Predictive Analytics in Healthcare!

How can big data and analytics affect healthcare?

Now, understanding how data warehouses work and how they can be used to generate business intelligence to help us make better and more informed decisions, data warehouses can also help us aggregate and analyze the mass amounts of data for predictive analytics. But specifically, what areas can predictive analytics help us with?

Source Article

analytics main picture

Image Source

Clinical Performance Management—The operational and communications related processes that can help support and measurable patient outcomes and experience across a continuum of care. Population health management is a subset of this category. One thing that I am interested are the focuses on the development of a mobile health technology approach that involves Fitbit devices for self-management of cardiovascular disease. Because of “connected” technology, the data can be used greatly in predictive analytics for pre-disposition and preventative care as well as management!

Financial Performance Management— Helping healthcare organizations navigate tumultuous market and regulatory shifts, and not be forced to scale costs exceeding their revenue. One of the biggest hassles and problems as of late are insurance and reimbursement claims, especially with the amount of paperwork that is invested into each one, it is common for things to slide or for things to be overlooked and thus get denied. With predictive analytics, it can help us answer “what is it about these rejected claims that we can fix?” or “What technology will allow us to bring greater patient health outcomes without increasing the expenditure cost?”

Image Source

Analytics in Financial Performance Management

Although these aren’t new developments, they have proven difficult because it relies on the users and organizations to understand what metrics to be looking for, but if they do, the traditional software can sometimes prove adequate. However, healthcare data is subject to change over time, it’s just dependent on how volatile it is. In our discussion regarding data warehouses, I mentioned how because healthcare data metrics change it’s important that they are bound so early on, thus the point of HealthCatalyst’s “Late Binding Model”. These traditional static rules that can apply to other sectors and industries do not apply in healthcare, and due to the volatility can quickly lose pace and lag further and further behind from staying up to date. Although it’s possible to update through manual purposes and analysis, if there is already the data and technology why not make the most of it?

Healthcare financial data is massive, complex, iterative, and volatile elements, so predictive analytics would allow an organization to sift through these massive and collected data sets, uncover patterns and trends that correlate the target data elements to some outcomes of interest. As a result, by having predictive analytics and modeling, it can empower organizations to ask fundamentally different questions instead of the typical statically and manually created rules of analysis in business intelligence. As a result, organizations can be smarter about addressing root causes of these patterns before they run into the hitch.

Analytics in Clinical Performance Management

Although money and business decisions are great, that’s not the meat of my interest nor the time for me to be worrying about that yet. All I want to do first is to practice medicine, integrate technology with better user interaction and experience, and then I’ll slowly slide my way into management and worry about the money then! My main interest is in how predictive analytics can be applied with connected technologies, one of the big topics in this year’s, 2016 HIMSS Conference. Although I wasn’t there, I stayed up to date with keynotes and associated blogs that I do follow and connected technology along with predictive analytics was quite the buzzword!

analytics connected technollogy

Image Source

Something I hope to do is how to develop a mobile health technology approach that involves Fitbit (or any other smart wearables) devices for self-management of cardiovascular disease. It’s not just limited cardiovascular diseases but anything where a patient is likely to shift into a less health, higher-risk category that would alert and help the provider to intervene either through averting or delaying the shift. Another application would be with diabetes maintenance, although there are already devices that do such a thing I urge research to take it one step further in connective with the healthcare provider to help with maintenance and monitoring. Although it could be elective, any patient that has a smart wearable could connect with the organization for preventative and maintenance care. But first, it’s important to find the patterns and commonalities that can create the greatest likelihood of a specific outcome occurring in the patient population.

terminator analyticsOnce those patterns have been found and validated, they can become the foundation for probabilistic modeling then refined through machine learning and human instruction. With proactive and preventative alerts, they can be used in maintaining and improving population health in a broader sense as well (nothing like Skynet I hope…)  but the overall point is that hospitals, health systems, and providers all can collect, record, store and share more data that is medically relevant to the patient! Now if the bureaucratic red tape was a bit better, there could be even faster innovations in removing the bottleneck of healthcare data due to lack of interoperability. However, other problems lie in insurance reform, health IT adoption, and better quality reporting problems.

Section Source

Regarding the difficulty in data sharing and interoperability, a while ago there was a report that was presented to the national committee that is overseeing the health IT adoption process as well as shared at the Healthcare Information Management Systems Society (HIMSS) conference that accused the industry for realizing the problem but refusing to fix it until there was enough revenue in it. Vendors had stated that they would keep software proprietary and prevent the exchange of data and used that as another cash opportunity from hospitals that had to pay additionally to set up a communication with another vendor’s EHR. Another method is blocking protocols within the software that would otherwise allow data exchange, all of which is done to corner their respective markets. In fact, there is no “healthcare interoperability crisis” just a “bureaucratic red tape” that is preventing interoperability from proceeding nationally. There are standards like the Health 7 International’s (HL7) Fast Healthcare Interoperable Resource (FHIR) standard that is seeing increased adoption. HL7 provides the framework but FHIR (pronounced f-eye-yuh… or fire) is the protocol that allows the data exchange.  Based on RESTful APIs and the HTTP/HTTPS protocol along with XML/JSON, it’s growing in popularity due to the simplicity and ease of use. In many cases and situations, EMRs are made and designed for hospital use and for a profit, not for the actual consumers that have to use the  product, they’re often times made for the hospital business infrastructure which can differ greatly from the physician’s’ own workflow and efficiency. But with predictive analytics and more of a patient focus in the software, maybe we’ll soon be hearing about EMR that actually has the workflow and patient care as a primary factor. 

In the end, providers already have data assets in the form of EHR and billing systems, the problem lies in the integration of these disparate sources to form patient-centered data sets. This is where data warehouses can help and through business intelligence tools and predictive analytics can lead to greater and improved population health outcomes! From past experience and a very limited perspective as an outsider, EMRs and Health IT are often viewed as a panacea in the medical community, and not as a supplement to key communication and collaboration.

Analytics gifThe resulting models are the core technology to compute and track the health and financial risk status of the patient population being served. Therefore, we can sit  back a little bit afterwards and just soak in the success, before we gotta get back out and hustle for more improvements!

 

Article by Sir. Lappleton III

I'm a happy-go-lucky college student that started a blog as a way to not only document my education and my experiences, but also to share it with whoever stumbles upon my site! Hopefully I can keep you guys entertained as well as learn about a few things from IT as well as from my time and experiences as I plunge deeper and deeper into healthcare! A couple of my areas of focus is data management, system security (cyber security), as well as information technology policy.