An 85-year-old man with advanced oxygen-dependent chronic obstructive pulmonary disease (COPD) presented to the emergency department (ED) with increasing shortness of breath and cough. Initial evaluation demonstrated worsening hypoxemia and a chest x-ray showing a new, large left-sided pleural effusion. A therapeutic thoracentesis was performed, which relieved the patient’s symptoms, but the etiology of the effusion remained unclear.

At the time of admission, the resident asked the patient about his advance directives, and he stated his wish to be DNR/DNI (do not resuscitate/do not intubate). The patient’s wife confirmed that her husband never wanted to be “shocked or placed on a breathing machine.” The resident placed a note in the chart to document the discussion.

A few days later, the patient was found unresponsive and pulseless. A code blue was called. The on-call resident (different from the admitting resident) responded and found the patient to be in ventricular fibrillation. Unaware of the patient’s advance directive, the resident successfully resuscitated the patient (with medications and shocks) and transferred him to the ICU. The resident then contacted the patient’s wife, who reiterated the patient’s wishes to not undergo such measures. Immediately following this discussion, the patient again became pulseless, and resuscitative efforts were appropriately withheld. The patient died within minutes.

At this particular hospital, the policy was that residents should both document any code status discussion and enter a DNR/DNI “order” (to be cosigned by the attending later) in the electronic medical record. In this case, the discussion with the patient regarding code status was appropriately documented, but a specific DNR/DNI “order” was not entered into the medical record. Had the order been entered, it would have triggered the placement of an easily visible wristband onto the patient by the nursing staff, who would have also documented the order in their nursing records (neither of which happened). Even though there was no formal DNR order in the electronic record, the nurses might have chosen not to “call the code” had they seen the record of the code status discussion in the resident’s progress note. Unfortunately, there was no computer terminal at the patient’s bedside (and there were no longer any paper medical records), and so the bedside nurses had no access to the record of the DNR discussion, which contributed to the error. The end result was that the patient was resuscitated when he explicitly told his providers that he wished not to be.

by Mary K. Goldstein, MD, MS

Communication among many actors is a crucial component of medical care. The first relevant communication in this case is the discussion between the resident and the patient about code status. An appropriate DNR discussion provides the patient with sufficient understandable information necessary for him to make a voluntary and informed decision about whether he would like CPR. This includes information about the nature of the patient’s medical condition and prognosis, as well as the nature of the procedure and its expected risks and benefits for this patient. Such discussions require exquisite tact and sensitivity to cultural issues and to the patient’s prior beliefs about resuscitation.(1-4) Informed consent also requires that the patient has capacity to understand the information presented and to communicate a decision and that he or she be free from coercion.(5) The physician who discusses code status with the patient should enter a chart note documenting the conversation, with enough information to communicate clearly to other health care professionals the plan of care based on the patient’s decision. Local policy may dictate the required elements, but in general the note should include the basis for the patient’s decision, the patient’s capacity to understand and communicate, the plan for management, and, if any factors complicate the discussion (eg, language barriers), how these were addressed. Note that a request for DNR should not be generalized, absent specific discussion, to limitation of other treatment.(6)

In this case, the code status order was not entered, essentially failing to transmit important information to other actors in the patient’s care. One study found “unwritten DNR orders” (situations in which the patient’s DNR status was known to the team but not recorded) in 2% of seriously ill poor-prognosis patients.(7) The same study found that 31% of patients who did have a DNR order lacked documentation of a consent conversation.

Common reasons for such omission errors include lack of knowledge and simple forgetting. An inexperienced resident may have assumed that nursing staff would act on his chart note without an order. Perhaps the resident was aware of the need for a specific order but forgot to enter it. The admission process involves juggling many pieces of information, and it is easy for the physician to become distracted by another aspect of the patient’s case or by an interruption. Physicians work in high-interruption environments, where some tasks may be dropped by even the most conscientious resident. Ideally, a code status discussion would include the nurse, who offers a continued presence after the initial conversation. This practice would provide an additional safety measure to ensure that important clinical decisions made with the patient at the bedside are communicated to others and the medical chart.

A final error in communication occurred in this case when the medical record was unavailable during the code. It is ironic that a system intended to make patients’ records legible and widely available—the electronic medical record (EMR)—was not available at the bedside.

Systems Solutions

What could potentially minimize the risk of omitting an order when a problem has been considered carefully? To address lack of knowledge as a potential cause, residents (and all physicians) often need training in hospital operations. When residents think about working as a team, they usually define their team by the attending, their fellow residents, and the medical students. Generally, they fail to recognize that they function as part of a larger group of care providers that includes nurses, pharmacists, therapists, clerical staff, and others. A thought-experiment of walking through the frequently nurse-driven activation of a code—including how the nurse checks the patient’s code status—may help the resident develop a “systems perspective.” It would allow residents and other physicians to see how their actions fit with the actions of other providers and within the overall organizational structure.

It also appears that the resident in this case failed to transmit information to the on-call team. With frequent transfer of care from one physician to another, it is essential that a health care system have consistent and standardized procedures for the “hand-off” of information.(8) When not done well, this is a source of potential error. The Joint Commission on Accreditation of Healthcare Organizations (JCAHO) made improving hand-offs one of its 2006 National Patient Safety Goals.

With respect to information technology (IT), effective implementation can certainly improve the availability of advance directive information. For example, at least one EMR system allows for a progress note created with an “Advance Directive” title to automatically generate a posting/alert. Alternatively, the growing use of standard order sets (eg, admission orders, preoperative orders, transfer orders) also offers an opportunity for intervention. This process might generate a prompt for a “code status order” on all hospitalized patients. A more sophisticated EMR could potentially prompt for a code status order when an advance directive note is entered (and vice versa), provided the system has a particular template or process for such notes.

Implementing technology successfully goes far beyond installing the software. Technologically elegant solutions may fail to improve care if not implemented wisely.(9-11) Introduction of new technology creates its own new sources of error.(9) For example, a careful observational study of the introduction of computerized physician order entry (CPOE) to an intensive care unit noted problems that arose with introduction of new technology intended to improve patient safety.(12) Many of the identified problems involve a mismatch between the workflow conceptualized in CPOE and the actual workflow, a disconnect that drives clinicians toward devising workarounds. More recently, a study in the pediatric literature suggested increased mortality after implementation of a commercial CPOE system in an intensive care unit.(13)

IT solutions are often tested with computer-oriented early adopters who may not respond to automated systems in the same manner as typical clinician-users. Such solutions should be developed in an iterative design process in which early implementations are conducted in actual patient care settings, followed by cycles of re-design until the system works well in the clinical setting with a wide variety of clinician-users. Refining important details such as the location of computers (eg, bedside portable terminals) and adapting systems to particular clinical settings (such as intensive care units) are critical details that cannot be overlooked. Finally, there must be adequate back-up systems for managing computer slowdowns and power failures.

IT provides a necessary and effective tool to allow providers to improve the quality and safety of care. However, IT systems should be viewed as tools rather than solutions. This case illustrates that even a computerized process to ensure communication of DNR orders will always rely, to some degree, on the actions of the care team to ensure that the patient’s wishes are respected.

Take-Home Points

• Proper identification of patients with DNR/DNI orders requires effective systems to support busy and easily interrupted providers and care processes.
• Physician hand-offs should include information about code status for all patients.
• IT provides an important tool for patient care but may also introduce new opportunities for error.

Mary K. Goldstein, MD, MS
Professor of Medicine
Geriatrics Research Education and Clinical Center, VA Palo Alto Health Care System
Center for Primary Care and Outcomes Research, Stanford University School of Medicine

1. Tulsky JA. Beyond advance directives: importance of communication skills at the end of life. JAMA. 2005;294:359-365.
[ go to PubMed ]

2. Ruhnke GW, Wilson SR, Akamatsu T, et al. Ethical decision making and patient autonomy: a comparison of physicians and patients in Japan and the United States. Chest. 2000;118:1172-1182.
[ go to PubMed ]

3. Mebane EW, Oman RF, Kroonen LT, Goldstein MK. The influence of physician race, age, and gender on physician attitudes toward advance care directives and preferences for end-of-life decision-making. J Am Geriatr Soc. 1999;47:579-591.
[ go to PubMed ]

4. Clarke DE, Goldstein MK, Raffin TA. Ethical dilemmas in the critically ill elderly. Clin Geriatr Med. 1994;10:91-101.
[ go to PubMed ]

5. Goldstein MK. Ethics. In: Ham RJ, Sloane PD, Warshaw GA, eds. Primary Care Geriatrics: A Case-Based Approach. St. Louis, MO: Mosby; 2002.

6. Beach MC, Morrison RS. The effect of do-not-resuscitate orders on physician decision-making. J Am Geriatr Soc. 2002;50:2057-2061.
[ go to PubMed ]

7. Sulmasy DP, Sood JR, Ury WA. The quality of care plans for patients with do-not-resuscitate orders. Arch Intern Med. 2004;164:1573-1578.
[ go to PubMed ]

8. 2005 National Patient Safety Goals. Oakbrook Terrace, IL: Joint Commission on Accreditation of Healthcare Organizations; 2005.

9. Goldstein MK, Hoffman BB, Coleman RW, et al. Patient safety in guideline-based decision support for hypertension management: ATHENA DSS. Proc AMIA Symp. 2001;:214-218.
[ go to PubMed ]

10. Ash JS, Berg M, Coiera E. Some unintended consequences of information technology in health care: the nature of patient care information system-related errors. J Am Med Inform Assoc. 2004;11:104-112.
[ go to PubMed ]

11. Berg M. Implementing information systems in health care organizations: myths and challenges. Int J Med Inform. 2001;64:143-156.
[ go to PubMed ]

12. Cheng CH, Goldstein MK, Geller E, Levitt RE. The effects of CPOE on ICU workflow: an observational study. Proc AMIA Symp. 2003;:150-154.
[ go to PubMed ]

13. Han YY, Carcillo JA, Venkataraman ST, et al. Unexpected increased mortality after implementation of a commercially sold computerized physician order entry system. Pediatrics. 2005;116:1506-1512.
[ go to PubMed ]