How to Improve EMR/EHR Using Predictive Analytics

Making predictions in the healthcare industry is nothing new. Based upon years of compiling data, for example, medical researchers have been able to determine a person’s risk for heart attack or stroke based on his lifestyle choices – smoking, alcohol abuse, high cholesterol diet, lack of exercise, and so on.

In reality, however, when medical professionals are treating individual patients, they are often making treatment decisions based on incomplete information. Certainly, they use their training, their past experience and the knowledge and research of other medical professionals, but ultimately, they are still acting upon incomplete information – information on the individual patient and information that exists out there somewhere but is not known.

Big data and predictive analytics in healthcare can close this knowledge gap and can result in far better treatment delivery and prevention.

So What is Predictive Analytics in Healthcare?

One of the key disruptive trends in medicine today is the use of predictive analytics for healthcare vendors. Specifically, this means the use of big data in much the same way that banks use it to predict types of loans best suited for individual customers.

But in medicine, big data can be used to determine, for example, if a patient should be discharged to home, sent to a rehab facility, should have follow-up phones calls, or a couple of home visits post-discharge.

Data analytics can thus improve patient care, patient satisfaction, and costs. It can search huge amounts of data that medical professionals do not have at their fingertips. And, even if they did, do not have time to analyze and integrate individual patients’ situations, in order to make better decisions.

How Are Predictive Analytics Used in Healthcare?

Using predictive analytics in healthcare just makes sense. In essence, it uses statistical methods and technology, developed by data scientists, gathers huge amounts of data and then, using techniques such as artificial intelligence, creates a prediction profile (called an algorithm) from past patients. That profile can then be applied to a new patient, given all of the variables that exist in that individual’s current and past medical history, lifestyle, etc.

And the real beauty of PA? Predictions can be made for individuals, not groups, which is what traditional research and statistics do.

A Simple Example of Predictive Analysis in Use

One of the best examples of predictive analytics in healthcare is the recent development of a predictive analytics algorithm by Kaiser Permanente, an insurance provider in California, relative to newborn care. About 0.05% of infants are born with infections that require antibiotic treatment at birth. But 11% of them have been receiving these antibiotics – clearly an overuse. Their algorithm could accurately predict the risk of infection based upon the individual mother’s clinical profile and the infant’s condition at birth. By using that algorithm, doctors could determine, which babies really need the antibiotics. About 250,000 newborns were thus spared unnecessary treatment and potential side-effects. Costs were also reduced.

Use cases such as this one make it easy to understand the benefits of predictive analytics in healthcare. And there are many beyond just the management of treatment protocols.

Here is just a short synopsis of the ways in which predictive analytics in healthcare using big data can benefit patients, treatment providers, and insurance companies.

More Accurate Diagnoses

Chest pain. It’s a common complaint. The first step in an emergency room visit will obviously be an EKG. There is no sign of active heart attack at the time of admission but there appears to have been some activity previously. If the attending physician is able to access that patient’s medical history through an electronic medical records system and is then able to plug that information into predictive analytics software for healthcare diagnoses, a better decision can be made regarding diagnosis and treatment. Should the patient be admitted for observation? Should an IoT device be utilized?

A major use of practical predictive analytics in medicine has been in the diagnosis and treatment of Alzheimer’s. Some people have a genetic risk factor, and, of course, patients can now be tested for that. However, when a primary care physician can then plug his patient’s data in a predictive analytics algorithm, he can receive all of the latest data regarding the potential progression of Alzheimer’s in his patient as well as the latest in treatment protocols.

Improved Public Health and Preventive Medicine

We now know how important early intervention is in preventing or reducing the severity of disease, especially in the area of genomics. When patients and their doctors know the risks, through healthcare informatics, lifestyle changes can be made and treatment/monitoring protocols can begin earlier. This will, in turn, save the medical costs.

The entire field of genomics, in fact, is a perfect place for predictive analytics use cases in healthcare in terms of prevention. Dr. Daniel Kraft, Chair of the Medicine and Neuroscience at Stanford University, states that until now we have only been involved in treating people once they get sick. We can now focus on avoidance of illness and disease. He predicts that in the future medications will be designed individually because predictive analytics tools for healthcare will place people into “similar subtypes and molecular pathways.”

Big Data and Predictive Analytics Cuts Down Healthcare Costs

Many employers provide healthcare to their employees as a benefit. Using a predictive algorithm, they can input the characteristics/data of their workforce and get predictions of future costs. They can also work with insurance providers who now have their own predictive analysis algorithms (much more accurate, actually, than former manual actuarial tables). Employers can use predictive analytics to make decisions about which insurance provider will best meet their needs.

A company along with hospitals can work with insurance providers to integrate databases and use predictive analysis to come up with better insurance products for a specific workforce. This is where predictive analytics use cases in healthcare can benefit everyone. For example, if it is discovered that the average employee visits his PCP twice a year, an urgent care facility three times a year, an ER once a year, and is hospitalized once in five years, these metrics can be fed into a model and a more tailored product can be devised. This may result in cost savings for both employer and employee. In fact, this is the basic principle behind the Accountable Care Organization (ACO) model that was incorporated in part by the U.S. Affordable Care Act in 2009.

Pharmaceutical Companies Can Benefit From PA Too

One of the coming trends in the pharmaceutical industry is to develop smaller amounts of specific drugs for specific populations, based on predictive analytics. Traditionally, a drug might be developed, approved, mass produced and then prescribed to a large population. When that drug failed to have the desired effect on a large population, the drug would then be discontinued. If research is able to predict more accurately the numbers of people who will benefit from a drug, then it could be lucrative for a pharmaceutical firm to revive that drug for that smaller population.

The other huge beneficiary of predictive analytics in the pharmacy industry is the patient. Traditionally, a “shotgun” approach to drugs has been the norm. A new drug might become a treatment protocol for 25,000 patients when only 10 in fact benefitted from it. This is a waste of money and, for the patient, could result in negative side effects.

The Ultimate Beneficiaries – Patients

The use of predictive analytics will ultimately transform patient care. People will receive treatment protocols that work for them specifically. They will use medications that are specifically targeted based upon their individual characteristics and will be made aware of health risks earlier so that preventive treatments and lifestyle changes can occur.  

In short, patients will be more informed. And recent studies show that patients, even senior citizens, are open to becoming a part of the technology and the data collection that is necessary to accomplish better, individualized, and more cost-effective healthcare.

The Clinical Predictive Analytics Market

There is an explosion of predictive analytics for healthcare vendors right now, all promising to have the solutions for healthcare providers, employers, and insurance enterprises. These are especially attractive for companies that do not have the budget for in-house data scientists to develop the software for solid analytics. Finding the right one to meet your unique needs will be critical because making a mistake can be costly indeed.

It will be important for any organization in the healthcare sector to carefully align its unique needs, current and future, and choose a big data as a service vendor that:

• Has significant experience in healthcare software solutions.
• Provides solid visualization layers for readability.
• Sets up security systems that are impenetrable.
• Offers a superior ETL (extract, transform, load) process to pull data from databases and place it in others.
• Gives you only what you currently need but provides ongoing support and expansion as those needs arise.

It may be time for you to get on board by choosing an analytics partner that has the expertise, the background and the processes to help you. Romexsoft may very well be that partner you are looking for. Give us a call, and let’s have a discussion.