Taming the tiger: Johns Hopkins Health System employs automation to tackle the organizational management of thousands of annual alerts and recall notices

Health Management Technology, August, 2004

The management of recalls and alerts in healthcare is a complex, challenging issue. It is often under-resourced and not clearly understood. Typically, it receives minimal organizational focus until an event impacts a patient. That was the case at Johns Hopkins Health System, and it spurred us into action.

PROBLEM

On Nov. 30, 2001, a recall was issued for a bronchoscope that had a serious design flaw, making it almost impossible to clean. The manufacturer sent the recall notice to the delivery address that it had on record for our organization--which, in fact, was a loading dock. The recall notice subsequently floated around our organization for two months, because no one person was responsible for the management of recalls and alerts, and also because we suffered from an internal lack of consensus about how to handle such notifications.

During this period, as many as 416 patients may have been exposed to the bronchoscope that could not be properly sterilized. It cost our organization more than $200,000 in labor expenses alone to track and rectify the broncho-scope situation. We recorded and treated this incident as a sentinel event.

As every healthcare organization knows, the problem with transmittal and receipt of alerts is that we don't know what we don't know. Alerts and recalls can be issued for pharmaceuticals, dietary supplements, supplies, medical devices and equipment, as well as for a host of non-patient related goods used by facilities management, information systems, etc. They can arrive at healthcare facilities via fax, phone call, e-mail or hard copy snail mail--and they can be sent to pharmacies, laboratories, radiology, materials management, medical directors, facilities management risk management or, as in our case, to loading docks.

Of course, this is in addition to FDA and other (ECRI, Medwatch, etc.) alerts and recalls notice sources, with information available on multiple Web sites, where users must visit and download files that can generate up to 80 printed pages, most of which they don't need.

Clearly, the problem of managing alerts is not simply an IT problem. It demands that the organization devote attention to internal processes and procedures, and that it assign responsibilities. Even the right alerts-management technology must have a defined, workable process in place at the enterprise level to efficiently push information where it needs to go. Timeliness is critical. But again, this is not a one-size-fits-all situation. There is no one internal process that will work for every healthcare organization.

Our internal root cause analysis identified many breaking points within the alert and recall process:

* Inaccurate contact information;

* Unclear policies and procedures;

* A manual, cumbersome, paper-based alerts-management process;

* Lack of accountability;

* Huge delays in posting of public notices (the bronchoscope recall did not appear on the manufacturers' Web site until May 2002);

* An astounding volume of alerts to manage, projected nationally at more than 8,000 per year.

SOLUTION

Johns Hopkins investigated whether we could develop our own information system to track and monitor recalls. A system, whether technology- or paper-based, is required to meet JCAHO standards and is useful for responding to other reviews from other regulatory agencies. As we analyzed the complexity of our requirements, we realized that we did not have the technical capabilities nor the organizational imperative to override other organizational priorities and make a technical solution a reality.

In 2002, we were approached by Mitretek Systems, a nonprofit science, engineering and information technology company in Falls Church, Va. Mitretek executives had also recognized the problem hospitals face in handling alerts, and they were exploring advanced system solutions to address the problem.

They suggested that we work with them to develop a solution that could meet the needs of all three hospitals in the Johns Hopkins Health System and also be used nationally by any hospital or healthcare system. Our decision to work with Mitretek on developing

RASMAS (Risk and Safety Management Alert System) was made, in part, because we saw no other solutions available in the marketplace, and because we wanted to influence the final product. We acted as both the alpha and one of three beta organizations to develop, test and recommend adjustments to the system.

IMPLEMENTATION

The "installation" of RASMAS within Johns Hopkins Hospital was simple and painless. Because the solution runs as an ASP on a subscription basis, there was no extraordinary outlay for technology or installation of new hardware or software. Most of our PCs were already operating with Internet Explorer 5.0 and could access the system. Any standard browser will work.

Here is how the system operates. An alert-processing subsystem in RASMAS collects alerts from a variety of sources including manufacturers and federal sources, and also allows subscribers to send information they have received. When we scan and send information to RASMAS, it is accessible within a 24-hour turnaround.