Prologue
During a midnight shift in an intensive-care unit (ICU), a patient suffering from an infectious disease was hooked on to a ventilator and was being monitored by a physio-logical monitor in an isolation room (with an anteroom). While this patient was asleep, the assigned caregiver decided to help transfer her other patient, who was
recovering well, back to a regular floor. The first patient managed to extubate himself while the caregiver was out of the isolation room. No one heard the ventilator alarm or the physiological monitor alarm. When the caregiver returned to the room, she immediately called a code and attempted to resuscitate the patient. However, the patient could not be revived.
What is the relevance of this story to proactive risk assessment and Health Care Failure Mode and Effect Analysis (HFMEA)™?A team examining this high-risk situation might have identified a number of vulnerabili-ties that could have been mitigated to avoid the conse-quent harm and personal tragedy that occurred. Table 1 depicts how this outcome may have been averted by using the HFMEA™proactive risk assessment model.
JCAHO Standards
The Joint Commission for the Accreditation of Health-care Organizations (JCAHO) drives much of the activi-ty within health care through its standards for attaining accreditation. Although these standards are voluntary, almost all hospitals choose to apply for JCAHO accred-itation. The new LD 5.2 JCAHO patient safety stan-dard reads as follows: “Leaders ensure that an ongoing, proactive program for identifying risks to patient safety and reducing medical/health care errors is defined and implemented.” The intent section requires that at least 1 high-risk process be examined annually. For this process, health care facilities are required to identify failure modes and the possible effects for each failure mode. For the most critical effects, facilities also must determine what events would lead to a failure and take steps to mitigate that possibility.
This selection is to be based, in part, on information published periodically by the JCAHO that identifies the most frequently occurring types of sentinel events. The National Center for Patient Safety (NCPS) will also iden-tify patient safety events and high-risk processes that may be selected for this annual risk assessment for its hospitals.
How to Make the Most of Failure
Mode and Effect Analysis
Erik Stalhandske, MPP, MHSA; Joseph DeRosier, PE, CSP;
Bryanne Patail, BS, MLS; John Gosbee, MD, MS
Current accreditation standards issued by the Joint Com-mission for the Accreditation of Healthcare Organizations (JCAHO) require hospitals to carry out a proactive risk assessment on at least 1 high-risk activity each year for each accredited program. Because hospital risk managers and patient safety managers generally do not have the knowledge or level of comfort for conducting a proactive risk assessment, they will appreciate the expertise offered by biomedical equipment technicians (BMETs), occupational safety and health professionals, and others. The skills that have been developed by BMETs and others while conduct-ing job safety analyses or failure mode effect analysis can now be applied to a health care proactive analysis. This article touches on the Health Care Failure Mode and Effect Analysis (HFMEA)™ model that the Department of Vet-erans Affairs (VA) National Center for Patient Safety developed for proactive risk assessment within the health care community. The goal of this article is to enlighten BMETs and others on the growth of proactive risk assess-ment within health care and also on the support docuassess-ments and materials produced by the VA. For additional informa-tion on HFMEA™, visit the VA website at
www.patientsafety.gov/HFMEA.html. (Biomedical Instrumentation & Technology 2003; 96–102).
From the Department of Veterans Affairs National Center for Patient Safety, Ann Arbor, Mich.
Address correspondence and reprint requests to Mr Stal-handske, Department of Veterans Affairs National Center for Patient Safety, 24 Frank Lloyd Wright Drive, Lobby M, PO Box 486, Ann Arbor, MI 48106. [email protected]
Erik Stalhandske, MPP, MHSA, et al
MANAGEMENT & TECHNOLOGY
The Approach of the Department of
Veterans Affairs to Failure Mode
Effect Analysis and Roll-Out
Health care failure mode and effect analysis was devel-oped in the summer of 2001 by the Department of Vet-erans Affairs (VA) National Center for Patient Safety (NCPS). Facilities could use this process to meet the new proactive risk assessment requirement contained in the JCAHO patient safety standard LD.5.2. HFMEA™ streamlines the hazard analysis steps of the traditional
failure mode and effect analysis (FMEA) process by combining the detectability and criticality steps of the traditional FMEA into an algorithm presented as a deci-sion tree. It also replaces calculation of the risk priority number with a hazard score that is read directly from the hazard matrix table. This table was developed by NCPS specifically for this purpose. The Department of Veter-ans Affairs introduced the program through 4 national videoconferences during the last week of August 2001. About a week before the videoconference, written mate-rials and a PowerPoint presentation on the HFMEA™ process were distributed to the participants to generate discussion during the training session. Approximately one fourth of the 163 Veterans Health Administration facilities (VHAs) participated on each call.
The videoconferences began with a 50-minute taped presentation covering the basics of HFMEA™and pre-sented 3 examples, 2 of which were health care–specific topics (bar code medication administration and prostate-specific antigen testing). There were 3 breaks for questions and answers during the broadcast. At the conclusion of the taped broadcast, a panel comprising of James P. Bagian, MD, PE, Joseph M. DeRosier, PE, CSP, and Erik J.
handske, MPP, MSHA, led a discussion on proactive risk assessment and addressed questions on the process.
In September 2001 a copy of the taped presentation was distributed to all VHAs to help train HFMEA™ team members. The HFMEA™technical manual, copies of slides used in the videoconference, and answers to fre-quently asked questions were distributed with the videocas-settes.
In December 2001 the NCPS started to receive requests for additional HFMEA™ training. Based on these requests, a 1-day HFMEA™training course was developed and has been conducted repeatedly. This course included more detailed explanation on how to use the HFMEA™ decision tree, along with large-group exercises and small-large-group breakout sessions. Some of this material was presented at the Association for the Advancement of Medical Instrumentation (AAMI) conference in June 2002.
Basics of HFMEA
™
Key Aspects of the HFMEA™Process
Health care failure mode and effect analysis is a 5-step
process that uses a multidisciplinary team to proactively evaluate a health care process. The team uses process flow diagramming, a hazard-scoring matrix, and the HFMEA™decision tree to identify and assess potential vulnerabilities. The HFMEA™ worksheet is used to record the team’s assessment, proposed actions, and out-come measures. Health care failure mode and effect analysis includes testing to ensure that the system func-tions effectively and that new vulnerabilities have not been introduced elsewhere in the system.
Step 1: Define the HFMEA™Topic.
• Define the topic of the HFMEA™ along with a clear definition of the process to be studied. Think about narrowing the scope so that the review is manageable and the actions are operationally sound (see Figure 1).
Step 2: Assemble the Team.
• The team should be multidisciplinary and should include subject matter expert(s) and an advisor (see Figure 1).
Erik Stalhandske, MPP, MHSA, et al
MANAGEMENT & TECHNOLOGY
Step 3: Graphically Describe the Process. • Develop and verify the flow diagram (this is a process
vs. chronological diagram).
• Consecutively number each process step identified in the process flow diagram.
• If the process is complex, identify the area of the process to focus on (take manageable bites).
• Identify all subprocesses under each block of this flow diagram. Letter these substeps consecutively (ie, 1a, 1b, …, 3e, etc).
• Create a flow diagram composed of the subprocesses. Letter these substeps consecutively.
(Hint: It is very important that all process and subprocess steps be identified before proceeding.)
Step 4: Conduct a Hazard Analysis.
• List all possible/potential failure modes under the subprocesses identified in HFMEA™Step 3. Num-ber these failure modes consecutively (ie, 1a(1), 1a(2), …, 3e(4), etc). Transfer the failure modes to the HFMEA™worksheet. (See Figure 2.)
(Hint: Failure modes include anything that could go wrong that would prevent the subprocess step from being carried out; they describe what could go wrong. For example, if logging onto a laptop computer is the process step, possible failure modes include not being able to log in and delayed login.This is the step in the process where the expertise and experience of the
team really pays off. Use vari-ous methods including the National Center for Patient Safety’s triage/triggering ques-tions, brainstorming, and cause and effect diagramming to identify potential failure modes.)
See www.patientsafety.gov. • Determine the sever-ity and probabilsever-ity of the potential failure mode and record these on the HFMEA™ worksheet. Look up the hazard score on the hazard score matrix and record this number on the HFMEA™worksheet. (See Figures 3–5.)
• Go to the HFMEA™ decision tree. Use the deci-sion tree to determine if the failure mode warrants fur-ther action. Record the action to “proceed” or to “stop” on the HFMEA™worksheet. If the action is to “stop,” proceed to the next subprocess identified in Item 2. (Note: If the score is 8 or higher, document the ration-ale for any “stop” decisions.) (See Figure 6.)
• List all the failure mode causes for each failure mode where the decision is to “proceed” and record them on the HFMEA™worksheet.(Hint: Each failure mode may have multiple failure mode causes. Failure modes include anything that could go wrong that would prevent the subprocess step from being carried out. For example, if logging onto a laptop computer is the process step, possible failure modes are not being able to log in and delayed login. Possible failure mode causes would include the computer not being available, no power, no login ID for the operator, etc.) Step 5: Actions and Outcome Measures
• Determine if you want to “eliminate,” “control,” or “accept” the failure mode cause. Record this decision on the HFMEA™worksheet.
• Identify a description of action for each failure mode that is to be eliminated or controlled.
(Hint: Place the control measure in the process at the earliest feasible point. Multiple control measures can be placed in the process to control a single hazard. A control measure can be used more than once in the process. Solicit input from the process owners if they are not represented on the team. Try to
simulate any recommended process change to test them before facility-wide implementation.)
• Identify outcome measures that will be used to ana-lyze and test the redesigned process.
• Identify a single, responsible individual by title to complete the recommended action.
• Indicate whether top management has concurred with the recommended action.
• Test to ensure that the system functions effectively and that new vulnerabilities have not been introduced elsewhere in the system.
How to Develop Reasonable and
Con-crete Failure Modes
Besides brainstorming, there are several techniques that should be used to develop reasonable and concrete fail-ure modes once the process diagrams are complete and the focus areas are chosen. Reviewing databases, such as Food and Drug Administration’s MAUDE, could identi-fy malfunctions and user-interface design issues such as inadvertent shutdown of external pacemaker machines. Usability tests performed by the HFMEA team or found
in the literature survey can be useful.1Usability testing is
a human factors engineering technique applicable to devices, work areas, or larger processes.2
Another technique is patient safety rounds. For exam-ple, on discussing the matter with ICU nurses and resi-dents, you may find that they are worried about missing important alarms several times in a week because of dis-tractions or noise levels. Another approach is to use the findings from routine safety assessment tool “audits.” For example, you might find that many of your hospital’s com-pressed-air wall outlets have green “Christmas tree” adap-tors attached—the VA’s approach to this issue has been to recommend that all facilities switch to a clear adaptor. This avoids the potential hazard of matching the wrong color adaptor and thereby providing the wrong color cue. In some complex and difficult HFMEA™s, videotaping can provide documentation of high-hazard areas. Analysis of the video provides data on close calls or adverse events and is one of the most reliable development tools for list-ing failure modes. For example, the University of Mary-land has used videotape documentation to research safety issues during respiratory arrest resuscitation.3
Erik Stalhandske, MPP, MHSA, et al
MANAGEMENT & TECHNOLOGY
Figure 5. Hazard-scoring matrix.
Figure 6.Decision tree.
Figure 3. Severity rating.
Conclusions
Failure mode effect analysis (FMEA), once the sole domain of engineers and production industries, is now finding a receptive audience within health care. Those who have FMEA experience and skills may find them-selves in demand to respond to the new JCAHO stan-dards and the need for proactive risk assessment within health care.
References
1. Welch DL.Human factors usability test and evaluation.
Bio-med Instrum Technol.1998;March/April:183–187. Volume, 32 No.2
2. Gosbee JW, Arnecke B, Klancher J, Wurster H, Scanlon
M. The role of usability testing in healthcare organizations.
Proceedings of the Human Factors Society 40th Annual Meeting. Santa Monica, Calif: Human Factors Society; 2001:1308–1311. Minneapolis, MN 2001
3. Xiao Y, Moss J.Practices of high reliability teams:
observa-tions in trauma resuscitation. Proceedings of Human Factors & Ergonomics 44th Annual Meeting.Santa Monica, Calif: Human Factors & Ergonomics Society; 2001:395. Minneapo-lis, MN 2001
The Role of BMETs and Clinical
Engineers in Risk Assessment and
Procurement
The Joint Commission for the Accreditation of Healthcare Organizations (JCAHO) Leadership Stan-dard LD.5.2 requires proactive risk assessment. The intent statement for this standard requires leaders to select a minimum of 1 high-risk process for proactive risk assessment each year. Clinical engineers and bio-medical equipment technicians are excellent candidates for participation in proactive risk assessment teams.
Proactive risk assessment is a systematic, step-by-step, methodical process that requires logical think-ing. Engineers, by the very nature of their education and experience, are perfectly suited to this task given their broad knowledge of how systems work and their in-depth knowledge of how medical devices function. Their experience in investigating medical device-related incidents/variances gives them a good under-standing of how medical equipment may be misinter-preted by users because of poor human factors design. This experience can be put to good use when trying to develop potential failure modes or failure mode caus-es, and for proactive risk assessment.
BMETs can also broaden their impact by assisting in procurement activities. This can increase the visi-bility of your department, help ensure the safety of the procurements, and positively affect the bottom line.
Engineers are taught to think in terms of systemat-ic, logical, objective-based, and well-supported con-clusions. Consider researching how procurement decisions are reached and convincing hospital officials about the added benefit of applying your skills to the procurement process.
• Provide a proactive approach to procurement— think about what might go wrong with the differ-ent models under consideration and provide a modified failure mode effect analysis to compare different models.
• Encourage end-user input into the procurement decision process.
• Include consideration of past experience (good or bad) with the particular manufacturer.
• Use existing data from your colleagues, ECRI, and Food and Drug Administration’s MAUDE database to support the safest purchase.
• Develop explicit criteria to enhance the objective evaluation of the equipment.
Definitions
Effective control measure: A barrier that elimi-nates or substantially reduces the likelihood of a hazardous event occurring.
Healthcare Failure Mode and Effect Analysis:
(1) A prospective assessment that identifies and improves steps in a process, thereby reasonably ensuring a safe and clinically desirable outcome. (2) A systematic approach to identify and prevent product and process problems before they occur.
Hazard analysis: The process of collecting and evaluating information on the hazards associated with the selected process. The purpose of hazard analysis is to develop a list of hazards that are of such significance that they are reasonably likely to cause injury or illness if not effectively controlled.
Failure mode: Different ways in which a process or subprocess can fail to provide the anticipated result.
Probability: See the probability rating scale in Fig-ure 4.
