To know a change is an improvement, we must develop measures for the outcomes of interest and create visually appealing and intuitive graphs that show how the measures are changing over time.
To identify opportunities for improvement, we must find associations between the measures of interest and the factors that affect their performance.
To make proactive decisions that anticipate improvements in quality, we must develop predictive models that show how the measure is likely to behave under various scenarios and base decisions upon the model’s predictions.
The Anderson Center data analytics team provides all these services and more to active improvement teams, the Cincinnati Children’s planning department and senior leadership, and health services researchers. Approximately half of the staff hold graduate degrees in quantitative disciplines, such as applied mathematics, quantitative analysis, statistics, biostatistics and epidemiology. Other staff include project managers, software application developers and data managers.
Over the years, this team has developed several hundred performance metrics in areas such as patient safety, capacity management, system flow, clinical outcomes and patient / family experience. The team produces monthly performance reports for these measures, which are published on the Cincinnati Children’s web site. The team also works closely with improvement leaders and investigators to analyze data so that reasons for performance are better understood, and with understanding, suggest opportunities for improvement.
The data analytics team has developed knowledge and expertise in electronic data sources, including the enterprise-wide electronic health record system. The team also continually improves its expertise in coding methods suitable for making optimal use of these data sources. It has built up technological expertise in developing customized software applications to support improvement work.
Examples of Services and Products
|Measure development in multiple domains along with operational definitions that give instructions for automatic population using electronic medical record data
||Run and control charts for hundreds of performance metrics, produced and published monthly and quarterly
|Statistical analysis suitable for questions in the delivery of health services, including factor analysis, analysis of variance, multivariate regression, proportional hazards analysis and network analysis, to show complex relationships
||Dashboards for clinical system improvement teams, planning department, divisions and institutes and nursing units that show the most recent values of performance metrics of interest
|Predictive modeling, particularly in capacity management and space utilization
||Software applications such as a real-time adverse-event detection system
One of the most comprehensive initiatives is the development of an integrated performance measurement and reporting system (PMRS) that will automatically calculate the values for performance metrics using the organization’s electronic data and give end-users the ability to track metrics of interest, in real time, on personalized measure “home pages.”
The system will also give users the ability to configure and produce dashboards simply by clicking a button. Users will be able to perform modest levels of analysis, such as showing the run charts of performance metrics stratified by areas of interest. This initiative is intended to allow master’s- and PhD-level analysts to perform more in-depth analysis and to put performance information directly in the hands of the end-users.
Other initiatives include the revision of methods to assess patient experience, additional development of software applications that track risk and identify intervention opportunities, predictive modeling such as understanding future needs for additional ICU beds, development and testing of an overall index of patient serious harm and support for a national network of children’s hospitals working toward improving patient safety.