Looking at HAI Case Records in Context of Your Physical Hospital Environment

How Cognitive Computing (AI) adds a different pair of trained eyes

The Challenge

A healthcare-acquired infection (HAI), also known as a nosocomial infection, is an infection that is acquired in a hospital or other health care facility. In the United States, the Centers for Disease Control and Prevention estimate roughly 1.7 million hospitalassociated infections, from all types of microorganisms, including bacteria and fungi combined, cause or contribute to 99,000 deaths each year. Nosocomial infections can cause severe pneumonia and infections of the urinary tract, bloodstream and other parts of the body. Many types display antimicrobial resistance, which can complicate treatment.

“Why does the system see that? And why doesn’t everybody see something that obvious?”

– Nurse in the Midwest

The Solution

4th-IR, in collaboration with a Midwest hospital and an established company in hospital construction & safeguarding, used Natural Language Processing to analyze over 100 HAI cases. The analysis was done against several IBM Watson corpora of knowledge specifically trained for this purpose. IBM Watson technology was used to understand the unstructured and structured information embedded in the HAI reports. The resulting information was combined with a limited set of structured hospital information to find the potential source of pathogens leading to HAI cases.

HAI Case Reports in Context

HAI Case Reports are full of information gathered by multiple departments. Much of the information is structured, however it lacks quality. By using IBM Watson’s Natural Language Processing (NLP) capabilities, the team was able to cut through the ambiguity by creating a concept space; thereby normalizing the language used in the reports. To do so, 4 th-IR and its partners trained two corpora. One by processing over 10,000 articles on HAI prevention research, and one by analyzing nearly 3000 documents on how to manage the containment of infections in hospital facilities.

Spatial context was added by digitizing hospital floor plans and adding the physical movement of patients for radiology exams.

“ You can’t just look at what happens in the OP, you have to understand what maintenance is doing as well. ”

 – Facility Manager

Timely, Evidence Based Action

  • Save Lives o Avoid physical space defects that lead to HAIs o Accelerate remediation of existing defects
  • Reduce Costs o Preventive, targeted maintenance vs big interventions o Reduce performance-based fines
  • Improve Quality o Reduce length of stay
  • Drive Innovation o Use existing data assets in new ways

Viewing your facility through a new lens

Microscopes are used to look at pathogens on a different scale. An MRI is used to look deep into the human body without physical trauma. In the same way Artificial Intelligence can be used to look at your existing documentation in a different way. For example, 4 th-IR’s solution allows you to look at your floor map with an understanding of how it relates to your documented HAI cases.

 

Analyzing the Connections

Microscopes are used to look at pathogens on a different scale. An MRI is used to look deep into the human body without physical trauma. In the same way Artificial Intelligence can be used to look at your existing documentation in a different way. For example, 4 th-IR’s solution allows you to look at your floor map with an understanding of how it relates to your documented HAI cases.

 

Adding cases and documents will make the solution more effective. This can be a timeconsuming task for care teams. AI is perfectly adapted to that. Combining the intuition of a human with the unrelenting insights of an AI will create amazing results

Seeing every new case in the context of all other past cases

The past may contain the answers to today’s instance of a HAI. By looking at each case through the lens of insights gained from processing these past events, one understands how to avoid future reoccurrence.

Looking Forward

Combining NLP with Graph Databases and the emerging Block Chain technology, will enable healthcare operators to establish a clear chain of evidence in each instance of an infection. This knowledge will be the first step towards prevention.

4th-IR combines machine learning with business and healthcare expertise to develop practical technology solutions that enable healthcare organizations to deliver premium care while enhancing their bottom line.