Healthcare organizations need to move through a torrent of data that is quickly growing in size and complexity, adapt to changes in regulation and patient behavior and continue to grow revenue. By leveraging analytics for provider scoring, fraud detection, preventative healthcare and cost containment, companies can improve financial and operational performance.
It’s no secret that the healthcare market is undergoing significant disruption. In addition to the deluge of new rules and regfulations, data size and complexity is exploding. The numbers can be staggering.
- ICD codes are going to grow from 18,000 to over 155,000.
- Fraud is estimated to be a $250 Billion problem that is growing rapidly.
- Waste is estimated to be over $760 Billion (some estimates are over $1.2 Trillion)
By leveraging analytics, you can greatly improve the performance of key programs that will drive improved quality and patient outcomes with significant ROI. Fuzzy Logix has proven solutions in key areas including preventative healthcare, provider scoring, fraud detection and cost containment enabling organizations to use analytics as a competitive edge within the market.
Preventative Healthcare Analytics
Leveraging in-database and in-GPU analytics for research and development in healthcare can vastly improve the ability of researchers to obtain results. Researchers have experienced multiple benefits including the ability to run models using much larger data and with many more variables than previously possible. They are also able to accelerate the time required for their entire end-to-end process from months to hours. One of the reasons we founded Fuzzy Logix was to develop solutions to benefit mankind and seeing our products make such tremendous contributions to preventative healthcare has been powerfully rewarding for our staff and customers and we look forward to future growth in this area.
Utilizing In-Database Analytics To Predict The Risks of Developing Chronic Diseases
Optum – Large Scale Predictive Models for Chronic Illness
See how historical data is used to build statistical models which help to identify individuals at risk of a certain disease, based on the similarity of their historical claims with others whose history and outcome is known. The objective is that early intervention for people at risk may delay or even prevent the onset of the disease. We present a case study of developing the models for type 2 diabetes mellitus. The entire process starting with raw claims data, building regression models with diagnostics, solving for coefficients and computing accuracy measures, is completed in 30 minutes for population sizes of over 2 million people with 2,000 predictors.
For pure academic researchers, we have special programs to grant access to these solutions and for commercial customers, we offer programs with proven and measurable ROI.
In today’s tough economic market, identifying and mitigating risk is critical for healthcare organizations to survive. The use of provider scoring models to review the quantity, quality and efficiency of care has become a common strategy throughout the industry. Properly engineered, these models can churn through hundreds of millions of lines of claims data in minutes and assign scores to providers as a result payers can find risk and best practices in their portfolios. Many payers run scoring models infrequently due to the processing time required. As a result the time between scoring runs means that risk, quality and cost issues go unnoticed for long periods of time and have severe financial implications for the organization. With Fuzzy Logix solutions, you can score providers as frequently as desired. For example, for one of the largest insurers in America, we scored 700 million rows of provider data to less than 20 minutes.
Nearly one of every ten dollars spent in healthcare goes to fraud. The fraudsters have become very sophisticated in building complex scenarios and as such Special Investigations Units (SIU’s) need the tools to better identify and stop fraud. Improving the way you identify and capture fraud, particularly prior to payment, will have a positive financial impact. Catching fraud requires a combination of rules-based and quantitative solutions. Rules based (pay and chase) solutions catch known fraud and quantitative solutions capture hidden or unknown fraud. Rules based fraud solutions often have performance challenges due to the amount of claims and complexity of the rules being processed. Sometimes, all rules cannot be processed until after payment is made. Quantitative solutions are just emerging, but are proving to be very effective in rooting out hidden fraud. Fuzzy Logix can help with both types of models. Our solutions allow the rapid processing of rules based fraud and include quantitative models to find hidden fraud. For example, we recently reduced the processing time of a fraud query from 29 hours to 20 seconds. On that same project, using our quantitative models, we expect to find millions of dollars in potential hidden fraud.
Models for preventative health solutions can take weeks to build and run, but the results of the research can lead to early intervention and delay or prevention of negative health outcomes. The opportunity is enormous. By some estimates, preventable causes of death, such as tobacco smoking, poor diet and physical inactivity, and misuse of alcohol have been estimated to be responsible for nearly 40% of total yearly mortality in the United States (Mokdad AH, Marks JS, Stroup DF, Gerberding JL. Actual causes of death in the United States, 2000). Using in-database analytics can increase the ability of researchers to get answers that save lives. One of our clients was spending many hours running calculations on small data sets (the end-to-end process took days) and by deploying in-database models, can now run models on 10X as much data in less than 30 minutes.
Variation is the enemy of cost containment. The challenge is to identify the variation. Many solutions can help companies find gross levels of variation, but in hospitals and care facilities, it’s the little things that add up to big numbers. Using pattern recognition and outlier analysis, we identify areas where variance leads to excess expenditure. Think of the millions of things that happen at hospitals. There is simply no way for human analysts to find all the variance. By using pattern recognition and other techniques, we can find outliers in treatment behavior, identify evidence-based best practices and quickly find waste, extra tests, medical errors and other issues. In addition, these models can be used for revenue recovery. PWC estimates the size of waste in healthcare at $1.2 Trillion. Clearly companies who try to manage cost containment are pushing traditional solutions to the limit with mixed results. By leveraging analytics (including techniques from other industries) healthcare companies can recognize large and immediate benefits.
Please contact us to discuss how our solutions improve business performance.