Beyond the M&M

Hospitals review every death to flag problems, jump-start quality improvement.

Hospitals have long worked to learn from their patients' deaths, conducting morbidity and mortality (M&M) conferences to systematically review and address the events leading up to inpatient deaths. But now a handful of hospitals across the country are trying new approaches to mortality review, with some success.

In contrast to M&M conferences, which have traditionally focused on isolated cases and individual decision making, these new approaches take a more systemic, educational approach that discourages finger-pointing.

Illustration by Sarah Ferone
Illustration by Sarah Ferone

“We've been trying to shift from looking at individual actions to looking at teams and systems,” said Wendy Simon, MD, associate director of quality for the hospitalist service at the University of California, Los Angeles (UCLA) Medical Center. “Hands-on providers have a wealth of knowledge about quality of care, symptom management, and efficiency of processes. That's allowing us to pick up on issues that don't make it into the medical record and that we might have missed in the past.”

Innovative approaches

At UCLA Medical Center, which launched its rapid mortality review program a little over 3 years ago, every death is reviewed by a clinical care team made up of an attending physician, a resident, and a nurse involved with the patient's care, along with a documentation specialist. The review team convenes within 48 to 72 hours of a death, said Dr. Simon.

The team looks beyond errors that directly contributed to the death, she noted. Instead, “we consider what we could have done better even if the death itself was not preventable.”

The results of the discussions are captured in a database, enabling Dr. Simon and her team to track trends and isolate the most pressing problem areas. In addition to identifying errors that directly contributed to deaths, the reviews provide a rationale for prioritizing quality improvement efforts, said Dr. Simon.

As of July, just under 6% of the 535 cases in the database were classified as potentially preventable, which is in line with rates published by other institutions, while 55% revealed opportunities for improvement. The latter fell into 4 main categories: advanced care planning, communication and teamwork, systems issues, and delays in recognition of deterioration.

“We've found that a small number of deaths are potentially preventable, but more than half have something to talk about and about one-third inspires action,” said Dr. Simon. “If we only looked at that small group of potentially preventable cases, we would miss a lot of opportunities for improvement.”

Similarly, Boston-based Brigham and Women's Hospital's comprehensive mortality review system relies on direct input from frontline clinicians. That's a departure from the traditional practice of hiring third-party reviewers to analyze cases long after the fact.

“We ask frontline providers to give us their opinions right after a death occurs,” said ACP Member Allen Kachalia, MD, JD, chief quality officer at Brigham and Women's. “Rather than rely on chart documentation, our clinicians can speak from firsthand experience and tell us how we could have done better from a systems perspective.”

Within a few hours of a death at Brigham and Women's, the attending physicians receives a link to a web-based questionnaire in his e-mail inbox asking him to describe any potential complications during the case or make suggestions for improvement, said Dr. Kachalia. Currently, nurses do not receive the questionnaire automatically but the program is working to add that functionality. Responses are typically returned within 18 hours, and any concerns are flagged for further review.

A multidisciplinary hospital-wide mortality review committee meets once a month to discuss the flagged cases and determine what is needed for improvement. The goal is to use data from the reviews to inform quality and safety efforts, said Dr. Kachalia.

“Follow-up is key,” he noted. “The mortality review committee decides what cases require action and what the next steps should be. They then refer the case to the appropriate place and provide oversight to make sure something actually gets done.”

Some data suggest that comprehensive mortality reviews can lead to reductions in inpatient mortality. At the University of Pennsylvania Health System (Penn) in Philadelphia, for example, observed mortality decreased from 2.45% to 1.62% over 6 years following the creation of a mortality review committee in 2006. During the same time period, the health system also reported decreases in deaths from severe sepsis and improvements in delirium management, according to a 2013 paper published in the Joint Commission Journal on Quality and Patient Safety.

The process at Penn starts with sending a “360” form to all clinicians involved in caring for a patient within 72 hours of their death, said Craig Umscheid, MD, MS, FACP, vice chair of quality and safety for the department of medicine and director of Penn's Center for Evidence-based Practice. The goal of the form is to provide a 360-degree view of what has happened in each case. Unit-based leadership teams made up of a physician, a nurse, and a quality manager meet weekly to review the completed forms and flag cases for further review by the mortality committee.

The process has historically been coordinated by a mortality nurse, and the hospital has worked to improve the process since the study ended. For example, the questionnaire was shortened to increase efficiency and maximize participation by clinicians, and the process is being automated so it will no longer rely on manual coordination by a nurse.

Reviewing every death seems to be having an impact on mortality at UCLA and elsewhere. However, such programs can be challenging for smaller hospitals to establish, said Dr. Kachalia. “Time and resources are frequently cited as barriers to conducting comprehensive reviews of deaths,” he said. “The biggest struggle for many hospitals is how to respond to concerns from frontline providers and make changes in a timely way.”

Burnout can also be a factor, note the authors of an article about Mayo Clinic's Mortality Review System, published in the March 2014 Journal of Patient Safety. Reviewing multiple cases that end in death can be discouraging, they note, and also time consuming for physicians and nurses.

To make it work, hospital administrators have to be fully on board, said Dr. Umscheid. That might mean building some nonclinical time into providers' job descriptions to dedicate to reviews. Hospitals should also keep close track of their progress on related quality initiatives and whether they are associated with improvements in mortality.

“Doing this well requires high-level leadership buy-in,” he said. “You need resources from the system level, and administrators have to perceive a return on investment.”

Feeding the quality pipeline

Quality improvement should be the principal goal of a mortality review system, several hospitalists said. If done right, the review process can lay the foundation for measurable improvements in patient care and outcomes.

At Penn, for example, delayed identification of sepsis emerged as a common theme in reviews. That triggered the creation of an automated early warning and response system that led to earlier recognition and transfer to the ICU, as well as decreased sepsis mortality, according to a report coauthored by Dr. Umscheid in the January 2015 Journal of Hospital Medicine.

Aspiration was identified as another major driver of preventable mortality, said Dr. Umscheid. In response, an evidence-based risk assessment bedside tool has been integrated into the nursing assessment at admission. Since the tool was adopted, the hospital has seen significant reductions in aspiration pneumonia, he said.

UCLA's rapid mortality review has also fueled major improvement initiatives, said Dr. Simon. In response to issues that arose during reviews, the hospital has realigned its clinician education efforts to focus on advance care planning as well as reducing preventable deaths through earlier recognition of deterioration due to sepsis and other potentially treatable conditions.

While not necessarily reviewing every death, other hospitals are also rethinking the structure and focus of mortality reviews. At Johns Hopkins Hospital in Baltimore, for example, M&M conferences are still the main vehicle for discussing inpatient deaths, but the way they are conducted has shifted, said Daniel Brotman, MD, FACP, director of the hospitalist program.

“The traditional value of surgical M&Ms was to assign blame—that's an outdated culture that persists, but it is changing,” he said. “We're now choosing cases for M&Ms that have the potential to offer global lessons based on systems issues.”

Consider a hypothetical case of a patient who develops complications because he continued to receive vancomycin despite worsening renal failure, said Dr. Brotman. The tendency in the past has been to point fingers, but such adverse events can be rooted in a variety of system failures, including flaws in computer alert systems.

“It's a healthy change for us to be asking what might help prevent errors from happening again rather than focusing on blame because most errors can't be traced to a single practitioner,” he said. “We need a system with enough redundancy and support so that we're not reliant on individuals' actions to double-check things.”