Operationalizing ACO data analytics, care coordination a tough task

Tales from the cutting edge of accountable care recount the difficulty of integrating health data analytics to improve care coordination for ACOs.

WASHINGTON, D.C. -- To healthcare CIOs struggling to analyze Medicare data for the purpose of operationalizing care coordination gains and getting reimbursements: Rest assured, you're not alone.

While vendors and consultants might sell seamless back-end data automation, real-world execution of health data analytics isn't so automatic. It can require a separate support staff to succeed, unless you want to add data science to the plates of already harried physicians and nurses.

That was the picture speakers drew at the start of the eHealth Initiative's Third Annual National Forum on Data and Analytics in Healthcare. Even Niall Brennan, head data champion at the Center for Medicare and Medicaid Services and acting director of the Offices of Enterprise Management, acknowledged that CMS is still in the beginning stages of figuring out how to loop the right data back to the right providers, at the right time, in the right format.

Crawl, walk, run

I subscribe to a 'crawl, walk, run' philosophy, because there's been a lot of big-bang [government] IT and analytics efforts that failed spectacularly.
Niall BrennanCMS acting director, Offices of Enterprise Management

Turns out the formatting is no small task. Accountable care organizations (ACOs) -- whose payoff comes from using analytics on claims data to improve patient health -- ask CMS for data in various ways, Brennan said. A more sophisticated ACO might want structured data on 400,000 claims monthly to feed into their own system "and do their own thing," he said, while others request the data in a more traditional report format.

However, the challenges aren't holding CMS back. The agency is putting "a lot of chips" on the power of granular patient data to improve care while simultaneously cutting healthcare costs. Brennan is convinced that taking it slow will help this movement succeed.

"I subscribe to a 'crawl, walk, run' philosophy, because there's been a lot of big-bang [government] IT and analytics efforts that failed spectacularly," Brennan said, adding later that CMS is also trying to optimize the data it sends to providers so that it can be used in the various EHRs providers use.

Texas ACO: New staffing models

The country's fifth-largest federally qualified health center, the nonprofit Lone Star Circle of Care, covers five Texas counties and serves a patient base that's 50% Medicare and Medicaid insured and 35% uninsured, said Chief Medical Information Officer Tracy Angelocci, M.D. It also is an ACO spanning 30 clinics and 140 employed physicians.

While this type of cash-strapped organization would seem to be the last place to implement cutting-edge health IT, Angelocci said that Lone Star uses analytics to find patients who need extra interventions to keep healthier, and employs a full-time staff of nurse care coordinators to manage those calls.

The nurses, who can see individual patient data sorted for quick views by quality measures (A1C for diabetics, vaccinations needed or current, mammograms skipped, etc.), can call patients for reminders and book appointments while they're on the line.

Measuring gains from actionable analytics

Employing care coordinators not only pays off by engaging patients and keeping them out of the hospital, but it also keeps follow-up work away from physicians and nurses who are busy caring for patients.

Lone Star measured progress in the first full year of using the nurses, and saw a big payoff: Care coordinators boosted depression screenings by 600%, blood-pressure screenings by 75%, and regular mammograms by 48%. Overall, Lone Star estimated it saved 707 patient visits to clinics via the analytics-empowered nurses. Other improved metrics included tobacco screenings and cessation initiatives, colorectal cancer screenings, and flu shots.

The care-coordination nurses, even though they must manage an average of 900 patients per day because of Lone Star's resource limitations, can see the person through the data, which drives them to succeed. "These nurses have become health coaches and advocates for the patients," Angelocci said. "There's a real sense of accountability the patients have [to them]."

It's not all analytics utopia, however. Some of those improvements brought measures up to a total of only 50% compliance among Lone Star's patient population. So, there are many patient interventions yet to be made. Angelocci said the care coordinators divide and conquer, sometimes concentrating on one care-quality metric for a certain period of time, for example setting up flu shots during flu season and moving on to another metric after it's over.

Counting costs for rural providers

Another emerging ACO is the Nevada, California-based National Rural ACO, which bands together providers across the country to benefit from Medicare Shared Savings Plan reimbursements. Ridgecrest [California] Regional Hospital and National Rural ACO medical director Earl Ferguson, M.D., said that health data interoperability between EHR systems presents a challenge to aggregate the data needed to report to CMS, never mind the analytics to prove care quality gains.

From the rural provider's perspective, big data's time and financial costs are daunting.

"[CMS] is forcing providers to concentrate on the data and not on the patient care, requiring a lot of things of providers in terms of administrative costs," Ferguson said. "We have to practice entirely differently than we used to do. [We need to] start getting providers focusing on healthcare and not on collecting data."

While medium and large-sized hospitals might eschew cloud vendors who provide health IT systems such as practice-management systems and EHRs, Ferguson believes cloud will be the main path that National Rural ACO members will eventually use to better their CMS reimbursements. He said that ACO participation can cost a provider up to seven figures initially, but he estimates that cloud software providers can reduce that to nominal hardware startup costs and about $10,000 per month.

Let us know what you think about the story; email Don Fluckinger, news director, or contact @DonFluckinger on Twitter.

Next Steps

[Video] Learn about a care coordination platform with an operating system approach

Dig Deeper on Clinical data analytics software and systems