What is predictive analytics?

In a world where screens can recognize our faces and cars can drive themselves, keeping up with technological advancements can be challenging. There are some inventions that are so ingrained in our daily lives that we can’t imagine living without them. There are other developments that are so ground-breakingly obvious, we can’t imagine how this wasn’t already thought of.

Well, predictive analytics is a little bit of both.

Predictive analytics is forecasted data based on machine learning from antecedent evidence and statistical algorithms.

That’s a lot of words to describe an easy concept.

In more basic terms- predictive analytics uses existing data patterns to predict better results.

This isn’t new either. Predictive analytics has been around for awhile, it just hasn’t been able to be utilized to its full potential until more recently.

This type of software is reliant on the access to clusters of fast, reliable data. It can’t put the puzzle together without the pieces. With technology constantly updating, the abilities of predictive analytics is expanding.

In today’s society, the data of our actions are constantly being tracked, recorded and inevitably predicted. While this does make it a little creepy when an ad for something you were just thinking about (but never said aloud) pops up on your facebook feed, it is definitively beneficial in settings like healthcare.

To better understand how predictive analytics is the solution within the healthcare sector, it’s important to fully understand the problem.

Let’s narrow down where the problem is.

One of the many things that Covid-19 has shown us is the holes within our healthcare system. Most specifically: post-acute care.

Before the global pandemic, there was already an issue in communication for patients once discharged from the hospital. Whether being sent home or admitted to a facility, there is a large gray area of who regulates and monitors patients’ care.

Is it the:

A. Patient? B. Caregiver? C. Facility? D. Therapist?

The answer is E. all of the above. 

With so many players on the field, there has to be clear, consistent communication. Everyone has a hand in the process of the patient’s recovery and all are accountable for doing their part. However, it is hard to do your part if you don’t know exactly where your role fits within the recovery plan.

Due to fragmented exchanges of information, recovery times are longer than they should be, feedback goes into the abyss and data is outdated. With all of their information scattered amongst different people, patients suffer through trial-and-error to find the right recovery plan.

This is a burden on all ends of the spectrum. Not only are time, money and resources being depleted for ambivalent treatments, but the patients’ health is put at higher risk of injury.

Lack of unity leads to misdiagnosis, inaccurate treatment plans and re-injury. There is a limit of recovery success, because the pathways are unclear. The system needs a proactive platform that combines all of the data and uncovers the steps to recovery in addition to the potential risks that might occur.

The answers are everywhere, we just need the resources to piece them together.

So, how can predictive analytics help mitigate this problem?

With access to patient data, predictive analytics in healthcare finds patterns amongst diverse patients for efficient treatment plans that can predict better recovery results. The algorithms can assess the data pool to evaluate which therapeutic regimens lead to effective recovery.

It sounds more complex than it really is- so let’s put things into perspective.

Bill, a 73 year-old patient, went into surgery to get a knee replacement. In the midst of his recovery, he could still sense something was wrong. When he went back to get it examined, his wound was reopened to assess the problem. Bill then contracted MRSA.

Weakened by the infection while still recovering from the knee replacement, Bill was unable to walk efficiently. This landed him in a wheelchair and made him more susceptible to falls. Which is exactly what happened.

1 in every 5 falls lead to serious injuries like broken bones- for Bill, it was his leg. And once you fall the first time, the chance of falling again is doubled. After 17 different surgeries over the span of 5 years, Bill eventually had his leg amputated.

His leg. Amputated. 

Think back to where we started- knee replacement!

And that’s not even the end. Because of falls risk, the constant surgeries, and the infection, Bill’s time in the wheelchair was extensively drawn out to the point that his other leg could no longer go straight.

A procedure that was supposed to improve his ability to walk instead took it away altogether. That’s not including the bacterial infection and the loss of time, money, appendage and, most importantly, hope.

Somewhere during the 17 surgeries, Bill understandably lacked the commitment to comply with his at-home care. Recovery doesn’t just take a physical toll on the patient, it affects their mental strength as well.

There are so many times within Bill’s situation that predicted pathways could have altered the outcome of his experience.

1. Following the knee surgery, the problems during his recovery could have been forecasted and essentially prevented. This would have potentially saved him from reopening the wound and contracting MRSA.
2. After getting the infection, the combination of all of his current medical needs in addition to his lifestyle could have been flagged. This would have put him on high-risk alert and possibly prevented the fall that broke his leg.
3. The multiple surgeries on his leg could have been significantly decreased if they were able to fully understand the problem and see what was restricting him from being able to heal. This could have saved Bill his leg and about 5 years of his life.
4. From spending so much time in the wheelchair, better data could have anticipated the effect this time would have on his other leg.

Predictive analytics can’t change the past, but it can help predict the future.