Tips for starting your healthcare analytics journey

Healthcare organizations launching their first data analytics program often make missteps early in the process that can set back the project just as it is getting started.

It's crucial to have a roadmap that gets an organization off to a good start, says Anand Shroff, CTO at Health Fidelity, an analytics firm primarily serving insurers and accountable care organizations.

The first step is picking the right people to lead the project. Depending on the focus, that can include physician and nurse clinical leaders, operational units such as the supply chain department, utilization professionals and a sponsor from the CFO office, because that’s where the budget originates.

After getting the right people in place to drive the project, it makes sense to seriously plan for how to incorporate findings into actions.

Too often, people "put analytics in place just to have it because it's cool," says Shroff. "But they haven't fleshed out use cases and talked to people that the analysis would affect."

Consequently, a smart early step is to identify the people who will be the users and who the analysis will influence, he advises. "Find and define use cases, get an idea of interventions analytics will drive, and then look to see if you have the right infrastructure and people."

Also See: Analytics vendors aim to boost population health

Infrastructure matters. Just because an organization has an electronic health records system doesn't mean that data can be readily pulled, Shroff notes. Data in an EHR is designed for clinical use, not analytics. The EHR can be a silo, and getting at the data may be difficult.

And often times, project participants haven’t fleshed out what types of clinical data they really have and need before fishing for it. Data from other IT systems, such as enterprise resource planning and financial systems, are easier to access.

The IT department also is important, because its personnel will help execute the analysis. An executive sponsor should work closely with IT to ensure the requested data is appropriate and available, Shroff advises, and to make sure that throughout the project the department doing the analytics and the IT department are on the same page.

"On the IT side you need a delivery team headed by a project manager, and staffed and resourced, depending on what the requesting department wants," he adds.

An organization should first assess whether any existing vendors have technology to address the use cases that will be analyzed, according to Shroff. If internal resources are not sufficient, then an advisory firm can help chose an outside vendor or put together a request for proposals. Getting an advisory firm depends on the scope of a project and its budget, because the firms don’t come cheap. But if an analytics project is on an enterprisewide scale, "you’ll probably want a firm," he says.

There are always trade-offs of functionality, how long to get to the point of readiness and cost to most any project. An organization might pick a vendor that's ready to go but at a higher cost, or another vendor may offer outlays of capital at a lower price if the organization decides the project can be done at a slower pace.

Regardless of the pace, Shroff says, a tight focus on project plans and timelines, which with key milestones established and highlighted, can keep a project on track. Finally, if training of pertinent personnel is not comprehensive and ongoing, "project success will be very questionable," he adds.