| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 1999;68:359-361
© 1999 The Society of Thoracic Surgeons
a Section of Cardiothoracic Surgery, Dartmouth-Hitchcock Medical Center, Lebanon, New Hampshire, USA
Address reprint requests to Dr Nugent, Section of Cardiothoracic Surgery, Dartmouth-Hitchcock Medical Center, One Medical Center Dr, Lebanon, NH 03756
e-mail: william.nugent{at}Hitchcock.org
Presented at the Thirty-fourth Annual Meeting of The Society of Thoracic Surgeons, Jan 26–28, 1998, New Orleans, LA.
Abstract
This report describes the activities of the Northern New England Cardiovascular Disease Study Group. The group consists of representatives from six institutions: Eastern Maine Medical Center in Bangor, and Maine Medical Center in Portland, ME; Optima Health Care in Manchester and Dartmouth-Hitchcock Medical Center in Lebanon, NH; Fletcher Allen Health Care in Burlington, VT; and Beth Israel-Deaconess Medical Center in Boston, MA. The Northern New England Cardiovascular Disease Study Group maintains a voluntary primary cardiac surgical database that has risk-stratified information on more than 60,000 consecutive patients who have undergone open heart surgical procedures in northern New England since 1987. In 1991, the group reported that significant variation in mortality rate existed between centers, a difference not explained by case mix. The finding led to a regional retrospective review of deaths in an effort to identify "mode of death." Mode of death is defined as that event that started the chain of events ultimately leading to the death of the patient. The most common mode of death was found to be low cardiac output syndrome. This information has led to a regional effort toward prevention, early recognition, and successful treatment of low cardiac output syndrome in the perioperative period.
The decision by the Health Care Financing Administration to make available to the public hospital-specific mortality data on patients having coronary artery bypass grafting (CABG) ushered in the era of accountability in health care. Ultimately, because of the limitations of administrative data sets, this effort was abandoned, though it stimulated the creation of a number of highly sophisticated primary cardiac surgical databases designed to level the playing field by attempting to provide accurate risk stratification.
Northern New England Cardiovascular Disease Study Group
The Northern New England Cardiovascular Disease Study Group was organized in 1987 as a consortium of hospitals performing open heart operations in northern New England. This group now comprises representatives from six institutions: Eastern Maine Medical Center in Bangor and Maine Medical Center in Portland, ME; Optima Health Care in Manchester and Dartmouth-Hitchcock Medical Center in Lebanon, NH; Fletcher Allen Health Care in Burlington, VT; and Beth Israel-Deaconess Medical Center in Boston, MA. The organization maintains a voluntary regional risk-stratified database designed to report outcomes on cardiac surgical patients by providing meaningful data to clinicians. From its inception, the intent was to use regional data to improve clinical outcomes in patients undergoing cardiac surgical procedures. The consortium maintains three registries with complete data on more than 60,000 consecutive patients. The six member institutions add approximately 8,000 cases annually.
Mortality results from database
In 1991, the Northern New England Cardiovascular Disease Study Group [1] reported on the first 3,055 patients entered into the database and found a twofold variation in mortality among the institutions (Table 1). Crude mortality rates between centers ranged from 3.1% to 6.3% and adjusted mortality rates, from 2.3% to 5.8%. These differences persisted after adjusting for severity and thus were not explained by case mix.
|
As this effort progressed, practice changes began to appear throughout the region. Examples of those changes included the development and sharing of critical pathways and protocols, the implementation of same-day admission programs for elective surgical procedures, the hiring of a full-time surgical assistant to help in the operating room, the altering of the timing and format of cardiac catheterization conferences, and a reduction in the number of personnel in the operating rooms. In part, as a result of these changes, the regional mortality rate dropped 24% over the ensuing 18 months despite the fact that the patient population seen was significantly older and sicker (Fig 1).
|
Mode of death
The mode of death was derived by first acknowledging that mortality after CABG was often the final consequence of a cascade of events. The goal was to identify the seminal event that began the downward spiral resulting in a mortality. A coding form was developed that provided a chronology of outcomes and then required the abstractor to choose a seminal event for each patient. Seven categories of possible modes of death were included: heart failure, dysrhythmia, respiratory failure, hemorrhage, infection, neurologic causes, and other causes.
Here are two examples of how to determine mode of death. The first example is that of a patient who undergoes an uneventful CABG. On postoperative day 2, he sustains a stroke (dense right hemiparesis). Pneumonia develops and the patient dies 2 weeks later of respiratory failure. The seminal event in this case was the initial stroke. The second example is that of a patient who has a difficult operation requiring an intraaortic balloon pump and high-dose pressor agents for separation from cardiopulmonary bypass. By the second postoperative day, it is evident that this patient has sustained a stroke (dense right hemiparesis), and the remaining course is that of pneumonia, respiratory failure, and death. In this case, the seminal event is the low cardiac output failure that was first noticed in the operating room.
Trained chart abstractors using predefined coding rules reviewed 8,641 consecutive patients having CABG between July 1987 and April 1991. They ultimately coded 387 deaths. All entries were blinded to center and surgeon, and discrepancies were resolved by an end-points committee. The results were subsequently validated and found to be both consistent and reproducible. Figure 2 shows the distribution of assigned modes of death. Of the 387 patients who died, the specific cause was heart failure in 65%, hemorrhage in 7%, dysrhythmia in 5.5%, respiratory failure in 5.5%, and neurologic causes in 7.3%. The remaining 9.9% fatalities were attributed to other specific causes or could not be unambiguously assigned.
|
Having found heart failure to be the most common mode of death in the region, the group decided to see how mode of death compared among patients of physicians with varying mortality rates. Thus, physicians were stratified into high-, medium-, and low-risk terciles. An examination of patient and disease characteristics revealed no difference in patient populations represented in each risk tercile. The surgeons were all operating on similar patients. However, when mode of death was compared with risk tercile, 80% of the difference between the highest-risk tercile and the lowest-risk tercile was explained by incidence of low cardiac output (Fig 3). Patients were equally likely to die of hemorrhage, dysrhythmia, respiratory failure, or stroke regardless of the surgeon but were more likely to die of low cardiac output if operated on by the high-risk group.
|
Conclusions
In this 11-year process, the Northern New England Cardiovascular Disease Study Group has learned that a multiinstitutional model using comparative data for reporting surgical outcomes is feasible and effective in improving performance. The members believe that this model can be similarly effective in other clinical settings and that the clinical workplace can and should become a learning laboratory. Although the group has continued to track institutional mortality rates, it has concentrated on examining the data for meaningful information that might help identify and understand those processes that truly have an impact on mortality rates.
I present this work as an example of what can be accomplished when regional outcomes data are collected and analyzed by clinicians in a safe environment. We must stop seeing the excellent work that has gone into creating surgical outcome databases as an end in itself. Instead, we must realize that these repositories of data are a means to a far more important end, namely, improving the outcomes for our patients. Only then will these databases become more than a report card for bureaucrats and become a tool for clinicians.
References
This article has been cited by other articles:
|
|
 |
|
  P J Bradshaw, K Jamrozik, M Le, I Gilfillan, and P L Thompson Mortality and recurrent cardiac events after coronary artery bypass graft: long term outcomes in a population study Heart, December 1, 2002; 88(5): 488 - 494. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
