Although analytics is generally seen as an organizational initiative with departmental responsibilities and objectives, both clinical and operational departments can identify specific use cases for which they would like to use data to identify gaps and inefficiencies that need to be reconciled to provide the highest value to the patient and the organization as a whole. It is important to make sure your executives understand how far-reaching a well-developed analytics framework can be.

2.) How to address data governance when defining an analytics strategy?


This may be the most critical question to discuss before even looking at analytic tools or resources.

Data governance is the process of establishing data ownership, including making sure the necessary data is accurate and available in a digestible form. Typically, I have seen this responsibility fall on the IT department, as most data resides in the systems IT supports. However, this does not mean other parts of the organization have no role in data governance. The data governance steering committee must have representation from the applicable business and clinical units to ensure the necessary data components are available to be reported and researched.

The fact of the matter is if your data is inconsistent, your analytics will reflect those inconsistencies. Data integrity, clarity and uniformity are critical to establish before any basic reporting can be performed or any decision can be made on collected data.

3.) How should my organization transition to predictive analytics?

Easy there, slugger. That's like a 20-year-old asking how soon he can retire. There are a lot of variables at stake in the transition from historical reporting to predictive analytics. Organizations need to approach the journey to predictive analytics as a journey with multiple maturity stages. If you ask any organization what the number one pitfall is to approaching a journey with multiple stages, they will tell you emphatically skipping a stage is the quickest route to catastrophe. Healthcare IT loves multiple stage models (as seen in the HIMSS EMR adoption model or Gartner hype curve). Analytics is no different. When moving from historical reporting to predictive analytics, there needs to be increased investment in the right technology, processes and resources to achieve the desired result. For example, suppose you had two choices on who should perform your tonsillectomy. Would you prefer a) me with the best surgical tools available or b) an experienced surgeon with a chainsaw and a rusty butter knife? I hope you would run screaming from either option. You're setting yourself up for a poor outcome each way. It's the same situation with analytics: not only do we need high quality data to work with; we need both the right tools and the skilled, professionally trained researchers to perform the work.

More Voices