Leonardo.lth.se

Balancing “No Blame” with Accountability in Patient Safety
To the Editor: Wachter and Pronovost (Oct. 1 try made a very big point of maintaining blood
issue)1 question the “no-blame” paradigm in pa- glucose target ranges of 80 to 110 mg per decili- tient-safety improvement and suggest the adoption ter (4.4 to 6.1 mmol per liter) in patients in crit- of explicit punitive approaches to poorly perform- ical care units. Many clinicians who opposed this ing physicians. We counsel caution. In a longitu- recommendation were told that they were simply dinal study over a 2-year period in a large facility,2 not keeping up with evidence-based medicine. The we found that penalties did not deter undesirable Normogylcemia in Intensive Care Evaluation–Sur- behavior. Rather, penalties drove evidence of non- vival Using Glucose Algorithm Regulation study compliance underground, encouraging people to (ClinicalTrials.gov number, NCT00220987) proved conceal it and thus perversely reducing account- quite definitively that this goal not only did not help, but actually caused excess deaths as com- Drawing the line between blameworthy and pared with looser glucose control. The flip-flop blameless acts was difficult and involved subjec- by the Centers for Medicare and Medicaid Services tive judgments of observers about the foresee- on beta-blocker use immediately after myocardi- ability of harm, reasonable care, and prudence. al infarction is another example in which what The question was: Who was permitted to draw was advertised as beneficial was actually harm- that line?3 And who reported “violations”? In the ful. A very clear definition of what is truly a pa- example of hand hygiene described in the article, tient-safety practice, scientific criteria, and cer- those difficulties are compounded by uncertain- tainty of the evidence are needed to mandate a ties in the evidence base about when and how clinical practice. If not, we will continue to vio- late the ancient creed of “do no harm” in mis- In our study, peer intervention was more ef- guided safety efforts.
fective in generating accountability and desired Harry Peled, M.D.
change than punitive administrative action; less St. Jude Medical Center blame led to more accountability. Our research Fullerton, CA clearly suggests that by demanding penalties, we [email protected] might stifle accountability rather than enhance it.
No potential conflict of interest relevant to this letter was re- Sidney W.A. Dekker, Ph.D.
Lund University Lund, Sweden The authors reply: We agree that we need to
proceed cautiously and err on the side of parsimony in choosing practices that are suitable for an ac- countability approach. Candidate practices should be relatively easy to follow, have a strong and en- No potential conflict of interest relevant to this letter was re- during evidence base, and be ones in which other approaches have been tried and not succeeded. 1. Wachter RM, Pronovost PJ. Balancing “no blame” with ac-
countability in patient safety. N Engl J Med 2009;361:1401-6.
We believe that infection-control practices such 2. Dekker S, Laursen T. From punitive action to confidential as hand hygiene easily meet these criteria and are
reporting: a longitudinal study of organizational learning from ideal practices to start with. However, at this point, incidents. Patient Saf Qual Healthcare 2007;5:50-6. clinical care standards such as tight glucose con- 3. Dekker S. Just culture: balancing safety and accountability.
Farnham, United Kingdom: Ashgate, 2008.
trol and the use of certain medications at appro- 4. Pittet D, Allegranzi B, Boyce J, et al. The World Health Orga-
priate times seem more suitable for approaches nization Guidelines on Hand Hygiene in Health Care and their consensus recommendations. Infect Control Hosp Epidemiol that use education, traditional audit and feedback, Our auditing methods will also need to mature. To the Editor: With regard to the criteria listed Although some auditing can and should be done
in Table 1 of the article by Wachter and Pronovost, by colleagues (promoted by team training and a who gets to decide what is a critical “patient- shared ethic of patient protection), it is human safety practice”? The Institute for Healthcare Im- nature for colleagues to avoid “ratting out” each provement and others in the patient-safety indus- other, particularly when there are penalties at n engl j med 362;3 nejm.org january 21, 2010 Downloaded from www.nejm.org at LUNDS UNIVERSITET on February 1, 2010 . Copyright 2010 Massachusetts Medical Society. All rights reserved. T h e n e w e ng l a n d j o u r na l o f m e dic i n e hand. The solution is not to abandon account- tices such as hand hygiene. Without minimizing ability, but rather to develop stronger auditing the challenges we face and fully recognizing the strategies with the use of methods such as video need to proceed slowly, it borders on magical surveillance, computerized triggers, and unan- thinking to believe that a strategy of “more of nounced, secret monitoring of compliance by hos- the same” will achieve the levels of safety and pital personnel. Clearly, we have much to learn reliability that our patients deserve.
here, and we agree that we must be careful to Robert M. Wachter, M.D.
preserve the collegial exchange and openness that University of California at San Francisco are so essential to organizational learning.
We recognize that finding the balance between Peter J. Pronovost, M.D., Ph.D.
accountability and “no blame” will be difficult. Johns Hopkins University But, a decade into the safety movement, we now Baltimore, MD know that our present strategy guarantees lack- Since publication of their article, the authors report no fur- luster adherence to a number of low-risk, univer- ther potential conflict of interest.
sally accepted, and evidence-based safety prac- Lovastatin in X-Linked Adrenoleukodystrophy
To the Editor: As reported previously in the 0.39 μmol per liter, respectively. However, even
Journal, lovastatin lowers levels of very-long-chain with this decrease, C26:0 levels remained above fatty acids in plasma in patients with X-linked the control level (mean ±SE) of 0.67±0.13 μmol adrenoleukodystrophy (X-ALD).1 Further studies per liter.4 Furthermore, the reduction in C26:0 did not reproduce this finding with the use of was no longer significant at 22 weeks. There was simvastatin in patients or with the use of lova- a decrease of 0.38 mmol per liter in the level of statin in X-ALD–knockout mice.2,3 Still, many pa- oleic acid (C18:1) at 8 weeks and a decrease of tients with X-ALD worldwide receive lovastatin.
0.44 mmol per liter at 22 weeks. There was no We conducted a randomized, double-blind, change in levels of C26:0 in erythrocytes or lym- placebo-controlled, crossover trial comparing phocytes at either measurement. Finally, the lovastatin at a dose of 40 mg once daily with levels of C18:1, C24:0, and C26:0 in LDL lipopro- placebo (Current Controlled Trials number, tein particles remained unchanged.
ISRCTN31565393). Outcome measures were lev- This trial was designed to investigate whether els of hexacosanoic acid (C26:0) in plasma, low- lovastatin has a biochemical effect in vivo in density lipoprotein (LDL) particles, lymphocytes patients with X-ALD and to provide pilot data for and erythrocytes, and plasma LDL cholesterol a possible large-scale trial with clinical outcome after 22 weeks of treatment. For some outcome variables. We conclude that lovastatin leads to a measures, an intermediary analysis at 8 weeks small decrease in levels of C24:0 and C26:0 in was performed. A total of 14 men with X-ALD plasma; this must be considered a nonspecific (adrenomyeloneuropathy phenotype) were enrolled result of the decrease in the level of LDL choles- in the study. Merck provided lovastatin for this terol. Since very-long-chain fatty acids are virtu- study but did not participate in the trial design, ally water insoluble, and only a small fraction data analysis, or drafting of the letter.
binds to albumin,5 most of the very-long-chain No patients dropped out of the study, and fatty acids in plasma are transported as choles- neither myopathy nor rhabdomyolysis or other terol esters in lipoprotein particles such as LDL. adverse events were observed. Data on all major This finding is corroborated by the finding that outcomes are summarized in Table 1. There were the level of C18:1 was also reduced, and it is significant decreases of 1.44 mmol per liter in further supported by the lack of an effect on the level of LDL cholesterol in plasma at 8 weeks C26:0 levels in peripheral-blood lymphocytes and and 1.35 mmol per liter at 22 weeks. At 8 weeks, erythrocytes and in the content of very-long- the levels of plasma tetracosanoic acid (C24:0) and chain fatty acids in the LDL lipoprotein fraction. C26:0 had decreased by 14.2 μmol per liter and Our data indicate that investment of substantial n engl j med 362;3 nejm.org january 21, 2010 Downloaded from www.nejm.org at LUNDS UNIVERSITET on February 1, 2010 . Copyright 2010 Massachusetts Medical Society. All rights reserved.

Source: http://www.leonardo.lth.se/fileadmin/lusa/Sidney_Dekker/articles/2010/NEJM.pdf