Feasibility of Using Digital Clinical Records to Identify Factors Associated with Nursing-Sensitive Outcomes During...

Background/Objectives: Nursing-sensitive outcomes are key indicators of nursing care quality and safety. Digitalization of clinical documentation enables prospective surveillance using routine data, but real-world evidence on integrated digital, clinical record-based approaches remains limited. This pilot study assessed the feasibility of using routinely collected digital data, combined with ward-level indicators, to support nursing-sensitive outcome surveillance and to preliminarily identify factors associated with these outcomes during hospitalization. Methods: An observational pilot study was conducted at a 400-bed Italian university hospital over two months (November 2025-January 2026) across ten wards. Clinical variables, including the Blaylock Risk Assessment Screening Score (BRASS) and ward-level indicators, were extracted from the electronic medical record system. The exploratory composite endpoint of documented nursing-sensitive outcomes was modeled using multivariable logistic regression with cluster-robust standard errors and Cox proportional hazards regression. Results: A dataset comprising 1552 hospitalizations and 20,415 patient-days was assembled, with 100% completeness for socio-demographic data and 93% for BRASS. The cumulative event rate of nursing-sensitive outcomes was 7.35% (6.08 per 1000 patient-days at risk; 5.58 per 1000 total patient-days). Pressure injuries were the most frequent event (n = 67), followed by falls (n = 39) and physical restraint (n = 14). In the adjusted models, the BRASS (OR = 1.12 per point, 95% CI 1.08-1.15; HR = 1.10, 95% CI 1.07-1.13) and admission to the hospital from the emergency department (OR = 4.41, 95% CI 3.02-6.45; HR = 3.46, 95% CI 2.48-4.83) were independently associated with nursing-sensitive outcome occurrence. In a 48 h landmark analysis restricted to hospitalized patients still event-free on day 2, the BRASS association was attenuated (OR 1.07, 95% CI 1.04-1.10). Conclusions: An integrated, digital, clinical record-based approach proved feasible for prospective nursing-sensitive outcome surveillance and supports the design refinements planned for longitudinal analyses.