Benchmarking saves hospital $1 million

Study finds drastic undercharging in pharmacy

Curiosity may have killed the cat, but it saved a Florida hospital $1 million. Sarasota Memorial Healthcare System (SMH), an 845-bed hospital on Florida’s west coast, had done internal benchmarking for years but recently decided to start external benchmarking as well. The hospital bought a software program that supplies data from other hospitals across the state and right away found a discrepancy.

"I was just curious," says Susan Swier, RN, BS, CPHQ, SMH’s quality improvement coordinator. "When we got the Statewide Module [from Bellevue, WA-based HBS International Inc.], I started running standard reports just to see how we compared to our peers."

Swier received a big surprise when she pulled a report that showed SMH was millions of dollars below the Florida statewide average for pharmacy department charges. "Of the five hospitals I looked at, our pharmacy charges were the lowest — $15 million below average," she says. "We really seemed to be losing money."

Within a few days, Swier pulled together a multidisciplinary team to uncover why the hospital’s pharmacy charges were so low and to come up with ways to fix the situation. The team used a four-step process — identification of the problem, analysis, recommendation, and follow-up — and in four one-hour meetings had the solution that would add $1 million to the hospital’s bottom line.

The group began by reviewing SMH pharmacy charges by service line. It learned that five service lines (vascular, cardiac, general surgery, pulmonary, and orthopedics) had lost pharmacy charges exceeding $1 million each, Swier says. With 16,000 drugs dispensed by the pharmacy, it was hard to know which ones were causing the problem. So the team decided to start with one of the most expensive drugs: urokinase, a thrombolytic agent used to dissolve blood clots in arteries and veins. Urokinase, an IV additive, costs nearly $650 per vial, and the average patient dose in 24 hours costs more than $15,000.

To see just how much money they were undercharging for urokinase, the team compared pharmacy charges for the diagnosis-related group (DRG) in which urokinase is most frequently administered — DRG 478-Other Vascular Procedures. For pharmacy charges in that DRG alone, SMH was $250,000 below the expected charges for the state of Florida. The team then double-checked its efforts by scrutinizing the pharmacy charges for Primary Diagnosis Code 99674-Thrombosis of Vascular Grafts, a vascular surgery in which urokinase is given. Of the five hospitals in the comparison group, SMH was nearly $40,000 below the benchmark in pharmacy charges, Swier says. (See graph, above.)

Next, the quality improvement (QI) team conducted a chart review of patients who received urokinase. An unsettling pattern emerged. In just a six-month span, less than half of the patients who were given urokinase were charged for it. "That’s when we began looking at the pharmacy charge process," Swier says. For any pre-mixed or standard medication dispensed by the pharmacy, a charge was generated before the drug left the department. But the hospital used a different system for pharmacy-mixed medications.

"We learned that for any drug the pharmacy has to mix, such as urokinase, the patient was billed only after a carbon copy record came back to the pharmacy department," Swier says. "We found a number of nursing documentation issues, which are now being examined by a separate QI team. Also, fewer than half of these carbon copies were making their way back to the pharmacy billing clerk."

Bette Brotherton, MBA, MSN, CPHQ, director of quality improvement, says they found about 10 reasons why the copies were not getting back to the pharmacy for billing, including simple human error such as failure to remove the carbon copy or failure to put it in the box for the courier to pick up each morning. Also, the computer system in the intensive care unit (ICU) didn’t link with the system used in the pharmacy. ICU nurses were documenting cases in their computer system, but unless they also put it on paper, it wasn’t making it to the pharmacy. "They did not need to wait until the end to do the charges. They could do it at the point of distribution, but we had to redo our computer system to make that happen," Brotherton says. "A lot of hospitals are probably losing money the same way we were."

The solution was to have the pharmacy charge at the point of distribution for all medications before they left the department. But then another problem emerged: the hospital’s mainframe computer. "It was a system that was designed in-house," Swier says. "It was designed 10 years ago, and in order to charge for the pharmacy-mixed medications upfront, we had to go in and make major programming changes." The alterations ate up 200 programming hours for four months and cost nearly $14,000 in programming time. "But the net gain is phenomenal," Swier says. "Because we now bill for all meds as they leave the pharmacy, we’ve added more than a million dollars in net profit to our bottom line each year."

The team did some follow-up to make sure the charges were being captured and found that some money was still being lost, Brotherton says. Further study found a system glitch. If a pharmacist mixed more than one bag of urokinase at a time and hit reprint to get the second label, the system didn’t charge for the second bag. Again, the solution was simple: Have the pharmacists key in each order separately.

"This really speaks to the value of comparative data," Brotherton says. "You can watch your own data over time and they may not change, but they may be way out of line with your peers. It really focuses your performance improvement activities when you have these benchmarks."