Checking in with residents on the discharge summary
By Gina Shaw
Several years ago, John Fitzgibbons, FACP, chair of the department of medicine at Lehigh Valley Hospital in Allentown, Pa., was caring for a woman admitted with lung symptoms. Her chart indicated a history of uterine cancer that had led to a hysterectomy, so physicians were suspicious that the symptoms might point to a metastasis.
"As I went through her chart, I noticed that the initial diagnosis of uterine cancer couldn't be found," he recalled. "Finally, I found a discharge summary where it was included in the past medical history. The resident who had done the summary had seen that the patient had had a hysterectomy, and didn't know why. So he assumed it was uterine cancer and entered that in the summary, and from there on the woman carried that diagnosis in her medical records, when there was no evidence she'd ever had it."
Ask any chief of service in internal medicine about discharge summaries, and chances are they will sigh, roll their eyes or laugh. These summaries, essential to quality patient care and mandated by JCAHO, are a frequent bottleneck in many hospitals—often too long, incomplete, inaccurate or missing altogether.
Benjamin Lebwohl, MD, (standing, left) and Nicholas Fieback, FACP, (standing, right) train residents at New York Presbyterian Hospital/ Columbia in writing discharge summaries.
"Discharge summaries are a challenge for house staff and increasingly so," said Nicholas Fiebach, FACP, director of the Medical House Staff Training Program at Columbia University Medical Center in New York City. "First, what happens during hospitalization now has become even more complex and more detailed, with more data generated during a shorter length of stay. In addition, both programs and house staff are confronted by duty hours limitations, with less available time to catch up on things like discharge summaries."
But, as the case of the mysterious uterine cancer demonstrates, a poorly written discharge summary can interfere with making effective treatment decisions down the road. As a result, some programs are trying to quantify the problem and define solutions.
At Columbia University Medical Center, which is part of New York Presbyterian Hospital, discharge summaries have long been a challenge for house staff. In one survey, only 30% of summaries were available by the first patient follow-up visit (JCAHO mandates that a complete, finalized document is signed by 30 days after discharge, and a preliminary document by seven days after discharge).
Knowing that a JCAHO site visit was coming up in the fall of 2005, hospitalist Peter Stetson, MD, assistant professor of clinical medicine and a specialist in patient safety and informatics, decided to build a Web-based tool for discharge summaries that would integrate with the medical center's in-house Web-based clinical information system.
"Previously, discharge summaries were done by dictation, which could only be done in a couple of locations in the hospital--not at night, not on the weekends, not when the residents were most likely to have down time," he said. "What's more, dictation is not a skill that comes naturally to most people. We thought that if residents had a tool to work on discharge summaries throughout hospitalizations, it would make it easier to have them done at the time of discharge."
Dr. Stetson developed DSUM Writer, a Web-based application with pre-filled section headers, decision support, the ability to generate a "working draft" that can be edited over time, and access to other elements of the patient's electronic record to "pull in" information as needed.
Rolled out in February of 2005, DSUM Writer is now the tool of choice for about 80% of discharge summaries done at Columbia. Turnaround time for discharge summaries is about two weeks faster than under the old system, and in 2005 there was a 67% increase in discharge summaries completed compared with the previous year.
But what about quality? That's the next issue Dr. Stetson plans to evaluate. "It's hard to measure, because there is no gold standard or validated tools to measure the quality of a discharge summary," he said. "So we had to back up a step and create a validated tool, which we're still doing."
That lack of national standards distresses Dennis M. Manning, FACP, assistant professor in the department of medicine at Mayo Clinic College of Medicine in Rochester, Minn., and the department's quality and patient safety director. "There just aren't that many articles in the literature on the whole discharge process, and on the summary in particular, which is incredible, because it's such an important piece of information."
Dr. Manning and Mayo were ahead of the curve on discharge summaries: when the hospital's electronic medical record was initiated nearly a decade ago, the summaries were the first application implemented. At Mayo, discharge summaries are handed to the patient and sent to the referring physician, literally at the moment the patient leaves the hospital. "If you're going to do that, you have to have started the document earlier, in a skeleton or preliminary fashion days ahead," he said. "Our practice is to begin the document almost on admission."
Mayo's residents can dictate their initial summary, and then self-edit it throughout the patient's hospitalization on a template that is part of the hospital's electronic medical record. (Last year, the option to write and edit the entire summary, without using dictation, became available.) The system allows the resident to "pull" information from elsewhere in the chart, such as allergies, EKG readings or pathology reports. "It's a multidisciplinary document," said Dr. Manning. "There are sections for nursing to add information, and PT and OT to add notes for patients who are going to continued care."
Columbia's Dr. Fiebach noted that the electronic discharge summary can create another set of problems. "It's very tempting to cut and paste in an uncritical and unedited way. If a profusion of technical details gets dumped into the discharge summary, that doesn't solve the problem," he said. "Through orientation and conferences, as well as advice built into the template itself, we're working with our house staff not just to transfer information but to edit it: to provide only that information that will be helpful to the person taking care of that patient afterward."
When done well, these electronic summaries are contributing to an improved quality of care, said Columbia's internal medicine Chief Resident Benjamin Lebwohl, MD. "We had a patient with AIDS who had a prolonged hospital stay during which there were many diagnostic twists and turns. The ultimate diagnosis was cerebral toxoplasmosis," he said. "When the patient was readmitted two weeks later, the discharge summary pointed us exactly toward the essential events that had happened and the follow-up plan and medications. We did not have to reinvent the wheel and were able to pick up where the recent team had left off."
Some institutions have tried a less high-tech, simple template approach. Monmouth Medical Center in Long Branch, N.J., developed and introduced a summary template in 2003, after physicians realized that the discharge summaries were both too long and missing critical information.
Information that should have been in the summaries, but often wasn't, included procedures performed during hospitalization, the names of consulting physicians, and most important, the medications patients were discharged with.
"Meanwhile, there was a lot of extraneous information, like lab data," said Pratibha Rao, MD, author of a study on the new template published in the November-December 2005 issue of the American Journal of Medical Quality. "They'd give us lots of numbers about BUN and creatinine, etc., where it would have been sufficient to say that the patient was having mild renal insufficiency."
The template Dr. Rao and her colleagues created had 14 fields, including presenting complaint, "MAJOR" laboratory, imaging or pathology results, past medical history, name and specialty of all consultants, discharge medications, and follow-up. Brevity and accuracy were stressed with comments like this one after Principal Diagnosis: "This must explain presenting complaint."
From July 2003 on, Monmouth's residents dictated all their summaries using the template, which was prominently displayed in the hospital. The follow-up study found that the summaries' quality increased at the same time that their length decreased. Before the template's introduction, summaries averaged 81 lines; with the template, they dropped to 35 lines, a 67% reduction. Mean quality scores, although more difficult to quantify because of subjective variations in the ratings by the three assigned reviewers, also improved.
"The residents were very excited about the template," Rao said. "It was a major education point for them, telling them exactly what they needed to include and saving time and effort."
Physician, teach thyself
Education can also be a key component in improving discharge summaries, reports Jennifer S. Myers, ACP Member, a hospitalist and patient safety officer at the University of Pennsylvania. Despite the availability of a dictation card with guidelines, she said, "we were getting very lengthy summaries that were," too much of the H & P and too little of the hospital course." As at Monmouth Medical Center, discharge medications might be left out while extraneous test results would be included. "There was no consistency.
During 2004 and 2005, Dr. Myers and her colleagues did a randomized study of an educational intervention on dischargesummary best practices. All interns received either no teaching (the usual practice), a lecture about how to dictate a quality summary which featured examples of good vs. poor quality summaries, or the lecture plus an individual review session where their own summaries were reviewed with a hospitalist attending present for feedback.
The summaries were then graded with an evaluation tool created by the researchers that examined such factors as the inclusion of all proper summary contents, as well as subjective items like readability and brevity. "The group that got the lecture plus feedback had the greatest improvement in summary content, ,organization, and overall readability" Dr. Myers said. "The piece that improved the most was organization: they actually put in headings or left them out as appropriate."
The hospitalists also surveyed residents on their comfort level with the discharge summaries. "Prior to the training, the majority of interns did not feel comfortable with the process, had not had previous training in discharge summary best practices, and wanted to learn to improve their skills," Myers says. "The groups that got the intervention liked the teaching, thought it was useful, and recommended it for future intern classes."
That's the sticking point in practice. Although the required lecture has been incorporated, as part of one of the monthly intern reports in lieu of a clinical case, systematizing the feedback element has proven harder.
"We have a group of 11 hospitalists. At any given time on service, they can pull summaries and review them with interns. That always gets high marks," said Dr. Myers. "We've not been able to implement a hard and fast rule that every intern has to have a summary reviewed, because there are so many competing educational initiatives, but I think we're doing better than we were before when the quality of these summaries got no attention whatsoever."
Will these initiatives and others like them solve the underlying problem of discharge summaries—how to improve the quality of patients' follow-up care by disseminating complete and accurate information? Lehigh Valley's Dr. Fitzgibbons isn't entirely sure.
"Discharge summaries have not always been something we've focused on," he said. "We ask residents to go back and try to find the source documents. If the person has been seen by a cardiologist or a hematologist during the hospitalization, looking at that consult is likely to yield more information. It can be complicated to collect, but if you're focusing on a particular problem, you're likely to know what file is going to be the appropriate one."
As hospitals transition to electronic medical records, there are other ways to retrieve information, he noted.
"The key elements are the reason for admission, discharge diagnosis, discharge meds and discharge plan," said Dr. Fitzgibbons. "If we said that those are the key things we want to have done properly and didn't have to worry about other information, we'd be more likely to do a better job of getting summaries that would be valuable to people."
Dr. Manning sees the discharge summary a little differently.
"To write an excellent discharge summary on the patient, you have to understand case thoroughly and fully—the pathophysiology, multi-system issues, time sequencing and all the nuances of what's important and less important," he said. "The discharge summary isn't just a chore: it is a precious and necessary patient safety device as well as a quintessential educational tool."
Internist Archives Quick Links
What will you learn from your Annals Virtual Patient?
Annals Virtual Patients is a unique patient care simulator that mirrors real patient care decisions and consequences. CME Credit and MOC Points are available. Start off with a FREE sample case. Start your journey now.
Internal Medicine Meeting 2015 Live Simulcast!
Unable to attend the meeting this year? On Saturday, May 2, seven sessions will be streamed live from the meeting. Register for the simulcast and earn CME credit after watching each session. Watch it live or download for later viewing.