July 21, 2009           

Medication Errors in Long-Term Care

Medication errors in long-term care (LTC) or skilled nursing facilities (SNF) are just as frequent as in acute hospitals. And the elderly patient population cared for at most long-term care facilities are particularly vulnerable to the adverse effects of medication errors. A variety of factors contribute to this vulnerability. First and foremost, they tend to be on larger numbers of medication, increasing the likelihood of drug-drug interactions and increasing the likelihood of additive effects of side effects. Both renal and hepatic function tend to decline in the aging population, interfering with clearance of many drugs. Impaired ability to chew or swallow may necessitate changes in the usual manner of medication administration. The elderly, particularly those with pre-existing dementia, tend to have a reduced “cognitive reserve” that may increase their likelihood of developing delirium when subject to certain drugs and other agents. And the cognitive impairment plus impaired vision or hearing may also prevent the patient himself from acting as a last-line defense against a medication error. The patient in SNF/LTC also receives much less direct attention from physicians and other medical professionals than do patients in acute care facilities. The provider ordering medications is often not even present at the SNF/LTC at the time the medication is ordered (i.e. the order is often entered remotely or called in by telephone). And often the nurse dispensing medications has multiple other duties and responsibilities that may lead to distractions and interruptions in the medication dispensing/administration stages. Undoubtedly, if one does thorough root cause analyses we would see workload, staffing, continuity and staff turnover issues as important root causes.

Gurwitz et al 2000 studied 18 community-based nursing homes and found a rate of adverse drug events of 1.89 per 100 resident months (plus another 0.65 per 100 resident-months of potentially adverse drug events). Many of these were potentially life-threatening or serious. And, most importantly, over 50% were considered to be potentially preventable. In fact, 72% of the more serious adverse events were potentially preventable. Gurwitz et al 2005 looked at adverse drug events in 2 academic LTC facilities and found an even higher rate of ADE’s (9.8 per 100 resident-months). Again, 42% of these were potentially preventable. The authors attributed the much higher rate in the second study to better and multiple methods for detecting adverse drug events. In both studies, most of the errors tended to occur in the stages of ordering and monitoring. Several drug categories were overrepresented: antipscyhotic medications, anticoagulants, diuretics, and antiepileptic medications.

But acute care hospitals may play a role in the medication errors that occur in LTC facilities. We’ve written a lot recently about the problem of Medicare readmissions. Readmissions from long-term care facilities are especially problematic for some acute care facilities. And many of these are related to medication errors and adverse medication events. Sometimes the acute care facility actually contributes to such errors. A recent study (Tjia et al 2009) looked at medication discrepancies in patients transitioning from acute hospitals to long-term care facilities (skilled nursing facilities or SNF’s). At least one medication discrepancy occurred in a whopping 71.4% of admissions to the SNF and in 21.3% of all medications. The discharge summary and the patient care referral form did not match in 52.3% of cases. In addition, there is often a delay in medication administration for those SNF patients transferred from acute care facilities, especially those transferred in the evening. One pilot study (Ward et al 2008) showed the mean delay from arrival in dose of an ordered medication was 12.55 hours and the mean number of omitted medications was 3.4.

On top of that are those cases where medication reconciliation on discharge from the acute facility is faulty and patients are inadvertently continued on medications (eg. proton pump inhibitors) that had been intended as prophylaxis only for a limited time while that patient had been acutely ill.

Very few SNF/LTC facilities have CPOE (computerized physician order entry), one useful mechanism to reduce medication errors. However, most do have pharmacy order systems that have the capability to set up alerts based on “rules”. So, for the example given above, a rule could be programmed that says “if a drug in the proton pump inhibitor class is prescribed and was not on the prior medication list, then check with the attending physician to confirm its necessity”.

All the other types of medication errors we see in acute care also occur in chronic care. We’ve emphasized the problem of workarounds with barcoding bedside medication verification systems (see our June 17, 2008 Patient Safety Tip of the Week “Technology Workarounds Defeat Safety Intent”). One study (Patterson et al 2006) showed that such workarounds occur significantly more frequently in LTC/SNF than in the acute care hospital. And, just as we have learned with CPOE in acute care and ambulatory care, physicians ignore the vast majority of alerts and reminders triggered during order entry. A study in a LTC facility (Judge et al 2006) showed that prescribers who received an alert were only about 10% more likely to take an appropriate action, though responses were higher for anticoagulant or CNS side effect warnings. Errors due to look-alike/sound-alike (LASA) drug pairs also obviously occur in the SNF/LTC setting (Walliser et al 2007). Errors in repackaging of medications (Gerber et al 2008) are also common. That study, done in LTC facilities in Germany where repackaged medications for each patient are put in pill organizers containing the entire day’s medications, showed errors in 7.3% of all pill organizers and affecting over half of LTC residents. Incorrect halving of the medications was the most frequent error encountered.

Medication errors related to administration of medications may occur despite good automated dispensing processes. One study (van den Bemt et al 2009), using a disguised observation technique, found errors in 21.2% of all observed medication administrations in 3 SNF’s. Administration technique errors such as those related to medication crushing were the most common errors identified. But lack of supervision of intake by the patient was also frequent. Wrong time of medication administration was particularly relevant for those medications supposed to be taken in relationship to meals.

ISMP Canada’s Medication Safety Self-Assessment (MSSA) program administers the assessment across all levels of the healthcare continuum. The tool is administered via secure internet access and for LTC facilities consists of 125 items by which LTC facilities assess their relative performance in medication safety. Though still in its infancy, the program has identified both strengths and weaknesses in medication safety systems in LTC facilities. On the positive side, LTC facilities tend to use computerized pharmacy information systems with resident medication profiles, have allergies listed on each page of the MAR, dispense by unit dose, use patient photographs to assist in correct patient identification, have standardized administration times, and have pharmacists available to work with care teams. But areas they have identified as being in need of improvement are: dealing with high-alert medications, minimizing interruptions during the medication administration processes, lack of CPOE/clinical decision support systems, use of dangerous abbreviations, and relative lack of use of patient safety learning tools.

Another study from Canada just published (Field et al 2009) looked at computerized alerts for renal dosing in a LTC setting. This was a randomized controlled trial in which the prescribing physician saw the alerts in the intervention arm and did not see them in the control arm. The alerts were successful for reducing the maximum frequency of medications, avoiding medications that should be avoided, and supplying missing information (such as a current creatinine level).

The bottom line is that we have a whole lot of opportunities to improve medication safety in the LTC/SNF setting and especially to improve our coordination of care between the LTC/SNF and acute care settings. Particularly when CMS initiates its new focus on reducing readmissions to acute care, increased collaboration between the acute and chronic settings to reduce adverse drug events will be crucial.

References:

Gurwitz JH, Field TS, Avorn J et al. Incidence and preventability of adverse drug events in nursing homes. American Journal of Medicine 2000; 109: 87-94

http://www.amjmed.com/article/S0002-9343(00)00451-4/abstract

Gurwitz JH. Field TS. Rochon JJ et al. The incidence of adverse drug events in two large academic long-term care facilities. American Journal of Medicine 2005; 118(3):251-8

http://www.amjmed.com/article/S0002-9343(04)00718-1/fulltext

Tjia J, Bonner A, Briesacher BA, McGee S, Terrill E, Miller K. Medication discrepancies upon hospital to skilled nursing facility transitions. J Gen Intern Med. 2009; 24:630-635

http://www.springerlink.com/content/62wm7n6368uk4567/

Ward KT, Bates-Jensen B, Eslami MS et al. Addressing delays in medication administration for patients transferred from the hospital to the nursing home: A pilot quality improvement project. American Journal Geriatric Pharmacotherapy 2008; 6(4):205-211

http://www.sciencedirect.com/science?_ob=ArticleURL&_udi=B7CWH-4V0KV6K-3&_user=10&_coverDate=10/31/2008&_rdoc=3&_fmt=high&_orig=browse&_srch=doc-info(#toc#18107#2008#999939995#724122#FLP#display#Volume)&_cdi=18107&_sort=d&_docanchor=&_ct=6&_acct=C000050221&_version=1&_urlVersion=0&_userid=10&md5=bcaf1feaa79371477cc7cc2e45cd635c

Patterson ES. Rogers ML. Chapman RJ. Render ML. Compliance with intended use of Bar Code Medication Administration in acute and long-term care: an observational study.

Human Factors 2006;. 48(1):15-22

http://hfs.sagepub.com/cgi/content/abstract/48/1/15

Judge J, Field TS, DeFlorio M. et al. Prescribers' responses to alerts during medication ordering in the long term care setting. Journal of the American Medical Informatics Association 2006; 13(4):385-90

http://www.jamia.org/cgi/content/abstract/13/4/385

Walliser G, Grossberg R, Reed MD. Look-alike medications: a formula for possible morbidity and mortality in the long-term care facility. Journal of the American Medical Directors Association 2007; 8(8):541-2

http://www.jamda.com/article/S1525-8610(07)00343-X/abstract

Gerber A, Kohaupt I, Lauterbach KW et al. Quantification and classification of errors associated with hand-repackaging of medications in long-term care facilities in Germany.

American Journal Geriatric Pharmacotherapy 2008; 6(4):212-9

http://www.sciencedirect.com/science?_ob=ArticleURL&_udi=B7CWH-4V0KV6K-4&_user=10&_coverDate=10/31/2008&_rdoc=4&_fmt=high&_orig=browse&_srch=doc-info(#toc#18107#2008#999939995#724122#FLP#display#Volume)&_cdi=18107&_sort=d&_docanchor=&_ct=6&_acct=C000050221&_version=1&_urlVersion=0&_userid=10&md5=addf464308c9d1eed97d481345bf2893

van den Bemt PMLA, Idzinga JC, Robertz H, Kormelink DG, Pels N. Medication Administration Errors in Nursing Homes Using an Automated Medication Dispensing System. J Am Med Inform Assoc 2009; 16: 486-492.

http://www.jamia.org/cgi/content/abstract/16/4/486

ISMP Canada. Medication Reconciliation and Medication Review:

Complementary Processes for Medication Safety in Long-Term Care. ISMP Canada Safety Bulletin 2008; 8: 1-3

http://www.ismp-canada.org/download/ISMPCSB2007-09MedRec.pdf

Field TS, Rochon P, Lee M et al. Computerized Clinical Decision Support During Medication Ordering for Long-term Care Residents with Renal Insufficiency. J Am Med Inform Assoc 2009; 16: 480-485.

http://www.jamia.org/cgi/content/abstract/16/4/480

Print “Medication Errors in Long Term-Care”

July 28, 2009           

Wandering, Elopements, and Missing Patients

The May 25, 2009 issue of Inside the Joint Commission Online has several articles dealing with wandering, elopements and missing patients. The issue was apparently triggered by an unfortunate sentinel event in Pittsburgh in which an elderly woman with dementia wandered onto a roof and froze to death. All types of facilities need to develop policies and procedures for:

1. doing an assessment for risk of wandering or elopement
2. implementing risk reduction strategies for those patients at risk
3. performing a prompt and thorough search when a patient is missing

Many of the risk factors for wandering and elopement are also risk factors for delirium. So if you have been following this column and begun screening for delirium risk factors, you are halfway there! Alzheimer’s disease or any dementia may predispose the patient to wandering. Most of you recognize the term “sundowning” which we apply to those patients with dementia who become more confused and disoriented in new unfamiliar settings such as the hospital. Such patients may be prone to wandering and elopement. But any patient with impaired cognition may be at risk. This includes patients with psychiatric disorders, developmental disabilities, and acquired neurological disorders.  But there are other risk factors or contributing factors as well. Many of the drugs we’ve talked about under delirium (particularly sedating agents) may contribute. A prior history of wandering or elopement (eg. at a long-term care facility prior to admission) should be a red flag.

Some standardized questions that appear on most wandering assessment tools are:

• Is there a history of dementia?
• Is there a prior history of wandering or elopement?
• Has the patient been legally committed?
• Do they have a court-appointed legal guardian?
• Are they a danger to self or others?
• Do they lack cognitive ability to make decisions?
• Do they have physical or mental impairments that increase risk of harm?
• Do they walk round aimlessly?
• Do they often get agitated doing simple tasks?
• Do they ask the same question over and over?
• Do they frequently ask where they are?

Others have emphasize “exit-seeking behavior” such as talking about going home or asking about things not available on the unit (typically something such as candy bars).

Just as with fall risk assessments or delirium risk assessments or even DVT risk assessments, things change during a hospitalization. Therefore, a single assessment for elopement risk on admission is not sufficient. That risk assessment must be repeated after surgery, at internal transfers of care, and any time there has been a significant change in the patient’s mental status or overall medical status. The same patients should have formal risk assessments for delirium and falls.

So what do you do when you identify a patient as being at-risk for wandering or elopement? It makes sense to put them in a room where staff would be more likely to see them exit the room (usually closer to the nurse’s station). Many floors have one or two rooms that are video monitored, a logical choice for such patients. Consider having the patient wear a gown that is a different color than the usual gowns so that all staff would recognize such patient as being “lost” if encountered in other parts of the hospital. Potential exit doors on the unit should be fitted with appropriate alarms (that are functioning correctly) and with appropriate signage to keep the door closed. Consider keeping the patient in a room with a roommate or have family members stay in the room. Attention to the patient’s physical needs (food, water, warmth, pain management, toileting) are important. Letting the patient walk or exercise under supervision may be useful. The references below (AHRQ Web M&M, Veteran’s Administration Wandering Resources, Spencer 2008, Evidence-based guideline: Wandering) all contain excellent points on care management of the patient at risk for wandering.

Internal patient transports may also be vulnerable events. You’ve heard us talk on several occasions about the “Ticket to Ride” concept in which a formal checklist is completed for all transports (eg. to radiology). Such checklists typically contain information related to adequacy of any oxygen supplies and medications needed but should also include information about things like suicide risk and elopement risk. These all need to be conveyed to the caregiver who may be accepting the patient in the new area. Just as we’ve talked about cases where a patient may attempt suicide in a bathroom in the radiology suite that is not suicide-proofed, a patient at risk for elopement may wander off easily while waiting in the radiology suite if not appropriately supervised.

The response to a missing patient is critical. It must be rapid, well-planned, and thorough in order to find the patient before he/she suffers any harm. Some facilities have chosen to merge their infant abduction policy with the elopement or missing patient policy since the procedures may be very similar. But be careful – you don’t want your staff searching for a newborn by mistake when they should be looking for a wandering geriatric patient.

Staff on the unit need to be notified as soon as a patient is missing. A very brief head count of patients and look in rooms on a unit is typically done but this should last no more than a couple minutes. At that point the hospital phone operator should be notified and “code yellow” (or whatever name you use at your facility) should be announced over the public address system. It should be announced with a brief description of the missing patient (age, sex, race, unit, etc.). Exits from the building need to be immediately locked (some doors may be locked from a central location) or manned by designated staff members. No one should be allowed to leave the building(s). The operator may notify visitors over the PA system that they are under no danger but need to avoid going near exits for the time being. Key assigned staff should immediately go to a designated “command center” from which they will direct the response. Each unit (clinical and nonclinical) will have a specific predetermined area they must search in a systematic fashion. The command center must have an overlay grid of the buildings and surrounding areas and be able to mark off areas on the grid that have been searched. The search teams must have keys to their search areas since sometimes patients lock themselves into rooms inadvertently.

We also recommend early outdoor search since a patient can easily stray far from the building (or into automobile traffic) in a very short period of time. We also recommend that the local police department be notified immediately by the operator when the “code yellow” is called (don’t forget to include them in your planning process). Many facilities also use many security video cameras that are monitored centrally. Security staff may be able to scan those quickly to look for a patient exiting the building.

Someone on the unit from which the patient disappeared should be designated to send out a general email to all saff, describing the missing patient and including a photograph if one is present on the information system. The patient’s physician should be notified by the operator or staff on the patient unit. Someone needs to be designated to be in communication with the family as well.

What do you do when you find the patient? First, be aware they are likely confused and be careful not to frighten them. Do a brief assessment as to whether they may have been injured. Notify the command center you have found the patient and either return them to their unit or to the emergency department. They should be evaluated by a physician at that time to determine whether any injuries have occurred. In the unfortunate circumstance where the patient is found dead, the scene should be left undisturbed because the authorities will treat it as a crime scene.

Drills are critical for any event that is likely to be rare but critical when it occurs. Just as we’ve hammered home in our discussions about surgical fires, it is important that all staff know what to do during certain emergencies and the best way to prepare for those is with drills. Yes, you can and should include education and training on missing patient alerts during orientation and annual reorientation but you have to periodically run a drill to see whether the responses are adequate and timely. During drills one may also see various nooks and crannies and other areas (eg. ventilation ducts) that a patient could get into, perhaps leading to some physical improvements to prevent such dangerous access. You might also consider using a “secret wanderer” (person dressed as a patient in one of the special colored gowns) to see if staff identify them as a wanderer.

Staff education obviously is important but should be ongoing rather than just being delivered at annual orientation sessions. The Bay Pines (Florida) VA Hospital uses a creative reminder device on patient care areas that is shaped like a stop sign and uses the mnemonic “DON’T GET LOST” which stands for:

            D            Determine at-risk patients

            O            Observe for wandering triggers

            N            No-fall environment

            T            Teach staff/nonclinical support

            G            Get patient involved in activities

            E            Exit control

            T            Talk to patient and provide reassurance

            L            Low patient to staff ration

            O            Offer food, drink and toileting

            S            Structure and routine

            T            TEAMWORK!!

Ongoing surveillance is also important. When we do patient safety walkrounds we also incorporate much of what traditionally has been termed environment of care rounds. We look to see that doors and other accesses to dangerous places are locked and appropriately alarmed. We also look at windows leading to rooftops and make sure no one could open them and exit onto a rooftop.

How can technology be utilized to minimize the risk of elopement? There are several technology devices that can be used to alert staff when at-risk patients are leaving their bed or their area of care and others that can be used to track and locate such patients. All agree that such systems should never be relied on as the sole means of monitoring such patients. And each day the system/device should be checked to ensure it is functioning properly.

In addition, our electronic medical records (EMR’s) and clinical decision support tools can be used to help identify at-risk patients and flag them as being at-risk. For instance, if a patient has wandering during one hospitalization (or wandering during a LTC stay), his medical record can have a flag set that identifies him as at risk for wandering during future admissions.

You probably will be unable to prevent every potential elopement. When one does occur, do a debriefing session as soon as possible to identify potential missed clues and other useful lessons. Then do a formal root cause analysis within a short timeframe. There are always valuable lessons learned that hopefully can prevent other elopements in the future.

References:

Elopement. Inside the Joint Commission Online 2009; 15:4-8. May 25, 2009.

Gerardi D. AHRQ Web M&M Case. Elopement. December 2007

http://www.webmm.ahrq.gov/case.aspx?caseID=164

Veterans’Administration. Wandering Resources.

http://www.google.com/url?sa=t&source=web&ct=res&cd=1&url=http://www1.va.gov/VISN8/PatientSafetyCenter/wandering/WanderingResources.doc&ei=erBnStTLEJWCtgevh9z8Dw&usg=AFQjCNGKEGzZyWiVb-Yzum84Fy4AxjIEJg&sig2=T0w7YphxCEnls_FA5SO9xg

National Guideline Clearinghouse. Evidence-Based Guideline: Wandering.

http://www.guideline.gov/summary/summary.aspx?doc_id=12992&nbr=006688&string=wandering

Spencer E. Policy for Assessment and Care Management of Patients who are at risk of Wandering in the Acute Care Setting. University Hospitals of Leicester. August 2008

http://www.google.com/url?sa=t&source=web&ct=res&cd=1&url=http://wanderingnetwork.co.uk/Policy for assessment and care management of patients at risk of wandering in the acute setting[1].pdf&ei=-9NnSoCIOZSCtgeL1eSuCw&usg=AFQjCNHEYoKeBfuwUJG5G0gzxS2LY0exWQ&sig2=wr8W-l1rJC22y5ufSvOmew

Print “Wandering, Elopements, and Missing Patients”

August 4, 2009           

Faulty Fall Risk Assessments?

One of the “under the radar” journals we like to read is “Nursing Times” out of the UK. We like it because they are not afraid to challenge conventional wisdom. We recently noted their position on bedrails (see our August 2009 “What’s New in the Patient Safety World” column).

They also recently had a great discussion on fall risk prediction tools (Oliver & Healy 2009). In particular, they discourage use of those prediction tools that have a “score” that places a patient in “high”, “moderate”, or “low” risk categories. Many of these “scores” use predictive risk factors that you cannot modify (eg. age, gender, prior falls). Instead, they favor those risk prediction tools that focus on modifiable risk factors that can be put in a checklist-style format (yes, our old favorite: the checklist!) so you can then focus your program on actually modifying those risk factors.

They do admit that some general risk reduction interventions may benefit all patients (eg. safer footwear, minimizing use of sedating drugs, etc.) but other interventions should be tailored to the individual patient’s modifiable risk factors. What can you modify? You can improve sensory deficits like visual and hearing impairment. You can improve orthostatic hypotension and, to a degree, postural instability. Bowel/bladder care and timed toileting may help since such a high proportion of falls occur on toileting activities. Knowing that a patient fell in the past is not modifiable but knowing the circumstances of the fall may help identify a modifiable risk factor. A history of falls shortly after assumption of the upright posture should lead to a formal assessment for orthostatic hypotension (another soapbox issue for us: see our April 16, 2007 Patient Safety Tip of the Week “Falls With Injury”).

They point out that most fall risk reduction programs reduce the risk of falls only about 20% and even the best only reduce falls by 30-40%. If a tool has a negative predictive value of 90%, that still means that of every 1000 patients you screened as “low” risk, 100 will still fall! Thus, some fall risk prediction tools may lead to a false sense of security. So focusing instead on things you can actually fix makes a lot of sense.

And, of course, Oliver & Healey note that like any other tools you need to also be sure that the tool has been validated for the type of setting and patient population in which you intend to use it. We certainly concur with that. A good example was in our October 7, 2008 Patient Safety Tip of the Week “Lessons from Falls....from Rehab Medicine” where we discussed the risk factors for falls in rehab medicine settings. Using risk assessment tools that had been validated in acute med/surg hospitals or SNF’s would have resulted in misleading predictions in the rehab setting.

We do like several of the concepts in the Oliver & Healey paper:

• Focus especially on those risk factors that a modifiable
• Beware that “scored” risk prediction tools may give you a false sense of confidence

But are there other risk prediction tools that “scoring” a patient as being high risk might be useful even if you cannot modify the individual risk factor? We think there are. For example, a preoperative screening tool for obstructive sleep apnea (OSA) would likely include obesity as a predictive factor for OSA. You obviously cannot modify obesity in the very short run. Yet if the patient scored “high risk” on such a tool, you would take precautions for appropriate postoperative monitoring of such patients and try to minimize use of those drugs that may further aggravate OSA. (But, even then, we’re sure Oiver & Healey may argue such tools will miss some patients who do have OSA and might produce a false sense of security.) Other tools, like the Pneumonia Severity Index (PSI) or the CURB-65 tool, have many risk factors that are not individually modifiable but still help us identify which patients should be admitted and which should probably be admitted to intensive care.

And even some non-modifiable fall risk factors may still lead to specific interventions. For example, male gender has been indentified in some tools as a risk factor for falls. While you obviously cannot modify that risk factor, you might look extra carefully at toileting needs of the male patient. As before, we don’t know how much of the male risk for falls is “macho” vs. “modesty”. But if that latter is a factor in raising the fall risk during toileting in males, you may need to consider having non-female staff assist the males in toileting activities.

So we don’t think you should ignore non-modifiable risk factors but we think there are some good lessons learned here. While general patient safety efforts are important, focus especially on those things you can modify for the individual patient. And make sure that a “score” on a predictive tool does not sway you from focusing on those modifiable risk factors in individual patients who score “low” risk.

References:

Oliver D, Healey F. Falls risk prediction tools for hospital inpatients: do they work? Nursing Times 2009;105: 18–21

http://www.nursingtimes.net/nursing-practice-clinical-research/falls-risk-prediction-tools-for-hospital-inpatients-do-they-work/1999146.article

Print “Faulty Fall Risk Assessments?”

August 11, 2009           

The Radiology Suite...Again!

We’ve talked about the radiology suite as being a site at high risk for adverse events and medical errors (see our Patient Safety Tips of the Week for October 16, 2007 “Radiology as a Site at High-Risk for Medication Errors”, February 19, 2008 “MRI Safety” and September 16, 2008 “More on Radiology as a High Risk Area”).

Our October 16, 2007 Patient Safety Tip of the Week “Radiology as a Site at High-Risk for Medication Errors” discussed multiple types of medication errors that can occur when patients are in the radiology suite. That article was based primarily on the United States Pharmacopeia’s 2004 MEDMARX® Data Report “A Chartbook of 2000–2004 Findings from Intensive Care Units and Radiological Services”.

Importantly, we’ve noted that most of the events have nothing to do with radiology, per se, but rather that you have sick patients with complex problems being brought to an area prone to handoff fumbles. We pointed out at least 14 reasons why such errors are common in radiology areas.

Last month’s Pennsylvania Patient Safety Authority “Patient Safety Advisory” added new data to previous reports on medication errors occurring in the radiology suite. The PPSA found almost 1000 medication events in radiology areas over a 5-year period in Pennsylvania’s incident reporting system. Though contrast agents and radiopharmaceuticals were the leading drugs involved, the vast majority of involved agents had nothing to do with radiology. Rather, they were drugs used in the medical management of the patient elsewhere. And those that frequently led to patient harm were high alert drugs like insulin and heparin.

Especially problematic in the radiology suite are drug infusions that must be interrupted for the radiology procedure to be performed. These infusions are often omitted, given at the wrong rates, or there may be failure to give a loading bolus after interrupted infusions. Connecting tubing to the wrong catheter site (catheter misconnection) has also been seen.

The use of moderate sedation (the politically correct name for what we used to call conscious sedation) is not unusual in the radiology suite and is fraught with all the same dangers we see with moderate sedation elsewhere. In fact, it may be even more risky since lack of ready access to the MAR (medication administration record) may obscure the fact the patients have already received before they came to the radiology suite sedating agents, narcotics, or other medications that may be additive with the drugs being used for moderate sedation.

One type of error we had not previously encountered that was noted often in the PPSA report is that due to confusion of various Technetium formulations. There are apparently over 60 different products containing various technetium formulations used in imaging.

The PPSA advisory also notes that ready access to the patient’s full medical chart is often lacking in the radiology suite. There may be lab data that would influence the decision to use contrast which is not reviewed prior to contrast use.

Furthermore, during some procedures verbal orders are given, adding an additional opportunity for communication errors such as giving a medication but not recording it.

When we are asked about good topics for a FMEA (failure mode and effects analysis), medication administration in the radiology suite is always near the top of our list. That is one recommendation the PPSA also makes in its report. They also recommend proactively discussing plans for managing infusion therapy prior to transport to radiology.

Our old friend, the checklist, is the perfect fit for radiology procedures. Yet we see very few radiology departments using them. They should be used to verify the correct identity of the patient and the correct procedure, check critical lab results before administering contrast, managing IV medications, managing oxygen, labeling all vials and basins, etc.

The “Ticket to Ride” concept (see our Patient Safety Tips of the Week for April 8, 2008 “Oxygen as a Medication” and November 18, 2008 “Ticket to Ride: Checklist, Form, or Decision Scorecard?”) is another tool ideally suited for the radiology suite. It was originally developed for patients on oxygen therapy needing transport to radiology since studies had shown over 50% of such transports resulted in patients running out of oxygen. However, this concept addressing handoffs has been expanded to include attention to medication management, suicide risk, wandering risk, etc. in patients transported to radiology or other sites within the hospital. And, indeed, another recent Pennsylvania Patient Safety Authority “Patient Safety Advisory” discussed development of a standardized handoff communication tool for intrahospital transports. It has an excellent discussion on the development of questions you’ll want to use in your own transport communication tool.

Again, the bottom line is that you must anticipate the dangers in patients being transported to the radiology suite and prepare for them ahead of time.

References:

United States Pharmacopeia. MEDMARX® Data Report “A Chartbook of 2000–2004 Findings from Intensive Care Units and Radiological Services”. 2007

http://www.usp.org/products/medMarx/

Pennsylvania Patient Safety Authority. Patient Safety Advisory. Medication Errors Occurring in the Radiologic Services Department.

Pa Patient Saf Advis 2009; Jun;6(2): 46-50.     

http://patientsafetyauthority.org/ADVISORIES/AdvisoryLibrary/2009/Jun6(2)/Pages/46.aspx

Pennsylvania Patient Safety Authority. Patient Safety Advisory. Safe Intrahospital Transport of the non-ICU Patient Using Standardized Handoff Communication.

Pa Patient Saf Advis 2009; Mar;6(1): 16-9. 

http://patientsafetyauthority.org/ADVISORIES/AdvisoryLibrary/2009/Mar6(1)/Pages/16.aspx

Print “The Radiology Suite…Again!”

August 18, 2009           

Obstructive Sleep Apnea in the Perioperative Period

In recent weeks we’ve had a lot of searches on our site for our articles on obstructive sleep apnea (OSA) in the perioperative period. We’ve done several articles on the hazards of both oxygen therapy and opiate/sedatives in the post-op patient with OSA (see our Patient Safety Tips of the Week for April 8, 2008 “Oxygen as a Medication”, June 10, 2008 “Monitoring the Postoperative COPD Patient”, January 27, 2009 “Oxygen Therapy: Everything You Wanted to Know and More!” and May 12, 2009 “Errors With PCA Pumps”). Unfortunately, OSA is often undiagnosed and unrecognized before postoperative complications occur. Therefore, several groups have attempted to develop screening tools to help identify the OSA patient pre-operatively (see our May 6, 2008 Patient Safety Tip of the Week “Preoperative Screening for Obstructive Sleep Apnea”). Another recent paper (Chung 2009) describes the Berlin Questionnaire, the STOP questionnaire, the STOP-Bang model, the ASA checklist, and other screening tools that may be used for preoperative screening for OSA.

So what do you do when you identify a patient who is a good candidate for OSA and you need to do surgery or an invasive procedure? An excellent article appeared recently in the Journal of Clinical Anesthesia (Bolden 2009) that describes a protocol for perioperative management of OSA used at the MetroHealth Medical Center/Case Western Reserve University School of Medicine in Cleveland.

You’ve often heard us say that the best patient safety lessons begin with stories. Well, the Bolden article begins with 3 case reports, one prior to implementation of the OSA management protocol, one where the protocol was not followed, and one where the protocol was followed. These provide graphic illustrations of the hazards of OSA in the perioperative period. One was an obese patient with recently diagnosed (but not treated) OSA who suffered a fatal cardiopulmonary arrest in the postoperative period after receiving opiates and sedatives and no formal oxygen saturation monitoring. The second was another obese patient with symptoms strongly suggestive of OSA who had refused polysomnography. A procedure was performed under moderate sedation in the radiology suite and the patient then received analgesia via PCA (patient-controlled analgesia) pump on a general nursing unit without monitoring via the OSA protocol. She was found unresponsive and apneic but fortunately was resuscitated and had a good outcome. The third was a patient with OSA managed via the OSA protocol postoperatively in whom the monitoring identified apneic episodes promptly so that appropriate interventions occurred.

The article has a good discussion of OSA and many of the perioperative factors that may accentuate OSA. They note that many of the drugs used during surgery or procedures (anesthetic agents, sedatives, opiates) not only worsen collapse of airways but may also blunt hypoxic and hypercarbic responses. In addition, the supine position that if often needed postoperatively may further aggravate OSA. They go on to discuss the guidelines suggested by both the American Society of Anesthesiologists (ASA) and the American Academy of Sleep Medicine (AASM) and the actual protocol developed at their institution for management of OSA in the perioperative period.

After identifying patients with OSA or at risk for OSA, these patients are admitted to designated beds on regular nursing floors that are equipped with continuous pulse oximeters that alarm both inside and outside the rooms. They stratify the risk based not only on the invasiveness of the procedure being done but also on the severity of the OSA and the likely need for postoperative IV opioids. Their protocol(s) have arms for both documented and undocumented (but suspected) OSA and for those cases where narcotics are likely to be IV or oral. They also have protocols for ambulatory procedures and separate these by whether narcotics are likely to be needed postop. And they give special consideration to patients undergoing tonsillectomy and adenoidectomy, patients who have already had uvulopalatopharyngoplasty, and pediatric patients.

A good recommendation deals with patients who receive CPAP/BiPAP at home. These patients bring their machines with them to the hospital on the day of the schedule procedure. The hospital Clinical Engineering department evaluates the machines before use in the hospital. The protocol(s) also stress use of anesthesia techniques other than general anesthesia where possible (regional or local anesthesia). They also stress the importance of involvement of anesthesiologists in the postoperative management both inside and outside the operating room and PACU because anesthesiologists are more likely to be familiar with choice and titration of the sedative and analgesic medications that may be more optimally used in the OSA patient. Specifically, shorter-acting agents are usually preferred so that there is less drug accumulation that might lead to delayed consequences. The protocol  provides guidance for choosing the site for the procedure to be done, monitoring, the period and venue of observation, and recommendations about discharge home. The article also includes a nice algorhithm that graphically illustrates many of the above recommendations.

The Chung article noted above had a couple additional clinical pearls on the preoperative assessment. They note that difficult intubation is eight times more common in patients with OSA so that appropriate evaluation and contingencies for intubation are important. They also mention that some studies show that CPAP prior to surgery may reduce the rate of serious complications and shorten hospital length of stay.

One other study (Gali 2009) showed a two-phase process to identify patients at higher risk for perioperative respiratory desaturations and complications may be useful to stratify and manage surgical patients postoperatively. In that study a combination of an obstructive sleep apnea screening tool preoperatively (SACS) and recurrent PACU respiratory events was associated with a higher oxygen desaturation index and postoperative respiratory complications.

Perioperative management of the pediatric OSA patient is a bit more complex. An excellent review on this (Schwengel et al 2009) notes that OSA may present differently in young children (eg. more often presenting with failure to thrive or behavior disorders) than in older children and adults. Also, while many of the risk factors for OSA are the same in children and adults, enlargement of the tonsils and adenoids and craniopharyngeal abnormalities are more likely to be predisposing factors for OSA in the pediatric population. They advocate screening all children for OSA and note that “Does your child snore?” remains the most important question since a history of nightly snoring is 91% sensitive for OSA (though only 75% specific). Other questions would deal with observations of restless sleep, nocturnal diaphoresis, behavioral problems, direct observation of apnea or other respiratory abnormalities, and family history of OSA or sudden infant death syndrome (SIDS). Physical examination needs to focus on the airway, especially the nose, pharynx and tongue and look for presence of craniofacial abnormalities (which may both predispose to OSA and present difficulties for intubation). Attention to muscle tone is also important. They have a nice discussion of the cardiopulmonary complications of longer-standing OSA and note when cardiac testing (such as echocardiography) might be important preoperatively. They then have an excellent discussion on pain management and monitoring. An algorhithm is also available in this article.

The Bolden article also highlights the need for all types of providers who are doing moderate sedation anywhere within the health system to be aware of the risks in the OSA patient. So your OSA management protocol should apply not just to patients going through the OR but also those in your endoscopy suites and your radiology area (see last week’s Patient Safety Tip of the Week “The Radiology Suite…Again!”).

All your sites and facilities should have some sort of mechanism for identifying patients who may be at risk for OSA before they have their invasive procedure or moderate sedation and then have in place a protocol like the Case Western one for management of such patients during the post-procedure period.

References:

Chung F, Elsaid H. Screening for obstructive sleep apnea before surgery: why is it important?. Current Opinion in Anaesthesiology 2009; 22(3):405-411

http://journals.lww.com/co-anesthesiology/toc/2009/06000

Bolden N, Smith CE, Auckley D. Avoiding adverse outcomes in patients with obstructive sleep apnea (OSA): development and implementation of a perioperative OSA protocol. Journal of Clinical Anesthesia 2009; 21(4): 286-293

http://www.jcafulltextonline.com/article/S0952-8180(09)00107-X/abstract

Gali B, Whelan FX, Schroeder DR, Gay PC, Plevak DJ. Identification of Patients at Risk for Postoperative Respiratory Complications Using a Preoperative Obstructive Sleep Apnea Screening Tool and Postanesthesia Care Assessment. Anesthesiology 2009; 110:869-877

http://journals.lww.com/anesthesiology/Abstract/2009/04000/Identification_of_Patients_at_Risk_for.29.aspx

Schwengel DA, Sterni LM, Tunkel DE, Heitmiller ES. Perioperative Management of Children with Obstructive Sleep Apnea. Anesthesia & Analgesia 2009; 109(1):60-75

http://www.anesthesia-analgesia.org/cgi/content/abstract/109/1/60

Print “Obstructive Sleep Apnea in the Perioperative Period”

August 25, 2009           

Interruptions, Distractions, Inattention...Oops!

Several of our columns devoted to analogies between medicine and aviation have discussed the “sterile cockpit” concept (see our Patient Safety Tips of the Week for October 2, 2007 “Taking Off From the Wrong Runway” and May 26, 2009 “Learning from Tragedies. Part II”). This, of course, means that during certain procedures that require meticulous attention there should be no competing activities or conversations that might distract the pilot or healthcare worker from the task at hand.

Two recent papers have dealt with the issue of distractions and interruptions and how these have continued to cause even the best healthcare workers to make some of the deadliest errors. The August 2009 issue of the ISMP Medication Safety Alert! NurseAdvise-ERR contains an outstanding article on “inattentional blindness”. It’s actually a good human factors article in terms understandable by those who do not have a Ph.D. in psychology.

It begins with real-life examples where intelligent, hardworking, dedicated healthcare workers do things like draw up and administer a 1000-fold overdose of heparin. The article describes how our brains filter innumerable bits of information to select just a few to keep in our consciousness and how the brain also “fills in the gaps” (a form of confirmation bias). It discusses the factors that shape what you attend to. Most importantly, it tells you what error-reduction strategies won’t work for this type of problem (education, training, rules) and encourages you to shift your efforts to decreasing diversions and increasing the conspicuity of critical information. Its reference list has some useful classic articles and also contains a link to the FAA Human Factors course online.

The second article (Redding & Robinson 2009) looked at “Interruptions and Geographic Challenges to Nurses’ Cognitive Workload”. The researchers actually followed nurses around during their shift and recorded all distractions and interruptions they had plus all physical and geographic challenges that interrupted or delayed their patient care tasks. The interruptions they found fell into the following categories:

1. Employees asking questions
2. Distracting peripheral conversations
3. Supplies not readily available (causing nurse to go to another area)
4. Phone calls
5. Family questions
6. Patient call lights
7. Miscellaneous

The interruptions from coworkers were very often irrelevant from a patient care standpoint or were inquiries for patient information that could have been obtained from the medical record. The peripheral conversations were often ones in which the nurse was not even a participant. The phone calls were often from physicians (for orders or getting updates) but also included personal phone calls. Questions from family members were frequent interruptions and included not only questions about the patient but also things like directions to other locations in the hospital.

The geographic obstacles and traffic patterns may have been more unique to the facility. However, the most salient point is that the disjointed traffic flow often led to other distractions and interruptions like questions from other caregivers or families.

Redding and Robinson had several practical recommendations to minimize interruptions and distractions. One that we often use deals with interruptions by patients’ family members. This may be especially problematic when families are large or geographically dispersed or even dysfunctional. That is having the family (with the patient’s permission, of course) designate one “point person” who will serve as the communication liaison. You can then meet in a scheduled manner with that liaison who, in turn, can talk to the other family members. (Just keep in mind the potentially unintended consequence of the story changing as it passes through the chain of family members.) Another option is using scheduled meetings with the whole family. A third more creative solution is use of a secure website where family with designated access can get regular updates.

They also had useful suggestions for managing phone calls, ranging from triage systems for phone calls to use of hands-free wireless devices actually carried by nurses (be wary of the latter – see the discussion below on synchronous vs. asynchronous communication). They also had practical recommendations about keeping certain commonly used supplies right in the patient rooms and other suggestions about physical plant design.

And, perhaps most important, are their recommendations for hourly patient rounds. During these rounds they anticipate problems and ask patients about toileting needs, pain control, position changes, food/water, etc. And simply telling the patient when they would be back next often reduced the frequent use of the nurse call buttons.

They did also note other research that recommended use of visual signals such as a hat or apron with the words “Please do not interrupt” during certain activities. One of the most interesting concepts we saw in John Nance’s book “Why Hospitals Should Fly” (see our June 2, 2009 Patient Safety Tip of the Week “Why Hospitals Should Fly…John Nance Nails It!”) was his description of a nurse preparing medications with a red towel over her left shoulder to signify “do not disturb”.

Redding and Robinson documented 244 interruptions while observing 32 nurses for one hour each (7.6 interruptions per hour). Prior research showed nurses experience an average of 3.4 interruptions per hour (Wolf 2006), 6.3 per hour (Ebright 2003) so all these studies are consistent.

Travel patterns may influence the nature of the interruptions. Ebright noted that a nurse whose patients were dispersed had about the same number of interruptions as one whose patients were clustered. However, whereas half the interruptions for the “clustered” nurse were from her own patients, almost all the interruptions for the “dispersed” nurse were from people other than her patients.

Remember, many interruptions are “two-way”. That is, one healthcare worker needs to interrupt his/her routine to ask another healthcare worker a question (thus interrupting that worker as well). So some of your solutions could have a “double” return on investment.

Our old friend the checklist is a good way to help you remember items to do in a structured process. However, one problem often noted in aviation is that pilots forget where they were in their checklist after an interruption. So sometimes you need to go through the whole checklist again.

If you are interested in learning further about cognitive psychology, one of the best articles on how interruptions and distractions interfere with working memory and result in errors was in JAMIA almost 10 years ago (Parker & Coiera 2000). That article nicely describes how “working memory” differs from long-term memory. Working memory is what we use in most daily tasks (though, as you know from many of our other columns some frequently performed tasks are actually done at a subconscious level and may never even reach working memory). But working memory has significant limiations. Intrusion of another plan or task can cause one to forget a task in as little as 10 seconds. In addition, without conscious attention, accurate memories in working memory only last about 20 seconds.

They also emphasize 2 other concepts: the primacy effect and the recency effect. The primacy effect means you tend to recall best those items that have been in working memory the longest. The recency effect means you tend to recall best those things that were added most recently to working memory. Some distractions may obliterate the recency effect but not the primacy effect. Also, because of these two effects, those items that were in the “middle” of your task list are most likely lost due to interruptions or distractions. And the more “new” plans you force into working memory displaces all but the oldest plans.

Parker and Coiera talk about the promise of asynchronous means of communication (e-mail, voice mail, etc.) as a means of reducing interruptions. Allowing noncritical questions to be answered at the recipient’s convenience certainly can result in fewer interruptions. But it may also cause other inefficiencies, as we all see our email inboxes getting bloated. But they caution about some other new technologies (wireless phones) that actually make individuals too accessible and thus even more prone to interruptions. And we all know about the rash of traffic accidents occurring as drivers are text messaging while driving. In healthcare, text messaging has the double jeopardy of causing both interruptions and distractions plus using abbreviations that may be ambiguous and dangerous.

At any rate, a cultural change is required if we are going to reduce the adverse effects of interruptions and distractions. Certainly, face-to-face conversations are important in fulfilling social needs and promoting teamwork and camaraderie. They may also be critical when one needs confirmation that the message was both received and understood (i.e. hearback). But the cultural change is really adopting the culture of safety in which everyone recognizes what activities and situations are high risk and respecting the “sterile cockpit” during those activities. There is a little bit of “selfishness” in all of us that fosters the culture of interruption. We all think our current activity is the most important one. And we are all guilty. Last week I interrupted the head nurse on a unit to do “catheter rounds”. I never thought to ask or observe what task she was doing when I interrupted her. It would make a lot more sense to schedule a time each day to do those catheter rounds. Then no one would have to be interrupted from other equally important tasks.

By the way, try the exercise of recording interruptions and distractions in your own life – whether it is the patient care part of your work or the administrative part of your work or simply the rest of your daily life. You’ll be surprised at how often you are interrupted or distracted! No wonder we all make mistakes! By documenting when, how often, by whom, the activity we were doing, and under what circumstances the interruption occurred we can develop our own strategies to minimize interruptions and distractions during critical periods.

For example, if you find that many of your interruptions occur because you must take a particular route to get certain equipment or supplies, you can redesign the physical space to put the supplies in a different location (but beware of unintended consequences of your redesign!). If your interruptions tend to occur while you are doing a handoff activity, you may have to move your handoff activity to a room with a “Do Not Disturb” sign on the door. Get creative! Put an “On the Air” sign (like a radio/TV studio uses) over your door and light it up when you don’t want to be disturbed. And if it is one colleague who you legitimately need to interact with that tends to interrupt the most, consider a daily scheduled meeting so you can still deal with all their needs but avoid interruptions at the wrong time. And many of your interruptions are for information that is inadequately made available elsewhere (“Hey, who’s on call for neurology today?” could be avoided if you kept an accurate list available in places everyone knows to look).

References:

ISMP. Inattentional blindness: What captures your attention? ISMP Medication Safety Alert! NurseAdvise-ERR August 2009

http://www.ismp.org/Newsletters/nursing/default.asp

FAA. Human Factors Course On-line.

http://www.hf.faa.gov/webtraining/Intro/Intro1.htm

Redding D, Robinson S. Interruptions and Geographic Challenges to Nurses' Cognitive Workload. J Nurs Care Qual 2009; 24: 194-200

http://journals.lww.com/jncqjournal/Abstract/2009/07000/Interruptions_and_Geographic_Challenges_to_Nurses_.6.aspx

Wolf LD, Potter P, Sledge JA, Boxerman SB, Grayson D, Evanoff B.Describing Nurses' Work: Combining Quantitative and Qualitative Analysis.
Human Factors 2006; 48: 5-14

http://hfs.sagepub.com/cgi/content/abstract/48/1/5

Ebright PR, Patterson ES, Chalko BA, Render ML. Understanding the Complexity of Registered Nurse Work in Acute Care Settings. Journal of Nursing Administration 2003; 33(12):630-638

http://journals.lww.com/jonajournal/Abstract/2003/12000/Understanding_the_Complexity_of_Registered_Nurse.4.aspx

Parker J, Coiera E. Improving Clinical Communication: A View from Psychology.

J Am Med Inform Assoc 2000; 7: 453-461

http://www.jamia.org/cgi/content/abstract/7/5/453

Print “Interruptions, Distractions, Inattention…Oops!”

September 1, 2009           

The Real Root Causes of Medical Helicopter Crashes

In the past year we have done several columns on the “epidemic” of EMS helicopter crashes (see our July 8, 2008 Patient Safety Tips of the Week “Medical Helicopter Crashes” and February 3, 2009 “NTSB Medical Helicopter Crash Reports: Missing the Big Picture”, and our October 2008 What’s New in the Patient Safety World column “More Medical Helicopter Crashes”). In all in 2008 there were either 7 EMS helicopter crashes, in which 28 people died or 13 helicopter crashes with 29 deaths, depending on your source of information and exact timeframe. Our previous columns looked at multiple aspects of the crashes but have centered on the issue of appropriateness of medical helicopter runs.

Last week the Washington Post ran a series on the safety issues involved in medical helicopter crashes and did its own investigation into the 2008 crash in Maryland that took the lives of a patient, 2 EMT’s and a pilot and left a second patient seriously injured. The lead article (Gaul 2009) focuses on many of the financial issues involved. The second article (Flaherty and Johnson 2009) does a root cause analysis of the Maryland crash and the subsequent search for the downed helicopter.

They point out that this $2.5 billion industry is now dominated by private, for-profit companies and continues to grow rapidly, especially fueled by generous reimbursement from government and private health insurers (air transport costs roughly ten times what ground ambulance services cost in any region). Growth of medical helicopter services especially increased after Medicare, the nation’s largest payor of healthcare services, changed its methodology of payment in the 1990’s and then added a 50% premium for rural flights in 2002.

The article further describes the administrators of some medical helicopter companies pressuring pilots to accept all flight (even in questionable weather conditions) and “fishing for the golden trout” (meaning finding all the air transports possible). And it describes how some of the companies cozy up to those most likely to refer patients for their services (the emergency medical community and the hospitals). But they also point out the financial attractiveness of helicopter transports to the receiving hospitals.

A primary root cause of the problem is money. Simply put, the medical helicopter industry today is a very profitable one. Competition for patients is heavy. One of the articles in the Washington Post series tells of an instance where two medical helicopters went to pick up and transport a trauma patient (Flaherty 2009b) from a small rural hospital to a large urban trauma center. One helicopter had been called in via the traditional EMS system. The other had been summoned via a direct phone call to the helicopter pilot by a nurse who was employed by the helicopter company but who happened to be working at the hospital that day.

Competition even entered indirectly into the Maryland crash, where direct financial reward does not come from each transport (the Maryland system is publicly funded and run by the Maryland State Police). In that crash, the pilot was initially reluctant to fly into a heavily fogged-in area but heard that one of the commercial medical helicopters in nearby Washington, D.C. had flown into a nearby area. He basically said “If they can do it, I can do it.” and decided to fly that fateful night.

Another root cause is “helicopter shopping” where EMS responders or hospitals call multiple helicopter services even after one has refused to fly because of bad weather.

Yet another root cause is an unintended consequence of a law passed many years ago to encourage competition among major airlines. The Airline Deregulation Act of 1978 prohibits the states from interfering with airline prices or routes or service. That, of course, was intended to apply to passenger (and perhaps freight) air travel and not really intended to cover medical helicopter services. Nevertheless, the wording of that Act has been used in the court system to fight any attempts by state or regional emergency planning organizations to oversee comprehensive management of the medical helicopter programs. So it is only the FAA that has had regulatory oversight of the medical helicopter industry. Just as in the case of the general airline industry, the FAA has been roundly criticized for many years for its cozy relationship with industry and its tendency to hope issues are corrected voluntarily rather than by mandate. Innumerable recommendations for safety improvement made by the NTSB (National Transportation Safety Board) have never been mandated by the FAA nor implemented by the general airline industry or the medical helicopter industry. The FAA often responds that the industry felt that the recommended changes were too expensive. A few years ago the NTSB had made recommendations about adding terrain warning alarms, night vision goggles, flight risk assessments, and others but the FAA neither mandated these nor monitored how many companies voluntarily adopted them. The FAA, following last year’s rash of medical helicopter crashes, has begun the process of developing the new safety rules (eg. terrain warning systems, formal flight risk assessments, and others) but we haven’t yet seen those actually implemented.

The root cause that we have been so concerned about is the use of medical helicopters for transport of patients who are more appropriately transported via ground ambulance. There are many anecdotal reports of patients with relatively minor injuries being transported by helicopter, only to be discharged from the hospital emergency room without requiring admission. Most significantly, there is no database nationally or at the state or regional level that addresses the appropriateness of the helicopter transport.

We previously noted a 2006 study done by Dr. Bryan Bledsoe and his colleagues that was a meta-analysis of helicopter transport of trauma patients. Using several widely-used injury severity or trauma scores, they showed that almost 2/3 trauma patients brought by helicopter to a trauma center had minor or non-life-threatening injuries and that 25% were discharged from the hospital within 24 hours. Some helicopter services apparently have rates as high as 20% of transported patients being discharged from emergency rooms shortly after arrival (Greene 2009). Even in Maryland, where the trauma system is a model and the medical helicopter system a public one, the post-crash hearings revealed that almost half of patients transported by helicopter to trauma centers were released within 24 hours (Dechter 2008).

Time of transport is also important. Often, responders in the field do not appreciate the relative transport times involved in air vs. ground transport. In our February 3, 2009 Patient Safety Tip of the Week “NTSB Medical Helicopter Crash Reports: Missing the Big Picture” we noted we had that very day chanced upon a medical helicopter evacuation scene that is exactly 29 minutes from the Trauma Center. Quite frankly, there is no way that the whole process of helicopter transport could have gotten that patient to the Trauma Center faster than ground transport would have.

Much of the argument centers on the need for explicit criteria to help decisions about helicopter transport. There have long been guidelines for what patients should be transported to level I trauma centers. These are the guidelines developed by the American College of Surgeons that are based upon both physiologic condition of the patient, physical signs of injury, and other predictors of injury such as mechanism of injury and some other factors related to the crash site. However, those guidelines do not address who should be transported by helicopter and who should be transported by ground ambulance. In the Maryland crash, it apparently was the factors not directly related to the patients but rather mechanisms of injury (intrusion of the rear end of the car more than one foot) that led to the decision to use air transport. The Maryland Medical Protocols for Emergency Medical Services Providers was subsequently revised with regard to helicopter transport for “C” and “D” trauma patients. Those categories are the categories related to mechanism of injury or other factors related to the crash (eg. ejection from the vehicle, high speed of crash, death of another occupant, etc.) and those related to patient characteristics such as age, history of dialysis or anticoagulation, etc. For those two categories, the trauma decision tree calls for those patients to be transported by ground if they are within a 30-minute drive to the nearest trauma center and requires consultation with the receiving trauma center for decisions about helicopter transport. Since those revisions were made, the number of medical helicopter transports has dropped by about half (Flaherty and Johnson 2009).

Multiple factors undoubtedly contribute to the high use of medical helicopter transport. Concern for the patient or the trauma victim obviously is the number one factor. We spent years developing good trauma systems that help get appropriate patients to the level I trauma centers within the “golden hour”. And the ACS criteria have proven very useful in triaging those patients to the appropriate site. But other considerations may also be driving the use of medical helicopters. Concerns about possible EMTALA violations (the federal “dumping” statute that also calls for hospitals to assess and ensure stability and safe transport of patients) and malpractice liability also enter into the decision-making process. The time pressures and complex nature of the trauma scene also affect the decision-making process. And there are varying degrees of medical and field experience in the first responders. Even “drama” comes into play (one of the passengers on the Maryland crash actually went along to see what a medical transport is like). And, unfortunately, financial considerations and the sometimes cozy relationships between the helicopter companies and the EMT or hospital personnel may also play a role.

Are there good systems? Yes. Despite the 2008 accident, the Maryland system remains a model system and has shown resiliency and a leadership role. The public system in Maine, also profiled in the Washington Post series, is also well run and coordinated well with the state’s ground ambulance and EMS system.

It is time for the federal government to either step up and undertake appropriate safety oversight of the medical helicopter industry or, better yet, change the law to allow state or regional emergency medical systems to do the oversight. Just as in all areas of healthcare, we need to be able to monitor outcomes and to do that you need good data collection. Inclusion of criteria for air transport into trauma guidelines would be a major step. We would hope that the ACS would use the Maryland experience and add such criteria to their guidelines so that the same approach applies regardless of what state a trauma victim might be in.

And if you are a small rural hospital and just use the medical helicopters for transferring patients with acute MI or acute stroke for definitive procedures, make sure you also know the outcome data. If you wait for the helicopter but never get anyone to the tertiary care hospital within the therapeutic window for primary angioplasty, you better rethink your strategy. Again, the lack of good outcome data related to transports is detrimental.

Our advice from prior columns: Even if your organization does not own its own medical helicopter, there are things you can do to help ensure the safety of your staff and patients. First and foremost, make sure the benefit of the helicopter trip is likely to outweigh the risks. Second, make sure the company that runs the helicopter has a culture of safety. If it uses standardized dispatch protocols, has night vision imaging equipment and terrain awareness and warning systems, does flight risk assessments, is meticulous in maintenance, has (and enforces use of) helmets and shoulder harnesses, and has good training programs for its pilots and any of your staff that may fly – that’s the sort of partner you are looking for. You should be participating in simulation exercises and other crew resource management drills with them. Also, the medical helicopter transport is another great process to consider for one of your FMEA (Failure Mode and Effects Analysis) activities. Get good data from the helicopter company about transport times and any issues that arose. Make sure you get the data you need from the receiving hospital on key measures like time to angioplasty. And beware of the old adage that new safety technology may simply push the envelope – there is a tendency to take more risks when the system is perceived to be safer. So a healthy dose of skepticism and vigilance is always a good thing.

References:

Lowy J (Associated Press). Safety board probes EMS helicopter crashes. Houston Chronicle. Jan. 15, 2009

http://www.chron.com/disp/story.mpl/side/6212321.html

Emmons N. NTSB to Issue Helicopter EMS Safety Recommendations.

February 9, 2009

http://include.nurse.com/article/20090209/NATIONAL01/102230015/-1/frontpage

Gaul GM, Flaherty MP. The Deadly Cost of Swooping In to Save a Life. Washington Post 2009; August 21, 2009

http://www.washingtonpost.com/wp-dyn/content/article/2009/08/20/AR2009082004500.html

Part One

The Deadly Cost of Swooping In to Save a Life

Medical helicopters are a $250 billion industry and growing fast. A closer look at what's really at stake.

By: Gilbert M. Gaul
and Mary Pat Flaherty

http://www.washingtonpost.com/wp-srv/special/nation/medical-helicopters/?sid=ST2009082100030

Flaherty MP, Johnson J. FATAL FLIGHTS Where's Trooper 2? Vanishing in Midair.

Washington Post 2009; August 23, 2009

http://www.washingtonpost.com/wp-dyn/content/article/2009/08/22/AR2009082202372.html?sub=AR

Flaherty MP. Audio Sidebar: Two Helicopters, One Patient, Mixed Signals.
Washington Post 2009; August 21, 2009

http://www.washingtonpost.com/wp-srv/special/nation/medical-helicopters/mixed-signals.html?sid=ST2009082100030

FAA Introduces New Medical Helicopter Rules

NewsInferno.com. Date Published: Monday, April 27th, 2009

http://www.newsinferno.com/archives/5800

Bledsoe BE. Wesley AK. Eckstein M. Dunn TM. O'Keefe MF. Helicopter scene transport of trauma patients with nonlife-threatening injuries: a meta-analysis. Journal of Trauma-Injury Infection & Critical Care 2006; 60(6):1257-65 http://www.jtrauma.com/pt/re/jtrauma/abstract.00005373-200606000-00015.htm;jsessionid=LzvDYgJNbkdJpBhDDCFtr3VBPJJ6WwQ1bvdXstQHvMNQ7Lk0Mygl!447927974!181195628!8091!-1?index=1&database=ppvovft&results=1&count=10&searchid=1&nav=search

Greene J. Rising Helicopter Crash Deaths Spur Debate Over Proper Use of Air Transport.

Annals of Emergency Medicine 2009; 53: A15-A17 (March 2009)

http://download.journals.elsevierhealth.com/pdfs/journals/0196-0644/PIIS0196064409000316.pdf

Dechter G, Jones B. Md. medevac crash raises question about trauma procedures.

The Baltimore Sun. October 1, 2008

http://www.volunteerfd.org/sogs/articles/430962

Maryland Institute for Emergency Medical Services Systems. The Maryland Medical Protocols for Emergency Medical Services Providers.

Effective July 1, 2009

http://www.miemss.org/home/LinkClick.aspx?fileticket=WKeNmP+DJ9w=&tabid=106&mid=534

Flaherty MP. Protecting the Lives of Those Who Save Others. LifeFlight of Maine Favors Practicality, Safety Over Growth. Washington Post. August 22, 2009

http://www.washingtonpost.com/wp-dyn/content/article/2009/08/21/AR2009082103839.html?sid=ST2009082103952

Print “The Real Root Causes of Medical Helicopter Crashes”

September 8, 2009           

Barriers to Medication Reconciliation

Healthcare Quarterly, published by Longwoods Publishing of Canada, is a little-known resource that contains a wealth of great information. Though every issue has some articles that are of general interest, at least once a year they publish a special edition with patient safety papers. One of the papers in the most recent special edition (Coffey et al 2009) looked at the implementation of admission medication reconciliation at two Canadian academic medical centers, one a general tertiary care center and the other a pediatric tertiary care center. The approaches taken at the two centers differed but both encountered similar barriers and themes during implementation and provide some very valuable lessons learned.

Both centers began with a medication reconciliation form to be filled out by physicians on admission. This form had spaces for the medications to be entered and then reconciled by checking either “continue”, “stop”, or “change”. The paper discusses the various strategies for compilation of the best possible medication history (BPMH) and reconciling the medications, such as using pharmacists or nurses for the various roles. The availability of resources, particularly clinical pharmacists, largely determined the roles at each hospital. However, the nature of the patient population was also important in the program design. Both hospitals developed screening processes to identify patients for independent BPMH and reconciliation by a pharmacist. The general hospital had previously identified that a high rate of medication discrepancies occurred in patients on four or more medications (Cornish et al 2005) so they focused their medication reconciliation efforts on patients meeting this criterion plus patients on a high alert medication or patients with an unclear medication history. Three quarters of their patients met those criteria, whereas only one quarter of those at the pediatric hospital met those criteria.

Both initially tried to have physician leaders orient all members of medical teams to use of the Med Rec forms. However, within a few months they realized that was not going to work. There was too much turnover of medical staff and the physician leaders were not always available. In addition, though the Med Rec forms were present at both hospitals, physicians often did not know where to look for them, forgot about them, or just left them blank. Most importantly, they were often looked upon as extra work or duplicate effort (since they were already listing the medications in their History and Physical). Recognizing these barriers, the general hospital added pharmacist staff and encouraged pharmacists to initiate the Med Rec process as early as possible in the admission process (often in the emergency department). At the pediatric hospital nurses were trained to do the BPMH and reconciliation. Interestingly, when nursing began doing these processes there was a precipitous dropoff in physician compliance with the Med Rec process. They use this as an argument against making Med Rec too dependent upon a particular individual. They subsequently relaunched their educational efforts with the physician leader and a resident champion plus began feedback with physician-specific compliance rates and saw an improvement physician compliance, though rates for physician compliance with both steps is in the 40-60% range.

They summarize several lessons lessons learned:

• Physicians don’t like to document medication history separately from the History and Physical.
• Constant turnover residents and staff physicians makes the educational effort difficult.
• Workload demands on both pharmacists and nurses are substantial.
• Data collection and auditing are labor-intensive and burdensome.
• Participation in the Safer Healthcare Now! Initiative in Canada was a success factor.
• The regulatory requirement for medication reconciliation in the hospital accreditation process was a significant accelerator of the process.
• Seeing the benefit of having the BPMH at the time of discharge improved acceptance.
• Having a strong physician leader and supportive management culture on some units (eg. cardiology) led to high rates of immediate compliance.
• When there is shared accountability, you must clarify which team members are accountable for which steps.

We’ll add some of our own comments. First, we fully agree that duplication of documentation is a real problem. Physicians (and everyone, for that matter) do not want to write out all the medications in their History and Physical, the Medication Reconciliation Form, and the admission order sheet/form. So why not merge them? One hospital we have worked with uses the medication reconciliation form as the admission medication order form. It is structured so that the physician checks which of the past medications should be continued on admission and then the physician adds any additional medications below those. Note that this concept can be used when order entry is either by paper or CPOE. Then, when writing or dictating their History and Physical, the physician can simply reference that list by stating “refer to the medication reconciliation form for medication history”. So that is a time saver. Does it have a downside? Yes, we find some physicians refer to the form even when it has not, in fact, been completed.

Second, time of last dose is a very important element during medication reconciliation that is most often overlooked. Failure to record it results in either omission of doses or duplication of doses. It is unrealistic to think that you will get this filled out for every medication and it will be of little consequence for many medications. However, especially for high alert medications (such as insulin, warfarin, cardiac medications, anticonvulsants, psychiatric medications, narcotics, sedatives, etc.) it is very important to record the time of the last dose.

Third, availability of the medication reconciliation sheet to all relevant clinical staff is critical. Frequent readers of this column know we have often pointed to the unintended consequences of healthcare information technology. One of those is that a paper-based medication reconciliation form is often located only in the paper chart. Other healthcare workers that should have easy access to this often now only use the computer system to access patient information. For instance, pharmacists may know that the Med Rec form is in the chart but if they have to interrupt their busy routine to walk up to the patient floor to see it they are not likely to do so. So if you don’t make your Med Rec form part of your EMR (electronic medical record), at least send or fax a copy of it to your pharmacist with the admission orders. Note that if you made the Med Rec form the admission order form as above, you wouldn’t have this problem.

Fourth, we like the concept of focusing the medication reconciliation process to high risk patients. That, in theory, could cut down considerably on the workload imposed by the need to do medication reconciliation. However, as noted in the Coffey paper, three quarters of the patients in a typical adult general hospital would likely meet the criteria for high risk. So we wonder how much work would be reduced by such an approach. We would prefer the universal approach to medication reconciliation and we’ll likely need to make better use of information technology to accomplish this and reduce workload at the same time. However, remember our caveats in our December 30, 2008 Patient Safety Tip of the Week “Unintended Consequences: Is Medication Reconciliation Next?”. Of interest in this light is the recent paper from Partners Healthcare in Boston (see Schnipper et al 2009) which used an IT solution and redesign of its medication reconciliation process. They demonstrated an overall 28% relative risk reduction in unintentional medication discrepancies with a potential for harm (though the reduction was seen for discharge but not admission). Interestingly, they found more success for the intervention in patients deemed to be at high risk for medications discrepancies (using a risk score developed from their control population but not yet validated). It will be most interesting to see if that risk score can be validated in other populations so that it could be used as a flag for patients needing more intensive efforts at medication reconciliation.

Fifth, one of our pet peeves is the failure to record indication for a drug (or reason for discontinuation of a drug). On admission, not knowing the reason a patient is taking a drug may lead to erroneous management of that drug. For example, a patient taking digoxin for heart rate control may require different digoxin dosing and target levels than someone with CHF taking it for its inotropic effect. Or the target INR levels in patients on warfarin may differ depending on the reason for anticoagulation.

Sixth, the answer to the recurring question “who is responsible for medication reconciliation?” is “everyone is”. While it’s easy to say that the physician is ultimately responsible, it is clear that almost everyone involved in a patient’s care (the physician, nurse, pharmacist, and patient himself) have important roles in medication reconciliation. When Partners Healthcare (see Schnipper et al 2009) in Boston redesigned its medication reconciliation process, physicians were assigned responsibility for taking preadmission medication histories and referring to this list when ordering medications. However, pharmacists were responsible for confirming the medication reconciliation process at admission and nurses were responsible for confirming the medication reconciliation process at discharge.

Seventh, we often forget to go back and refine our medication reconciliation process. Your BPMH (best possible medication history) or PAML (preadmission medication list) is often incomplete on the day of admission and should be further refined on subsequent days by contacting family, primary care providers, etc. Yet we often forget to do this. Note that the same applies to allergy lists, etc. In one study of medication reconciliation in trauma patients (Schenkel 2008), admission medication lists were highly inaccurate. The median duration of the medication reconciliation process was 2 days. And 25% required 3 or more days. So you need some sort of flag or tickler to tell you to go back and complete this step. For instance, there is a big difference between “no known allergies” and “no adequate historian available to provide allergy information”. Particularly with electronic medical records, you can set a flag based on the latter that will prompt a caregiver on Day 2 (or later) to go back and attempt to get this information. We recently saw a discharge medication reconciliation sheet that started with the comment “unable to get medication history” despite the fact that multiple medications were prescribed during the hospital course that obviously were being taken prior to admission. So don’t forget to go back and capture this key information if you could not get it at the time of admission.

Eighth, we agree that shared accountability creates problems. We see that when physicians co-manage patients they often assume the “other” physician will do something and both fail to do what is necessary. The same applies to a Med Rec process where responsibilities are shared. Therefore, you need to make it clear up front what you expect the physician, the nurse, and the pharmacist to do in the process.

Ninth, demonstrating value of the BPMH list downstream is an important factor in getting physician buy-in to do this on admission. So when a physician is discharging or transferring a patient from one service to another and sees how the BPMH and Med Rec form can facilitate those events, they are more likely to do a BPMH themselves. However, often the emergency department physician is the first to begin a BPMH and they may not perceive the same value in this list.

Tenth, doing medication reconciliation well is resource-intense. But it has a significant return on investment not only from a patient safety perspective but also from a financial perspective. The Northwestern Memorial Hospital website has an excellent page “Making the Case for Medication Reconciliation”. Our October 23, 2007 Patient Safety Tip of the Week “Medication Reconciliation Tools” also provides links to some of the ROI tools.

See also our previous columns on medication reconciliation:

• October 23, 2007 Patient Safety Tip of the Week “Medication Reconciliation Tools”
• May 13, 2008 Patient Safety Tip of the Week “Medication Reconciliation: Topical and Compounded Medications”
• December 30, 2008 Patient Safety Tip of the Week “Unintended Consequences: Is Medication Reconciliation Next?”

Referenes:

Coffey M, Cornish P, Koonthanam T, Etchells E, Matlow A. Implementation of Admission Medication Reconciliation at Two Academic Health Sciences Centres: Challenges and Success Factors. Healthcare Quarterly, 12(Sp) 2009: 102-109
http://www.longwoods.com/view.php?aid=20719&cat=604

Cornish PL, Knowles SR, Marchesano R, Tam V, Shadowitz S, Juurlink DN, Etchells EE. Unintended Medication Discrepancies at the Time of Hospital Admission. Arch Intern Med. 2005;165:424-429

http://archinte.ama-assn.org/cgi/reprint/165/4/424

Schnipper JL, Hamann C, Ndumele CD, et al. Effect of an electronic medication reconciliation application and process redesign on potential adverse drug events: a cluster-randomized trial. Arch Intern Med. 2009; 169: 771-780

http://archinte.ama-assn.org/cgi/content/abstract/169/8/771

Northwestern Memorial Hospital. Making the Case for Medication Reconciliation.

http://www.nmh.org/nm/making+the+case

Schenkel  S. The Unexpected Challenges of Accurate Medication Reconciliation. Annals of Emergency Medicine 2008; 52(5): 493-495

http://www.annemergmed.com/article/S0196-0644(08)00613-6/abstract

Print “Barriers to Medication Reconciliation”

September 15, 2009           

ETTO's:

Efficiency-Thoroughness Trade-Offs

We are a week late with our typical holiday book review! In our module on root cause analysis we spend a deal of time discussing efficiency-thoroughness trade-offs (ETTO’s). That concept, best associated with Erik Hollnagel, is well known to everyone in the human factors and safety fields. We’ve not discussed it in our Patient Safety Tip of the Week columns before since the literature has been somewhat arcane and not easy to wade through. But the concept is very important for those who need to understand how systems work and Hollnagel has now come out with a new book (The ETTO Principle: Efficiency-Thoroughness Trade-Off. Why Things That Go Right Sometimes Go Wrong) that attempts to explain ETTO’s in a simpler way. It’s a paperback edition, written without footnotes so that it is a relatively easy read, though it still contains much of the jargon that dominates the human factors literature. As in any good patient safety or human factors presentation, he lets stories get your attention and bring home points.

He begins with the simple story of a worker who cuts off his thumb while using a circular saw. The worker knows the dangers and has the safety equipment needed (a push stick that is to be used to push the wood piece through the saw blade). But the worker realizes that he can cut many more pieces of wood without taking the time to set up the safety device (i.e. he is much more efficient) and since he has done it already many times without accident, he perceives it as relatively safe so he trades off thoroughness for the sake of efficiency. Hollnagel stresses that for most recurrent work situations people will naturally choose the more efficient mode as long as they perceive it to be safe based on their past experience.

Sound familiar? We’ve discussed workarounds in several of our columns. The workaround is the ultimate ETTO. The worker uses a much more efficient means of getting a task done while he perceives little or no safety risk.

When you think about it, almost everything we do in our lives, whether work or play, involves the ETTO principle. We are constantly trying to balance efficiency and thoroughness. In some cases we (sometimes unfortunately) tend to weigh efficiency more heavily whereas in others we weight thoroughness more heavily. You’ve heard us quote James Reason “correct performance and systematic error are two sides of the same coin”. Hollnagel notes that the ETTO principle is really saying the same thing. It is the very things we do regularly to accomplish our goals efficiently that, under slightly different circumstances, comes back to bite us. And it is human nature that, for the most part, we tend to favor efficiency over thoroughness.

Hollnagel provides numerous ETTO “rules” that serve as signals such a trade-off is being made. You’ll notice most of them from our discussions on workarounds and root cause analyses:

“It looks fine”

“It is not really that important”

“It is normally OK, there is no need to check”

“I’ve done this a million times before, so trust me…”

“It normally works”

“It’s been checked earlier by someone else”

“It’ll be checked again later by someone else”

“It’s good enough for now”

“This way is much quicker”

“There’s no time to do it now”

“We must not use too much of X”

“I can’t remember how to do it (and I can’t be bothered to look it up)”

“It looks like a Y, so it probably is a Y”

“We must get this done”

“It must be ready in time”

“If you don’t say anything, I won’t either”

“I’m not the expert on this, so I’ll let you decide”

Hollnagel provides lots of real-life examples of ETTO’ing in multiple industries, all of which have applicability to healthcare. But he also provides some healthcare-specific examples. One of the most compelling stories is that of a patient who received inappropriate chemotherapy. The patient was a farmer who had delayed seeing a physician because he did not want to skip the “calving” season on the farm. He was diagnosed as having gastric carcinoma and scheduled for chemotherapy. He was anxious to being the chemotherapy so he would be ready for the upcoming hay-drying seasone on the farm. The oncologist was anxious to start it because he was going to a conference for two weeks. The chemotherapy was begun and no one ever noticed that the final pathology report had shown a lymphoma rather than gastric carcinoma (the treatment, response to treatment, and prognosis for a lymphoma are much different than for a gastric carcinoma). He died after 5 months, receiving the wrong chemotherapy. Hollnagel outlines several of the ETTO’s obvious in the case, all of which favored efficiency over thoroughness. The patient was first seen on a day the oncologist was seeing 35 other patients, a practice that oncology clinic had obviously adopted for the sake of efficiency. The chemotherapy was scheduled prior to receipt of the final pathology report, again for efficiency’s sake because both the physician and patient wanted to get started for the above-mentioned time pressures. And the pathology report (which apparently did return prior to onset of the chemotherapy but was never noticed) had the final diagnosis typed in small font on the second page compared to the oncologist’s clinical diagnosis which appeared in large font on the front page. That was another ETTO favoring efficiency (that the pathologist did not make a separate phone call to the oncologist when a disparity between clinical and pathological diagnoses was present).

Another with good lessons learned was a near-miss aviation situation. A vintage plane used by a sight-seeing company had inaccurate fuel gauges so the crew had a policy of always dipping the fuel tanks to verify fuel before flights. The plane usually held most of the fuel in the main tanks and enough reserve fuel for 45 minutes of flight in the auxillary tanks. On one particular day, the situation was reversed – the auxillary tanks held all the fuel needed for that day’s planned flight and the main tanks held the reserve fuel. During the briefing, the information about the reversed fuel situation was mentioned but never emphasized. When the pilot asked the technician if the main tanks had been dipped, the answer was “yes” (but, again, no discussion about the reversal of where the main fuel source for today was). Shortly after the flight began, they had engine problems and could not restart the engines despite attempts to switch back and forth between main and auxillary tanks. They made a successful emergency landing. Lots of ETTO’s in this one! The failure to emphasize the reversal of the fuel allocation led the pilots to assume all was as usual. Because they knew the fuel tanks had been dipped by someone else, they did not dip them themselves each time the plane flew that day. Flying with known inaccurate fuel gauges was tolerated and the workaround was dipping the tanks prior to takeoff to verify the tanks were filled. You may recognize the latter phenomenon as “normalization of deviance” where an organization comes to accept a deviation from best practice as the new standard practice (i.e. a workaround becomes the norm). How many of you know of examples in your organizations where there is a workaround around a faulty piece of equipment or a faulty procedure that has now become accepted “standard” practice? If you don’t start looking today – you’ll find them!

Hollnagel even talks about how the ETTO principle applies to root cause analyses. In an RCA we continually ask the question “why” on and on again. He notes that in any investigation you must apply a “stop rule” in which you establish criteria for ending your investigation. It is really the ETTO principle: you are really trading off throroughness vs. efficiency. You have to expend resources (time, personnel, money) to continue the investigation but you also have a need to fully understand the causes of the accident in order to prevent future recurrences. He also points out that our bias to investigate cases with bad outcomes (our “failures”) rather than our successes is also an ETTO.

Hollnagel’s book has good discussions about the strengths and flaws of the various theories of accidents. He points out the complexities that make finding simple causes unlikely and the fallacies of those approaches looking for such. He also emphasizes some of the inherent biases and faults of root cause analysis. In particular, he talks about how most accident theories apply best to “tractable” systems (meaning those systems that are readily predictable and manageable) and fail to take into account novel factors that occurred (which, by definition, could not have had a root cause since they did not exist prior to the event!). And he makes the case that the systems most prone to such unpredictabilities are those that use RCA’s the most: hospitals and healthcare!

He emphasizes that rather than focusing on failures when we review cases, we should look at what normally should take place. Then we should review what ETTO’s took place and why such things that usually go right went wrong in the particular case. To illustrate his point, Hollnagel desribes in detail a trauma case admitted to a hospital in which multiple physicians caring for a patient all focus on the patient’s head injury, assume that one and other are addressing other issues, only to ignore a hip dislocation which led to loss off blood supply and eventual need for a hip replacement.

But we are still left wondering where ETTO fits into the patient safety world. We liken it to PET scanning. PET scanning was a great technology that spent years looking for a clinically useful application. We think the ETTO principle is the same: a great concept looking for a useful application. While an understanding of the ETTO principle is very useful in helping us to understand behaviors in analyzing events that have already happened, we think that its greatest utility will be in prospectively managing processes. In designing new processes, applying the ETTO principle to predict responses in various scenarios. Hollnagel apparently feels the same way and discusses potential applicability of the ETTO principle in forward-looking methods like Probabalistic Risk Assessment (PRA) or Human Reliability Assessment (HRA). Those of us who have been involved in design of clinical computing systems readily recognize the importance of understanding how individuals are likely to react in certain situations. We know that ETTO’s are likely to occur and that needs to be an integral part of such planning and design.

References:

Hollnagel E. The ETTO Principle: Efficiency-Thoroughness Trade-Off. Why Things That Go Right Sometimes Go Wrong. Burlington, VT: Ashgate Publishing Company, 2009 http://erik.hollnagel.googlepages.com/theettobook

Print “ETTO’s: Efficiency-Thoroughness Trade-Offs”

September 22, 2009           

Psychotropic Drugs and Falls in the SNF

Our interest in this topic was kindled by the anecdotal finding of a drop in fall rates at an SNF that coincided with a change in the psychiatric consulting pattern at the facility. Use of psychotropic drugs in nursing homes is common. In the late 1980’s and 1990’s there was great concern that psychotropic drugs were often being inappropriately used, often for convenience of staff rather than to improve the functional status of patients. Hence, considerable regulatory oversight at both the state and federal levels was introduced in attempt to improve appropriateness of use of such drugs. CMS (known as HCFA at that time) issued guidelines that required documentation of the reason(s) for use of these drugs, regular reviews, and periodic attempts to wean patients off these drugs.

A review by the Office of the Inspector General in 2001 concluded that psychotropic drug use in nursing homes was generally appropriate. It was deemed medically appropriate and within Medicare guidelines for 85% of patients. In 8% it was felt that use of psychotropic drugs was inappropriate (which could mean that an appropriate indication was not documented, dose was improper, periodic tapering not done, etc.) and in the remaining 7% appropriateness could not be determined because of insufficient records. However, a new study published this month (Agashivala 2009) shows that inappropriate use of psychoactive medications continues to contribute to falls in the nursing home setting. That study analyzed data from the large 2004 National Nursing Home Survey database and performed logistic regression methods to look at potentially inappropriate psychoactive medications (PIPM’s), using Beer’s criteria, and the risk of falls (see our January 15, 2008 Patient Safety Tip of the Week “Managing Dangerous Medications in the Elderly “ and our June 2008 What’s New in the Patient Safety World column “Potentially Inappropriate Medication Use in Elderly Hospitalized Patients” for discussions on Beer’s List). They found a significantly increased risk of falls in patients on PIPM’s compared to those patients taking no psychoactive medications or other psychoactive medications not designated as PIPM’s on Beer’s List.

Concerns have also arisen over the past 10 years about the increased mortality rates seen in nursing home patients treated specifically with antipsychotic drugs, both the newer “atypical” antipsychotics and the older antipsychotics (Gill 2007). A recent study (Kamble 2009) showed that about a third of nursing home residents with dementia are treated with such drugs. Another recent study (Bronskill 2009), done on Canadian nursing homes, found considerable variation in the prescribing rates for antipsychotic drugs. They found that rates of prescribing such drugs varied from 11.6% in SNF’s in the lowest quartile to 30% in the highest. More importantly, they showed that 30-day and 120-day mortality rates were higher in those facilities prescribing more antipsychotic drugs despite similar clinical characteristics at all facilities. Their work supports using antipsychotic drug prescribing rates as a measure of quality of SNF care.

There are, of course, many reasons why nursing home residents are at greater risk of falls. First and foremost, they tend to have physical disabilities that impair their mobility. That includes not only neurological conditions that may impair strength and balance but also arthritic and orthopedic conditions. Second, a big contributor to falls in the elderly is what we refer to as the “multiple sensory deficit syndrome”. This means that impairments of sensation, vision, hearing, etc., each of which may be mild and not enough to cause falls by itself, are additive and collectively increase the risk of falls. Third, orthostatic hypotension is common in the nursing home population. This may be a side effect of several medications a patient may be taking or it may be part of their underlying condition (eg. diabetic autonomic neuropathy) and it may be aggravated by “deconditioning” from lying in bed for prolonged periods. Fourth, confusion or dementia may lead to behaviors (eg. wandering) in which the patient is more likely to fall. Fifth, incontinence may lead to falls both because it increases the risk of slipping on a wet surface and because a patient may be hurrying to the bathroom and slip.

But the biggest contributor to the increased fall risk in the elderly is medication. Polypharmacy is problematic in the elderly, whether they are in nursing homes or the community. And drugs may increase the risk of falls both individually and in combination with other drugs. Drugs most likely to contribute to falls are long-acting benzodiazepines, tricyclic antidepressants, antipsychotic drugs, and any drug that causes sedation as a primary or side effect. An epidemiologic study done on elderly ambulatory nursing home residents (Thapa 1995) controlled for many other variables and concluded that the attributable risk for recurrent falls for psychotropic drugs was 36%.

So what should you do in your nursing home to reduce the risk of falls? Obviously a good fall prevention program looks at both intrinsic and extrinsic fall risk factors. Attention to the environment to ensure adequate lighting, safe floors (eg., no carpet edges to trip on), timed toileting, assisted toileting, grab bars in bathrooms, etc. are very important. Proper implementation of restraint policies is beyond the scope of today’s column and the relationships between restraints and falls are very complex. The same applies to the bedrail issue we have previously discussed (see our December 18, 2007 Patient Safety Tip of the Week “Bed Rails” and our August 2009 What’s New in the Patient Safety World column “Bed Rails: Taking Emotion Out of the Debate”).

But review and regular re-review of medications is essential. When drugs in the high-risk categories noted above are used, there should be clearcut indications and usually also documentation that less invasive interventions have not sufficiently helped. The desired goals of the drug therapy should be made clear to all involved and progress toward these goals clearly documented in the chart. The medications generally should be started in low doses and titrated as needed, monitoring very carefully for side effects or other unintended consequences. Your monitoring tool should include not only general assessments (such as ability to ambulate) but also side effects that are specific to the drug being used (eg. orthostatic hypotension or drowsiness from tricyclic antidepressants, extrapyramidal symptoms and signs from antipsychotic drugs, etc.). The goals should clearly specify how long it is anticipated the patient will need to be on the drug and then attempts to taper the patient off that drug should take place. Most states have regulations regarding how often such evaluations and attempts to taper patients off these drugs must be made. Regular review of the medication regimen by a consulting clinical pharmacist, consulting psychiatrist, or the SNF medical director are important. Since in many cases the patient was admitted from an acute care setting already on such drugs, good medication reconciliation is important and the same questions should be asked (eg. why?, how long?, goals? etc.) at the time of transfer to the SNF.

Falls are also a frequent reason for SNF patients requiring readmission to acute care hospitals. So as CMS and other payors move to penalize acute care hospitals for readmissions, suddenly it is important that you pay attention to fall rates, prescribing patterns, and other quality factors at SNF’s even if your facility does not own its own SNF.

References:

Department of Health and Human Services Office Of Inspector General. Psychotropic Drug Use in Nursing Homes. November 2001

http://oig.hhs.gov/oei/reports/oei-02-00-00490.pdf

Agashivala N, Wu W. Effects of Potentially Inappropriate Psychoactive Medications on Falls in US Nursing Home Residents+: Analysis of the 2004 National Nursing Home Survey Database. Drugs & Aging 2009; 26: 853-860

http://adisonline.com/aging/Abstract/2009/26100/Effects_of_Potentially_Inappropriate_Psychoactive.5.aspx

Gill SS, Bronskill SE, Normand S-LT, et al. Antipsychotic Drug Use and Mortality in Older Adults with Dementia. Annals of Internal Medicine 2007; 146: 775-786

http://www.annals.org/cgi/content/abstract/146/11/775?maxtoshow=&HITS=10&hits=10&RESULTFORMAT=&fulltext=gill+ss&searchid=1&FIRSTINDEX=0&resourcetype=HWCIT

Kamble P, Chen H, Sherer JT, Aparasu RR. Use of Antipsychotics among Elderly Nursing Home Residents with Dementia in the US: An Analysis of National Survey Data. Drugs & Aging 2009; 26: 483-492

http://adisonline.com/aging/Abstract/2009/26060/Use_of_Antipsychotics_among_Elderly_Nursing_Home.5.aspx

Bronskill SE, Rochon PA, Gill SS, et al. The Relationship Between Variations in Antipsychotic Prescribing Across Nursing Homes and Short-Term Mortality: Quality of Care Implications. Medical Care 2009; 47: 1000-1008

http://journals.lww.com/lww-medicalcare/Abstract/2009/09000/The_Relationship_Between_Variations_in.10.aspx

Thapa PB, Gideon P, Fought RL, Ray WA. Psychotropic Drugs and Risk of Recurrent Falls in Ambulatory Nursing Home Residents. American Journal of Epidemiology 1995; 142: 202-211

http://aje.oxfordjournals.org/cgi/content/abstract/142/2/202

Print “Psychotropic Drugs and Falls in the SNF”

September 29, 2009           

Perioperative Peripheral Nerve Injuries

You’ve heard us several times in the past recommend the surgical “timeout” or any other preoperative “huddle” be used for things other than merely identifying correct patient and surgical site (see our Patient Safety Tips of the Week for April 9, 2007 “Make Your Surgical Timeouts More Useful” and December 9, 2008 “Huddles in Healthcare” and March 10, 2009 “Prolonged Surgical Duration and Time Awareness”). One recommendation is to discuss whether the patient will be at risk for nerve injury and at what point in a procedure would repositioning of the patient be considered. This month’s issue of the journal Anesthesiology contains an article on perioperative peripheral nerve injuries (Welch 2009). They reviewed all perioperative peripheral nerve injuries at one institution over a 10-year period and found a nerve injury rate of 0.03% per anesthetic given, a rate similar to that found in previous literature. That article and the accompanying editorial (Prielipp 2009) both acknowledge that this is probably an underestimate of the true incidence of perioperative nerve injuries because of the methods used to ascertain cases. 65% of those injuries involved the upper extremities. They found three underlying medical conditions to be statistically associated with nerve injuries: diabetes, hypertension, and tobacco use. They found nerve injuries associated with general and epidural anesthesia but none with monitored anesthesia care.

Perioperative nerve injuries are the second or third most common reason for malpractice claims against anesthesiologists and the number of closed claims for perioperative nerve injuries has remained quite consistent over time (Cheney 1999).

Ulnar nerve injuries are the most commonly reported upper extremity nerve injuries (Barner 2003). Almost all series have demonstrated a strong male preponderance for these, though the reason is very unclear. Patients either excessively thin or obese appear to be at greater risk. Often the symptoms and signs are delayed in onset and there appears to be a correlation with lengths of hospitalization greater than 14 days. This, plus the finding that some patients admitted to medical services develop ulnar neuropathy, have suggested that prolonged bedrest in the supine position may be a contributory factor (Warner 2000). Though the mechanism most often speculated about is pressure on the nerve at the elbow, cases have occurred even in circumstances where adequate padding to the elbow has been utilized (Cheney 1999).  The nerve appears to be particularly vulnerable at the elbow, where it passes throught the ulnar groove and an aponeurotic band of tissue. Various movements and positions of the elbow may increase compression of the nerve in this area. Flexion of the elbow and pronation of the forearm and internal rotation of the shoulder, in particular, are movements that increase pressure in the ulnar groove. This may be aggravated by subluxation. Anatomical or arthritic changes in the region may also accentuate the pressure. These dynamic factors may help explain why patients lying supine with elbows flexed and forearms pronated may be at risk for developing ulnar neuropathy, regardless of whether they are medical or surgical patients.

Though we know of no solid evidence base for preventive measures, most recommendations include careful padding and intraoperative positioning (eg. keeping forearms supinated), avoiding hard surfaces or edges near the elbow, and not leaning against the patient (Barner 2003).

The same authors (Barner 2002) reviewed perioperative nerve injuries in the lower extremities as well. They discuss the complexities and challenges, both clinical and electrophysiological, involved in differentiating dysfunction of individual peripheral nerves from dysfunction more proximally or at the lumbosacral plexus level. Though they stress the vulnerability of the peroneal nerve to external pressure at the level of the fibular head in procedures done in the lateral decubitus position or in the lithotomy position (particularly when external devices such as stirrups or leg straps are used) but again stress that localization of the site of the compression may be misleading.

And it’s not just the OR that carries the risk for perioperative nerve injuries. A recent report noted peroneal nerve injuries associated with colonoscopy in children (Bong 2009). They found 3 cases of peroneal nerve injury in a cohort of 746 children undergoing colonoscopy, suggesting that such injuries are not uncommon. In addition to risk factors such as weight loss, malnutrition, chronic systemic inflammation, and steroid use, positional factors and relatively long duration of anesthesia were considered to be contributing factors. Duration of anesthesia was 120 minutes in two cases and 70 minutes in the third. All patients were temporarily in the supine position with hips flexed and externally rotated and knees flexed and supported laterally (“frog legs” position) and then in a lateral decubitus position. Pressure on the peroneal nerve may have occurred directly while in the lateral decubitus position or from the lateral knee support in the “frog legs” position.

The risk factors and contributing factors to these perioperative nerve injuries have not been well defined in the literature. Compression of or stretching of the nerve, ischemia, and direct trauma are often mentioned as possible mechanisms of injury but in most cases a mechanism is not clearly demonstrated.

We do know that the existence of some underlying polyneuropathies may render nerves more vulnerable to the effects of pressure or other trauma. It’s well known that entrapment neuropathies are more common in patients with underlying polyneuropathies. So a history of known polyneuropathy, such as a diabetic polyneuropathy, should clearly be considered as a risk factor for perioperative nerve injury. Patients with hypertrophic neuropathies, whether hereditary or acquired, are especially susceptible to the effects of compression.

Weight loss is another predisposing factor to some compressive neuropathies. Peroneal nerve injuries are particularly likely to occur in individuals who have lost the typical subcutaneous fat pad that protects the nerve near the head of the fibula. So it might be expected that cancer patients might be particularly susceptible because of weight loss and the frequent occurrence of polyneuropathy (either as a remote effect of cancer or a result of chemotherapy).

As above, we know of no solid evidence base for preventive measures. But the American Society of Anesthesiologists (ASA) did issue a practice advisory for prevention of perioperative peripheral neuropathies in 2000. They made recommendations on specific positioning to use or to avoid in attempt to prevent various nerve or plexus injuries, plus recommendations about use of protective padding in certain cases and equipment such as automated blood pressure cuffs. The specifics can be found in that document. However, what is most striking is that virtually all of the recommendations are made based upon opinions of consultants, task force members, and a random sampling of ASA members. This practice advisory points out the dearth of scientific literature on interventions to minimize the risk for perioperative nerve injuries.

Strikingly, even the data from the medicolegal databases lacks information about total duration of anesthesia, duration a patient was in a certain position, etc. You don’t have to be a neurologist to know that positional changes may affect nerve function. All of you have experienced your arm or leg “falling asleep” when it is in a certain position too long. What do you do in response?  You change position and maybe shake the limb around a little bit and the sensation returns in a few seconds. You can often then return it to the original position. It is amazing that the literature on perioperative nerve injuries does not even comment on the issue of periodic or intermittent repositioning of limbs.

There have been at least a few attempts at better understanding the relationship between positional changes and nerve damage. Kamel et al (Kamel 2006) retrospectively looked at patient undergoing spinal surgery who had intraoperative monitoring of somatosensory evoked potentials (SSEP’s). The amplitudes of SSEP’s are affected by multiple factors, such as mean arterial blood pressure, depth of anesthesia, manipulation of the spine, etc. but they may also be sensitive to ischemia, compression, stretching or other dysfunction of peripheral nerves or plexuses. In their study, they found that about 6% of changes in upper extremity SSEP’s during spine surgery responded to changes in position of the affected upper extremity. They were thus able to determine which body positions were more likely to be associated with such changes during spine surgery. The study did not include assessment of individual nerve function so information cannot be generalized to make recommendations for protection of specific peripheral nerves. But such methodology is promising and needs to be studied prospectively in a variety of surgical settings. Since patients under anesthesia cannot sense symptoms related to peripheral nerve dysfunction, SSEP’s do have promise at uncovering reversible changes during a variety of surgical procedures, not just spinal surgery.

Even though the rate of 0.03% per anesthetic given may sound quite low, the fact is that translated to about one case a month at the medical center where Welch et al. did their retrospective study. We agree with the term used in the Prielipp editorial: “scream”. The problem is really “screaming out” for legitimate scientific investigation into etiological factors related to these nerve injuries and, more importantly, effectiveness of various preventive intervention strategies.

We wish we had some firmer recommendations for you in this significant patient safety issue. However, all we can say at this time is that you should attempt to identify patients at highest risk, try to minimize the total duration of anesthesia or the duration they are in certain positions, and use the ASA recommendations for position, padding, equipment, etc. A lot more needs to be done before we have any definitive recommendations.

References:

Welch MB, Brummett CM, Welch TD et al. Perioperative Peripheral Nerve Injuries: A Retrospective Study of 380,680 Cases during a 10-year Period at a Single Institution. Anesthesiology 2009; 111(3): 490-497

http://journals.lww.com/anesthesiology/Fulltext/2009/09000/Perioperative_Peripheral_Nerve_Injuries__A.12.aspx

Prielipp RC,Warner MA. Perioperative Nerve Injury: A Silent Scream? Anesthesiology 2009; 111(3): 464-466

http://journals.lww.com/anesthesiology/Fulltext/2009/09000/Perioperative_Nerve_Injury__A_Silent_Scream_.3.aspx

Cheney FW, Domino KB, Caplan RA, Posner KL. Nerve Injury Associated with Anesthesia: A Closed Claims Analysis. Anesthesiology 1999; 90(4): 1062-1069 http://journals.lww.com/anesthesiology/Fulltext/1999/04000/Nerve_Injury_Associated_with_Anesthesia__A_Closed.20.aspx

Barner KC, Landau ME, Campbell WW. A Review of Perioperative Nerve Injury to the Upper Extremities. Journal of Clinical Neuromuscular Disease 2003; 4(3): 117-123

http://journals.lww.com/jcnmd/Abstract/2003/03000/A_Review_of_Perioperative_Nerve_Injury_to_the.4.aspx

Barner KC, Landau ME, Campbell WW. A Review of Perioperative Nerve Injury to the Lower Extremities: Part I. Journal of Clinical Neuromuscular Disease 2002; 4(2):95-99

http://journals.lww.com/jcnmd/Abstract/2002/12000/A_Review_of_Perioperative_Nerve_Injury_to_the.9.aspx

Warner MA, Warner DO, Harper CM, Schroeder DR, Maxson PM.. Ulnar Neuropathy in Medical Patients. Anesthesiology 2000; 92(2): 613-615

http://journals.lww.com/anesthesiology/Fulltext/2000/02000/Ulnar_Neuropathy_in_Medical_Patients.47.aspx

Bong CL, Seefelder C. Peripheral Neuropathy Following Colonoscopy Is Not Uncommon. Anesthesia & Analgesia 2009; 108(1): 384-385

http://www.anesthesia-analgesia.org/cgi/content/full/108/1/384?maxtoshow=&HITS=10&hits=10&RESULTFORMAT=&author1=bong&andorexactfulltext=and&searchid=1&FIRSTINDEX=0&sortspec=relevance&resourcetype=HWCIT

American Society of Anesthesiologists Task Force on Prevention of Perioperative Peripheral Neuropathies. Practice Advisory for the Prevention of Perioperative Peripheral Neuropathies: A Report by the American Society of Anesthesiologists Task Force on Prevention of Perioperative Peripheral Neuropathies. Anesthesiology 2000; 92(4): 1168-1182

http://journals.lww.com/anesthesiology/Fulltext/2000/04000/Practice_Advisory_for_the_Prevention_of.36.aspx

Kamel IR, Drum ET, Koch SA, Whitten JA et al. The Use of Somatosensory Evoked Potentials to Determine the Relationship Between Patient Positioning and Impending Upper Extremity Nerve Injury During Spine Surgery: A Retrospective Analysis.

Anesthesia & Analgesia 2006; 102(5): 1538-1542

http://www.anesthesia-analgesia.org/cgi/content/full/102/5/1538

Print “Perioperative Peripheral Nerve Injuries”

Click here to leave a comment on any of these tips.

To get "Patient Safety Tip of the Week" emailed to you, click here and enter "subscribe" in the subject field.

Go to Tip of the Week Archive a patient safety resource solution loaded with tips, tools, and techniques you can use in your patient safety and quality improvement initiatives.

Click here to see the consulting services and patient safety solutions that we provide.