Healthcare Consulting Services
Healthcare Consulting Services with a Focus on Patient Safety Solutions and Quality Improvement Across the Health Care Continuum. Your Patient Safety Resource Solution.
April 19, 2016
Independent Double Checks and Oral Chemotherapy
Our May 5, 2015 Patient Safety Tip of the Week “Errors with Oral Oncology Drugs” pointed out that, with the recent proliferation of oral chemotherapy agents, we are beginning to see more and more patient safety issues arising from their use. One facet we only briefly touched upon in that column was the use of independent double checks.
One of our most frequently accessed columns is our October 16, 2012 Patient Safety Tip of the Week “What is the Evidence on Double Checks?”. Usually when we are talking about independent double checks we are referring to such checks being done at the same stage of the medication continuum (eg. during ordering or during preparation). But medication safety also includes another type of independent double check at several levels. For example, in a hospital medications ordered by a physician or other prescriber are typically checked by a pharmacist and nurse at different times. In our November 3, 2015 Patient Safety Tip of the Week “Medication Errors in the OR – Part 2” we noted one of the reasons for more medication errors in the OR is that the anesthetist is often the one choosing a medication, preparing it, and administering it, thereby lacking those independent checks that might occur elsewhere.
While we’ve pointed out the limitations of double checks in several columns, we still recommend independent double checks for high alert drugs. In the hospital, chemotherapy agents are high alert drugs that get such attention. But what happens outside the hospital? Particularly as more and more oral chemotherapy agents are being developed and used, do they get the same degree of scrutiny that is typically done for IV chemotherapy agents? A new study (Griffin 2015) suggests no. Griffin and colleagues in Canada compared the performance of independent double checks for oral vs. intravenous chemotherapy and found striking differences. Using direct observation by two human factors experts and semi-structured interviews they found 57 systematic checks for IV chemotherapy compared to only 6 for oral chemotherapy. Systematic checks required the check be done clearly as a part of a qualified healthcare provider’s (HCP) role, be done consistently, and be done with all the information necessary to inform such a check. They defined “partial” checks if one of those elements was missing. Even when they included “partial” checks, the results were 64 checks for IV chemotherapy vs. only 17 for oral chemotherapy.
For IV chemotherapy there were checks at multiple levels (planning, ordering, and dispensing). So for IV chemotherapy there were several opportunities for a qualified HCP to identify and mitigate an unintentional error before it reached the patient. Not so for oral chemotherapy. Often the only qualified HCP with an opportunity to intercept an error was the community pharmacist and these were in the ordering and dispensing phases only. And those community pharmacists were often lacking critical information. They often did not even have access to critical information such as the type of cancer the patient had.
Though this was a Canadian study, the authors cite studies from the US (Weingart 2007) that showed few of the safeguards routinely used for infusion chemotherapy have been adopted for oral chemotherapy at US cancer centers and lack of consensus at these centers about safe medication practices for oral chemotherapy.
In our May 5, 2015 Patient Safety Tip of the Week “Errors with Oral Oncology Drugs” we discussed an analysis of incidents involving oral chemotherapy agents in Canada (ISMP Canada 2015). They analyzed 516 incidents over a 12 year period and while patient harm or death occurred in only a small minority of cases there were multiple lessons learned. They were able to identify 3 major themes: (1) lack of specialized knowledge by care providers (2) medication name mix-ups and (3) lack of safe medication handling processes. Under the first theme ISMP Canada noted that chemotherapy regimens can be very complex and that community pharmacists may not have a good understanding of chemotherapy cycles, side-effect profiles, etc. The same applies to many other non-oncologist healthcare professionals. As a result, patient education on these complex issues may be insufficient. The example they provide is a patient who was on a protocol-defined capecitabine cyclical dosing regimen who was mistakenly given capecitabine daily when admitted as an inpatient and then the prescription given the patient at discharge also erroneously called for daily capecitabine. You may recognize this problem of prescribing a drug daily rather than on a different regimen because we’ve written about very similar problems with methotrexate in the past (see our What’s New in the Patient Safety World columns for July 2010 “Methotrexate Overdose Due to Prescribing Error” and July 2011 “More Problems with Methotrexate” and February 2016 “Avoiding Methotrexate Errors”).
As a subtheme of the “lack of specialized knowledge” they mention that failure to include some critical information on the prescriptions may contribute. Such information might include diagnosis, patient height and weight, duration of the chemotherapy cycle, etc.
We also cited a classic article on medication errors related to oral chemotherapy (Weingart 2010). Weingart and colleagues warned that the use of oral chemotherapy was expanding and that few of the safeguards that apply to prescribing, dispensing and administering IV chemotherapy have been applied to oral chemotherapy. They identified issues with oral chemotherapy from a variety of sources. They identified over 500 such errors, including 99 actual adverse drug events. While many of the remaining events were near-misses they did provide the opportunity for patient harm and thus were helpful in identifying potential vulnerabilities. Importantly, they identified that over half the errors (55.9%) were intercepted, preventing actual patient harm. Of those intercepted, this was most often by a pharmacist (69.5%), followed by patient or family (10.2%) or nurse (6.9%). Wrong dose errors were most frequent but wrong drug and missed dose errors also occurred. But one particularly salient problem was supplying the wrong number of days. Whereas the majority of errors involving wrong dose, wrong drug, or missed dose resulted in near misses, 39.3% of errors involving wrong number of days supplied resulted in adverse drug events. We went on in that column to discuss several other instances of pharmacies supplying the wrong number of days or doses for oral chemotherapy.
So now that we’ve identified lack of double checks as a vulnerability for oral chemotherapy, how do we address it? One is to ensure that a pharmacist with access to all critical clinical information and a thorough understanding of both the particular cancer and the chemotherapy regimen is involved as a potential double checker. Unfortunately, this may mean that some community pharmacists and mail order pharmacies are not the most appropriate for patients on oral chemotherapy regimens. Anyone who has picked up a prescription at a community pharmacy will recognize how busy the pharmacist is, further increasing the vulnerability to errors. And very often there is only a single pharmacist working. So there is no opportunity for a double check. The only barrier left to intercept the error is the patient him/herself (or their family/caregiver). So one consideration would be to ensure that prescriptions for oral chemotherapy are only dispensed in certain pharmacies that meet all the safety requirements. Such pharmacies would most likely hospital- or clinic-based but some community pharmacies would be eligible if they met those criteria. That may also require various payors to ensure that such special pharmacies are included in their networks. Since the criteria should include not only ability to perform double checks but also have full access to relevant clinical information, knowledge about the various chemotherapy regimens, and ability to sit down with the patient to thoroughly go over the instructions, we doubt that mail order pharmacies would meet the criteria.
Secondly, recognizing that complex regimens and cycles are the most vulnerable to errors involving wrong number of day supplies, we need to look at how we prescribe. We are always balancing patient convenience against patient safety. In several of our previous columns on home infusion chemotherapy we noted no one seemed to be asking “what is the highest dose that a patient could tolerate in one day (or less) if there was inadvertent administration of the infusion?”. A safety culture would design the protocol with sublethal dosages that would protect the patient in the event of “what can go wrong will go wrong”. It also would not put the healthcare workers at the “sharp end” in a situation none of us would want to be in. The same question should apply to oral chemotherapy regimens and be “What would be the highest aggregate dose a patient could tolerate over a specified period?” and avoid prescribing more than that inadvertently. Yes, the patient might be inconvenienced by having to do another physician or clinic visit to get a prescription for the next cycle or the second part of a complex regimen. But isn’t that preferable to receiving a chemotherapy overdose due to an avoidable error?
Our prior columns related to chemotherapy safety:
Griffin MC, Gilbert RE, Broadfield LH, et al. Comparison of Independent Error Checks for Oral Versus Intravenous Chemotherapy. Journal of Oncology Practice 2015;
Published online before print September 29, 2015
Weingart SN, Flug J, Brouillard D, et al. Oral chemotherapy safety practices at US cancer centres: questionnaire survey. BMJ 2007; 334: 407
ISMP Canada. Analysis of Incidents Involving Oral Chemotherapy Agents. ISMP Canada Safety Bulletin 2015; 15(4): 1-4, April 22, 2015
Weingart SN, Toro J, Spencer J, et al. Medication errors involving oral chemotherapy. Cancer 2010; 116(10): 2455-2464
April 26, 2016
Lots More on Preventing Readmissions But Where’s the Beef?
Busy month for articles on preventing readmissions! Two on tools for predicting readmission and more on programs looking to prevent readmissions.
The HOSPITAL score is a readmission prediction score originally derived and internally validated on medical patients at the Partners HealthCare Network in Boston in 2013 (Donzé 2013). It is based on 7 independent factors:
H hemoglobin at discharge
O discharge from an Oncology service
S sodium level at discharge
P procedure during the index admission
IT index type of admission (urgrent)
A number of admissions during the last 12 months
L length of stay
Donzé and colleagues now have validated the HOSPITAL score in an international multicenter study that demonstrated the score could identify patients at high risk of 30-day potentially avoidable readmission with moderately high discrimination and excellent calibration in medical patients (Donzé 2016).
Another study (Siracuse 2016) showed that risk-stratification models, such as the Readmission After Total Hip Replacement Risk (RATHRR) Scale, can identify high-risk patients for readmission and permit implementation of patient-specific readmission-reduction strategies to reduce readmissions and health care expenditures. The variables (from total hip replacement patients in a large administrative database) found to be associated with increased risk of readmission after total hip replacement were: being older than 71 years, African American, in the lowest income quartile, revision replacement, liver disease, congestive heart failure, chronic pulmonary disease, renal failure, diabetes, fluid and electrolyte disorder, anemia, rheumatoid arthritis, coagulopathy, hypertension, and obesity. These were used to create the RATHRR Scale, which was applied to the validation cohort and explained 89.1% of readmission variability in that cohort. The authors suggest that the RATHRR Scale could be used preoperatively to identify patients at greatest risk for readmission and pay special attention to their needs before and after discharge.
The RATHRR score weighs comorbidities heavily. Another recent study (Havens 2015), looking at readmissions following emergency general surgery, also found that a Charlson Comorbidity Index score of 2 or greater was a predictor of readmission, along with public insurance and leaving against medical advice.
Both the HOSPITAL and RATHRR scores do identify patients at highest risk for readmission. However, we’re not sure they are really of much use in implementing care plans to reduce the risk of readmission. We’ve had this discussion before in our many columns on fall prevention. Fall risk “scores” are not of much benefit if they just identify patients at highest risk without directing focus to those factors that are most amenable to interventions.
So what are those factors that might be amenable to intervention and prevent readmissions? Auerbach and colleagues (Auerbach 2016) reviewed in detail 1000 cases where general medicine patients (average age 55 years) were readmitted to academic hospital within 30 days of discharge and found that about one-quarter (26.9%) of readmissions were potentially preventable. Perspectives of patients, physicians and intense case review were used. Some factors identified but that did not prove to be independent variables were failure to adequately treat symptoms, failure to monitor for medication adverse effects or noncompliance, failure to schedule timely followup visits, and need for additional or different home services. Of factors independently associated with preventable readmissions, premature discharge was a factor in some cases but other factors were equally or more important. Such included inability of patients to keep followup appointments, patient lack of awareness of whom to contact after discharge, failure to relay important information to outpatient healthcare professionals, and lack of adequate discussion about care goals for patients with serious illnesses. But another big factor identified was decision making by the emergency department (admitting patients that others felt did not need admission), though we concur with the accompanying editorial (Atkins 2016) that hindsight bias may be influencing that perception. The authors also note some factors that were somewhat surprisingly not found to be contributing factors. These included patient functional status and patient reports of and satisfaction with care. Significantly, in about half the admissions deemed preventable, gaps in care during the index admission were felt to have been present.
And what is the impact of interventions designed to prevent avoidable readmissions? Another new study reported results of a multicomponent intervention in which Yale-New Haven Hospital and an affiliated community hospital partnered with community resources in attempt to prevent readmissions (Jenq 2016). This was part of a special Medicare program to promote programs aimed at preventing readmissions. The teams screened inpatients, using a tool from the BOOST project (see our May 10, 2011 Patient Safety Tip of the Week “Preventing Preventable Readmissions: Not As Easy As It Sounds”) to identify patients at high risk for readmission. The intervention consisted of patient/family education, use of community care “consultants”, followup phone calls aimed at making sure patients were engaged in their care, medication management, compliance with discharge instructions, and home visits if necessary. There was also coordination between the care “consultants” and staff if the patient had been discharged to an SNF or long-term care facility.
The adjusted readmission rate for the target group decreased from 21.5% to 19.5% (that for the control population went from 21.1% to 21.0% during the same period). The relative risk reduction was 9.3%, a modest reduction at best and less than the target Medicare had set. In fact, the number needed to treat (NNT) to avoid one readmission was 50, a substantial number. Figuring in the costs of the program, the authors calculated it took $7000 of resources to prevent one readmission. The authors do note, however, that the average cost to Medicare for readmission is about $12,000 so there may still be some incentive for Medicare to support such intervention programs. We note that does not take into account the additional services and costs on the outpatient side so any potential savings to Medicare is even smaller.
The study, one of the few done on sizeable populations, illustrates the challenges and somewhat frustrating results seen in this endeavor.
Almost all the studies done on preventing readmissions focus on fiscal issues. Not surprising since hospitals may now penalized by CMS and other payors for readmissions. But lost in all of this is what happens to the patient in terms of outcomes, patient safety, mortality, and quality of life. Almost no studies on readmissions deal with those issues. We concur with the editorialists (Atkins 2016) that perhaps the most important lesson from the Auerbach study is that it highlights the fragmented healthcare system and gaps in communication and continuity of care. We can’t tell you how often we see instances where primary care physicians (PCP’s) are never even informed that their patient was in the ER or hospitalized. Or instances where patients show up in the PCP office after discharge and no records of the hospitalization or ER visit are accessible. Or, worse yet, results of pending tests done during the hospitalization are never received and acted upon by anyone. We’ve even seen instances where on admission to the hospital the patient’s PCP’s name gets replaced in an EMR field with the name of a hospital-based physician and any information that might get routed to a PCP now gets routed elsewhere. And while the ER took some blame in the Auerbach study, there are likely reasons the ER was put in the position of making those decisions to admit. Often the patient went to the ER because the physician cross covering for the PCP (or other physician) did not know the patient and took the easy route, telling the patient “Go to the ER”. Particularly in academic settings such as those in the Auerbach study it is common for patient calls after hours to be taken by residents who do not know the patient and also take the “go to the ER” route.
So, yes, readmissions deserve attention but for the right reasons. We always recommend that whenever a patient is readmitted (or has an unplanned admission) a mini-RCA (root cause analysis) should be done. While the PCP should usually be the one to initiate that process, he/she may meet the same barriers that caused the problem in the first place. So if you are part of a larger organization or an ACO there should be one or more individuals who should do the legwork and oversee a thorough investigation of the root causes.
Donzé J, Aujesky D, Williams D, Schnipper JL. Potentially avoidable 30-day hospital readmissions in medical patients: derivation and validation of a prediction model. JAMA Intern Med 2013; 173(8): 632-638
Donzé JD, Williams MV, Robinson EJ, et al. International Validity of the HOSPITAL Score to Predict 30-Day Potentially Avoidable Hospital Readmissions. JAMA Intern Med 2016; 176(4): 496-502
Siracuse BL, Chamberlain RS. A Preoperative Scale for Determining Surgical Readmission Risk after Total Hip Replacement. JAMA Surg 2016; Published online March 09, 2016
Havens JM, Olufajo OA, Cooper ZR, et al. Defining Rates and Risk Factors for Readmissions Following Emergency General Surgery. JAMA Surg 2015; :1-7 Published online first November 11, 2015
Auerbach AD, Kripalani S, Vasilevskis EE, et al. Preventability and Causes of Readmissions in a National Cohort of General Medicine Patients. JAMA Intern Med 2016; 176(4): 484-493
Atkins D, Kansagara D. Reducing Readmissions—Destination or Journey? JAMA Intern Med 2016; 176(4): 493-495
Jenq GY, Doyle MM, Belton BM, et al. Quasi-Experimental Evaluation of the Effectiveness of a Large-Scale Readmission Reduction Program. JAMA Intern Med 2016; Published online April 11, 2016
To get "Patient Safety Tip of the Week " emailed to you, click here and enter "subscribe" in the subject field.
If you don't see the search term you expected to see here, its probably because that tip already went to our Tip of the Week Archive. We do a new tip every week. Click here to search the entire site or you can 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. Or it may have moved to our What's New Archive.
Click here to see the consulting services and patient safety solutions that we provide.
Click on the "Contact Us" button at the left to send us your comments on our "What's New in the Patient Safety World" columns.
To get "Patient Safety Tip of the Week "emailed to you, click here and enter "subscribe" in the subject field.
Pharmacologic interventions to prevent or treat delirium have been elusive. Antipsychotic drugs have sometimes been touted to be successful but the evidence has not been very convincing. A meta-analysis of antipsychotic treatment in patients with delirium (Kishi 2015) suggested that second generation antipsychotics have a benefit for the treatment of delirium with regard to efficacy and safety compared with haloperidol but emphasized that further study using larger samples is required. But another recent systematic review and meta-analysis found that current evidence does not support the use of antipsychotics for prevention or treatment of delirium (Neufeld 2016). Those authors found antipsychotic use was not associated with reduction in delirium incidence, change in delirium duration, severity, or hospital or ICU length of stay.
For several years now there has been interest in the use of dexmedetomidine, an α2-adrenoreceptor agonist, as a sedation agent in the ICU because it might be associated with less delirium. In our February 10, 2009 Patient Safety Tip of the Week “Sedation in the ICU: The Dexmedetomidine Study” we discussed the SEDCOM (Safety and Efficacy of Dexmedetomidine Compared With Midazolam) Study, which concluded that dexmedetomidine was as effective as midazolam at keeping patients in the desired sedation range and was associated with a reduced prevalence of delirium and reduced time to extubation (Riker 2009). However, we urged caution in interpreting the conclusions of that study because of several methodological and other concerns outlined in our column. We again discussed dexmedetomidine in our June 16, 2015 Patient Safety Tip of the Week “Updates on Delirium”.
Now another study has addressed the use of dexmedetomidine in intubated ICU patients with delirium (Reade 2016). The Dexmedetomidine to Lessen ICU Agitation (DahLIA) study was a double-blind, placebo-controlled, parallel-group randomized clinical trial in 15 ICU’s in Australia and New Zealand. Subjects were ICU patients who were deemed to be ready for extubation except that they had delirium. Dexmedetomidine increased ventilator-free hours at 7 days compared with placebo (median, 144.8 hours vs 127.5 hours, respectively). Among several secondary outcome measures they also found that dexmedetomidine reduced time to extubation (median, 21.9 hours vs 44.3 hours with placebo), and accelerated resolution of delirium (median, 23.3 hours vs 40.0 hours).
Again, this sounds encouraging, particularly since our pharmacologic armamentarium for managing delirium is so limited. Yet there are again some red flags that urge us to be cautious in recommending widespread use of dexmedetomidine. First of all, this study applies only to a very select group of patients – those who were already well enough to be being considered for extubation except for their delirium. The authors note that they screened 21,500 patients to recruit just the 74 patients randomized in the study! That small sample size (actually only 71 patients after 3 withdrawals for various reasons). Even more importantly, the study was terminated before its planned recruitment of 96 patients. Studies with early termination typically show more exaggerated effect sizes. Early termination was apparently done because the funding source ceased funding beyond the originally defined period. The authors note that the funding source had no role in the design of the study and had no access to study data during the study, and the authors performed sensitivity analyses suggesting the abbreviated sample size was unlikely to alter the primary conclusion. Nevertheless, such occurrences always raise our “hype radar” or “spin radar” (see our February 16, 2010 Patient Safety Tip of the Week “Spin/Hype…Knowing It When You See It”).
So while we are somewhat encouraged by the results of the DahLIA study, we’re not yet ready to jump on the dexmedetomidine bandwagon for more widespread use. Remember, this was a very narrow patient population and it would be premature to extrapolate the results to patients with delirium earlier in their ICU course (i.e. before they were deemed otherwise ready for extubation). The good news, though, is that the dexmedetomidine seemed to be well tolerated in this study and adverse events were rare. We therefore look forward to further studies on the use of dexmedetomidine for either prevention or treatment of delirium.
Some of our prior columns on delirium assessment and management:
Kishi T, Hirota T, Matsunaga S, Iwata N. Antipsychotic medications for the treatment of delirium: a systematic review and meta-analysis of randomised controlled trials. J Neurol Neurosurg Psychiatry 2015; Published online first September 4, 2015
Neufeld KJ, Yue J, Robinson TN, et al. Antipsychotic Medication for Prevention and Treatment of Delirium in Hospitalized Adults: A Systematic Review and Meta-Analysis. Journal of the American Geriatrics Society 2016; published online 23 March 2016
Riker RR, Shehabi Y, Bokesch PM, et al for the SEDCOM (Safety and Efficacy of Dexmedetomidine Compared With Midazolam) Study Group. Dexmedetomidine vs Midazolam for Sedation of Critically Ill Patients. A Randomized Trial. JAMA. 2009; 301(5):489-499. Published online February 2, 2009
Reade MC, Eastwood GM, Bellomo R, et al. Effect of Dexmedetomidine Added to Standard Care on Ventilator-Free Time in Patients With Agitated Delirium: A Randomized Clinical Trial. JAMA 2016; Published online March 15, 2016
When confronted with patients having delirium our first approach is to look for and remove any precipitating or contributing factors. One such factor we tend to forget about is the use of antibiotics. Given the high prevalence of delirium in the ICU and post-operative settings, it would not be surprising to find antibiotic use frequent in such patients.
A recent review of antibiotic-associated encephalopathy (Bhattacharyya 2016) is very timely and identifies 3 unique clinical phenotypes: encephalopathy commonly accompanied by seizures or myoclonus arising within days after antibiotic administration (caused by cephalosporins and penicillin); encephalopathy characterized by psychosis arising within days of antibiotic administration (caused by quinolones, macrolides, and procaine penicillin); and encephalopathy accompanied by cerebellar signs and MRI abnormalities emerging weeks after initiation of antibiotics (caused by metronidazole). Those clinical features of each phenotype can and should lead to recognition of the pathogenetic role being played by the antibiotic and lead to its discontinuation.
The phenotype characterized by myoclonus and/or seizures (Type 1 AAE) is often due to penicillin or cephalosporins and often occurs in the setting of renal insufficiency. It usually appears within days of antibiotic administration. Seizures associated with cephalosporin-associated encephalopathy were frequently nonconvulsive. EEG may show generalized slowing but often shows periodic discharges with triphasic morphology or epileptiform discharges. MRI is normal in these cases. The encephalopathy usually resolves within days of discontinuation of the offending antibiotic.
Type 2 AAE also typically begins within days of antibiotic initiation and is characterized by frequent occurrence of psychosis and resolution within days of discontinuation of the offending antibiotic. Seizures are rare in this type and the EEG is more likely to be normal (or show nonspecific findings). MRI is usually normal. This phenotype may occur with procaine penicillin, sulfonamides, fluoroquinolones, and macrolides.
The third type (Type 3 AAE) occurs with metronidazole begins weeks after initiation and is characterized by frequent occurrence of cerebellar dysfunction. Seizures are rare and EEG usually shows only nonspecific abnormalities but the MRI is typically abnormal, showing a typical pattern of T2 hyperintensities in the dentate nuclei of the cerebellum
with variable involvement of the brainstem, corpus callosum, or other regions.
The authors also note that isoniazid (INH) may cause an encephalopathy that does not fit nicely into one of the 3 above phenotypes. Onset is weeks to months after INH initiation. Psychosis is common but seizures are rare and EEG may just show nonspecific abnormalities.
The Bhattacharyya paper acknowledges the issue of strength of association with antibiotic use in each phenotype and also has a nice discussion on the possible pathophysiologies of each phenotype and the pharmacokinetic and patient-related factors that are important.
Overall this is an important contribution to the clinical management of the patient with delirium and a key reminder to evaluate all aspects of care.
Some of our prior columns on delirium assessment and management:
Bhattacharyya S, Darby RR, Raibagkar P, et al. Antibiotic-associated encephalopathy. Neurology 2016; published online before print February 17, 2016
Our What's New in the Patient Safety World columns for February 2015 “17% Fewer HAC’s: Progress or Propaganda?” and January 2016 “HAC’s Have Declined Since 2010” discussed interim data sets from AHRQ which showed that there was a 17% reduction in hospital-acquired conditions (HAC’s) between 2010 and 2014. Over the 4-year period the biggest reductions in HAC’s percentage-wise were seen for CLABSI’s (-72%), CAUTI’s (-38%), and post-op venous thromboembolism (-43%).
AHRQ’s Chartbook on Patient Safety summarizes patient safety measures, including overall hospital-acquired conditions (HAC’s) and hospital-associated infections (HAI’s), highlighting the trends between 2010 and 2014 (AHRQ 2016). Similarly, CDC has reported substantial reductions in CLABSI’s, CAUTI’s, surgical site infections (SSI’s), hospital-onset C. difficile infections, and hospital-onset MRSA bacteremias over a roughly similar time frame (CDC 2016).
While we’ve had some degree of skepticism in interpretation of the data, overall we’ve felt comfortable that true progress is being made.
One of the many interventions cited as contributing to the apparent improvement in HAI and HAC rates is the financial penalty hospitals pay for poor performance in these rates (primarily for CMS/Medicare patients but also for some other insurers). We’ve always been concerned about how coding changes have obfuscated some quality parameters. For example, we’ve always been concerned about changes in sepsis coding may have artificially lowered mortality rates for both sepsis and pneumonia (see our
March 2016 What's New in the Patient Safety World column “Finally…A More Rationale Definition for Sepsis”). Now a study from the Stanford Graduate School of Business questions whether coding practices have similarly impacted quality reporting for the HAI’s reported to CMS as well (Bastani 2015).
Bastani and colleagues note that CMS does not directly monitor the occurrence of the various HAI’s. Rather it collects administrative (billing) data from hospitals and does chart reviews of a small sampling to assess validity. The Stanford researchers used more sophisticated techniques to assess how rampant “upcoding” might be. In particular, one form of upcoding would be assigning a designation present-on-admission (POA) to an infection when it was, in fact, a hospital-acquired infection. Upcoding would be financially beneficial to hospitals either by increasing reimbursement or avoiding penalties.
They compared rates of HAI’s in states that require strict reporting of HAI’s to those in states that have weaker reporting requirements. Overall, they found hospitals in the more weakly regulated states reported lower rates of HAIs and higher rates of POA infections. They estimate there are more than 10,000 upcoded infections annually, resulting in an added costs of $200 million to CMS.
Bastani and colleagues are careful to not impute a motive to such “upcoding” While such could be intentional in attempt to avoid the CMS penalties, they also note it might reflect lack of clinical knowledge by “coders” or lack of communication between clinicians and coders (talk about being tactful and politically correct!).
They conclude that their findings suggest, contrary to widely-held beliefs, increasing financial penalties alone may not reduce HAI incidence and may even exacerbate the problem. They make several policy recommendations based on their results, including a new measure for targeted HAI auditing and suggestions for effective adverse event reporting systems.
AHRQ (Agency for Healthcare Quality and Research). Chartbook on Patient Safety. March 2016
CDC (Centers for Disease Control and Prevention). National and State Healthcare Associated Infections Progress Report. 2016
Bastani H, Goh J, Bayati M. Evidence of Strategic Behavior in Medicare Claims Reporting. Stanford Graduate School of Business 2015; Working Paper No. 3396; July 13,,2015
Our July 7, 2009 Patient Safety Tip of the Week “Nudge: Small Changes, Big Impacts” reviewed the book “Nudge” by Richard Thaler and Cass Sunstein. Yes, that’s the one that leads in with the story about how painting a picture of a fly in a male urinal resulted in 80% decreased spillage! The theme obviously is that small changes which cost little or nothing (i.e. nudges) can result in big impacts. The book is full of examples of how nudges can help steer people to make better choices in their personal life (savings, investments, healthcare, etc.) or from a societal perspective (improve the environment, improve organ donations, etc.).
In that 2009 column and in our February 18, 2014 Patient Safety Tip of the Week “Nudged, But Who Nudged Who?” we gave examples of how such small changes or “nudges” may lead to desirable changes in behavior in healthcare.
Hand hygiene is one area in which nudges may be helpful and that applies not only to healthcare personnel but also to visitors. A new study looked at factors related to use of alcohol-based hand sanitizers by visitors to a hospital (Hobbs 2016). The key finding was that when the hand sanitizers were placed in the middle of the lobby (with limited landmarks or barriers) visitors were 5.28 times more likely to use them. But the other key finding was that group behavior is important as well. In the Hobbs study individuals in a group were 39% more likely to use alcohol-based hand sanitizers. We’ve often viewed the same scenario with healthcare workers. A team in a teaching hospital (attending, several residents and students, and maybe a nurse or two) is doing rounds. If the attending stops to do hand hygiene before interacting with the patient, the whole team does hand hygiene. If he/she does not do hand hygiene, no one does. That’s a “nudge” that has a powerful impact.
So that addresses healthcare workers and visitors. What about patients themselves? After publication of a study last month (Cao 2016) we may need a “nudge” for them, too. Cao and colleagues did cultures of the hands of patients being admitted to post-acute care facilities from acute care hospitals. They found that 24.1% had at least one multidrug-resistant organism (MDRO) on their hands. Of course, other body parts may be colonized with hospital-acquired organisms but the patients’ hands are most likely to have been in contact with environmental surfaces, health care workers’ hands, or even other patients. Clearly, further studies need to be done to see how to intervene and prevent spread of such organisms in patients being discharged. Adding hand hygiene to patients being admitted to long-term care facilities would make sense. But adding hand hygiene to the discharge checklist of patients being discharged from acute care hospitals may make more sense since even those going home may be spreading MDRO’s. So a little “nudge” may be needed at discharge. Maybe putting another alcohol-based hand sanitizer in the lobby facing the other way will get both patients and visitors to perform hand hygiene on the way out, too!
“Nudges” do have positive impacts and we need to learn how to better deploy them.
Some of our other columns on hand hygiene:
January 5, 2010 “How’s Your Hand Hygiene?”
December 28, 2010 “HAI’s: Looking In All The Wrong Places”
May 24, 2011 “Hand Hygiene Resources”
October 2011 “Another Unintended Consequence of Hand Hygiene Device?”
March 2012 “Smile…You’re on Candid Camera”
August 2012 “Anesthesiology and Surgical Infections”
October 2013 “HAI’s: Costs, WHO Hand Hygiene, etc.”
November 18, 2014 “Handwashing Fades at End of Shift, ?Smartwatch to the Rescue”
January 20, 2015 “He Didn’t Wash His Hands After What!”
September 2015 “APIC’s New Guide to Hand Hygiene Programs”
November 2015 “Hand Hygiene: Paradoxical Solution?”
Thaler RH, Sunstein CR. Nudge. Improving Decisions about Health, Wealth, and Happiness. New Haven: Yale University Press, 2008
Hobbs MA, Robinson S, Neyens DM, Steed C. Visitor characteristics and alcohol-based hand sanitizer dispenser locations at the hospital entrance: Effect on visitor use rates.
Am J Infection Contol 2016; 44(3): 258-262
Cao J, Min L, Lansing B, Foxman B, Mody L. Multidrug-Resistant Organisms on Patients’ Hands. A Missed Opportunity. JAMA Intern Med 2016; Published online March 14, 2016
Go to the "Whats New Archive"
To get "What's New in the Patient Safety World"emailed to you,click here and enter "subscribe" in the subject field.
To get "What's New in the Patient Safety World"emailed to you,click here and enter "subscribe" in the subject field.