How predictive analytics can transform your facility – a guest post by David Wayne and Gus Gilbertson
Clinicians and hospital administrators are faced with numerous complex questions on a daily basis:
• What can we expect to happen if we continue to operate as we have?
• What sort of changes should we make in order to improve clinical and operational performance?
• How can we accurately measure the results of our performance improvement initiatives?
Data that is assembled, organized, and evaluated to support predictive analytics can help answer those questions and more – providing a sound and forward-thinking basis for decision making.
Data allows you to prioritize goals – and meet them
In any initiative or program, it is critical to use data to identify and prioritize goals. Predictive analytics allows you to gather information, analyze that information, predict outcomes, improve processes, and then measure results. For example, harnessing data for predictive analytics can help hospitals forecast a patient’s potential for readmissions and provide preventative treatment – thereby improving a patient’s health outlook and reducing costs.
With both clinical and business concerns in mind, clinicians and hospital administrators must take steps to use data for meaningful predictions – and to meet their facility’s overall goals.
Step 1: Get a complete picture of clinical and operational performance
Data management allows you to view your facility’s performance from a macro perspective and draw meaningful conclusions. To do this most efficiently, determine which measures are of critical importance to your facility when you first engage with a performance improvement team. Once you have determined which measures you want to monitor, make an effort to capture as much data as possible so you can generate fully informed insights. This overall approach enables you to take a systems view of processes, and to identify implications for both patient and business.
Step 2: Drill down into key metrics that are important to your business
Remember: without data, even the best guides get lost.
Analyzing and organizing data in a way that allows for the prediction of future performance in specific measures (within statistical limits) makes detecting performance changes easier. Monitoring data within selected key metrics is what enables the identification of assignable causes of variation – in other words, it lets you see when things start to trend in the wrong direction so that you can identify the problem and resolve it quickly. By the same token, the same application of this methodology also allows for rapid detection of change in the data caused by improvement.
Step 3: Use data to support small-scale testing of different changes
Once you have measured your key metrics, the next step is to track changes, develop processes for improvement, and test the new processes on a small scale. After you have performed small-scale tests, you can implement these new processes on a larger scale.
For example, using the appropriate data elements and data segmentation, you can test rapidly for 30, 14, or 7 day readmission rates for different patient population segments. The charts below establish a baseline with historical data for readmissions rate in order to enable prediction. Once the baseline is established, you can plan, test, executive, and evaluate improvement cycles. In this example, we see the results of two Plan-Do-Study-Act (PDSA) cycles and the improvement in readmission rates.
Step 4: Continue to measure your facility’s productivity and make tweaks as needed
It is important to continue to analyze your data even after you have implemented new processes and improved the performance of your facility. For example, in order to understand the productivity of a cath lab, clinicians and administrators need to track both volumes and average cath procedure length. These metrics have clinical and financial ramifications for cardiovascular service lines. Monitoring cath volumes and procedure length on an ongoing basis allows you to detect change, determine causes of variation (a changing market, patient population, your lab’s operations, etc.) and take action before major problems arise.
Accelerating learning and change
Data analytics accelerates learning and spreads effective change. Analyzing key metrics can and will drive improvements in your facility’s clinical and financial outcomes; it is not as difficult as it may seem. Harnessing the power of data is critical every step of the way – from the assessment phase, to testing, to implementation.
David Wayne and Gus Gilbertson work as consultants for the LUMEDX Consulting Group. They assist clients in transforming all aspects of their cardiology business: operational, clinical and financial.