Predicting Re-admissions Moves from Fiction to Fact

Predicting Re-admission Moves from Fiction to Fact

Over the last 45 years I have watched ‘technology’ gain an increasingly stronger foothold on the data and information we contend with on a daily basis. Focusing technology on making ever better decisions is of course critical. Perhaps in no sector are the benefits of computer-aided decision making becoming more demonstrable than in healthcare. Lives depend on such tools.

The historic technology landscape has seen us traverse the disciplines of knowledge management, in the creation of the experienced-based organization and the emergence of the knowledge-driven organization all inspired by computer assisted tools. Most recently we have been drenched in the world of ‘big data’. First came the attention to wealth management, now comes healthcare.

The rising importance of reducing hospital re-admissions has many ‘motivators’. Government penalties against hospitals for poor re-admission rates are paying a huge role. Reputational damage to hospitals for poor re-admission rates are debilitating. The difference that I now see, is that ‘knowing’ is one thing but the ability to now also do something about has become the real difference. Like the weather, which remains outside of our control, knowing about it is important but what can you do about it? Well, re-admissions probability management has reached a state where you can not only know about it you can do ‘something’ about it. Actually, you can now do many things about it.

What has happened is that specific use case computer models can exert a proactive discipline on care giver decisions and staffing time allocation. The conventional model is that a designated number of nurses arrive to address the residents on the assigned patient floor. This may not always be the best approach. What if floor five has more critical patients at risk for re-admission than floor two? What tools are available to instantly reallocate nursing resources to the most critical re-admission probabilities? With today’s predictive analytics such actions can be taken in real time. And the cost of adopting this technology has plummeted. No longer is ‘big data’ the exclusive domain of a few data scientists. Health care professionals, nursing floor staff can do ‘big data’.

The transformation has been in creating a usable tool that links the ‘data’ to the ‘intervention’ choices at the care giver level and in real time. This is a very dramatic transformation.

Extending this on-going ‘knowledge link’ to the hone care or skilled nursing service provider only increases the likelihood that a re-admission can be avoided. With probability based successful identification rates exceeding 85% health care can now know exactly where to look and take action. Commercial businesses have been applying this model for years, out of necessity, and with years of success. If I operate a fleet of trucks or airplanes I need to have a pretty good idea of where the potential problems are and which vehicles or planes are likely to surface with a problem next. Taking this exact model into the health care sector means I can now know which patients are most likely headed back to the hospital if I don’t not take action becomes a manageable agenda. Now, I don’t mean to equate matters that pertain to people with machines but the case is hard one to argue. Both need maintenance. and I can now know which ones are going to break down.

With the oncoming onslaught of ‘boomer’s’ needing care, unless these new systems are put into place, emergency rooms will be over-run with avoidable re-admissions. Predictive analytics in healthcare has now made what was previously deemed fiction, fact.

Contact me at mike@chartacloud.com  if you would like to learn more. I'd be happy to share what I have learned about this technology with you or your team.