Solutions

Health Care Applications:

Pattern Analysis of Hospital Patient Costs

A major issue facing the country today is the rising cost of health care. Health care costs are rising exponentially, but there is still poor understanding around the interplay of factors that are driving these unprecedented cost increases. At the same time, more and more patient level data is being gathered and archived in health care databases.
Quantum Leap is collaborating with a major metropolitan medical center to evaluate Quantum Leap Pattern Based Discovery for analysis of their patient database to discover patterns involving the interaction of multiple factors that associate with highest patient costs. The discovery of such patterns along with their constituent factors provides a basis for understanding key cost drivers as a basis for developing policies and strategies for effective cost control.

Discovering Patterns in Health Care Fraud

Health care costs in the US exceed one trillion dollars per year. Health care fraud and abuse contributes ~10% to these costs, resulting in approximately $100-$170 billion per year based on various estimates. Identification of fraud patterns can provide a basis for proactive monitoring and control of fraudulent behavior. From a data analysis perspective, the primary challenges for fraud pattern discovery relate to consistency of data representation, data size as well as a significant amount of missing data. Discovering fraud patterns efficiently under these constraints can provide a significant capability as part of a more comprehensive strategy by which to reduce health care costs.
Quantum Leap Pattern Based Discovery is being evaluated by a state Medicaid agency to analyze their databases to identify fraud patterns. This project can provide an exemplary framework for adoption by other states to provide a systematic pattern based capability for fraud discovery and prevention.

Analysis of Cancer Registries

As the amount of patient health data being stored in electronic data registries continues to grow rapidly, there is an increasing need to mine that data to discover informative patterns or relationships in the data that map to health outcomes of interest. For example, many cancer centers are maintaining and growing their cancer patient registries. These registries contain a wealth of patient specific information around demographics, treatment regimens, genetic and other medical information related to their condition. An important characteristic of such data is the prevalence of missing data. It is often difficult to obtain information on all the factors that are present in the database. The challenge is to identify informative patterns that relate patient factors with disease status even in the presence of missing data.
Quantum Leap Pattern Based Discovery is being used by a major regional cancer center to analyze their cancer registry to discover patterns that associate with an increased likelihood of getting specific types of cancer. In this project, Quantum Leap Discovery is being integrated with an underlying database to enable pattern discovery directly from relational data. The resulting patterns can provide critical insight into combinations of factors that associate with disease as a first step to developing strategies for potentially improving outcomes.

Drug Safety Patterns

Pharmacovigilance has emerged as an important new area of health care analysis. Analysis of drug safety after a drug has entered the marketplace has become especially important in light of several high profile cases recently reported by the media. Identification of patterns of appropriate drug usage for different patient segments is an essential part of this new and growing field.
Quantum Leap Pattern Based Discovery is being used by a major technology services provider as part of its offering in health care informatics to the pharmaceutical industry. Patterns of drug usage coupled with patient demographics provides a data driven approach to formulating appropriate guidelines for drug usage and safety.

To discuss these or other similar applications, please contact
info@quantumleapinnovations.com or 302-894-8000

Want to try out Quantum Leap?
Contact support@quantumleapinnovations.com to be alerted to our next release.

Solutions

Healthcare

pattern based Analytics