Reducing unnecessary care: The SCAMPs manifesto

Can we reduce health care costs without rationing? (Image: Fibonacci Blue via Flickr)

We all know the problem: The cost of health care needs to come down. About five years ago, pediatric cardiologists at Children’s Hospital Boston realized it was critical to practice more cost-effectively. “Most of the money that is going to be removed from the federal budget to reduce budgetary deficits is going to come from health care in one fashion or another,” cardiologist-in-chief James Lock told an audience of senior Children’s physicians last month. “There’s no question we were under a tremendous amount of pressure.”

Seeking to eliminate unnecessary care and testing, Lock’s team first turned to clinical practice guidelines, or CPGs, a tool meant to standardize “best practices.” But it soon became clear that CPGs were ineffective, giving no insight into how to improve care or how to deal with unexpected findings. Even worse, over time, many mandated CPGs have been shown to be wrong by subsequent data.

“There is no such thing as best practice — there’s only sound practice that is constantly being refined.”

Randomized controlled trials (RCTs), the classic gold standard, aren’t fool-proof either. A group of Canadian researchers looked at 100 RCTs to see if they were still valid based on new information in the literature. Within 5.5 years, 50 percent of the trials’ conclusions needed updating, based on new data. And by their nature, RCTs provide answers to narrow questions in narrowly defined groups of patients. They don’t capture the messy variability of actual patient populations.

Rising costs (click to enlarge)

Lock and colleagues decided to take a look at the clinical decisions they themselves were making. They dispatched a cardiology fellow to record, over seven days, every decision the faculty cardiologists were making. He tallied 1,188 decisions, mostly based on training, prior experience, first principles, anecdote, instinct – not data.

What the cardiologists came up with in response is called SCAMPs, or Standardized Clinical Assessment and Management Plans. Now going hospital-wide, supported by Children’s Program for Patient Safety & Quality, SCAMPs has three maxims:

•    To provide cost-effective care, you need reliable, relevant clinical data you can act on.

•    Progress can be made even if the data aren’t bullet-proof.

•    Changes in care based on the data are never final.

“There is no such thing as best practice,” Lock told the audience. “There’s only sound practice that is constantly being refined.”

For any medical condition or set of symptoms, doctors can create a SCAMP – an algorithm with a decision tree guiding them on how to manage each patient, but one that’s also designed to explore one or more clinical questions. These questions, and all plausible patient outcomes, are identified in advance and built into the SCAMP, allowing data collection to be focused around them, supported by nimble, dedicated software.

“The task of the people writing the algorithms is to define what needs to be captured,” cardiologist Steven Colan, who oversees the information technology that underlies SCAMPs, told his colleagues.

“The reasons that doctors choose not to follow a SCAMP are extremely important information, and fuel improvement at a rapid clip.”

The ideal SCAMP topic is one where people aren’t sure what constitutes the best care, and where there aren’t enough data to go by, elaborated cardiologist Rahul Rathod in a session called How to Create a SCAMP. “A SCAMP doesn’t need the rigor of a prospective, hypothesis-driven study, but it should generate data to help answer a question.”

Deviations from what’s prescribed by a SCAMP are not only OK, they’re expected. All are captured by the software and can potentially be used to adjust the SCAMP.

“If you want to do something different than what the SCAMP calls for, there’s only one requirement,” Lock said. “You have to explain why you’re doing it differently.  The reasons that doctors choose not to follow a SCAMP are extremely important information, and fuel improvement at a rapid clip.”

Total patient encounters by SCAMP as of July 2011 (click to enlarge)

As of July 20 of this year, more than 2,600 Children’s patients have been enrolled in 14 different pediatric cardiology SCAMPs. Five of these SCAMPs have already been modified based on findings in clinical practice.

One SCAMP, for example, addresses the management of children who present with chest pain – a common symptom generating parental anxiety, a lot of practice variation among doctors, and a lot of unnecessary testing. Based on a review of all patients seen for chest pain in 2009, cardiologist Kevin Friedman and colleagues projected that the SCAMP would have enabled a 20 percent reduction in the use of exercise stress tests, echocardiography and Holter monitoring. More recently, a historical comparison found a 20 percent actual decrease in hospital charges after the SCAMP was in place.

The chest pain SCAMP's effect on utilization. Patient care charges have been reduced about 20 percent since the SCAMP was implemented. (Click to enlarge)

More than 150 patients have now been enrolled in cardiac SCAMPs at other institutions around New England, and the Program for Patient Safety & Quality is asking every department and division throughout Children’s to embark on the SCAMPs process by June 30, 2012. A new era of rational — not rationed — care is coming.

Shojania KG, Sampson M, Ansari MT, Ji J, Doucette S, & Moher D (2007). How quickly do systematic reviews go out of date? A survival analysis. Annals of Internal Medicine, 147 (4), 224-33 PMID: 17638714

Rathod RH, Farias M, Friedman KG, Graham D, Fulton DR, Newburger JW, Colan S, Jenkins K, & Lock JE (2010). A novel approach to gathering and acting on relevant clinical information: SCAMPs. Congenital Heart Disease, 5 (4), 343-53 PMID: 20653701

Friedman KG, Kane DA, Rathod RH, Renaud A, Farias M, Geggel R, Fulton DR, Lock JE, & Saleeb SF (2011). Management of pediatric chest pain using a standardized assessment and management plan. Pediatrics, 128 (2), 239-45 PMID: 21746719