Critical Path Network

'Re-engineered' discharge uses checklist

Project RED incorporates 'virtual advisor'

When a patient discharged on a Friday is back in the hospital the following Monday, there can be many reasons. A few years ago, Boston Medical Center took on a project aimed at finding out whether discharge processes contributed to adverse events and rehospitalizations, and whether something could be done to turn that around. The result — re-engineered discharge.

Project director Brian Jack, MD, says while the key components of Project RED (re-engineered discharge) are things case managers and nurses have known all along, using them in a new way is paying off for patients and the hospital. It's now included as one of the 30 practices necessary in a safe hospital, as deemed by the National Quality Forum.

Project RED is a trial that redesigns the discharge workflow process and improves patient safety, Jack explains, through the use of a set of 11 components that are put to work as soon as a patient is admitted.

Most anyone involved in hospital discharge who reads the 11 components of RED will recognize that there's nothing new involved.

"Not at all — none of this is new, and there's nothing earth-shattering about any one of the components," Jack agrees. "But what's surprising is that no one had ever actually made the list before. It's not unlike what an airplane pilot does — the pilot and co-pilot go through a checklist every single time they get ready for takeoff, and when people are leaving a hospital, not only do [hospital staff] not go through a checklist, there really has been no agreed-upon checklist."

And with an estimated one out of every five of the 38 million hospital discharges each year resulting in an adverse medical event and preventable rehospitalization, the extra time it might take to work through the Project RED checklist is a small price, Jack adds.

What leads to rehospitalization?

Lack of standardization in the discharge process is a big contributor to adverse events and rehospitalizations within the 30 days after discharge, according to a report written by Jack and colleagues at Boston University School of Public Health.1

Nationally, studies indicate one-quarter of hospital-discharged patients are readmitted within 90 days, writes lead author Lee Strunin, PhD, usually as a result of "discontinuity and fragmentation of care at discharge."

In other words, Strunin says, the lack of coordination in the handoff from hospital to home or community care, gaps in social support, and inadequate physician follow-up can set in when a comprehensive discharge process isn't in place and followed closely during the initial hospitalization.

Sometimes, the breakdown is in transferring information about the hospitalization from a hospitalist who cared for the patient and the primary care physician who assumes care after discharge.

Jack described hospital discharge as "the poster child of patient safety" in comments to the Agency for Healthcare Research and Quality (AHRQ), which funded Project RED. Patients leave the hospital with discharge summaries that don't include important information; those summaries don't always wind up in the hands of primary care physicians; lab tests not received by the time of discharge can be overlooked; and follow-up with social supports can fall by the wayside, he explains.

The main thrusts of re-engineered discharge aim to eliminate four key contributors to rehospitalization:

  • Waiting until the discharge order is written before beginning the discharge process;
  • Case management staff hours — and, therefore, discharges — cover only the 7 a.m. to 3 p.m. shift;
  • Discharge information not always in the patients' language or, more commonly, not at their literacy level;
  • Discharge processes lacking benchmarking and continuous quality improvement.

Having a specific plan in place and persisting in making sure it's followed to the letter, has proved successful at Boston Medical Center, Jack says.

"We have done a randomized controlled trial of 750 patients, half in a usual care group and half in an RED group, and that study shows that the RED discharge is highly effective in decreasing ED visits and rehospitalizations within 30 days after discharge," says Jack. (Editor's note: The research paper he refers to had not been published at press time, but will be available at the Project RED web site,

Making Project RED hospital-friendly

While Jack acknowledges that more safe and effective discharges from the hospital, with a lessened likelihood of rehospitalizations, will save institutions money, he says a standardized discharge plan is, above all else, a safety issue.

"In the average hospital, we find that about 8 minutes is spent, on average, giving discharge instructions to patients," Jack explains. "During that time, the person's ready to go, the car's waiting, maybe there is a sick person needing to be moved on to the floor, and so the discharge instructions aren't communicated and understood as well as they should be, and that's a paradigm we need to change altogether."

The challenge, he says, "is to determine ways to make it work in busy hospitals."

The RED process involves not only the checklist for discharge, but also the designation of a discharge advocate whose job it is to begin working with the patient as soon as he or she comes into the hospital, so that discharge is an ongoing process rather than something saved until the last minute.

Staff education for re-engineering discharge is not too time-consuming, Jack explains, because it's what nurses, case managers, and social workers already know how to do. A training manual for staff and a booklet explaining the after-hospital care discharge plan (AHCP) to patients have been developed by Project RED, and are available as a free toolkit to download at

Meet 'Louise'

Still in trials is another aspect of the RED process, an "embodied conversational agent" named "Louise." Louise is an animated, interactive, patient education computer program created by computer scientists to teach the AHCP to patients.

Not a substitute for a human discharge manager, Louise is designed to talk (the program can produce speech in response to patient questions) to a patient at the time of discharge to explain the AHCP and answer questions about medications and the patient's individual discharge instructions (which are loaded into the program for each patient).

The program can answer questions about lab work, provide information on more than 300 medications, and — importantly — assess by the patient's responses whether he or she understands the AHCP.

If the patient doesn't understand, or has questions the program can't answer, a printed report goes to a nurse who can clarify the plan.

Unlike a nurse whose time is at a premium, the computer program can spend as much time with a patient as is necessary for him or her to feel comfortable and sure about the AHCP.

"A large percentage of people [for whom Louise has been used] like it," Jack says. "Some say they feel less intimidated about asking questions [of a computer] than they do of another person, and others like it just because they can control the pace."

Jack says the hope is that with easy training, a standard checklist, and the Louise program to augment the work of nurses and case managers, hospitals will find re-engineering their discharge processes to be cost-effective and a worthwhile investment in patient safety.

"We've found that 30 days after discharge, when we asked patients how prepared they were for discharge and whether they understood their AHCP, the difference between patients in the RED group and the non-RED group were significant," says Jack.


  1. Strunin L, Stone M, Jack B. Understanding rehospitalization risk: Can the hospital discharge be modified to impact recurrent hospitalization? J Hosp Med 2008; in press.