a b s t r a c t

automated external defibrillators (AEDs) emerged in the 1980s as an important innovation in pre-hospital emer- gency cardiac care (ECC). In the years since, the American Heart Association (AHA) and the International Liaison Committee for Resuscitation (ILCOR) have promoted AED technology for use in hospitals as well, resulting in the widespread purchase and use of AED-capable defibrillators. In-hospital use of AEDs now appears to have de- creased survival from cardiac arrests. This article will look at the use of AEDs in hospitals as a case of “medical re- versal.” Medical reversal occurs when an accepted, widely used treatment is found to be ineffective or even harmful. This article will discuss the issue of AEDs in the hospital using a Conceptual framework provided by re- cent work on medical reversal. It will go on to consider the implications of the reversal for in-hospital resuscita- tion programs and emergency medicine more generally.

(C) 2017 The Author. This is an open access article under the CC BY-NC-ND license

(http://creativecommons.org/licenses/by-nc-nd/4.0/).

Background

AEDs were originally developed to speed defibrillation in cases of out- of-hospital cardiac arrest due to ventricular fibrillation (VF) or pulseless Ventricular tachycardia by enabling more rescuers, including lay- people, to defibrillate. With the support of the AHA and ILCOR, emergen- cy medical services systems began adopting AEDs to enable more providers to defibrillate, with ease of training as the main rationale [1].

The problem of delayed defibrillation in hospitals was brought to the attention of several AHA leaders in the mid-1980s. They in turn began a concerted push for AEDs in hospitals, based on the recognition that de- fibrillation was often delayed-and the assertion that AEDs could miti- gate the problem by making it possible to train more caregivers, particularly bedside nurses, to defibrillate. AHA/ILCOR formally en- dorsed the idea in the 2000 ECC Guidelines, in a chapter titled “The Au- tomated External Defibrillator: Key Link in the Chain of survival [2].”

Several small studies in the next few years suggested some improve-

ment in hospital survival resulting from the addition of AEDs and re- placement of manual defibrillators with dual-mode (AED/manual) devices. But doubts about the technology began to surface: Studies ac- cumulated showing decreased shock efficacy with AEDs due to the hands-off time required for Rhythm analysis [3].

? JAS reports no conflicts of interest.

E-mail address: [email protected] (J.A. Stewart).

In 2010 a large cohort study (hereafter referred to as the “Chan study”) directly compared in-hospital use of AEDs with manual- defibrillator controls, analyzing resuscitation data from the AHA’s Na- tional Registry for Cardiopulmonary Resuscitation (NRCPR, now Get With the Guidelines-Resuscitation (GWTG-R)). Analyzing 11,695 ar- rests in 204 hospitals, the study found that AED use had no significant effect on survival from Shockable rhythms and decreased survival from non-tachyarrhythmic arrests [4]. Two single-center studies cor- roborated this finding [5,6].

The AHA’s current position appears to be that the recommendation

for in-hospital use of AEDs will stand, with a call for randomized con- trolled trials (RCTs) to evaluate and optimize use of AEDs in the hospital [7]. The most recent Emergency Cardiac Care (ECC) Guidelines (2015) do not address the issue.

Examining this history using the conceptual model of medical rever- sal may aid understanding of how AEDs in the hospital came to be ac- cepted as “standard practice,” the prospects for change in light of the reversal, and some wider implications for emergency medical research.

Main text

As defined by Vinayak Prasad MD and Adam Cifu MD, “[M]edical re- versal occurs when a new clinical trial — superior to predecessors by vir- tue of better controls, design, size, or endpoints — contradicts current clinical practice [8].” Their recent book showed that medical reversal occurs with troubling frequency: A review of 10 years of articles in the

https://doi.org/10.1016/j.ajem.2017.11.035

0735-6757/(C) 2017 The Author. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

872 J.A. Stewart / American Journal of Emergency Medicine 36 (2018) 871-874

New England Journal of Medicine (2001-2010), found a 40% incidence of medical reversal (146 of 363 articles) [9]. The authors discuss at length and build upon a conceptual framework originated by the medical sociol- ogist John McKinlay in his article “From “promising report” to “standard procedure”: Seven stages in the career of a medical innovation [10].”

McKinlay’s seven stages are:

• Promising report
• Professional and organizational adoption
• Public acceptance and state (third party) endorsement
• “Standard procedure” and observational reports
• Randomized controlled trial (RCT)
• Professional denunciation
• Erosion and discreditation.
  

  Promising report

  As described by McKinlay, treatments that are later reversed are typ- ically adopted following positive reports, often in the lay press, of results from individual cases or small patient samples. In contrast, the push for AEDs in hospitals began without even this limited evidence. AHA training materials began promoting the idea in the 1990s, based on an extrapolation from out-of-hospital practice. Evidence was then begin- ning to show some Survival benefit from deploying AEDs in out-of- hospital settings where Early defibrillation had not been possible before. In hospitals, where manual defibrillators were already readily available, the argument turned on the presumed ease of training and skill reten- tion for nurses who rarely dealt with cardiac arrest-and on the addi- tional assumption that it was not feasible to train those nurses to use manual defibrillators [11]. The present case is therefore somewhat atyp- ical in that the innovation began with no direct evidence, but instead be- cause of advocacy from the AHA, a public-service organization and quasi-regulatory body that develops the ECC Guidelines.

  Professional and organizational adoption

  The AHA/ILCOR formally endorsed use of AEDs in hospitals in the 2000 ECC Guidelines [2]. That endorsement followed the recommendations of a task force on AEDs, which came despite their acknowledgment that no di- rect evidence existed supporting use of AEDs in hospitals [12], and also in spite of studies beginning to show decreased shock efficacy with AEDs [13]. The recommendation was followed by the adoption of AED technol- ogy by most American hospitals over the next decade and a half.

  Public acceptance and state (third party) endorsement

  Of McKinlay’s seven stages, this is the least relevant to the current case. AED defibrillation was seen asa variation on an accepted emergency treat- ment (external defibrillation), so there were no third-party endorsement (i.e., insurance coverage) issues. Nor was public acceptance a particular concern, since the change affected hospital emergency response systems largely out of the public eye and did not involve an elective procedure.

  “Standard procedure” and observational reports

  The years since the formal endorsement in the guidelines have seen AED technology widely adopted by hospitals, typically by replacing man- ual defibrillators with dual-mode (manual and AED) devices, to the ex- tent that it is now considered standard practice. Observational reports provided support for the change. McKinlay says that at this stage “obser- vational studies usually suffer crippling methodological limitations,” and that is true of the studies that appeared to support in-hospital use of AEDs. Two single-center observational studies reported good survival from VF/VT arrests with AEDs [14,15], but both lacked a control group. One study with a historical control did show an improvement in survival to discharge from cardiac arrests in which the patient was defibrillated.

  After an intensive hospital-wide education program and placement of AEDs and dual-mode defibrillators throughout the hospital, survival to discharge for shockable rhythms improved to 15.6% (10 of 64 patients) compared to 2.2% (4 of 176 patients) before the program [16].

  Beyond the usual problems with single-center studies using historical controls [17] (or no controls, in the case of two of the studies), these stud- ies share a major weakness. The introduction of AEDs to the hospital is a systems intervention, “a change in how health care is delivered, one that af- fects how groups of caregivers act or behave in their professional role.” Systems interventions are most often evaluated by “before-and-after” studies. Such studies are highly susceptible to the “Hawthorne effect,” a well-accepted phenomenon affecting subjects who know their perfor- mance is being studied-they tend to perform better on the tasks under study regardless of the intervention [18]. The authors of the latter study, to their credit, acknowledged the possibility that “the improvement in survival was the result of a Hawthorne effect where increased education and awareness rather than the actual experimental conditions were re- sponsible for the outcome [16].”

  Randomized controlled trial (RCT)

  McKinlay regards a properly designed and conducted RCT as the best evidence and firmest basis for medical reversal, and Prasad and Cifu concur [19]. Even Chan et al. made a perfunctory call for RCTs [4]. But RCTs are hard to do in general and are especially difficult, and therefore rare, in the realm of resuscitation medicine [20]. Recruitment for studies that test an accepted standard of care is particularly challenging because of possible ethical objections [21,22].

  Beyond these concerns, other considerations would appear to make an RCT on in-hospital use of AEDs practically impossible. Randomizing be- tween manual and automated defibrillation at the event level is clearly unworkable: blinding would not be possible, and unless First responders were trained in manual as well as AED defibrillation, randomization would require delay in delivering an accepted lifesaving treatment to the manual group-something no institutional review board would likely approve. The only option would seem to be a cluster RCT [23], which would require participation by hospitals not using AEDs-but the wide- spread adoption of AEDs by hospitals in the wake of the AHA’s endorse- ment makes enrollment of non-AED hospitals virtually impossible. Even if non-AED hospitals could be found and enrolled, they would necessarily be extreme outliers with respect to resuscitation practices.

  It is difficult, then, to regard calls for RCTs as anything more than pro forma, given the enormous problems that would be involved in conducting an RCT of reasonable size comparing AEDs to manual defibrillators-much less one equivalent in size and reach to the Chan study. If RCTs are not a realistic option, we are left with the present evidence for the foreseeable future. That evidence shows that use of AEDs decreases overall survival from cardiac arrest in the hospital setting.

  Professional denunciation

  In the emotionally charged atmosphere of medical care, the momen- tum of a new technology too often puts the burden of proof on those who question the evidence for it, rather than on those who propose it. The result is that the technology quickly becomes the accepted thing to do [and] further attempts to test it and subject to the charge of being unethical… [22].

  And per McKinlay:

  The issue of double standards is perhaps most evident during this sixth stage involving professional denunciation. The many defective observational studies conducted up to this point seldom receive adequate methodological and statistical scrutiny, whereas RCTs are subjected to the most stringent criticism, employing standards that

  J.A. Stewart / American Journal of Emergency Medicine 36 (2018) 871-874 873

  are almost never invoked during the earlier stages…The nature of the standards invoked, and the force with which they are applied, seem to vary depending on whether or not the results are supportive of what has become standard procedure.

  If we substitute “RCTs” in the above quote with “a large multicenter registry-based controlled cohort study,” the AHA/ILCOR’s only official response to the Chan study provides a good example of McKinlay’s ob- servations. That response came in a 2013 Scientific Statement on in- hospital resuscitation, in a section on AEDs [7]. Within that section, 62% of the text is devoted to making the argument for AEDs in the hos- pital and summarizing the positive observational reports discussed pre- viously (with the addition of one that did not involve inpatients [24]), while failing to mention two observational studies that corroborated the Chan study [5,6]. The remainder dealt with the Chan study, but most of it (82 0f 143 words, or 57%) was used to cast doubt on the neg- ative results, including the claim that “it is likely that AEDs were placed in areas less well served by the cardiac arrest team, representing a po- tential selection bias”-despite the study’s finding that “relationships between AED use and survival were consistent in monitored and nonmonitored hospital units for each rhythm type.” Only 61 words (15% of the section total) were devoted to a straightforward account of the study. The authors wrote that one of the single-center observa- tional studies [16] “showed” an improvement in outcomes, but that the Chan study of 11,695 arrests from 204 hospitals only “suggested” that there was no association with increased survival. The discussion concluded that: “additional randomized clinical trials are required to evaluate and optimize use of AEDs in the hospital,” with the question apparently becoming how best-rather than whether-to use AED tech- nology in hospitals.

  Erosion and discreditation

  McKinlay says of this stage, “After some period of time (often more than a decade), and ever so gradually, an erosion of support begins to set in.” The 2015 Guidelines have dropped the 2010 Guidelines’ discussion of AED use in hospitals without comment and do not other- wise address the issue. Elsewhere, in describing the “important insights” provided by the GWTG-R registry, the Guidelines omit the 2010 Chan study while citing three studies that are arguably less clinically significant [25]. The 2015 European Resuscitation Council (ERC) Resuscitation guidelines acknowledge, mainly on the basis of the Chan study, that AED use “may cause harmful delays in starting CPR, or interruptions in chest compressions in patients with Non-shockable rhythms,” and recom- mends manual defibrillation when available-but continues to recom- mend AED defibrillation in areas where “first responders do not have skills in manual defibrillation [26].” That stance continues to reinforce the assumption undergirding the rationale for AEDs in the hospital: that nurses outside of specialty areas cannot learn to recognize VF without continuing training in rhythm recognition-an assumption that is unprov- en and may be unfounded [27].

  Prasad and Cifu say that too frequently “[t]he contradicted practice does not stop immediately…[but] continues for years to come” [28]. ECC Guidelines may remain in place for decades despite thin or nonex- istent evidence for adoption and the subsequent publication of negative studies [20,29]. Systems interventions may be especially hard to reverse due to entrenched financial and/or intellectual interests [18]. McKinlay says, “The power of such interests is also evident in their ability to im- pede the development of alternative practices (for which there may also be considerable observational support) that could conceivably threaten an activity in which there is already considerable investment.” An additional factor in the present case may the ubiquity of AED- capable defibrillators in hospitals today: if the AHA/ILCOR were to re- tract their endorsement, those devices would serve for many years as concrete reminders of the mistake.

  Conclusions

  Financial conflicts of interest (COIs) are a major problem in the de- velopment of clinical guidelines [30], and it can be assumed that indus- try influence was a factor in the present case: the recommendation for in-hospital AEDs represented a windfall for the defibrillator industry, with hospitals adding AEDs and/or replacing their manual defibrillators with dual-mode devices [31]. COI protections are clearly important in formulating the ECC Guidelines, and since 2000, the AHA has strength- ened its COI policies [32].

  Large quasi-regulatory bodies/public service organizations such as the AHA, together with government agencies, may have the resources and influence to support and encourage large RCTs of new therapies before they are widely adopted. An example is the study of public- access defibrillation with AEDs supported by the AHA and the National Heart, Lung, and Blood Institute [33]. A recent article makes a similar recommendation for the study of extracorporeal cardiopulmonary resuscitation [34].

  Recognition of the problem of delayed in-hospital defibrillation co- incided with the rising enthusiasm for AED technology, and promotion by the AHA ensured that the problem was closely linked with that pro- posed solution. That history has perhaps led to the assumption that if AEDs do not provide a remedy, nothing can. Additionally, because the problem has never been quantified and published time-interval data are artificially short [35], some have even suggested that delayed defi- brillation in hospitals may not be a serious problem [9]. A sense of res- ignation (“Nothing can be done about it”) or complacency (“Perhaps the problem was exaggerated”) may color attitudes for years.

  One of the harmful effects of this case of medical reversal is of course the millions of dollars hospitals have spent to adopt AED technology. More concerning is the probability that it has led to many deaths of pa- tients presenting with a non-shockable rhythm. A still more serious harm is that it has prevented and continues to obstruct consideration of other approaches to delivering rapid defibrillation, thereby forfeiting the chance to rescue many “hearts too good to die.”

  References

  1. Joyce SM, Davidson LW, Manning KW, Wolsey B, Topham R. Outcomes of sudden cardiac arrest treated with defibrillation by emergency medical technicians (EMT-Ds) or paramedics in a two-tiered urban EMS system. Prehosp Emerg Care 1998 Jan-Mar;2(1):13-7.
  2. The American Heart Association in Collaboration with the International Liaison Committee on Resuscitation. Guidelines 2000 for cardiopulmonary resuscitation and emergency cardiovascular care, part 4: the automated external defibrillator: key link in the chain of survival. Circulation 2000;102(8 Suppl):I-60-76.
  3. Nolan JP, Soar J. Defibrillation in clinical practice. Curr Opin Crit Care 2009 Jun;15(3):

     209-15. https://doi.org/10.1097/MCC.0b013e32832931cb.

     Chan PS, Krumholz HM, Spertus JA, et al, American Heart Association National Reg- istry of Cardiopulmonary Resuscitation (NRCPR) Investigators. Automated external defibrillators and survival after in-hospital cardiac arrest. JAMA 2010;304(19): 2129-36. https://doi.org/10.1001/jama.2010.1576.

  4. Forcina MS, Farhat AY, O’Neil WW, Haines DE. Cardiac arrest survival after imple- mentation of automated external defibrillator technology in the in-hospital setting. Crit Care Med 2009 Apr;37(4):1229-36.
  5. Smith RJ, Hickey BB, Santamaria JD. Automated external defibrillators and in- hospital cardiac arrest: patient survival and device performance at an Australian teaching hospital. Resuscitation 2011 Dec;82(12):1537-42.
  6. Morrison LJ, Neumar RW, Zimmerman JL, et al, on behalf of the American Heart As- sociation Emergency Cardiovascular Care Committee. Strategies for improving sur- vival after in-hospital cardiac arrest in the United States: 2013 consensus recommendations. A consensus statement from the American Heart Association. Cir- culation 2013;127:1538-63.
  7. Prasad VK, Cifu AS. Medical reversal: why we must raise the bar before adopting

     new technologies. Yale J Biol Med 2011 Dec;84(4):471-8.

     Prasad VK, Cifu AS. Ending medical reversal: improving outcomes, saving lives. Johns Hopkins University Press; 2015.

  8. McKinlay JB. From “promising report” to “standard procedure”: seven stages in the career of a medical innovation. Milbank Mem Fund Q 1981;59:374-411.
  9. Textbook of advanced cardiac life support. Dallas: American Heart Association; 1994.
  10. Atkins DL, Bossaert LL, Hazinski MF, et al. Automated external defibrillation/public access defibrillation. Ann Emerg Med 2001;37(4):S6-S67.
  11. Sato Y, Weil MH, Sun S, et al. Adverse effects of interrupting precordial compression during cardiopulmonary resuscitation. Crit Care Med 1997 May 25;5:733-6.

      874 J.A. Stewart / American Journal of Emergency Medicine 36 (2018) 871-874

      Hanefeld C, Lichte C, Mentges-Schroter I, Sirtl C, Mugge A. Hospital-wide first- responder automated external defibrillator programme: 1 year experience. Resusci- tation 2005;66:167-70.

  12. Gombotz H, Weh B, Mitterndorfer W, Rehak P. In-hospital Cardiac resuscitation out- side the ICU by nursing staff equipped with automated external defibrillators: the first 500 cases. Resuscitation 2006;70:416-22.
  13. Zafari AM, Zarter SK, Heggen V, et al. A program encouraging early defibrillation re- sults in improved in-hospital resuscitation efficacy. J Am Coll Cardiol 2004;44: 846-52. https://doi.org/10.1016/j.jacc.2004.04.054.
  14. Olson CM, Fontanarosa PB. Advancing cardiac resuscitation: lessons from eternally controlled trials. JAMA 1999;28(13):1220-2.
  15. Prasad VK, Cifu AS. Chapter 5: systems failure. Ending medical reversal: improving

      outcomes, saving lives. Johns Hopkins University Press; 2015.

      Prasad VK, Cifu AS. Chapter 9: a primer on evidence-based medicine. Ending medical reversal: improving outcomes, saving lives. Johns Hopkins University Press; 2015.

  16. Kern KB, Valenzuela TD, Clark LL, et al. An alternative approach to advancing resus- citation science. Resuscitation 2005 Mar;64(3):261-8.
  17. Jacobs IG, Finn JC, Jelinek GA, Oxer HF, Thompson PL. Effect of adrenaline on survival in out-of-hospital cardiac arrest: a randomised double-blind placebo-controlled trial. Resuscitation 2011;82:1138-43.
  18. Russell LB. Technology in hospitals: medical advances and their diffusion. Washington, DC: Brookings Institute; 1979.
  19. Moberg J, Kramer M. A brief history of the cluster randomized trial design. JLL bulletin: commentaries on the history of treatment evaluation. and available at http://www.jameslindlibrary.org/articles/a-brief-history-of-the-cluster- randomized-trial-design/; 2015, Accessed date: 31 January 2017.
  20. Friedman FD, Dowler K, Link MS. A Public access defibrillation programme in non- inpatient hospital areas. Resuscitation 2006;69:407-11.
  21. Kleinman ME, Brennan EE, Goldberger ZD, et al. 2015 American Heart Association guidelines update for cardiopulmonary resuscitation and emergency cardiovascular care. Part 5: adult basic life support and cardiopulmonary Resuscitation quality. Circulation 2015;132:S414-35. https://doi.org/10.1161/CIR.0000000000000259.
  22. Perkins GD, Handley AJ, Koster RW, et al. European resuscitation council guidelines for resuscitation 2015 section 2. Adult basic life support and automated external de- fibrillation. Resuscitation 2015;95:81-99. https://doi.org/10.1016/j.resuscitation. 2015.07.015.
  23. Stewart AJ, Lowe MD. Knowledge and attitudes of nurses on medical wards to defi- brillation. J R Coll Physicians Lond 1994;28:399-401.
  24. Prasad VK, Cifu AS. Chapter 8: the harms of medical reversal. Ending medical rever- sal: improving outcomes, saving lives. Johns Hopkins University Press; 2015.
  25. Morales-Cane I, Valverde-Leon MD, Rodriguez-Borrego MA. Epinephrine in cardiac arrest: systematic review and meta-analysis. Rev Lat Am Enfermagem 2016; 24:e2821. https://doi.org/10.1590/1518-8345.1317.2821.
  26. Cosgrove L, Shaughnessy AF. Conflicts of interest in clinical practice guidelines. Counseling and school psychology faculty publication series. Paper 4. and available at http://scholarworks.umb.edu/counseling_faculty_pubs/4; 2011, Accessed date: 31 January 2017.
  27. Fowler L. Bad shock: automated devices for jolting hearts may save fewer lives in hospitals. Fair warning. and available at http://www.fairwarning.org/2011/11/bad- shock-automated-devices-for-failing-hearts-may-save-fewer-lives-in-hospitals/; 2011 Nov., Accessed date: 31 January 2017
  28. Billi JE, Eigel B, Montgomery WH, Nadkarni VM, Hazinski MF. Management of con- flict of interest issues in the activities of the American Heart Association Emergency Cardiovascular Care Committee, 2000-2005. Circulation 2005;112:IV-204-5. https://doi.org/10.1161/CIRCULATIONAHA.105.170810.
  29. Hallstrom AP, Ornato JP, Weisfeldt M, et al. Public access defibrillation trial investi- gators. Public-access defibrillation and survival after out-of-hospital cardiac arrest. N Engl J Med 2004 Aug 12;351(7):637-46.
  30. Singal R, Singal D, Bednarczyk J, et al. Current and future status of extracorporeal car- diopulmonary resuscitation for in-hospital cardiac arrest. Can J Cardiol 2017 Jan; 33(1):51-60.
  31. Kaye W, Mancini ME, Truitt TL. When minutes count-the fallacy of accurate time documentation during in-hospital resuscitation. Resuscitation 2005;65(3):285-90.