Comput Inform Nurs. 2025 Jul 3. doi: 10.1097/CIN.0000000000001348. Online ahead of print.
ABSTRACT
This study investigated clinically significant association rules within pressure injury (PI) data from adult hospitalized patients to inform evidence-based prevention and management strategies. A retrospective cohort analysis was performed on a multicenter sample comprising 2386 PI cases (January 2018 to October 2023) across two tertiary hospital districts in Zhejiang Province, China. The analytical framework incorporated five patient-level demographic/clinical variables and six PI-specific characteristics. Association rule mining was conducted using the Apriori algorithm (minimum support = 10%, confidence threshold = 80%, lift >1), yielding 579 preliminary rules. Subsequent validation via χ2 testing retained 540 statistically significant associations (P < .05), of which 11 clinically actionable rules were established through Delphi consensus by a multidisciplinary expert panel. The results corroborate existing epidemiological evidence: advanced age (≥65 years), hypoalbuminemia (<35 g/L), and comorbid respiratory/neurological disorders constitute predominant risk factors for PI development. This study demonstrates the methodological rigor of association rule mining in identifying high-risk patient profiles, facilitating targeted early interventions to reduce PI incidence in inpatient populations.
PMID:40638211 | DOI:10.1097/CIN.0000000000001348
