Worsening heart failure (HF) arising in out-of-hospital settings is common and contributes to rehospitalisation, premature death, reduced quality of life and increased healthcare costs. Patients often delay seeking medical care for hours or days after symptoms escalate. Longer delays are associated with more severe clinical status at presentation and worse outcomes. Although multiple quantitative and qualitative studies have explored potential determinants of patient delay, findings are fragmented and inconsistent, in part because studies have evaluated differing sets of candidate predictors and have been conducted in varied populations and healthcare settings.
To address these gaps, this protocol specifies a systematic review and meta-analysis that will synthesise determinants of patient delay for worsening HF using a theory-driven framework. The authors plan to organise determinants according to Andersen’s Behavioural Model of Health Services Use, aiming to clarify which predisposing, enabling and need factors relate to delayed help-seeking and, where data permit, to quantify those associations.
The primary objective is to synthesise determinants of patient delay in seeking care for worsening HF symptoms that begin outside hospital among adults (≥18 years). The review will consider the time interval defined as the period from first awareness or onset of worsening HF symptoms to the decision to seek care or the first contact with a healthcare facility.
Secondary objectives include grouping candidate determinants using Andersen’s model into predisposing, enabling and need factors and, where possible, estimating pooled effect sizes for specific determinant–delay relationships.
Included study designs are observational analytical studies: cohort, case–control and analytical cross-sectional studies. Eligible studies must report quantitative measures of patient delay as defined above.
Population: adults aged 18 years and older with worsening HF symptoms occurring outside hospital settings. The protocol limits included reports to full-text articles published in English or Chinese.
Definitions: patient delay is operationalised as the elapsed time from initial symptom awareness or onset to the decision to seek care or the first healthcare contact. Determinants will be classified as predisposing (e.g., demographic or psychosocial characteristics), enabling (e.g., access or resources) or need factors (e.g., symptom severity or comorbidity), in line with Andersen’s Behavioural Model.
If specific definitions, measurements or thresholds for delay vary across studies, the review will account for those differences when grouping and synthesising findings.
The review will search the following databases from inception to 25 November 2025: PubMed, EMBASE, Web of Science Core Collection, CINAHL, Cochrane Library, CNKI, WanFang Data and SinoMed. Both English- and Chinese-language full-text articles will be eligible for inclusion. The protocol indicates duplicate screening and data extraction to ensure accuracy.
Titles, abstracts and full texts will be screened in duplicate. Data extraction will also be performed in duplicate, with predefined forms to capture study characteristics, population details, definitions and measures of delay, candidate determinants and effect estimates or other quantitative measures relating determinants to delay.
Where studies report heterogeneous delay definitions or different operationalisations of determinants, the review will document these differences and consider them in synthesis and subgroup analyses.
Risk of bias for cohort and case–control studies will be appraised using the Newcastle–Ottawa Scale. Analytical cross-sectional studies will be assessed with the JBI critical appraisal tool. These tools will inform judgments about study quality and guide sensitivity analyses.
Where studies report comparable definitions of delay and the same determinant with compatible effect estimates, random-effects meta-analyses will be conducted to pool effect measures. The protocol specifies prespecified subgroup and sensitivity analyses to explore sources of heterogeneity, such as differences in populations, settings or delay definitions.
If meta-analysis is not appropriate due to heterogeneity or insufficient data, the review will provide a structured narrative synthesis describing the direction and magnitude of associations reported across studies and grouping findings according to Andersen’s model.
The certainty of evidence for key determinant–delay relationships will be graded using the GRADE (Grading of Recommendations Assessment, Development and Evaluation) approach. This grading will inform confidence in pooled estimates or inferences drawn from narrative synthesis.
Results will be reported in accordance with standard systematic review practices, and findings will be organised to highlight which predisposing, enabling and need factors are most consistently associated with patient delay in worsening HF.
Because this work involves secondary analysis of published, anonymised data, formal ethical approval and informed consent are not required. The authors plan to disseminate findings via peer-reviewed journals, academic conferences and knowledge-translation activities to inform clinical practice and patient education about timely help-seeking for worsening HF symptoms.
The protocol has been registered in PROSPERO with registration number CRD420251242808.