The Journal of hospital infection · 2026
Artificial intelligence for hospital infection prevention and control: real-world implementation, impact, and the gap beyond model development.
СтатьяОбзорНазвание и аннотация: Перевод в очереди
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Sirago G, Zotti F, Mele F, Misceo F, Solarino B, Dell'Erba A, Ferorelli D
Аннотация
BACKGROUND: Artificial intelligence (AI) and machine learning (ML) may strengthen hospital infection prevention and control (IPC) through automated surveillance, early warning, and decision support, but the evidence base is fragmented and often limited to retrospective model development.
METHODS: We conducted a Preferred Reporting Items for Systematic Reviews and Meta-Analyses 2020 systematic review to synthesize studies of AI/ML in acute care hospital IPC, distinguishing implemented systems integrated into workflows from non-implemented development/validation studies. On January 17, 2026, we searched PubMed/MEDLINE, Scopus, Web of Science Core Collection, and Cochrane CENTRAL. Two reviewers independently screened records and assessed full texts.
RESULTS: Of 7089 records identified, 2061 duplicates were removed; 5028 titles and abstracts were screened; 177 full texts were assessed; and 59 studies were included. Only 9 of 59 studies reported real-world implementation (6 pilot and 3 routine), whereas 50 of 59 remained development/validation only. Conventional ML predominated (42/59), followed by natural language processing (8/59), deep learning (5/59), and rule-based or expert systems (4/59). Implemented studies more often reported process and operational outcomes than standardized infection outcomes and rarely quantified unintended consequences such as unnecessary isolation or alert fatigue. Risk of bias and applicability concerns were common, particularly regarding transportability and confounding in uncontrolled before-after evaluations.
CONCLUSIONS: AI/ML for hospital IPC shows promise, but translation into routine practice remains limited. Safer adoption requires implementation-focussed study designs, standardized IPC-relevant outcomes, and lifecycle governance with continuous monitoring after deployment.
METHODS: We conducted a Preferred Reporting Items for Systematic Reviews and Meta-Analyses 2020 systematic review to synthesize studies of AI/ML in acute care hospital IPC, distinguishing implemented systems integrated into workflows from non-implemented development/validation studies. On January 17, 2026, we searched PubMed/MEDLINE, Scopus, Web of Science Core Collection, and Cochrane CENTRAL. Two reviewers independently screened records and assessed full texts.
RESULTS: Of 7089 records identified, 2061 duplicates were removed; 5028 titles and abstracts were screened; 177 full texts were assessed; and 59 studies were included. Only 9 of 59 studies reported real-world implementation (6 pilot and 3 routine), whereas 50 of 59 remained development/validation only. Conventional ML predominated (42/59), followed by natural language processing (8/59), deep learning (5/59), and rule-based or expert systems (4/59). Implemented studies more often reported process and operational outcomes than standardized infection outcomes and rarely quantified unintended consequences such as unnecessary isolation or alert fatigue. Risk of bias and applicability concerns were common, particularly regarding transportability and confounding in uncontrolled before-after evaluations.
CONCLUSIONS: AI/ML for hospital IPC shows promise, but translation into routine practice remains limited. Safer adoption requires implementation-focussed study designs, standardized IPC-relevant outcomes, and lifecycle governance with continuous monitoring after deployment.