Up-to-date study results on PASC are vital to inform government and health policy during a fast-moving global health crisis; therefore, availability of preprints and currency of searching play a key role in PASC systematic reviews (SR) production. Our previous PASC SR emphasised the importance of searching preprint sources to identify unique included references in a timely manner.2
Objectives: To assess the inclusion of preprint servers and databases in SRs of PASC and to examine the currency and potential limitations of their search methods.
Methods: A sample (n=150) of SRs published between 2020-2023 were selected from Epistemonikos3, the largest database of healthcare SRs. We assessed:
• Inclusion of preprint servers and range of databases searched;
• Time between last search and publication of review;
• Limits applied and implications for potential bias. Review protocols, individual patient meta-analyses, scoping reviews, umbrella reviews, and SRs lacking clear search methods were excluded.
Results: We included 150 SRs in the main investigation. Assessment of an initial pilot sample of 50 found 84% searched at least one source containing preprints. Of these, 22% specifically searched a preprint server, and 12% searched only Google or Google Scholar for preprints. A total of 78% of SRs in the pilot sample applied restrictive language limits.
Conclusions: Preprint sources should be searched to ensure SRs of PASC are comprehensive and current. Given the global nature of the pandemic and resulting cases of PASC internationally, language limits risk introduction of bias and omission of potentially relevant research. Patient, public, and/or healthcare consumer involvement: Identification and evaluation of the latest PASC evidence and implementation of evidence-based decision-making benefits patients and the public during this evolving health crisis.