Using Data To Plan Care

DIA image Mar2015 Banner EN (jpg)

Heather Richards, program lead for Case Mix, says it's one of the most interesting projects she has ever worked on, but one of the most difficult to explain. "It's called a population grouping methodology, and it's all about using data to help inform health care planning at a population level," she explains. "It essentially combines the clinical information available for an individual over multiple years and across health sectors and uses this to assign risk indicators. This data can then be aggregated to help understand the burden of disease across a region or province."

The population grouping methodology uses several CIHI clinical databases that include information on day surgery, inpatient, emergency visit and physician billing. For each person, the methodology creates a clinical profile and predictive indicators. Person-level data is then rolled up to a region level to provide an idea of population-level needs—for example, expected costs based on the clinical characteristics of individuals. Provinces can later incorporate non-clinical determinants of health, such as education and poverty levels.

"It's really like a giant juicer," adds colleague Greg Zinck, manager of Case Mix at CIHI. "We put all kinds of data in and then apply an algorithm to reveal incredibly useful results."

For example, a jurisdiction could look at a group of patients with similar clinical risk factors and see how they are using the health system. It can then determine the likelihood of hospital admissions, long-term care needs and even death.

"Population grouping methodologies can be used to align health care resources to a population's needs," explains Ms. Richards. "Perhaps additional services, or different types of services, are needed in a region. The methodology can be used to monitor population health, understand care utilization patterns, explain variations in resource use and even provide a foundation for funding models."

Ms. Richards and Mr. Zinck are part of a team working on the project, but many other CIHI colleagues are also involved. "There are several people working with us to get the data we need for this project and to help us understand how to use it," says Ms. Richards. "Everyone wants to be part of this," notes Mr. Zinck. "They know how important it is and how useful it will be. It's great to have such a diverse group at the table with great ideas and viewpoints."

The project also includes an expert panel, which includes partners from provincial ministries, the Manitoba Centre for Health Policy, the Saskatchewan Health Quality Council, the Institute for Clinical Evaluative Sciences and even a health expert from Bonn, Germany.

In April, CIHI will release the alpha version of the methodology.  This will be a limited release for health system planners. The full pan-Canadian release is planned for March 2016.

Ms. Richards says the future applications of the methodology will be exciting to follow.  The final product will be open source and available to all health care providers. "Our customers have been asking for this, and now we can offer a product built in Canada using Canadian data. It's been exciting to work on a project that is taking full advantage of the wealth of health care data available through CIHI."

At a recent international conference in Qatar, Heather Richards was recognized with an Innovation Award for her paper describing the initial planning stages of the population grouping methodology. Pictured (left to right) are Kevin Murphy, Greg Zinck, Heather Richards, Lori Kirby and Jeff Hatcher.