Can Data Analytics Lower Hospital-Acquired Condition Incidence?

Hospital-Acquired Conditions (HACs) continue to burden our healthcare system and patient health. These events arise during medical stays and are often avoidable complications of care – including pressure ulcers, air embolisms, falls and trauma, and surgical infections following certain procedures.

According to a recent report by IBM Watson Health, HACs increased the average length of hospital stays by eight days in 2016, raised mortality risk by 72%, and accounted for more than $2 billion in excess hospital costs. Hospitals and payers could take a much more proactive approach to avoiding these conditions by applying evidence-based guidelines and using data to identify opportunities to create sustained avoidance of HACs.

Hospital acquired conditions have had such a huge impact on the U.S. health system that payers and providers have no choice but to prioritize addressing the situation.

CMS has rightly started to take notice. In fiscal year 2015, CMS mandated the Hospital-Acquired Condition (HAC) Reduction Program, reducing payments by 1% for hospitals that rank among the lowest-performing 25% with regard to HACs. CMS has also included specific measures related to HACs in its other Hospital Quality programs, such as the Inpatient Quality Reporting (IQR) Program and its Value-Based Purchasing (VBP) Program.

Related article: The Impact of Big Data on Medical Decisions

Noncompliant hospitals therefore lose twice-with poorer standards and therefore reputational risk, and also from regulatory fines.

What can payers and providers do to address the issue and improve their numbers while improving health outcomes?

With increased emphasis on reducing HACs and associated medical costs, one approach many healthcare payers and providers are taking involves the use of predictive analytics, which can shed light on current conditions and also suggest future actions to take to reduce the HAC incidents.

There are big savings, as well, given that providers are responsible for the cost of treatment including extended inpatient stays under pay for performance and value based arrangements. Using data analytics to report more accurately on these incidents has been shown to reduce HAC spend by as much as 15 to 30%.  

Challenges identifying and addressing HACs

What are some of the challenges inherent in taking this approach?

We’ve found that both payers and providers have trouble identifying episodes of care that are related to HACs. Using the correct data, and using data correctly, can be difficult to do when it comes to determining the specific cause of an episode and whether it was driven by a patient’s health status or by insufficient care at the facility.

There’s also a great deal of manual input that’s necessary to form a workable strategy for HAC surveillance-and with manual input comes human error. Making the switch from subjective standards to more reliable automated or digitized standards leads to inevitable disconnects between what the data show and what is happening on the hospital floor.

Gaps in knowledge of what causes these conditions, and determining effective prevention strategies, can create very human barriers to implementing a data analytics approach. 

Lastly, tying an episode of care related to HACs to value based care objectives requires not only a deep understanding of the data, but subject matter expertise to understand future contracts as it pertains to the data.

The nuts and bolts of a data-driven approach

As digital care comes more and more to the forefront, some of these difficulties can be overcome by employing machine learning or artificial intelligence (AI) along with predictive analytics. Using “smart” machines with access to much more data than any person’s memory could store can help in generating insights for patient safety teams, allowing providers to become more proactive and action-oriented in their approach to HACs.

By identifying common threads among HACs, and by determining risk factors through data analysis, providers can intervene much earlier and develop effective strategies to reduce adverse events. To give one example, by linking EHR data, bed alarm data, and nurse call light data to analytics, patient teams can much more accurately predict when a patient is at risk of an imminent fall.

Providers and health plans using dashboards with drill-through capabilities find it much easier to present their data analytics findings in a way that all users are able to interpret. This is where measurable savings can come in-as mentioned, these analytical reports can help healthcare organizations reduce avoidable HAC spend by 15 to 30% on average.

More important to the health of our populace, this use of data analytics produces insights that can help to improve HAC-related policies, support the development of long-term prevention strategies, and ultimately improve health outcomes.

Michael Kim is a VP at AArete, a global consultancy specializing in data-informed performance improvement, and heads its Center of Data Excellence.  He can be reached at mkim@aarete.com. John Marchisin is a Managing Director at AArete and can be reached at jmarchisin@aarete.com.