A recent study found that a serious complication occurs in approximately 1 in every 3000 obstetric anesthetics.1 Drug errors in anesthetic practice are common, occurring in an estimated 1 in every 133 anesthetics.2
We performed a literature review of published case reports of drug administration errors into the neuraxis occurring during obstetric anesthetic practice with the aim of estimating the frequency and nature of these errors. We also sought to identify the mechanisms and human factors involved in these errors and identify the potentially useful strategies for minimizing the incidence of such errors in the future.
We searched MEDLINE (1950 to May 2014) using the following terms “epidural analgesia” or “epidural anesthesia” or “obstetrical anesthesia” or “obstetrical analgesia” or “spinal anesthesia.” All records were retrieved by a [TiAB] search. Articles containing the terms “medication errors” or “medical errors” or “drug labeling” or “drug labelling” were then retrieved using the same method. Articles containing any of the first search terms and any of the second search terms were included for analysis. Case reports or series involving the perinatal period were included. We excluded nonobstetric case reports, IV drug errors, errors where the wrong dose of local anesthetic or opioid was administered to the neuraxis, and equipment-related drug errors. Reference lists were independently searched to identify further reports. One author also contacted the US Food and Drug Administration, Anesthesia Patient Safety Foundation, and the Institute for Safe Medication Practices to identify other reports.
The clinical circumstances of each error were recorded on a data collection form. To analyze the human factors, we used a technique used in the aviation industry and classified errors according to the Human Factors Analytics and Classification System (HFACS)3 (Supplemental Digital Content 1, http://links.lww.com/AA/B191), which is based on James Reason’s “Swiss Cheese” model of human error. This classification describes 4 levels of failure. Failings or holes occur at each level: organizational influence, unsafe supervision, preconditions for unsafe acts, and unsafe acts. Holes lining up in a coincidence of events lead to the latent failures that facilitate an active failure. The authors independently studied each case report and classified any factors believed to be involved according to the HFACS. Consensus was used to resolve differences in opinion.
One author summarized each drug error into a 1-line description using inductive reasoning. We surveyed the obstetric anesthetic department heads in the North West of England and asked whether, in their opinion, a practice recommendation applied in isolation at the time of the error category would have likely prevented it. Strongly recommended and recommended procedures relating to avoidance of drug administration error by IV bolus in anesthesia practice proposed by Jensen et al.4 were used as the proposed practice recommendations together with the National Patient Safety Agency (NPSA) recommendation5 for the use of non-Luer lock neuraxial devices (Supplemental Digital Content 2, http://links.lww.com/AA/B192). (The NPSA is an authority within the National Health Service in England. It monitors patient safety incidents and issues recommendations, reports, and guidance to prevent further incidents.) If a majority (>50%) of the obstetric anesthetic department heads agreed that the recommendation was likely to prevent the error category, then it was deemed a likely successful intervention.
Our literature search yielded 27 reports containing a total of 29 cases.6–32 The drug, dose, volume, route, setting, source of error, signs and symptoms, management, and sequelae of these errors are summarized in Table 1. No direct effects on the course of labor, mode of delivery, or neonatal outcome were reported. A range of hemodynamic and neurologic signs and symptoms was noted, but the most commonly reported complication was the failure of the intended neuraxial anesthetic technique. Four deaths were reported, all resulting from the accidental intrathecal administration of tranexamic acid.17,20,23,24 Each of these 4 cases was caused by an ampule error (misidentification of the ampule), with each case being found to have organizational influences and preconditions to act as circumstances of the error. Nine error categories were identified (Table 2).
The human factors from the HFACS identified in each report or case series are summarized in Table 3.
Our analysis identified 6 reports that described organizational process or climate as contributory factors in drug errors.15,17,18,24,25,27,30 The individual factors were inappropriate storage, labeling, or preparation of drugs15,17,18,23,24,27; time pressure24; lack of trained personnel30; and lack of timely communication and information about planned surgery to anesthesia provider.24
Although no reports explicitly mentioned a lack of supervision, 3 reports did comment on the inexperience of the providers committing the error.13,15,24 The finding that the case report authors described the inexperience of the providers suggests that they felt that the lack of experience may have been contributory to the error. Because of the lack of available trained personnel, 1 report30 describes the use of an untrained assistant as contributory to an error.
Preconditions to Acts
Three reports identified an adverse psychological condition, fatigue or stress, as contributory to the error.13,15,24 Several reports cited technical environmental factors as contributing to the error,6,17,18,20,22,23,25,27,30 including similarity in drug packaging, container, or infusion device. The only physical environmental factor was poor environmental lighting.15
We classified all cases as skill-based errors, because they were attention failures; however, examples in which clinicians confused similar looking packaging/vials/infusion devices were also classified as perception errors.6,17,18,20,22,23,25,27,30 There were 2 cases of exceptional violations;17,30 in both cases, a drug was incorrectly stored against current protocol.
Potential System Fixes
After receiving feedback from 5 of the 12 obstetric anesthetic department heads within the North Western deanery (England, United Kingdom), it was suggested that 4 of the 9 practice recommendations would have prevented the error scenarios had they been applied at the time of these errors. These are summarized in Table 2.
Various factors contribute to medication errors. Of the 9 identified error categories, none was believed to be immune from prevention according to our qualitative analysis. Diligent reading, checking, and labeling of ampules and syringes would likely prevent the majority of errors, because errors are often related to drug storage, labeling, or similar looking drugs and equipment. These findings concur with earlier work on the reduction of IV drug administration error in anesthetic practice.4 Qualitative analysis of the case reports suggests that the universal implementation of a non-Luer connection system may be of benefit in preventing some errors, but not all forms of obstetric neuraxial anesthetic drug error.
Drug administration errors can have devastating consequences to patients because anesthesia providers administer drugs IV or adjacent to critical structures when performing neuraxial anesthesia and analgesia. The consequences for health care providers can also be devastating. Skegg33 has documented several cases in New Zealand in which criminal prosecutions were brought against doctors and nurses. Several of these examples concerned anesthesiologists and an error in epidural drug administration. In Belgium, a case has been described in which both the anesthesiologist and midwife were found liable for an obstetric epidural drug error (wrong drug). The court of appeal found the midwife negligent for performing the injection, and the anesthesiologist was found guilty for not “directly controlling the midwife.”34
Drug errors in neuraxial anesthesia and analgesia have been quantitatively analyzed previously. In 2003, Hew et al.35 summarized 37 cases of nonepidural medications administered through the epidural route; 97% of these cases were secondary to syringe swap, ampule error, and epidural/IV line confusion. In 2012, Beckers et al.36 analyzed the reports of IV drugs administered into the epidural space. The authors found that syringe swap and ampule error and epidural catheter–IV line confusion were common causes of error and suggested that non-Luer lock epidural injection ports may increase safety.
The incidence of drug errors associated with neuraxial obstetric procedures cannot be identified from the published case reports, but our analysis likely represents the tip of the iceberg. It has been suggested that reporting bias, publication bias, and a lack of will for the physician to publish their “failures” and “complications” limit publication.37 It would be incorrect to conclude that drug errors are not a real problem merely because we only identified a small number of reports. Yentis and Randall38 conducted a survey of UK obstetric anesthetic department heads and found that of 179 respondents, 20 were aware of neuraxial drug errors within their department during the previous year. A similar but relatively small survey undertaken in Japan39 found that 54% of respondents had experienced wrong drug administration by the epidural route.
The task of providing obstetric anesthesia is complex and involves several steps. These steps require robust processes (e.g., storage, preparation, administration of drugs) and specific skilled performance (e.g., safe epidural and spinal procedure and subsequent monitoring). Situational awareness and vigilance are essential during all of these steps. The fact that 29 cases could be summarized into 9 categories illustrates that similar error mechanisms are being repeated. Our analysis suggests that each error has at least 1 practice recommendation that, if implemented at the time, would have likely prevented it. This finding suggests that the implementation of these recommendations could significantly reduce neuraxial drug administration errors. The introduction of non-Luer devices would likely have prevented 5 of 9 error categories. The careful checking of a drug before drawing it up, either independently or with another individual or device, before labeling the syringe is suggested to prevent errors in all but one (epidural/IV line confusion) of the error categories.
The validity of applying this type of human factors analysis to case reports to outline a plan to reduce practice errors, and as a result patient injuries, was demonstrated in the original description of the impact of continuous basic intraoperative monitoring on reducing anesthesia catastrophes.40 By analogy, development and promulgation of a formal protocol to help prevent the errors described in this review would be projected to improve patient outcomes.
Our analysis has identified latent conditions at all levels of Reason’s human error model.
This category refers to the decisions of upper level institutional management. These organizational influences directly affect the conditions and actions of health care professionals through resource management, policies, or culture. The storage, labeling, and preparation of drugs were contributory factors frequently encountered within our analysis. These case reports often described new systems that were implemented after the error. These systems included excluding certain medications from specific clinical areas, a prohibition against storing similar appearing medications in the same locality, preparing medications off site, or changing medication labeling to help differentiation. However, there was no evidence to suggest that these practice alterations reduced drug errors.
Unsafe supervision relates to causation being followed up the chain of command, from the operator to his or her superiors. We were unable to identify the level of supervision in many reports, because this was not frequently discussed. Previous work from the MEDMARX patient safety program41 has identified inexperience as a principle contributory factor in medication error in 78% of cases, but it is unclear whether closer supervision would have reduced the errors we analyzed.
Preconditions to Unsafe Acts
If errors are symptoms, then the preconditions are the underlying pathology. Similarities in the appearance of equipment such as drug packaging and infusion devices were technologic environmental factors. Three of the interventions we suggest seek to directly affect this misidentification of drugs: reading the label on any drug ampule or syringe carefully before drawing up or injecting a drug, labeling syringes, and checking labels specifically with a second person or a device (such as a barcode reader linked to a computer) before a drug is drawn up or administered. The fourth recommended intervention would prevent errors from similar appearing and functioning infusion devices. A NPSA project requires the use of non-Luer lock connectors for all epidural, spinal, combined spinal-epidural needles, catheters, and syringes.5 However, these systems are not yet available in the United Kingdom and other countries, including the United States because of precautions, complexity, testing, and cost.
In an Australian study, Abeysekera et al.42 found that haste and fatigue were present in 50% of anesthetic medication errors. Haste was described in one of the cases we analyzed, and fatigue was described in 3 reports. Although none of our suggested interventions sought to decrease anesthesia provider workload to combat errors, this has been assessed previously. Landrigan et al.43 found that medication errors on an intensive care unit were significantly reduced after a reduction in the hours worked per week and the elimination of extended intern shifts.
Unsafe acts can be considered errors or violations: the error is a mistake that occurs within the rules or standard procedures. A violation is a willful disregard for the rules and policies already in place. We identified 1 violation: an incorrect ampule was placed in the wrong location against working policy. Although the interventions mentioned earlier would seek to remove latent conditions and prevent the active failures of skill-based errors and perception errors, eliminating violations is much more complicated.
Of course, reading the label on an ampule or syringe, whether prelabeled or labeled by the practitioner, is simple to do and is recommend by the NPSA in the United Kingdom44 and the Institute for Safe Medical Practices in the United States.45 Various labels are available to aid this process with a similar color-coding system present in North America, Australasia, and the United Kingdom and Ireland. The reality and practicality of second person verification together with the expense and likely complexity of a computer verification system make these particular interventions more difficult, especially for private practitioners working alone. In 2010, the NPSA, in collaboration with the Royal College of Anaesthetists and the Association of Anaesthetists of Great Britain and Ireland, produced a qualitative feasibility study46 investigating the introduction of these systems into the National Health Service in the United Kingdom. This study suggested that both methods led to a perceived contribution to the prevention of drug errors, but they were limited by either the lack of a continuous presence of a second person or the practical and educational issues surrounding electronic barcode technology.
Our analysis has several limitations. First, there can be little doubt that by analyzing only published case reports, we are only sampling a tiny fraction of the neuraxial drug administration errors that occur during obstetric anesthesia practice. Our analysis is subject to large reporting and publication bias. The levels of evidence that this analysis uses and generates are low. We have analyzed the published case reports (level IV evidence), and our recommendations are based on expert opinion (level V evidence). For our recommendations to be properly assessed, they need to be evaluated in robust trials.
We found that drug errors during obstetric neuraxial anesthesia or analgesia are multifactorial. Our analysis found failures at all levels of Reason’s model of human error. We suggest that the implementation of the following processes may decrease the risk of these types of drug errors:
- Careful reading of the label on any drug ampule or syringe before the drug is drawn up or injected,
- Labeling all syringes,
- Checking labels with a second person or a device (such as a barcode reader linked to a computer) before the drug is drawn up or administered, and
- Use of non-Luer lock connectors on all epidural/spinal/combined spinal-epidural devices.
According to our expert opinion analysis, 1 or more of these interventions has the potential to prevent all of the errors identified in the published case reports. Although this analysis was conducted exclusively for obstetric anesthesia practice, we suspect that the factors leading to error would be common to those seen across anesthetic practice. Similarly, the practice recommendations (with exception of the non-Luer connectors) could be applied more widely across anesthetic practice to reduce all drug errors. Further study is required to determine whether the routine use of these processes will reduce drug error.
Name: Santosh Patel, MD, FRCA.
Contribution: This author helped design the study, conduct the study, analyze the data, and write the manuscript.
Attestation: Santosh Patel approved the final manuscript.
Name: Robert Loveridge, MBChB, FRCA.
Contribution: This author helped design the study, conduct the study, analyze the data, and write the manuscript.
Attestation: Robert Loveridge approved the final manuscript.
This manuscript was handled by: Cynthia A. Wong, MD.
The authors thank Dr. Matthew Beardmore for his help translating a French reference article.
1. D’Angelo R, Smiley RM, Riley ET, Segal S. Serious complications related to obstetric anesthesia: the serious complication repository project of the Society for Obstetric Anesthesia and Perinatology. Anesthesiology. 2014;120:1505–12
2. Webster CS, Merry AF, Larsson L, McGrath KA, Weller J. The frequency and nature of drug administration error during anaesthesia. Anaesth Intensive Care. 2001;29:494–500
3. Human Factors Analysis and Classification System. Available at: http://www.hfacs.com/hfacs-framework.html
. Accessed June 9, 2015
4. Jensen LS, Merry AF, Webster CS, Weller J, Larsson L. Evidence-based strategies for preventing drug administration errors during anaesthesia. Anaesthesia. 2004;59:493–504
5. National Patient Safety Agency. Patient Safety Alert—Safer spinal (Intrathecal), Epidural and Regional Devices—Part B (NPSA/2009/PSA004B). 2009 London, UK National Patient Safety Agency
6. Lejuste MJ. Inadvertant intrathecal administration of magnesium sulfate. S Afr Med J. 1985;68:367–8
7. Dror A, Henriksen E. Accidental epidural magnesium sulfate injection. Anesth Analg. 1987;66:1020–1
8. Tessler MJ, White I, Naugler-Colville M, Biehl DR. Inadvertent epidural administration of potassium chloride. A case report. Can J Anaesth. 1988;35:631–3
9. Mappes A, Schaer HM. Accidental injection of ether into the epidural space. Anaesthesia. 1991;46:124–5
10. Olmez G, Yalinkaya A. Accidental epidural injection of ephedrine. Int J Obstet Anesth. 2004;13:58–9
11. Sidi A, Froelich MA. Inadvertent epidural injection of ephedrine in labor. J Clin Anesth. 2004;16:74–6
12. Leão EB, Barros GA, Castiglia YM, Ganem EM. [Accidental spinal metoclopramide injection: case report.]. Rev Bras Anestesiol. 2004;54:663–7
13. Balestrieri PJ, Hamza MS, Ting PH, Blank RS, Grubb CT. Inadvertent intrathecal injection of labetalol in a patient undergoing post-partum tubal ligation. Int J Obstet Anesth. 2005;14:340–2
14. Huang JJ. Inadvertent epidural injection of ondansetron. J Clin Anesth. 2006;18:216–7
15. Goodman EJ, Haas AJ, Kantor GS. Inadvertent administration of magnesium sulfate through the epidural catheter: report and analysis of a drug error. Int J Obstet Anesth. 2006;15:63–7
16. Parodi G, Antoniucci D. Transient left ventricular apical ballooning syndrome after inadvertent epidural administration of potassium chloride. Int J Cardiol. 2008;124:e14–5
17. Veisi F, Salimi B, Mohseni G, Golfam, Kolyaei A. Accidental intrathecal injection of tranexamic acid in caesarean section: a fatal medication error. APSF Newslett. 2010;25:9
18. Merino I, Pérez J. Inadvertent epidural infusion of acetaminophen during labour. Int J Obstet Anesth. 2011;20:192–4
19. Loderer J, Suppan P. Accidental injection of ephedrine into the epidural space. Anaesthesia. 1979;34:78–9
20. Sabzi F, Teimouri H, Zokai A. Myoclonus, seizure, and ventricular fibrillation after intrathecal injection of tranexamic acid. J Teh Univ Heart Ctr. 2009;4:253–5
21. Peduto VA, Gungui P, Di Martino MR, Napoleone M. Accidental subarachnoid injection of pancuronium. Anesth Analg. 1989;69:516–7
22. Ross MJ, Wise A. Accidental epidural administration of Syntocinon. Int J Obstet Anesth. 2012;21:203–4
23. Raghu K, Shrevanni P, Haneef M, Ramyakumar, Chenna Rao M. Accidental intrathecal injection of tranexamic acid in a term gestation. Int J Neu Spi Sci. 2013;1:1–3
24. Antwi-Kusi A, Sam Awortwi W, Serwaa Hemeng A. Unusual complication following spinal anesthesia for caesarean section. Open J Anesthesiol. 2013;3:275–7
25. Quality and Governance Branch NSW Department of Health. Safety notice 010/10. Available at: http://www.health.nsw.gov.au/sabs/Documents/2010-sn-010.pdf
. Accessed June 11, 2015
26. Brownridge P. More on epidural thiopentone. Anaesth Intensive Care. 1984;12:270–1
27. Sigg TR, Leikin JB. Inadvertent epidural gentamicin administration. Ann Pharmacother. 1999;33:1123
28. Cay DL. Accidental epidural thiopentone. Anaesth Intensive Care. 1984;12:61–3
29. Ianov I, Manimekalai N, Wasiluk I, Vinjirayer A. Accidental epidural administration of oxytocin in a laboring patient: case report and root cause analysis [abstract]. In: Society of Obstetric Anesthesia and Perinatology 45th Annual Meeting. 2013:S59 Available at http://soap.org/display_2013_abstract.php?id=S%2059
. Accessed June 11, 2015
30. Monsel A, Ibrahim B, Mercier FJ. [Accidental epidural administration of distilled water: systemic analysis of one case]. Ann Fr Anesth Reanim. 2010;29:242–4
31. Chowdhury P, Dabrowska D, Said H. “Eternal vigilance is the price of safety”: a case of inadvertent administration of oxytocin via epidural catheter: 11AP1-8. Eur J Anaesthesiol. 2014;31:178
32. Gilani MT, Zirak N, Razavi M. Accidental intrathecal injection of magnesium sulfate for cesarean section. Saudi J Anaesth. 2014;8:562–4
33. Skegg PDG. Criminal prosecutions of negligent health professionals: the New Zealand Experience. Med Law Rev. 1998;6:220–46
34. Heylen R. The anesthesiologist and the law: the criminal liability of midwives and anesthesiologists for epidural anesthetic reinjection in obstetrics. Acta Anaesthesiol Belg. 1994;45:5–12
35. Hew CM, Cyna AM, Simmons SW. Avoiding inadvertent epidural injection of drugs intended for non-epidural use. Anaesth Intensive Care. 2003;31:44–9
36. Beckers A, Verelst P, van Zundert A. Inadvertent epidural injection of drugs for intravenous use. A review. Acta Anaesthesiol Belg. 2012;63:75–9
37. Niesters M, Overdyk F, Smith T, Aarts L, Dahan A. Opioid-induced respiratory depression in paediatrics: a review of case reports. Br J Anaesth. 2013;110:175–82
38. Yentis SM, Randall K. Drug errors in obstetric anaesthesia: a national survey. Int J Obstet Anesth. 2003;12:246–9
39. Kasaba T, Uehara K, Katsuki H, Ono Y, Takasaki M. [Analysis of inadvertent epidural injection of drugs]. Masui. 2000;49:1391–4
40. Eichhorn JH. Prevention of intraoperative anesthesia accidents and related severe injury through safety monitoring. Anesthesiology. 1989;70:572–7
41. Wolf ZR, Hicks R, Serembus JF. Characteristics of medication errors made by students during the administration phase: a descriptive study. J Prof Nurs. 2006;22:39–51
42. Abeysekera A, Bergman IJ, Kluger MT, Short TG. Drug error in anaesthetic practice: a review of 896 reports from the Australian Incident Monitoring Study database. Anaesthesia. 2005;60:220–7
43. Landrigan CP, Rothschild JM, Cronin JW, Kaushal R, Burdick E, Katz JT, Lilly CM, Stone PH, Lockley SW, Bates DW, Czeisler CA. Effect of reducing interns’ work hours on serious medical errors in intensive care units. N Engl J Med. 2004;351:1838–48
44. National Patient Safety Agency. Patient Safety Alert 20: Promoting Safer Use of Injectable Medications. A Template Standard Operating Procedure for: Prescribing, Preparing and Administering Injectable Medicines in Clinical Areas. Available at: http://www.nrls.npsa.nhs.uk/EasySiteWeb/getresource.axd?AssetID=60098&type=full&servicetype=Attachment
. Accessed June 11, 2015
45. Institute for Safe Medication Practices. Principles of Designing a Medication Label for Injectable Syringes for Patient Specific, Inpatient Use. Available at: http://www.ismp.org/Tools/guidelines/labelFormats/Injectable.asp
. Accessed June 11, 2015
46. Evley R, Russell J, Mathew D, Hall R, Gemmell L, Mahajan RP. Confirming the drugs administered during anaesthesia: a feasibility study in the pilot National Health Service sites, UK. Br J Anaesth. 2010;105:289–96