FMEA Can Help Prevent Errors

Most error-reduction effortsbegin in response to a seriouserror. Individuals and practicesites have become accustomed to reactingto an error after it has occurred.Unfortunately, only then do they considerhow to prevent the same error frombeing repeated.Would it not be useful tobe able to predict the types of errors thatcould occur and proactively institute preventivemeasures?

All too often, error potential is notincluded in decisions about which medications,devices, or technology to purchase.Instead, decisions are guided bycost, third-party formulary restrictions,contractual agreements with purchasinggroups or vendors, and pharmaceuticalmarketing efforts. Input and evaluationfrom those who will be using the productsmay not be sought, and error potentialmay not be considered ahead of time.The ultimate result is unforeseen safetyissues and errors once the product is inthe hands of clinical users or health careconsumers.

Failure Mode and Effects Analysis(FMEA) systematically identifies areas ofpotential failure and gauges what theeffects would be before an error actuallytakes place. This proactive process canbe employed to examine the use of newmedications and products, as well as thedesign of new services and processesthat may affect work flow. FMEA is bestemployed prior to purchase and implementation,so that preemptive action canbe taken.

To cite just one example, FMEA couldbe used to assess the potential for errorwith new medications when they are firstmarketed or before they are prescribedor added to your inventory:

Step 1: A process-flow diagram wouldbe designed. Then the ways the intendedproduct would be prescribed would beexplored. Who would prescribe it and forwhich patients? What clinical patientinformation would be important beforethe product is prescribed? How would itbe procured, stored, and used, fromacquisition through dispensing andadministration? Who would prepare anddispense it? What information would needto be given to the patient or caregiver?How would the product be administered?

Step 2: Potential failure modes (ie, howand where systems and processes mayfail and what can go wrong) would beidentified while considering how the productwould be used. Questions to be askedwould include the following: Does thedrug name look or sound like anotherdrug name? Would a similarly spelled drugname be listed in close proximity to theintended product on computer orderentry screens? Does the package labelclearly express the strength or concentration?Would the product be stored near, orcould it be mistaken for, another similarlypackaged product? Are dosing parameterscomplex?

For example, an FMEA performed onquetiapine (Seroquel) might have predicteda high likelihood of mix-ups with nefazodone(Serzone). These 2 medicationsare likely to be prescribed by the sametype of physician and for patients withsimilar diagnoses; are available in overlappingdosage strengths; often areadministered at the same frequency; andare likely to be stored together and toappear on computerized lists in closeproximity, alphabetically by brand name.Although Serzone is no longer marketed,these types of errors are still likely,because generic formulations of nefazodoneare available and may be prescribedas "Serzone."

Step 3: For each failure mode, staffmembers would then determine the likelihoodof making an error, as well as thepotential consequences. What wouldhappen to the patient if the drug weregiven at the wrong dose, at the wrongtime, or by the wrong route? What wouldhappen if a patient received the wrongmedication or if the wrong patient receivedthe medication?

Step 4: Staff members would considerthe severity of the outcome and identifyany preexisting processes in place thatcould help eliminate or detect the errorbefore it reached the patient. Eachprocess would then be evaluated for itseffectiveness, based on what was learnedin previous steps. For example, wouldobtaining additional patient information orusing computer alerts, bar coding, or adouble-check process catch these errorsevery time? Numerical values could beassigned to determine the likelihood ofthe occurrence, its severity, and thechance that it would be detected beforecausing patient harm.

Step 5: If failure modes reveal errorswith significant consequences, actionswould be taken to prevent the error,detect it before it reached the patient, orminimize its consequences. Suchactions might include improved communicationof orders by listing the indicationon prescriptions or differentiatinglook-alike products by ordering themfrom different manufacturers or by storingthem separately.

Although industries outside of medicinehave developed elaborate FMEA scoringsystems to rank items for action, a simplifiedFMEA process as described abovecan be an efficient, proactive risk managementtool.

Dr. Kelly is the editor of ISMPMedication Safety Alert!Community/Ambulatory CareEdition.