In the hospital setting, errors in the actions of doctors and other caregivers are a matter of life and death. We use root cause analysis (RCA) and failure mode and effects analysis (FMEA) to prevent errors. This should not be performed to assign blame, but rather to evaluate the system. In this case, we evaluate the death of Mr. B, a 67-year-old man with a dislocated hip. The details are described elsewhere, and will be addressed here in our RCA and description of items covered in the FMEA.
Root cause analysis provides a structure with which to determine causative and contributing factors, with the goal to improve future outcomes. Both individual and interconnecting systems should be examined (AHRQ, 2015). First in this case, there was no followup of the patient’s presenting tachypnea. Granted that rapid breathing occurs with pain, but respiratory rate higher than 27 is a predictor of cardiac arrest; over half of all patients with a sentinel event could have been identified up to 24 hours sooner based on respiratory rate with better than 95% accuracy (Cretikos et al, 2008). Second, Mr. B’s medications were not checked in detail. Many chronic users of pain medications are on relatively high doses due to increased tolerance (Ullrich, 2012). Oxycodone may have added to the effect of the hydromorphone. Nurses should have asked the patient and pharmacist about medication types, dosages, and schedules. With the ED being so busy, the doctor may have been in too much of a rush to achieve adequate sedation. Warnings for IV diazepam include a possibility of apnea and/or cardiac arrest; and concurrent narcotic medication dosages should be reduced by one third and given in small increments (RxList Inc, 2015). Hydromorphone warnings state emphatically that patients should be monitored closely for respiratory depression (Drugs.com, 2015). Particularly given this patient’s tachypnea on admission, when the alarm sounded for oxygen saturation at 85%, immediate action should have been taken. Actions taken during the code were appropriate; it was simply all too late.
Therefore, a process improvement plan would include these elements: First, standard close monitoring of respirations of any patient with a high respiratory rate or on narcotics (Cretikos, 2008). As this information is entered into the computer system, it might be programmed to send a message to RT for these conditions. Secondly, a standard admission process to obtain reliable details of the patient’s medication usage from the patient’s pharmacist and verified with the patient, or the family or caretakers if the patient is unresponsive, with this information relayed to the Hospital Pharmacy, with new medication doses to be determined by them. Thirdly, nurses should be asked if they felt constrained against questioning the doctor if they perceived any risks. Can physicians be more communicative with nursing staff? Answers would determine whether this should be part of the improvement plan. Fourthly, low oxygen saturation was ignored. This is an area where LPN’s should be trained in the conscious sedation protocol as Nurse J has been. Conscious sedation training might be modified to include monitoring of respiratory status beyond pulse oximetry. Fifthly, staffing was too low. The ACEP (2014) recommends a ratio of one doctor for every 1.8-2.8 patients per hour; although if there is a physician’s assistant or nurse practitioner available, either can typically handle about 30% of the typical ED patient load. In this case, staff was overwhelmed, but they did not call for available backup. A protocol must be established as to when to call for more staff, who and from where.
The change theory that I would use for implementation is Kurt Lewin’s of unfreeze-move-refreeze (CurrentNursing, 2013). First we unfreeze by increasing driving forces and decreasing restraining forces. This death will drive the desire for change, as will JCAHO mandates. A major restraining force will be fear of blame. Nurse leaders can consult with Administration and then reassure on this issue. Participants can also be assured that the process will be kept confidential. Another restraining force will be concerns about new procedures taking up too much time, which automating as many of the new processes as possible will address. Another would be costs; but they are minimal compared to the costs of lawsuits and damage to reputation. During the moving stage, computers and devices would be reprogrammed to make many processes automatic. Then protocols would be written and all staff trained on them. Reminder sheets with steps in the protocols would be posted where easily visible, perhaps one in each patient room. Finally refreezing occurs when the new behavior becomes a habit. New behaviors could be practiced regularly and/or trainers could visit the ED regularly and quiz staff on “What would you do if” scenarios. Training and retraining, quizzing and practicing should continue on a regular basis, along with constant process improvement.
A failure mode and effects analysis (FMEA) improves the likelihood that the plan will not fail (NCPS, 2002). Ideally, this analysis would be run before any adverse events occur in order to prevent them; indeed, the JCAHO requires that at least one high-risk process be identified each year, analyzed, and redesigned to protect patients from risk. In the current case, however, we will review the steps to conduct the FMEA now to prevent further serious adverse events in the future:
1. The members of the interdisciplinary team would include an ED doctor, an ED RN familiar with the conscious sedation protocol, an ED LPN, a representative from the IT Department to advise on the possibilities of automated processes, a hospital pharmacist and a respiratory therapist.
2. The steps for preparing for the FMEA would include determining the individuals who will join the team; reviewing the patient’s chart and any other written records of the events regarding Mr. B; interviewing those involved and perhaps requesting their written reports of events; also discussing the events with hospital administration to establish that the goal of these analyses is not punishment.
3. The next steps of the FMEA are severity, occurrence, and detection. Our five recommendations were: (1) That any patient with tachypnea or on narcotics such be monitored by Respiratory Therapy. (2) That all prior medications be detailed thoroughly and information relayed to and from the hospital pharmacist. (3) To improve nurse-doctor communication when nurses have concerns. (4) To improve monitoring of respiratory distress, including more knowledgeable response to pulse oximetry and monitoring beyond pulse oximetry. (5) To increase ED staffing.
The FMEA team may be able to add to these and refine these recommendations. For each recommendation, they need to determine the best methods for detecting when the threshold has been reached to implement the new processes, the likely frequency of the occurrence of each, and the potential severity of errors of each, as well as any new potential errors introduced by each recommendation, which further refinement may avoid. So for example, for Respiratory Therapy to monitor each patient with tachypnea, is 24 breaths per minute the best threshold? Certainly the potential consequences are most severe, as these may precede cardiac arrest or respiratory depression. RT can provide further guidance regarding the best monitoring devices and Pharmacy can provide a full list of medications to be included. IT can identify any problems that might arise with automatic notification. Would possible hackers, viruses, and/or power failure necessitate a backup protocol for notification of RT in these cases? The doctor can help us determine if there would be unacceptable delays with more standard consultation of the Pharmacy. Perhaps medications could be scanned before delivery so that problematic dosages or drug interactions would trigger an automated warning.
The interventions from the process improvement plan would first be tested by discussion and brainstorming among the FMEA team, seeking potential failures and remedies. A feedback system should be put in place once improvements are launched, so that unanticipated problems can be immediately addressed. Someone in each professional category should be responsible for monitoring the literature and standards of professional groups, to keep processes continually updated.
The professional nurse may function as a leader in promoting quality care and influencing quality improvement activities in several ways. S/he may report any errors that come to her attention, and initiate analyses such as we have done here. S/he may volunteer for FMEA teams and similar projects, serving as liaison to involve key participants. S/he may identify and ameliorate causes of resistance to change. S/he may write protocols and training materials for improved processes. S/he may work with IT to determine where processes may be automated for decreased error rates. S/he must speak up regarding understaffing and its hazards. Any area that promotes patient safety and improves hospital outcomes is appropriate for nurse leadership.
- Agency for Healthcare Research and Quality (AHRQ). (2015). Root cause analysis. Patient Safety
Network. Retrieved from https://psnet.ahrq.gov/primers/primer/10/root-cause-analysis
- American College of Emergency Physicians (ACEP). (2014). Staffing an ED appropriately and
efficiently. Clinical & Practice Management. Retrieved from http://www.acep.org/Clinical—
- Cretikos, M.A., Bellomo, R., Hillman, K., Chen, J., Finfer, S., & Flabouris, A. (2008). Respiratory rate:
The neglected vital sign. Med. J. Aust., 188 (11), 657-659. Retrieved from
- CurrentNursing. (2013). Change theory, Kurt Lewin. Nursing Theories. Retrieved from
- Drugs.com. (2015). Hydromorphone dosage. Hydromorphone Dosage Guide with Precautions.
Retrieved from http://www.drugs.com/dosage/hydromorphone.html
- National Center for Patient Safety (NCPS). (2002). The basics of healthcare failure mode and effect
analysis. NCPS. Retrieved from http://www.patientsafety.va.gov/docs/hfmea/FMEA2.pdf
- RxList Inc. (2015). Warnings. Diazepam injection. Retrieved from http://www.rxlist.com/diazepam-
- Ullrich, P.F. (2012). Narcotic drugs for the treatment of chronic pain: A double edged sword. Spine-
health. Retrieved from http://www.spine-health.com/blog/narcotic-drugs-treatment-chronic-