Development of a Core Confounder Set for Real-world Essential Hypertension Studies.

Abstract

Introduction: This study aimed to establish a core set of confounders for real-world essential hypertension studies to improve the reasonable control of confounding bias.

Methods: Controlled clinical studies of essential hypertension published between January 2001 and June 2020 were retrieved from PubMed. Matched or adjusted confounders from these studies were compiled to form a pool of potential candidates. The importance of each confounder was assessed through Delphi expert consultation, considering its statistical significance for impacting the prognosis of essential hypertension in published studies and its applicability in real-world essential hypertension studies. The most essential confounders were ultimately selected to constitute the core confounder set.

Results: Following a comprehensive literature review and group discussion, a total of 50 confounders were included in the Delphi questionnaire. Twenty-nine cardiologists from across China were invited to participate in the Delphi consultation, and consensus was reached after two rounds of consultation. As a result, a core set of 13 confounders was established comprising three confounders in the demographic characteristic and lifestyle domain, four in the baseline hypertension status domain, three in the comorbidity domain, one in the cointervention domain, and two in the common factor domain (study center and time). Additionally, a recommendation regarding the prioritization of controlling these confounders was formulated.

Conclusion: This study established a core set of confounders for real-world essential hypertension studies that can effectively control confounding bias and are readily accessible. These findings can serve as valuable references for protocol design and data analysis in real-world essential hypertension studies.