Recent study lauds automated surveillance

It may be superior to voluntary reporting

In a recent study conducted by researchers at Duke University, in Durham, NC, the use of an automated surveillance system for identifying adverse drug events (ADEs) proved far superior to voluntary reporting. (The study was published online ahead of print in the Journal of the American Information Association.) The system was implemented at Duke University Hospital, an academic medical center, and at Raleigh Community Hospital. Potential events identified by the computer were reviewed by medication safety pharmacists and scored for causality and severity; findings were compared between the two hospitals and with voluntary reports from nurses and pharmacists.

Over the eight-month study period, 25,177 patients were admitted to the university hospital and 8,029 to the community hospital. There were 1,116 ADEs in 900 patients at the university hospital for an overall rate of 4.4 ADEs per 100 admissions.

At the community hospital, 399 patients experienced 501 ADEs for a rate of 6.2 events per 100 admissions. The computerized surveillance detected ADEs at a rate 3.6 times that of voluntary reporting at the university hospital and 12.3 times that at the community hospital.

"We believe it's important to know the incidence of ADEs in order to figure out how to reduce the number and to be able to show we've made a difference," says Peter M. Kilbridge, MD, assistant professor of pediatrics at Duke Medical School, associate CIO for Duke University Health System, and lead author of the study.

Grant funds study

The research team received a $1.4 million grant from the Association for Healthcare Research and Quality (AHRQ) to fund the study. "AHRQ has spent a lot of time over the past few years looking at different technologies for health care and ranking effectiveness, and there are a number they consider highly effective," Kilbridge notes. "This particular approach of using computers to detect adverse drug events in patients was actually invented about 15 years ago and is very underutilized, yet it falls into AHRQ's list of 'highly effective use of information technology to improve safety.'"

He reiterates that "if you don't know how many ADEs you have, you can't show you've made a difference." Most hospitals, says Kilbridge, use a voluntary system to report ADEs.

"The problem with voluntary reporting is that it is totally anecdotal," he says. "It depends on the reporter recognizing that an event happened and having both the time and the inclination to report it. It has absolutely no bearing on the actual number of ADEs."

While conceding that there often is a reluctance among staff to report such events, Kilbridge says that is not the biggest issue with voluntary reporting. "The literature shows that most clinicians do not recognize an adverse drug event — or an adverse event at all," he says. "Lots of times people are preoccupied by something else, or they figure something bad happened to the patient because of disease and do not see a cause-and-effect relationship with an ADE."

System developed in-house

So, armed with the AHRQ grant, Kilbridge and his team set out to create the surveillance system to "try to get a handle on the actual number of events." While there were a few pieces in place, he says he basically developed the system from scratch, with the help of "a large army of programmers."

In day-to-day, real-world terms, the system looks at patient data in the hospital systems – lab data, medications that the patients are on – and then looks for indications in that data that the patient may have had harm from a drug. "For example, they may have been given an antidote to a drug, or had a very high level of a toxic drug, or a lab value changed that suggested a drug they were on was the cause," Kilbridge explains.

The system's computer is linked to the hospitals' computer systems and flags these potential indicators. Then, the pharmacist goes and investigates each one to see if the "flag" represents a true adverse event.

As it determines the total number of ADEs, the system also provides baseline data, so that in the future researchers can see if progress has been made in reducing the number of events. "In one particular area [he was not free to identify] we have already been able to show such a difference," says Kilbridge.

Is it practical for many hospitals to have such a system, or would it be too costly without a grant such as the one Duke received from AHRQ?

"My best answer is, compared with other kinds of technology systems people are implementing, this is very inexpensive," Kilbridge asserts. "If you do CPOE [computerized physician order entry], for a 400-bed hospital it could cost you at least $7 million or $8 million. A system like this, if you do it from scratch, is in the order of maybe $100,000 to $200,000, depending on how good your IT department is."

In fact, says Kilbridge, "AHRQ gives ADE surveillance systems a higher effectiveness rating than it gives CPOE — and it is clearly a much smaller dollar investment."

For more information, contact:

Peter M. Kilbridge MD, Assistant Professor of Pediatrics, Duke Medical School, Associate CIO, Duke University Health System, Durham, NC. Phone: (919) 668-2166.


  1. Kilbridge PM, Campbell UC, Cozart HB, Mojarrad MG. Automated Surveillance for Adverse Drug Events at a Community Hospital and an Academic Medical Center. J Am Med Inform Assoc. 2006 Apr 18; [E-pub ahead of print].