Methods

The study was approved by the University of Calgary (REB22-0691) and carried out in partnership with the Institute for Safe Medication Practices Canada (ISMP Canada). As the submission of medication error reports was entirely voluntary, written consent was not required from individual clinic staff members. Written, informed consent was provided by the head veterinarian or veterinary medical director of each participating clinic. Information provided when requesting consent included an emailed description of the study goals and how confidentiality of submitted error reports would be protected, as well as a formal consent form, both of which were reviewed and approved by our institutional human ethics board. This same information was provided to clinic staff during the first Continuing Education session (CE1). All medication error reports were submitted online through an existing platform hosted by ISMP Canada. Access to reports was limited and restricted to ISMP Canada analysts. Data collated from reports were shared through an online spreadsheet (Microsoft Excel; Microsoft) that was password protected and managed by an ISMP Canada IT security administrator. Data were analyzed within the spreadsheet, and the file was never downloaded, as per ISMP Canada policy. Submitted reports were screened by ISMP Canada analysts for the inclusion of personal information (eg, names), and any such information was removed before preparing the spreadsheet made available for analysis.

A total of 6 veterinary clinics within the city of Calgary were recruited to participate. Clinic recruitment was based on personal contacts through the University of Calgary’s distributed teaching network. The CE1 was presented to each clinic individually, between October and December 2022, as decided by clinic schedule and staff availability. The CE1 took approximately 1 hour and provided information on error theory and medication errors and a description of how to access and complete the online reporting system. Attendees were informed of how submitted reports would be analyzed and used for research and how submitted information would be treated as confidential.

The reporting system used was designed for human health care. Accessed online, it consists of a combination of mandatory and optional fields distributed among the following sections: incident, outcome, medication, follow-up, patient, option for the reporter to provide contact information, and contributing factors. Within these sections, certain fields were mandatory, including incident description; medication-use stages involved; type of incident; who discovered the incident; care area type, outcome, and medication(s) involved; if the incident occurred due to confusing drug name, label, or packaging; and if the reporter wanted or not to be contacted regarding the incident.

For this study, because certain fields did not include options relevant to veterinary medicine (“discovery by,” “care area type,” and “patient information”), users were instructed to use the open field in the first section of the report (“incident”) to provide a complete description of the ME event. The following information was requested: date and time of incident, patient information (species, breed, age, sex, weight, and temperament), medication information (name, dose, and route), a detailed narrative (what, when, why, and how the ME happened), and who discovered the incident and how. Reporters were asked not to provide patient, owner, clinic, or staff identifiers and told that any such information would be removed. Participating staff received the link to the ISMP Canada reporting system. Attendees were encouraged to report all errors, including both those resulting in harm as well as near misses (ie, errors discovered before reaching a patient). A description of how confidentiality and anonymity were protected was also given. To identify submitted reports from the study, participants were asked to include a code when submitting reports. The code did not identify individuals. As the same code was used by all individuals and clinics, it was not possible to identify any individuals or clinics. A user guide on completing and submitting an error, showing a fictitious example, was provided in 3 formats: a printed, laminated sheet (several copies distributed to each clinic), an electronic copy of the sheet, and a video illustrating how to complete and submit a report. The user guide included a QR code that linked to the ISMP Canada reporting site.

In March 2023, a one-page ME fact sheet was circulated to each clinic by electronic mail. This included a reminder that the study was ongoing (the study QR code was included) and examples of ME from human anesthesia. In a planned interim analysis, error reports submitted between October 18, 2022, and February 28, 2023, were analyzed, and aggregate data were presented in a second Continuing Education presentation (CE2) to each clinic (between April 18 and 24, 2023). During CE2, clinics were reminded that the study was ongoing. On June 16, 2023, a second fact sheet with information about look-alike and sound-alike medications, and the QR code to access the reporting system, was sent to each clinic. The last day of data collection was June 30, 2023.

Results

During the study period for each clinic, 2,728 companion animals (1,214 dogs and 1,514 cats) were anesthetized or sedated in the 6 participating clinics. A total of 49 reports were submitted. One report was excluded, as it described a duplicate of the same error. With this exclusion, the calculated ME incidence was 1.8%.

The completeness of reporting varied, and information was analyzed based on the narrative description and the mandatory reporting fields. Patient species, type of error, medication, and outcome were included in all reports, while patient temperament, age, weight, anesthesia stage, people involved with ME, how the ME was discovered, and contributing factors were not described by all reporters.

Age was described in 83% of reports (2.5 [0.3 to 16] years old). Body mass was included in 79% of reports (4.6 [1.7 to 44.2] kg).

Reporting of temperament was low (42% [20/48]), with positive temperaments (eg, “nice,” “lovely,” “good,” and “friendly”) described in 45% (9/20) of reports, negative temperaments (“aggressive,” “anxious,” “fractious,” and “grumpy”) in 35% (7/20) of reports and unclear descriptors (eg, hyperactive) provided in 20% (4/20) of reports.

The proportion of dogs and cats involved in MEs was similar (cats; 54% [26/48], dogs; 46% [22/48]). The results, according to ME classification are presented (Table 1). In more than two-thirds of cases, errors did not reach the patient (near misses). Approximately half of MEs occurred during premedication/sedation and the act of prescribing accounted for most MEs. Around two-thirds of MEs resulted from wrong dose errors and most of these resulted from calculation errors (80% [24/30]), followed by errors in filling a syringe (13% [4/30]) and other causes (6% [2/30]). After wrong dose errors, errors were classified as overdose in 53% of cases (16/30) and as underdose in 37% (11/30) of cases. Where dose errors could be quantified, the most common errors were as follows: 10-fold overdose (23% [7/30]), 2-fold overdose (20% [6/30]), and 10-fold underdose (17% [5/30]; Table 2). In 10% of wrong dose reports (3/30), insufficient information was provided to allow quantifying the dose error.

Opioids, sedatives, and NSAIDs were the drug classes most frequently involved in MEs, representing 27%, 23%, and 21% of reports, respectively (Table 3). Meloxicam was the most commonly reported single drug, involved in 19% of all MEs, followed by dexmedetomidine (15%).

All MEs involving meloxicam resulted in near misses and occurred primarily when prescribing (78% [7/9]) and most of the time for the postoperative (56% [5/9]). Most of the errors were wrong dose and wrong time MEs (33% [3/9]; Table 4).

The person(s) involved in the ME and the manner of discovery are presented (Table 5). Approximately two-thirds of MEs involved were discovered by registered veterinary technicians (technicians/nurses). In most of the cases, errors were discovered by a coworker. The method by which an ME was discovered was described in approximately half of the reports, with errors discovered predominantly during an independent double check of drug calculations. Other methods by which errors were discovered included awareness of common doses (ie, knowledge of common doses/expected volumes and observing an unusually large/small dose or volume) and reviewing the anesthesia chart. Contributing factors were selected from the ISMP Canada reporting system options and also described in the open narrative field. Workload was reported as a contributing factor slightly more often than interruptions, environment, staffing, or workflow problems (Table 6).

Discussion

The results of this study reveal that a spectrum of MEs occurs during the perianesthetic period in general practice small animal veterinary clinics. The reporting highlights several areas where action could be taken to improve perianesthetic safety, including medical maths, prescribing process, and use of NSAIDs.

Previous studies on perianesthetic errors have been performed in university teaching hospitals^10,11; however, this study concentrated specifically on perianesthetic ME in general practices. The decision to focus on general practice clinics in this study resulted from fundamental differences in how they generally function in contrast to academic teaching hospitals. Furthermore, the majority of pet dogs and cats will be treated in general practice clinics rather than in academic hospitals. Differences include the presence and participation of trainees in teaching hospitals, with a reasonable expectation of an increased risk of errors.¹² This may be balanced by the presence of staff dedicated to anesthesia delivery (anesthesiologists and technicians). A further contrast may exist between a larger variation in anesthetic medication protocols in teaching hospitals versus those encountered in general practice, where the variety of drugs available and protocols used may be narrower. Interestingly, the incidence of MEs observed in this study in general practices (1.8%, 48/2728) is higher than that reported in veterinary (0.5% to 1.2%)^10,11 and human (0.49%)¹³ teaching hospitals. However, any direct comparison of reporting rates should be performed with caution as reporting of MEs may be influenced by several factors. These include the seasonal presence of new trainees/residents in teaching hospitals,¹² variation in safety culture between countries,¹⁴ hospital size,¹⁵ voluntary or mandatory nature of reporting,¹⁶ and the relatively small datasets involved.

Of the 6 participating clinics, 69% (33/48) of reports were near misses. In veterinary medicine, reports of near misses (not restricted to anesthesia or medication errors) varied between 15%⁴ and 19.1%² of all errors reported. In human medicine, the rates of ME near misses are more inconsistent and variable, ranging from 11.8% (21/178 MEs) in an emergency department (when identified by direct observation)⁵ to 85.8% (280/330 MEs) in pharmacies (when identified by self-reports).¹⁷ It is recognized that reporting and analyzing near misses are the first steps in reducing MEs.^18,19 However, there remains the mistaken belief that near misses are not errors but rather successful outcomes.²⁰ For example, a survey¹⁸ of pharmacists found that 35.5% of those involved in an ME did not classify near misses as MEs. This highlights the importance of education in complete reporting. The large proportion of near misses reported in the current study, greater than any existing veterinary study, suggests that the CE strategy used was successful. Analyzing near misses is fundamental to understanding how patient harm can be avoided in some cases when an error occurs: specifically successful strategies employed to avoid harm.²¹ In the current study, most near misses did not cause patient harm as a result of simple, feasible, and practical strategies, such as independent double-checking calculations. The importance of this strategy is further reflected by most reported MEs being wrong dose errors.

Wrong dose errors are a well-recognized source of MEs, representing the most reported type of ME in veterinary medicine^2,4 and pediatric anesthesia.^22,23 In the current study, wrong dose errors arose primarily from 2 causes: calculation errors and syringe filling errors. Calculation errors occur mainly from unit conversion errors (eg, mg to mcg) and errors when multiplying or dividing.²⁴ Evidence from physician prescriptions identified calculation errors as stemming from insufficient basic mathematical skills.²⁵ This may highlight a critical gap in healthcare practitioner course curricula: the need for teaching and assessing drug calculation skills. Syringe filling errors are a particular concern when working with 1-mL syringes, where inaccuracy increases as the targeted volume decreases.²⁶ This is particularly problematic with volumes ≤ 0.04 mL. Calculation errors represented half of all reports submitted, and most of them resulted in an overdose, with potentially more serious consequences than an underdose. Together, these results highlight medical maths as an area for targeted education. Difficulties in calculating doses and constant rate infusions are not exclusive to veterinary medicine. In human pediatric practice, it is reported that 85% (35/41) of anesthetists were unable to calculate a dopamine infusion for a pediatric patient.²⁷ As a strategy to reduce calculation errors, smartphone applications have been described as being successful in facilitating anesthesia management and reducing errors by veterinary students²⁴ and human anesthesia providers.²⁸ Additionally, automated dose calculation tools exist and may represent an accessible option for veterinary patients.²⁹

The medication-use stage and anesthesia stage are other targets to be considered for safety improvement. In the current study, most MEs occurred during prescribing. Anesthesia is unusual among the specialties as prescribing, dispensing, and administration are often performed by the same professional, which facilitates the occurrence of MEs.³⁰ This reinforces the importance of double checks as a method of improving safety when handling medications. The role of technicians during the perianesthetic period, in conjunction with veterinarians, contributed to the timely discovery of prescription errors before they became administration errors. Most MEs occurred during the premedication and sedation stages. The risk of MEs during these stages is potentially heightened when considering that following premedication or sedation, many patients are not under continuous surveillance or physiologic monitoring and are unlikely to have an IV catheter in place. Consequently, the adverse effects of MEs may go unnoticed for some time and lack of IV access may delay treatment.

Opioids, followed by sedatives and NSAIDs, were the most commonly reported medication class involved in MEs. This is unsurprising as they are widely used during the perianesthetic period. Opioid overdose may lead to excessive sedation and cardiorespiratory depression. An ISMP Canada analysis of harm incidents identified that the top 5 medications involved in severe harm or death incidents consist of 3 opioids³¹ including opioids in 3 out of the top 5 medications. Medication error studies^8,23 in human pediatric anesthesia identified opioids as the most commonly involved medication.

Meloxicam was the single most represented medication in the present study. While it is not possible to explain the prevalence of meloxicam, various factors may play a role, including (1) meloxicam is a popular NSAID in veterinary medicine in Canada,³² including widespread perioperative use, and (2) meloxicam is available in several concentrations, presented in similar packaging (lookalike).

The current study found technicians to be most commonly involved with MEs. In Canada, technicians have significant responsibility for multiple steps in medication processing; therefore, this result is unsurprising. Technicians were also the professionals who most frequently discovered errors, demonstrating their important role in patient safety. Similarly, in human hospitals, nurses play a fundamental role in patient safety, and they are also much more likely to self-report an error when compared to physicians (OR, 2.8; 95% CI, 1.3 to 6.0).³³ In veterinary medicine, the role of technicians in reporting errors is unclear. Within the same study, technicians were much more likely to report an error when compared to a veterinarian in a small animal university teaching hospital (technicians: 65%; veterinarians: 27%), but the opposite was observed in a large animal university teaching hospital (technicians: 27%; veterinarians: 53%), and small animal specialty referral clinic (technicians: 37%; veterinarians: 41%).² Coworkers were more likely to identify an error as compared to self-identification. This reflects the importance of teamwork and communication but also the presence of a safety culture where staff can notify a colleague/supervisor about a possible error.³⁴

Excessive workload was the most common contributing factor identified for MEs in the current study. Similarly, a human study reported that 91.2% (104/114) of ICU nurses identified excessive workload as a contributing factor in MEs.³⁵

In veterinary ICUs, reducing the patient:technician ratio to less than 4:1 was associated with a reduction in errors.³⁶ Furthermore, excessive workload has been associated with burnout in veterinary technicians, underlining the value of addressing staff workload.

Limitations of the present study include incomplete reporting and a reliance on voluntary reporting. For example, patient temperament, how the ME was discovered, and contributing factors were reported in 42% (20/48) to 71% (34/48) of reports. It is likely that relying on completing an open field for much of the information collected contributed to this deficit. A veterinary reporting system with specific, relevant questions, in addition to open fields, could promote uniformity and consistency in data collection. Self-reporting is the predominant method for collecting ME reports in both veterinary and human health care.^2,4,37 It is well recognized that this method underestimates the actual incidence of all errors (not just MEs). A human hospital study identified 456 errors by direct observation, with only one self-report submitted over the same period.³⁸ Although our study provided rates based on the total number of procedures, it is not possible to infer the probability of incidents based on voluntary reports. The prevalence of near misses in the current study identifies areas for further research and education. Unfortunately, due to differences in items reported in other veterinary studies, it is not currently possible to know if similar near-miss rates may occur in other clinics/institutions. A further limitation is that medication handling practices were not collected from participating clinics; therefore, we do not know what safety measures, if any, were taken by individual clinics.

In conclusion, several different ME types were observed, with predominant features including occurrence during the premedication/sedation stage, an association with the act of prescribing, wrong dose errors arising from calculation errors, and technician involvement in both ME occurrence and discovery. This pattern and the high incidence of near misses highlight the value of active error prevention, such as independent double-checking calculations, and identify areas for improvement through education and potentially technology, such as the use of dose calculation applications. Widely disseminating ME data is not currently practiced in veterinary medicine, with collected error reports largely retained by private organizations. Sharing such information would raise awareness of the incidence and consequences of errors, promote a safety culture, identify common causes of errors more rapidly, and drive quality improvement.

Disclosures

The authors have nothing to disclose. No AI-assisted technologies were used in the generation of this manuscript.

Funding

Funding was provided by the University of Calgary (Clinical Research Fund), Alberta Graduate Excellence Scholarship–International, and a Zoetis Investment in Innovation Fund.