The Impact of Real-Time Data Analytics on Infection Prevention and Control Practices in Ophthalmology: A Nursing Perspective on Patient Outcomes and Cost-Effectiveness.

The study was a single-center retrospective before-after comparison, not a randomised trial, so differences could reflect changes over time rather than the system. The authors state that confounding variables were not controlled and a causal relationship cannot be established, even though they also report multivariate adjustment. The paper did not explicitly name a primary endpoint; QALY was used for sample-size calculation and infection, visual acuity, pain, and visual function were called secondary endpoints. Follow-up was only 1 month, so long-term infection, visual prognosis, complications, drug resistance, and system costs were not assessed. The cost-effectiveness analysis did not include long-term system upgrade, nursing training, or maintenance costs, and primary hospitals may not afford the initial investment or staffing. Infection counts were small, and the complication comparison was reported as p = 0.057 while the text also calls it significant. Patients with immunocompromise, severe systemic disease, pre-existing ocular infection, emergency treatment, multiple treatments, or blindness were excluded, so results may not apply to more complex ophthalmic patients. The intervention was a bundle of basic nursing procedures, routine postoperative antibiotics, data collection terminals, AI analysis, early warning terminals, and IoT support, so the effect of any single component cannot be separated.

The authors reported no conflicts of interest and no funding. The real-time analytics system terminals and intelligent IoT solution were constructed with assistance from Suzhou Zhenqu Technology Information Co., Ltd.

Do not conclude that real-time data analytics caused fewer infections or saved money. Patients were grouped by time period rather than randomised, the intervention included basic nursing procedures, routine postoperative antibiotics, and multiple system components, follow-up was only 1 month, infection counts were small, and the complication comparison had p = 0.057.