June 17, 2008
Technology Workarounds Defeat Safety Intent
During a Patient Safety Walk Rounds session, we witnessed an interesting but troublesome workaround. A nurse was using a new barcoding system to perform medication administration. The bar code on one unit-dose medication would not scan properly because of a crinkle in the barcode. The nurse was to then manually input the barcode from the package onto the system computer. However, the print on the barcode was too small for her to read. She then prepared to cut and paste the medication number from the computer screen into the manual entry input box. That, of course, would have bypassed the whole patient safety concept of a barcoding system, which is to verify that the medication being given is the same as the one on the computer screen. Simply typing in those same numbers seen on the computer screen would have also bypassed the safety mechanism involved in barcoding. While we intercepted this instance to prevent a potential error, there are undoubtedly many similar workarounds being used with barcoding systems.
Workarounds, or similar intentional violations of a rule or procedure, are very common in healthcare (see our Septemeber 4, 2007 Patient Safety Tip of the Week “Workarounds as a Safety Issue”). They are usually performed with the best of intentions and are usually an indication of a flawed or poorly designed underlying process or system. In the case at hand, the nurse would have had to go search for a magnifying glass to read the barcode or contact the pharmacy to send up another unit dose. Attaching a magnifying glass (and it would have to be lighted to help during night medication administration) to the barcode scanner would also be a workaround solution. Designing the software to prevent “cut and paste” would be smart but that wouldn’t prevent simply reading the drug number from the computer screen and entering it into the manual entry input box. The only real solution for the root cause here would be to ensure the printed numbers on the barcode label are big enough for someone to read in the event that the barcode becomes unscannable. So, again, this is a system design flaw that needs to be fixed.
We wondered how often this type of barcoding system workaround occurred and whether there were other similar workarounds we needed to watch out for. Fortunately, Ross Koppel (Koppel et al 2008) and his colleagues had just published an article identifying the multiple types of workarounds in barcode systems and their underlying causes.They identified 15 types of workarounds and 31 types of causes for the workarounds in barcoding medication administration systems.
They further classified the types of workarounds into three broader categories. Under “omission of process steps” they included things like failure to scan patient identification barcode, failure to scan the medication barcode, failure to review any alerts that popped up, failure to perform visual checks, failure to perform double checks for high-risk medications, etc. Under “steps performed out of sequence” the included things like documenting the medication as administered before they actually administered it or observed the patient take it. The third category “unauthorized process steps” included things like disabling the audio alarms or removing the scanner from the computer cart (which makes it impossible to hear the alert on the computer). It would also include things like carrying medications for more than one patient at the same time, scanning the same medication package several times when the patient is to receive multiple packages, and scanning patient identification barcodes from places other than on the patient. The latter phenomenon has popped up at many different hospitals and reproductions of the patient ID bracelets are made and attached to things like an RN’s clipboard, nurses pockets or belts, the computer-on-wheels, room doorjamb, etc. and then scanning those reproductions rather than the patients’ actual ID bracelets.
They group the 31 types of causes for workarounds into groups: (1) technology-related causes (2) task-related causes (3) organizational causes (4) patient-related causes and (5) environmental. Their table of the causes is comprehensive and they provide excellent examples of each type of cause. The technology-related causes include not only things like equipment failure but also both negative and postivie attitudes about the barcoding technology (so some distrust it and others put too much reliance on it). Task-related causes include time pressures, emergency cases, etc. Organizational causes include things like poor training, medications without barcodes, barcodes obscured by other stickers, and others. Patient-related causes include circumstances such as patient isolation and environment-related causes include things like loud noises obscuring the auditory alarms. We strongly encourage you to read the full article to see all the observed workarounds and the identified potential causes and their recommendations.
Equally important is the methodology Koppel et al used in their study. They used multiple methods to investigate: (1) direct observation (shadowing) (2) interviews with care providers (3) failure mode and effects analysis (FMEA) (4) attending meetings with staff on the barcoding administration system and (5) review of the barcode system override data log. While we have always been huge advocates for looking for workarounds when we do patient safety walk rounds, we also advocate using multiple methods to identify problems. For example, if we are writing a new “rule” for an alert or reminder on a CPOE system, we always employ multiple measurements of the use and impact of that alert/reminder. Those measurements should include both objective parameters (eg. number of times triggered, number of overrides, etc.) and subjective parameters (asking physicians what they like or don’t like about that particular alert). In addition, we try to measure the impact that alert had on the bigger picture (eg. did it reduce the occurrence of ADE’s or nosocomial infections).
The Koppel article is a must-read for anyone using or contemplating a barcoded medication administration system. It also has an excellent bibliography.
Another excellent paper (Cochran et al 2007) reviews both errors prevented by and those caused by barcoded medication administration systems, based on error reports submitted to the MEDMARX® database. About 10% of the errors caused by barcoding systems were related to workarounds. The two most common workarounds were scanning the medications at the nursing station instead of bedside or scanning the patient identification from something other than the patient’s wrist. Of course, there are other cases where the wrong ID bracelet may be placed on a patient such as the near-miss reported by McDonald (McDonald 2006) in which 2 patients were admitted around the same time had their wrist ID bracelets transposed and one of them almost received what could have been a fatal dose of insulin as a result.
There are several good articles on workarounds in general. The December 2005 Advisory from the Pennsylvania Patient Safety Authority has an excellent and practical. That article includes a list of over 70 “at-risk behaviors” compiled by ISMP relating to medication safety. Many of the at-risk behaviors involve workarounds and apply to multiple activities besides medication safety. The list is a good place to begin your look at potential error-producing behaviors that may be amenable to change. For those interested in some of the more technical and esorteric aspects of workarounds, see the series of articles by Halbesleben and colleagues (Halbesleben et al 2008a, Halbesleben et al 2008b, Vogelsmeier et al 2008).
Workarounds are virtually always a symptom of a flawed system. Hunt them down and you will identify numerous opportunities to improve patient care and avoid untoward outcomes.
Koppel R, Tosha Wetterneck T, Telles JL, Karsh B-T. Workarounds To Barcode Medication Administration Systems: Their Occurrences, Causes, And Threats To Patient Safety. JAMIA 2008; PrePrint: Accepted Article. Published April 24, 2008 as doi:10.1197/jamia.M2616 http://www.jamia.org/cgi/reprint/M2616v1
Cochran GL, Jones KJ, Brockman J, Skinner A, Hicks RW. Errors Prevented by and Associated with Bar-Code Medication Administration Systems. Joint Commission Journal on Quality and Patient Safety 2007; 33: 293-301
McDonald CJ. Computerization Can Create Safety Hazards: A Bar-Coding Near Miss. Ann Intern Med. 2006;144:510-516 http://www.annals.org/cgi/reprint/144/7/510.pdf
Pennsylvania Patient Safety Authority. Workarounds: A Sign of Opportunity Knocking. PA-PSRS Patient Safety Advisory—Vol. 2, No. 4 (Dec. 2005) http://www.psa.state.pa.us/psa/lib/psa/advisories/v2n4december2005/vol_2-4-dec-05-article_g-workarounds.pdf
Halbesleben RB, Rathert C. The role of continuous quality improvement and psychological safety in predicting work-arounds. Health Care Management Review 2008; 33: 134-144
Halbesleben JR, Wakefield DS, Wakefield BJ. Work-arounds in health care settings: Literature review and research agenda. [Review] [54 refs] Health Care Management Review 2008; 33: 2-12
Vogelsmeier AA, Halbesleben JR, Scott-Cawiezell JR. Technology implementation and workarounds in the nursing home. Journal of the American Medical Informatics Association 2008; 15: 114-9