AI and Automation: New Tools in the Fight Against Healthcare-Associated Infections

New technologies are providing hospitals with fresh and improved tools for preventing healthcare-associated infections (HAIs).

An article from Innovation News Network lays out several advancements that have been made in so-called “smart hospitals.” Those advancements include “automated disinfection systems employing robots and UV-C light solutions…that have been proven effective in killing a wide array of pathogens, including bacteria and viruses.” Smart sensors and wearables for staff members are another tool increasingly being used to “track patient and staff hygiene compliance.”

Artificial intelligence (AI) has also seen its use increase in hospital settings. Certain “predictive analytics, powered by machine learning and AI algorithms, play a crucial role in infection modeling and outbreak prevention.” When AI analyzes “vast amounts of data from various sources,” it can help to “identify patterns and trends that indicate a potential outbreak before it happens.”

Another tool that smart hospitals are using to try to enhance their infection control measures is real-time monitoring and data analytics. Using technologies that can track infection control measures in real time can provide “immediate insights into healthcare environments’ pathogen levels.” Emerging issues can be identified quickly, which can be leveraged to contain outbreaks before they become more problematic. Potential risks can be anticipated and proactive measures can be implemented to “safeguard patients and staff.”

In fact, using real-time monitoring tools is backed by research, including one study published in Frontiers of Public Health. This study was specifically meant to “investigate the potential risk factors for developing HAI in the ICU using real-time automatic nosocomial infection surveillance systems (RT-NISS) to surveil, and analyze the effectiveness of RT-NISS coupled with comprehensive interventions on HAI prevention and control in the ICU.” All data from inpatients in an ICU from January 2021 to December 2022 was pooled; comprehensive interventions were implemented in this ICU in 2022, the effects of the “RT-NISS application combined with comprehensive interventions on HAI prevention and control” could be evaluated.

By “implementing comprehensive interventions depending on infection surveillance by the RT-NISS in 2022, the prevalence proportion of HAI was reduced from 12.67% in 2021 to 9.05% in 2022,” and the “prevalence proportion of hospital-acquired multidrug-resistant organisms was reduced from 5.78% in 2021 to 3.21% in 2022.” The authors of the study concluded that “the adoption of an RT-NISS can adequately and accurately collect HAI case information to analyze the relative high-risk factors for developing HAIs in the ICU. Furthermore, implementing comprehensive interventions derived from real-time automation surveillance of the RT-NISS will reduce the risk and prevalence proportions of HAIs in the ICU.” Studies like these show that automating certain processes in hospitals and operating rooms can lead to actual tangible results and reductions in HAI rates.

Healthcare Purchasing News was able to speak with Evan Sylvester, MPH, CIC/LTC, WFR, MT(ASCP), senior director of Infection Prevention, North Division, at Providence Health Services: Puget Sound & Alaska in Seattle; and Joanna Mills, RN, MSN, CNS, CIC, system director of Infection Prevention at Walnut Creek, Calif.,-headquartered John Muir Health, about the promise and potential of automation and artificial intelligence in clinical settings.

How do you feel about increasing AI technology in operating rooms? Is AI automating any tasks that were previously burdensome to employees?

Sylvester: AI technology in operating rooms is a game-changer. It enhances efficiency by automating tasks such as scheduling, which optimizes OR times and reduces the burden on staff, allowing them to focus more on patient care. As AI continues to evolve, healthcare organizations are moving beyond the initial phases of adoption and are now focusing on leveraging AI to address staffing shortages and automate mundane tasks.

What is the potential for AI technologies to act as a tool for preventing healthcare-associated infections?

Mills: There are models already developed to identify risk associated with sepsis, utilizing numerous clinical variables. Predictive analytic models are also currently being utilized in data mining software to identify potential risk of infections. AI technology would be a great tool to determine individual patient risk and treatment tailored to the specific physiological factors and evidence (research) for validated protocols.

Sylvester: AI has significant potential to reduce healthcare-associated infections. It can be used in risk rating models, patient chatbots to guide them through their surgical journey, and frontline healthcare worker chatbots that provide quick access to internal policies and practices. Additionally, AI can enhance surveillance of hospital-acquired infections by identifying patterns in large datasets more efficiently than humans.

Is automation benefitting staffing? What about costs?

Mills: Any and all automation related to staffing and cost is beneficial. There is a tremendous amount of time and effort dedicated to scheduling and determining patient acuity in acute care facilities. AI could significantly benefit this process by utilizing historic data related to seasonal trends and types of patient populations to determine safe staffing levels.

Sylvester: Automation can benefit many tasks, particularly those that are repetitive and time-consuming. In terms of staffing, automation helps alleviate burnout and address shortages by taking over mundane tasks, allowing healthcare workers to focus on more critical aspects of patient care where AI may not be able to be deployed. While the initial investment in AI may be substantial, the long-term benefits include reduced costs and improved efficiency.

Let’s talk about data. Are there any automated tasks in IP departments collecting data analytics? If so, what can the data be used for?

Sylvester: Currently, there are limited AI tools specifically designed to simplify the infection preventionist's job. However, at Providence Swedish, we are developing a Central Line Blood Stream Infection (CLABSI) Machine Learning (ML) tool to identify patients at risk of infection 48 hours in advance. We are also exploring ML for more precise detection of Surgical Site Infections (SSI).

Are there any tools that are being used in IP departments to help with automation that are of particular note?

Sylvester: To my knowledge, there are no widely used AI tools in IP departments yet. However, the development of tools like the CLABSI ML tool shows promise for the future.

Are there any problem areas in general in the IP space that are on the rise / require intervention soon?

Mills: Antibiotic resistance and emergence of virulent pathogens are critical aspects of Infection control and prevention that threaten large patient populations.

Sylvester: Staff turnover has led to education gaps in Infection Prevention practices. Some organizations are using AI to create online educational modules for new hires and annual training. A more pressing concern is the rise in infectious diseases and the need for rapid detection of pathogens and their transmission.