Risk management methodology is considered a specific problem-solving methodology for quality improvement projects that focus on safety or prevention. Documenting and analyzing potential risks proactively are essential for improved patient safety.1 Accomplishing this goal requires an effective method to identify risks. In contrast to root cause analysis, which is carried out after an adverse event occurs, health care failure modes and effect analysis (HFMEA) is used prospectively to identify possible system failures and to fix these problems to make the system more robust before an adverse event actually occurs.2
Health care failure mode and effect analysis is a systematic method of identifying and preventing process problems before they occur. The HFMEA includes 6 steps: (1) team selection, (2) process identification, (3) process flow diagram preparation, (4) failure mode identification and scoring based on risk priority numbers (RPNs), (5) identification of root causes, and (6) determination of an action plan. A failure mode is an area where the process can break down and cause poor outcomes.3
Point-of-care testing (POCT) is defined as performance of diagnostic testing occurring at or near the patient’s bedside. Near-patient tests provide results that enable clinicians to rapidly diagnose, evaluate, or monitor patients. Time savings is related to savings in specimen delivery time, specimen requisition, processing, and communicating results.4
Laboratorians possess a wealth of scientific and technical knowledge that is used daily within the laboratory. In contrast in the case of POCT, specimens are procured, analyzed, and resulted at or near the patient’s bedside, which puts the POCT operator in the position of being responsible for all stages in the specimen workflow path.5
In this article, the author described the use of HFMEA to identify sources and amplifiers of POCT error, to categorize this sort of medical error, to suggest strategies to prevent such errors, and to describe monitors that assess and reduce the frequency of errors.
Saudi German Hospital-Aseer is a tertiary care hospital with 300 beds. There are 25 glucose meters distributed throughout the hospital. There is an electronic connectivity system to connect all glucose meters to a main server and workstation, which is located in the clinical laboratory and operated by a POCT coordinator.
To conduct HFMEA on the POCT system, we used the 6-step model developed by the Joint Commission.3
A multidisciplinary team was assembled, consisting of a diabetology consultant, a clinical laboratory director, a POCT coordinator, a manager of quality and patient safety and 2 nurses from intensive care units, 2 nurses from the emergency department, one diabetic specialist nurse, as well as 4 ward staff nurses. The team met once a week for 3 months to perform the HFMEA, develop a flow chart of the POCT process, analyze the process that was currently in place, and address all problems that could occur in the process.
Process Flow Diagram Preparation
The team mapped all the steps involved in the POCT process; for each of the steps, subprocesses also were identified. A detailed flow chart was created (Fig. 1). The team divided the POCT process into 3 major phases: pre–analytic, analytic, and post–analytic phases.
The pre–analytic phase includes (a) test ordering, (b) patient identification, (c) specimen collection, and (d) specimen evaluation.
The analytic phase includes (a) turning on the meter; (b) entering the operator identification (ID); (c) quality control (QC) performance, documentation, and review; (d) entering the patient identification number; (e) finger stick and applying blood to the test strip; (f) insertion of the test strip in the test strip holder; (g) result generation; as well as (h) result validation.
The final post–analytic phase includes (a) report formatting, (b) critical value reporting, as well as (c) report recording and retrieval.
Failure Mode Identification
Once all the steps in the process were identified, the team identified failure modes, or what could go wrong and how, using brainstorming and nominal group technique.6 Each of these failure modes was listed. Numeric values and rating scales were created and assigned to each category, severity of failure, and likelihood of failure (occurrence rate). These scales were developed using the nominal group technique to gain group consensus and, on the basis of HFMEA, scoring or ranking methodology (Fig. 2). For severity of failure, a rating from 1 to 4 was assigned, indicating a situation categorized from minor event (1) to a catastrophic event (4). Similarly, the occurrence rate was scaled from 1 to 4 to the estimated frequency of failure ranging from unlikely probability of occurrence or rare (1) to high probability of occurrence or frequent (4). An RPN was then calculated as the product of severity and occurrence scores for each failure mode identified.7
The failure mode with the highest RPN was selected for action on the basis of the Pareto principle (80% of consequences stem from 20% of the causes).8
Table 1 shows the different failure modes and their RPN.
Failure Mode Identification
The first opportunity for error at test ordering step is excessive ordering, which can lead test interpreters to be confused by the plethora of laboratory results. The second opportunity for error at the order step is mistimed testing in which test results lose temporal connection with therapeutic interventions. Patient identification errors in POCT are different from those in laboratory-based testing in that patient identification may be impaired in patients in whom identification bands were removed during surgery or not matched before leaving the operating room.
Specimen collection is another step with error potential (eg, inadequate amounts of blood applied to POCT glucose test strips). The fourth and last step of the pre–analytic phase is the assessment of specimen attributes other than those which POCT attempts to measure. These attributes may cause errors in test results (eg, anemia or hematocrit that may affect the POCT glucose results of the patient).
Checking of battery life on the status screen upon turning on the meter is the first failure mode in this phase. Entering the operator and patient identification number is an important step with error potential and one of the key performance indicators in the POCT procedure, namely, the percentage of tests done by non–certified operators and the percentage of tests done with invalid patient IDs. Running QC should be done daily as per hospital policy; otherwise, the device will be locked out automatically. In the result generation step, errors can occur when results are produced outside the POCT method measurement range. Such results may not be reliable.
The fourth error relevant to the analytical step is result validation. Lack of any QC, operator failure to recognize out-of-control QC, and the absence of a lockout device monitor can all lead to acceptance of invalid results. The basic role of QC is to assure that the testing device is reporting results that correspond to correct or expected values.9 Methods of POCT require calibration and calibration verification, which is performed by the POCT coordinator. Calibration is the process of setting the appropriate relationship between an instrument’s response to the amount of an analyte present in a particular sample. Calibration of POCT glucose meters is required every 6 months.9Calibration verification is the process of verifying that the amount of analyte measured by the instrument is accurate. Not calibrating, deviating from calibration protocol, or misreporting calibration data are the types of errors that may affect test results.
In the final phase of the POCT testing process, defects can occur at the step of report formatting. The result may lack units of measure or may use inappropriate ones. Similarly, reference intervals may be missing or incorrect. All of these defects may lead to misinterpretation of results, particularly in critical care situations. The second step with error potential is critical value reporting. First, the criticality of the POCT result may not initially be recognized. Second, the critical value may not be verified by the testing operator, or third, this critical result may not be reported to the health care provider to take the appropriate actions. Finally, the critical value may not be documented for subsequent retrieval. The critical value step was scored very high by our team. The RPN was calculated as 16, which is the highest score that can be assigned. Therefore, this step was considered the highest error-provoking step in the POCT testing process.
The error potential in the last steps of the post–analytic phase includes result recording and retrieval including failure to correctly generate result records. This error makes the operator unaware of discrepancies between different results and prevents comparison of results or delta check. Similarly, issues occur with results on which the clinician has acted but which have subsequently disappeared from the record.
Corrective Action Plan
The highest RPN (16) was assigned to critical value reporting in post–analytic phase of the POCT testing process. Corrective actions were discussed and implemented, which took the form of the following:
- 1.) Adding a critical value reporting policy and procedure to the policy and procedures for POCT, which was located in every POCT testing site.
- 2.) Specifying steps for critical value reporting, namely, (a) repetition of the test with a new finger stick, (b) notification of the responsible health care provider within 30 minutes, and (c) documentation of notification in the nursing critical value notification records. All of these steps were included in the training procedure including initial training and ongoing recertification training.
- 3.) Adding the previous steps to a checklist used by supervisors in structured observation of POCT operators.
- 4.) Direct observation of operators or nurses performing POCT glucose testing, along with testing them on scenarios for critical test results and the actions required in response to those results, which was added to competency assessment to certify or recertify the operators.
- 5.) Lockout device monitor: We asked POCT glucose devices manufacturers to apply a lockout option for all meters when any critical value result appears and was not acknowledged by the operator.
- 6.) Implementing a monthly report from the POCT coordinator for the laboratory director including the number of critical values and the percentage of reporting and compliance with the policy.
Three months after implementation of the recommended changes, the risk was reassessed as well as re-estimated and the RPN was recalculated.
There was an overall reduction in the risk for failure by 50% as the frequency of occurrences decreased by 50%. The recalculated RPN for critical value reporting decreased to “8.”
We conducted an HFMEA for the process of point-of-care glucose testing by glucose meters in our hospital setting.
This included the analysis of the testing process. Such an analysis is best accomplished by a broad multidisciplinary team to identify sources of risk. Documenting and analyzing potential risks proactively are essential for improved patient safety.10 We applied numerical values to the risks we identified. The numerical value attached to each risk is naturally subjective. If, however, group consensus is used and the relative values to given risks are adequately evaluated, the resulting numerical values should be valid.
We selected a quality tool known as the nominal group technique to collect information and process the data within the group.6
This tool is a structured method of brainstorming that encourages equal participation from everyone and provides a process for constructive dialogue. Before the first meeting, an educational packet was distributed to all team members to introduce members to the tool and provide a framework for how the meetings were going to be conducted. The packet outlined rules for participation in the group. During the first meeting, further education was provided and the HFMEA began once all members understood the process and agreed to the rules. Nominal group technique was extremely helpful to the team, leading to more time of open discussion and clarification rather than conflicts or tension among group members.
We identified sources and amplifiers of POCT error, categorized this sort of medical error, as well as suggested strategies to prevent and indices to monitor such errors. A preventable adverse event is defined as an adverse event due to medical error. Medical error in clinical laboratory testing including POCT results mostly from a failure in the testing process.11
There are 3 sources of POCT error, namely, test operator incompetence, operator nonadherence to test procedures, and the use of uncontrolled reagents.
Rapid availability of POCT results and the immediate therapeutic implications of POCT amplify the likelihood that erroneous results from these tests will cause preventable adverse events.
Critical value reporting has become a focus of medical error prevention. We calculated RPN for this failure mode as 16. After implementation of changes and recommendations to decrease failure risk and observation during a 3-month period, the factors were reassessed and the hazard analysis was repeated, resulting in a new RPN of 8. These results indicate that the implementation of the action plan recommended by the team resulted in a significant decrease in risk associated with critical value reporting in POCT.
In summary, HFMEA can be applied to the POCT process to identify and address possible weaknesses within the POCT system. Implementation of changes and then reassessment of risk can provide numerical changes of value and can be used for quality assurance processes.
1. Reeson J. Managing the Risks of Organizational Accidents. Cornwall, UK: MPG Books Ltd; 1997 .
2. Welborn C. Using FMEA to assess outsourcing risk. Qual Prog. 2007; 40: 17–21.
3. Failure Modes and Effects Analysis in Health Care: Proactive Risk Reduction. Chicago, IL: Joint Commission Resources; 2005 .
4. Price CP, Hicks JM. Point of care testing: an overview. In: Price CP Hicks JM, eds. Point of Care Testing. Washington, DC: AACC Press; 1999; 3–4.
5. Kost GJ. Prevention problems, medical errors, and biohazards in point-of-care POC. 2003; 2: 78–88.
6. Tague NR . The Quality Toolbox. 2nd ed. Milwaukee, WI: ASQ Quality Press; 2005 .
7. EP 18-3p. Risk Management Techniques to Identify and Control Laboratory Error Sources. Wayne, PA: Clinical Laboratory Standards Institute; 2009 .
8. Krouwer JS . Managing Risk in Hospitals: Using Integrated Fault Trees and Failure Mode Effects and Criticality Analysis. Washington, DC: AACC Press; 2004 .
9. Westgard JO, Barry PL . Cost-Effective Quality Control: Managing the Quality And Productivity Of Analytical Process. Washington, DC: AACC Press; 1986 .
10. Sharp GH, Gibson DJ, Tindle BH. Physician office testing: the importance of ongoing monitoring. POC. 2003; 2: 179–182.
11. De Rosier J, Stalhandske E, Bagian JP, et al. Using health care failure mode and effect analysis: the VA National Center for Patient Safety’s prospective risk analysis system. Jt Comm J Qual Improv. 2002; 28: 248–267.