Overlap of primary studies on an overview of the effectiveness of pharmacological interventions for the prevention of adverse events

Article type
Authors
Carrera M1, Pantoja E2, Parellada D3, Muñoz J3, Amparo N4, Barajas L5, Suclupe S6, Salas K6, Parise J7, Requeijo C8, Diaz Y9, Salvador J6, Torres M6, Solà I10, Bracchiglione J10, Martínez-Zapata M10
1Cochrane Iberoamerica. Autonomous University of Barcelona, Spain. Unidad de Paciente Crítico en Red de Salud UC Christus, Santiago de Chile, Chile
2 Vall d’Hebron University Hospital, Autonomous University of Barcelona, Barcelona, Spain
3Autonomous University of Barcelona, Barcelona, Spain
4Public Health Unit, ACES Alentejo Central, Alentejo, Portugal
5Research Unit in Evidence-based Medicine, Cochrane Center, Federico Gómez Children Hospital, Ciudad de Mexico, Mexico
6Quality Unit, Vall d’Hebron University Hospital, Barcelona, Spain
7Public Health and Clinical Epidemiology Research Center (CISPEC). Ecuador’s Associated Cochrane Center, Iberoamerican network. Faculty of Health Sciences Eugenio Espejo, UTE University, Quito, Ecuador
8Clinical Epidemiology and Public Health Service HSCSP, IR Sant Pau, Barcelona, Spain
9Epidemiology Resident, Heath Programs Department, National Institute of Respiratory Diseases Dr. Emilio Coni, ANLIS Malbrán – Ministry of Health of Argentina, Argentina
10Cochrane Iberoamerica– Clinical Epidemiology Service HSCSP, IR Sant Pau, CIBERESP, Barcelona, Spain
Abstract
"Background
Overlap of primary studies is a key methodological challenge for overviews. There are limited reports of methods used to address overlap, and there is no detailed assessment of the corrected covered area (CCA) of an overview of systematic reviews (SR).
Objectives
To estimate the overall and pairwise CCA of an overview of SR.
Methods
We assessed the CCA with the Graphical Representation of Overlap for Overviews (GROOVE) tool in an overview of systematic reviews evaluating pharmacological interventions to prevent adverse events in the intensive care unit. We calculated overall and pairwise CCA by outcome (delirium, nosocomial infections, ventilator-associated pneumonia [VAP], and gastrointestinal bleeding) and intervention. The study team was divided to evaluate and included systematic reviews in pairs.
Results
Sixty-five systematic reviews with a total of 1242 primary studies were included. The pairwise analysis showed a CCA of 1.2% (slight overlap). However, when assessing the overlap of studies according to the outcome, we found slight overlap for delirium (CCA=3.3%) and nosocomial infections (CCA=1.0%), and moderate overlap for VAP (CCA=5.7%) and gastrointestinal bleeding (CCA=9.1%).
Upon closer examination of overlap in primary studies according to the intervention evaluated, for the prevention of VAP, reviews assessing the effectiveness of topical oral chlorhexidine showed very high overlap (CCA 19.6%), and among reviews evaluating probiotic use, the overlap was moderate (CCA 9.9%). Conversely, reviews studying the efficacy of dexmedetomidine for delirium prevention exhibited a high degree of overlap (CCA 11%). In comparison, those assessing proton pump inhibitors used to prevent gastrointestinal bleeding showed very high overlap at 50%.
Conclusions
In our overview, the pairwise approach for assessing the CCA revealed highly overlapped pairs of SRs when considering the overlap by outcome and intervention. We encourage authors to complement the overall CCA assessment with a pairwise approach by outcome and intervention.
Statement on the relevance and importance to patients: This work will improve overview methods, resulting in more robust evidence production. Furthermore, it contributes to identifying redundant investigations in SR assessing pharmacological interventions to prevent adverse events in the intensive care unit.

"