Audit of Anesthetic Equipment in Veterinary Clinics in Spain and Portugal

Abstract

The objective of this retrospective study was to review the results of a 4-year audit performed on anesthetic machines and vaporizers used in veterinary clinics in Spain and Portugal. Data was collected between July 2016 and April 2020. Inspections were carried out by a team of seven veterinarians, using a human-modified system of checks that was adapted to a veterinary practice. The evaluation of each item was noted as "correct" or "incorrect". The vaporizers' performance was evaluated using a self-calibrating gas analyzer. The vaporizer was classified as "correct" or "incorrect" when the vaporization error was less than or equal to, or more than 20%, respectively. The anesthetic machine was classified as "conforming" if all its components were noted as "correct" and no leaks were detected, or as "non-conforming" if any of the components was noted as "incorrect" or if a leak was detected. If the inspector was able to repair on-site the item malfunctions detected and the machine was fit for use, they issued a final report as "conforming." On the contrary, if such malfunctions persisted, the final report was "non-conforming," and a recommendation to remove the machine from service until its final repair was provided. To perform statistical analysis, each inspected item was used as predictor, classification and regression trees were built, and a random forest analysis was performed. A total of 2,001 anesthetic machines and 2,309 vaporizers were studied. After inspection, 42.7 and 26.4% of the machines were non-conforming and conforming, respectively, whereas 30.9% could be repaired in situ. A total of 27.1% of the isoflurane vaporizers and 35.9% of the sevoflurane vaporizers were incorrect. Machine learning techniques showed that the most important variables in the classification of the anesthetic machines as conforming or non-conforming were mostly the scavenger system and the canister, followed some way behind by the APL valve, source of oxygen, reservoir bag, vaporizer, and connections.

Keywords: anesthesia; anesthetic machine; audit; equipment malfunction; retrospective study; safety; vaporizer; veterinary.

Figure 1

Distribution of the anesthetic machines that were audited in Spain and Portugal.

Figure 2

Number of anesthetic machines repaired separated by the number of malfunctions found and their final classification as conforming or non-conforming.

Figure 3

Percentage of malfunctions detected on each component of the anesthetic machine at initial inspection and their classification as correct or incorrect.

Figure 4

Classification tree of malfunctions detected at inspection of anesthetic machines. The classification tree represents the different selection criteria or “decision nodes” used to predict the most correct classification of the total number of cases (represented at the root of the tree as a 100%). As the data is classified in subsets, the percentage value represents the probability of a case of belonging to that data subset.

Figure 5

Random forest showing the importance of each variable in the inspection of anesthetic machines. The Mean Decrease Accuracy plot expresses how much accuracy the model losses by excluding each variable. The more the accuracy suffers, the more important the variable is for the successful classification. The variables are presented from descending importance. The mean decrease in Gini coefficient is a measure of how each variable contributes to the homogeneity of the nodes and leaves in the resulting random forest. The higher the value of mean decrease accuracy or mean decrease Gini score, the higher the importance of the variable in the model.

Figure 6

A box-and-whisker diagram (median, interquartile range, and extremes) of percentage errors as a function of the vaporization percentage (V; 0.5, 2 and 3%) and the gas flow (F; 0.5, 1, 2 and 3 L/min) for isoflurane vaporizers.

Figure 7

A box-and-whisker plot (median, interquartile range, and extremes) of percentage errors as a function of the vaporization percentage (V; 0.5, 2, and 3%) and gas flow (F; 0.5, 1, 2, and 3 L/min) for sevoflurane vaporizers.