Outcomes from an intervention study implementing fee‐free culture and susceptibility urine testing in dogs and cats

Abstract

Background

Urine culture and susceptibility (C&S) testing is important to confirm the presence of infection and aid in antimicrobial selection to ensure that the most appropriate drug is used. Veterinarians often cite the cost of C&S as a prohibitive factor limiting its uptake.

Methods

From January to December 2022, 10 general practice veterinary clinics in Melbourne, Victoria, Australia, were provided with access to free urine C&S and a decision support tool. At the end of the study period, a survey was distributed to veterinarians at participating clinics.

Results

A total of 480 urine C&S submissions were received in the 11‐month study period. Several clinics had an increase in submission rate, with an overall 2.4‐fold increase in the submission rate. Most survey respondents (80%) reported submitting more urine samples for C&S during the trial, compared to before. Many of the veterinarians (74%) reported delaying prescribing antimicrobials while awaiting C&S results.

Limitations

The small number of practices involved in this study may not be representative of all types of clinics in Australia.

Conclusion

Cost is a barrier to susceptibility testing being undertaken in most cases. Removal of the cost barrier and provision of a decision support tool resulted in positive changes to reported antimicrobial prescribing behaviour.

INTRODUCTION

Antimicrobial resistance (AMR) is a global health emergency with significant impacts on both human and animal welfare. Antimicrobial use is a key factor driving the emergence of AMR. ¹ In Australia, treatment of lower urinary tract infections (UTIs) is a common reason for veterinary administration of antimicrobials in both cats (6.3% of pet insurance claims that involved an antimicrobial prescription) and dogs (1.7% of pet insurance claims that involved an antimicrobial prescription). ² Studies also found that 38% of canine and 35% of feline urinary tract disorder cases were treated with antimicrobials. ³

The International Society of Companion Animal Infectious Diseases (ISCAID) guidelines for the diagnosis and management of bacterial UTIs recommend that all patients with clinical signs of UTI (stranguria, haematuria, pollakiuria, dysuria or a combination of these signs) should have urine culture and susceptibility (C&S) performed. ⁴ It is important to confirm the presence of bacteria in the urine, as many other medical conditions can also present with similar clinical signs, such as idiopathic cystitis in cats. ⁴ , ⁵ , ⁶ Susceptibility testing also aids in antimicrobial selection to ensure effective antimicrobial therapy.

However, empirical prescribing is common due to the challenge of obtaining reliable samples and the high cost associated with microbiological testing. ⁷ , ⁸ In Australia (and most other countries), the owner of the pet is fully responsible for all costs incurred in the veterinary visit. The ISCAID guidelines recommend the use of amoxicillin or trimethoprim‒sulphamethoxazole for the empirical treatment of sporadic UTIs. This is supported by susceptibility studies in Australia, where the occurrence of acquired AMR in UTIs is low. ⁹ , ¹⁰

In Australia, antimicrobials are categorised into ‘low’, ‘medium’ and ‘high’ importance by the Australian Strategic and Technical Advisory Group on Antimicrobial Resistance. These ratings reflect the use of antimicrobials in the Australian context, taking into account the availability of alternative antimicrobials for the treatment of human infections if AMR develops. A high importance rating indicates the limited or lack of treatment options available for preventing or curing infections in people. ¹¹

This study aimed to investigate whether an intervention that removes the cost of C&S and introduces decision support influences the veterinarian's decision to collect and submit samples for C&S testing. A secondary aim was to assess the effect of the intervention on prescribing behaviour.

METHODS

Ten general practice veterinary clinics (located in metropolitan Melbourne, Victoria, Australia) from a large corporate group voluntarily enrolled in the study. C&S testing on urine samples from dogs and cats was provided free of charge during the study period (4 January to 3 December 2022).

Veterinarians were invited to submit urine for C&S testing from patients with any urinary clinical signs, including but not restricted to stranguria, dysuria, polyuria, polydipsia, haematuria or a combination of these signs. The method of collection was requested on submission.

The workflow and time to reporting of test results for veterinarians were unchanged during the study, and the turnaround time of results was the same. At the diagnostic laboratory, urine samples were cultured, isolates were identified using matrix‐assisted laser desorption/ionisation time‐of‐flight mass spectometry (MALDI‐TOF) and the antimicrobial susceptibility results were determined using the VITEK 2 system (bioMérieux).

The frequencies of C&S submissions for each clinic were recorded on a monthly basis for the 12 months prior to the trial (January to December 2021) and during the 11 months of the trial. The control data were collected from 11 clinics not enrolled in the intervention study, from the same large corporate group and geographical area, during the study period.

The percentage change in C&S submissions was calculated using the following formula:

$$\begin{array}{ccc}& & \mathrm{Percentage}\mathrm{change}(\%)\hfill \\ \hfill & =& \left(\frac{\mathrm{No}.\mathrm{of}\mathrm{submissions}\mathrm{in}2022}{\mathrm{No}.\mathrm{of}\mathrm{submissions}\mathrm{in}2021}-1\right)\times 100\hfill \end{array}$$

The number of veterinarians working at the clinics involved in the study was estimated using VetCompass Australia data. ¹² All unique staff member codes associated with the consultation billing during the study period were tallied. Any staff member codes with fewer than 10 ‘consultations’ associated were excluded. This was to account for any accidental charges under a staff member who was not a veterinarian. The number of consultations each clinic performed was also estimated with VetCompass data using consultation billing items.

The rate of C&S submission was calculated using the formula:

$$\mathrm{Rate}(\%)=\frac{\mathrm{No}.\mathrm{of}\mathrm{C}\&\mathrm{S}\mathrm{submissions}}{\mathrm{No}.\mathrm{of}\mathrm{consultations}}\times 100$$

$$\mathrm{Ratio}\mathrm{of}\mathrm{submission}\mathrm{rate}=\frac{\mathrm{Submission}\mathrm{rate}2022}{\mathrm{Submission}\mathrm{rate}2021}$$

The submission data were compiled and recorded in Microsoft Excel, and statistical analysis was performed using Microsoft Excel and R studio (2023.09.1). The test of proportions for comparing two groups was used.

A decision support tool was designed to provide participating veterinarians with guidance on empiric treatment recommendations for sporadic UTIs based on the local prescribing guidelines. ⁴ , ¹³ The tool was distributed as a physical poster and digitally (Figure 1).

FIGURE 1.

Decision support tool for the treatment of sporadic urinary tract infections.

An online questionnaire (Qualtrics) was distributed to participating veterinarians via email and by a QR code left in clinics at the end of the study period (Supporting Information S1). The questionnaire took approximately 10 minutes to complete and included questions on the respondent's attitudes towards free C&S testing, the decision support tool and antimicrobial stewardship (AMS).

Descriptive analysis of closed questions was performed in R studio (2023.09.1). Open questions were analysed using content analysis and an inductive approach.

RESULTS

C&S results

A total of 480 submissions were received from 360 unique patients. There were 78 (22%) patients with multiple urine submissions, ranging from two to five submissions in the study period. Submissions were predominantly from dogs (62%). Of the 480 submissions, growth was reported in 171 (36%), ‘not significant’ growth was reported in 60 (13%) and no bacterial growth was detected in the remaining 249 (52%). Urine samples were collected via cystocentesis for 60% of cats (109/183) and 32% of dogs (94/297) (Supporting Information S2). There was no difference in the frequency of returning a positive culture between dogs and cats, so the results were combined (Supporting Information S3).

Most submissions were from female neutered animals (285/480, 59%), followed by male neutered animals (129/480, 27%). Samples positive for bacterial growth also mostly originated from neutered females (115/171, 67%) (Table 1).

TABLE 1.

Urine culture and susceptibility submissions by species, sex and neutering status.

Sex	Neutering status	Overall submissions (%)	Positive samples (%)	Dogs (%)	Cats (%)
Female	Neutered	285 (59)	115 (67)	181 (61)	104 (57)
	Entire	41 (8.5)	21 (12)	36 (13)	5 (2.5)
Male	Neutered	129 (27)	30 (18)	59 (21)	70 (35)
	Entire	25 (5.2)	5 (2.9)	21 (7.1)	4 (2.2)

Escherichia coli was the most frequently isolated pathogen for dogs (66/124, 53%) and cats (40/63, 64%). Proteus mirabilis and Staphylococcus spp. were also commonly isolated from dogs, while Enterococcus spp. and Staphylococcus spp. were commonly isolated from cats (Table 2). P. mirabilis (p = 0.007) and Streptococcus canis (p = 0.028) were more likely to be found in canine than feline samples.

TABLE 2.

Organisms isolated from urine sample submissions.

Organism	Overall (%)	Dogs (%)	Cats (%)
Escherichia coli	106 (62)	66 (53)	40 (64)
Proteus mirabilis	27 (16)	24 (19)	3 (4.8)
Staphylococcus spp.	18 (11)	13 (11)	5 (7.9)
Enterococcus spp.	17 (10)	8 (6.5)	9 (14)
Streptococcus canis	9 (5)	9 (7.3)	0 (0)
Pseudomonas spp.	4 (2.4)	2 (1.6)	2 (3.2)
Citrobacter spp.	2 (1.2)	2 (1.6)	0 (0)
Klebsiella spp.	2 (1.2)	0 (0)	2 (3.2)
Candida spp.	1 (0.6)	0 (0)	1 (1.6)
Unidentified	1 (0.6)	0 (0)	1 (1.6)
Total	187	124	63

Most submissions had a method of collection recorded (470/480, 98%). Voided samples were most common (269/480, 56%), followed by samples collected by cystocentesis (202/480, 42%).

P. mirabilis (p = 0.03) was more likely to be detected in voided samples, while Enterococcus spp. were more likely to be detected in cystocentesis samples (p = 0.03). While S. canis (p = 0.05) was more frequently detected in voided samples compared to other sample types, this was not statistically significant (Table 3).

TABLE 3.

Organisms isolated by method of urine sample collection.

	Voided (%), n = 269	Cystocentesis (%), n = 202	Unknown (%), n = 5
Escherichia coli	66 (25)	37 (18)	3 (60)
Proteus mirabilis	21 (7.8)	6 (3.0)	0 (0)
Staphylococcus spp.	9 (3.3)	9 (4.5)	0 (0)
Streptococcus canis	8 (3.0)	1 (0.5)	0 (0)
Enterococcus spp.	5 (1.9)	11 (5.4)	1 (20)
Pseudomonas spp.	3 (1.1)	1 (0.5)	0 (0)
Citrobacter spp.	1 (0.4)	1 (0.5)	0 (0)
Unidentified	1 (0.4)	0 (0)	0 (0)
Klebsiella spp.	1 (0.4)	0 (0)	1 (20)
Candida spp.	0 (0)	1 (0.5)	0 (0)

Multiple bacterial species were cultured in 37 samples (23% of positive cultures and 7.7% of overall submissions) and were most common in voided samples (28/37, 76%). While E. coli was isolated at similar rates in mixed and pure infections, enterococci were more frequently isolated in mixed infections (27% of mixed infections and 5.2% of pure infections) (Table 4). The most common combinations of organisms were E. coli with Enterococcus spp. (5/37, 14%) and E. coli with Staphylococcus spp. (3/37, 8.1%) (Supporting Information S4).

TABLE 4.

Number of polymicrobial urine samples where each organism was isolated.

Organism	Overall (%)	Canine (%)	Feline (%)
E. coli	22 (60)	16 (57)	6 (67)
Enterococcus spp.	10 (27)	6 (21)	4 (44)
Proteus mirabilis	8 (22)	7 (25)	1 (11)
Staphylococcus spp.	6 (16)	5 (18)	1 (11)
Streptococcus canis	4 (11)	4 (14)	0 (0)
Pseudomonas spp.	2 (5.4)	2 (7.1)	0 (0)
Klebsiella spp.	1 (2.7)	0 (0)	1 (11)
Total isolates	53	40	13
Total urine samples	37	28	9

C&S submissions

For the purposes of calculating the rate of C&S submission by each clinic, the submissions in December 2021 and 2022 (five samples) were excluded as the study period only included 3 days in December 2022. Overall, there was a 2.4 times (1.57/0.65, Table 5) higher submission rate among the 10 clinics in 2022 compared to 2021. Clinic A had the largest increase in submission rate, with a 24.6 times increase, while clinics I and J had a slight reduction in submission rate (Table 5). Clinics F and G were assessed together as the same pool of veterinarians worked in both clinics.

TABLE 5.

Number of C&S submissions (No. Subs), number of consultations (No. Cons) and C&S submission rate (SR) by each clinic in 2021 and 2022.

Clinic	2021			2022			Ratio SR 2022/2021
	No. Subs	No. Cons	SR (%)	No. Subs	No. Cons	SR (%)
A	3	2347	0.13	69	2190	3.15	24.6
B	53	7099	0.75	156	6449	2.42	3.2
C	41	5641	0.73	82	4188	1.96	2.7
D	24	5851	0.41	57	5195	1.10	2.7
E	12	1905	0.63	27	1683	1.60	2.5
F and G	36	3827	0.94	34	2814	1.21	1.3
H	25	3539	0.71	23	2725	0.84	1.2
I	4	1731	0.23	4	1904	0.21	0.9
J	30	2959	1.01	23	3101	0.74	0.7
Overall	228	34899	0.65	475	30249	1.57	2.4

The clinics in the control group had a submission rate of 0.21% in 2022, where there were 83 C&S submissions from a total of 39,559 consultations. The submission rate from the intervention group (1.57%) was 7.5 times higher than the control group.

Post‐intervention study survey

There were 25 full‐time equivalent veterinarians represented by approximately 95 veterinarians working at the 10 clinics during the study period. As the workforce incorporated permanent and casual staff, it was not possible to quantify the exact number of veterinarians working in the study practices during the study period.

Responses were received from 27 participating veterinarians (response rate 28%). Most respondents (96%, 26/27) reported submitting urine for free C&S testing during the study period and 81% (21/26) reported that they submitted urine C&S tests more frequently than before the study, whereas the remaining 19% (5/26) reported that their frequency of submission stayed approximately the same. During the study, 74% (20/27) of respondents indicated that they delayed prescribing antimicrobials while waiting for C&S results in line with the decision support tool.

The most commonly reported barrier to C&S submission was obtaining a cystocentesis sample. Challenges to obtaining a cystocentesis sample included patient compliance, where some patients may require sedation, and obtaining the owner's consent. Other factors mentioned included the patient's signalment, fears of inaccurate results if the sample was contaminated, storage and handling of the urine sample for laboratory courier pick up (especially late in the evening and on weekends), and whether the C&S testing would change the outcome of the case. Nine (38%) responses said cost was the only factor.

Most respondents (22/23, 96%) found the decision support tool user friendly and 65% (15/23) reported that they followed the recommendations found in the guide most of the time or always (Figure 2). Three quarters (17/23, 74%) of the respondents prescribed empirical antimicrobials based on the decision support tool. More than half the respondents (14/23, 61%) felt that the guide changed the way they prescribed antimicrobials. Many (18/22, 82%) would also like to see similar tools developed for other clinical conditions, particularly dermatological conditions such as pyoderma, otitis externa and wounds.

FIGURE 2.

Frequency with which respondents followed the recommendations on the decision support tool.

All respondents agreed that AMS was moderately (3/22, 14%), very (7/22, 32%) or extremely important (12/22, 55%) in veterinary general practice. They also felt that the aids in this study helped to improve AMS awareness (22/22, 100%). The respondents also wanted to see more antimicrobial prescribing protocols developed in their clinics for various medical conditions and more training for veterinarians and pet owners.

DISCUSSION

This study is unique in evaluating the effects of the cost of C&S testing and prescribing support on C&S submission rates and perceived antimicrobial prescribing in veterinary practice. Due to the cost of C&S testing, veterinarians commonly prescribe antimicrobials empirically for sporadic cystitis. ⁸ , ¹⁴ , ¹⁵ C&S samples are reported to be more commonly submitted if empirical therapy has failed to resolve the infection or there is recurrence of the infection. ⁸ Therefore, isolates in diagnostic laboratory surveillance studies are likely to have had previous antimicrobial exposure, which may bias the results and reduce their applicability in first opinion general practice. Complicated or recurrent UTI cases are also likely overrepresented in studies that include referral specialist centres. ¹⁶ This intervention study likely captured more animals with first infections or no previous exposure to antimicrobials. However, the percentage of samples positive for bacterial growth in this study (36%) is similar to other studies. ¹⁷ , ¹⁸ Approximately one‐third of submissions (38%) were from cats and two‐thirds were from dogs, which is also consistent with other studies looking at C&S submissions. ⁹

Although cost is not the only barrier to C&S testing, seven out of nine clinics demonstrated an increase in their rate of submission for 2022 compared to 2021. Other barriers to C&S testing include staff factors, challenges obtaining a urine sample, socioeconomic status of the clinic's location and case load.

The reliance of the study practices on casual staff also likely affected the outcomes, but this reflects the realities of veterinary general practice. Casual staff may not have been made aware of the study, and hence, submit fewer C&S samples. In the survey, workload and concerns around obtaining samples were also found to hinder the veterinarians’ decision to perform C&S. Greater availability of technical support through veterinary technicians or nurses may enable urine collection and should be investigated.

Throughout 2021, Melbourne experienced very strict COVID lockdowns with movement restrictions and work‐from‐home orders. Veterinary clinics continued to operate during this time, so pets requiring medical care could still be seen. This could have resulted in varying effects on submission numbers. During this time, pet ownership numbers increased dramatically. ¹⁹ , ²⁰ Pet owners were also home with the pets for long periods of time and were more likely to notice any changes in the pets. This could have resulted in the 2021 submission numbers being artificially inflated.

In addition to the cost of diagnostic testing, previous studies have shown that lack of education/training and lack of antimicrobial use guidelines were also barriers to AMS efforts. ⁷ , ⁸ Previous evaluation of AMS strategies showed that many respondents (33%) found antimicrobial prescribing posters to be an effective intervention. ²¹

The intervention applied in this study encouraged the use of low‐importance antimicrobials for empirical treatment of sporadic UTIs, and most participants reported changing their prescribing. Decision support tools such as these could be useful in future AMS efforts as a method of disseminating guidelines and knowledge. However, actual prescribing behaviour was not evaluated in this study and should be investigated in future studies.

The method of urine sample collection was well recorded in this study compared to similar studies where the collection method was unknown in many of the cases. ⁹ , ²² This allowed us to investigate differences in the organisms isolated by method of collection. In this study, the differences were not marked, possibly due to the small sample size of the organisms isolated resulting in type 2 error. Cystocentesis samples may be more likely to culture Enterococcus spp., as they are a common commensal organism of cats and dogs and can become an opportunistic pathogen.

The proportion of mixed infections in this study was similar to other reports, comprising 3.2‒6% of overall submissions ¹⁷ , ¹⁸ and 23% of positive samples. ²³ , ²⁴ We confirmed that mixed infections are overrepresented in voided samples, likely due to contamination, which can occur at multiple points during urine collection. Therefore, cystocentesis should be used whenever feasible, to ensure that the organisms cultured are from the bladder or urinary tract and not environmental or skin contamination. However, barriers to cystocentesis reported in this study are consistent with previous research and include factors relating to the client, patient, or the process of obtaining a urine sample for submission. ⁷ , ⁸

One limitation of this study is that the veterinary practices involved were all small animal general practice clinics located in metropolitan Melbourne. The barriers facing rural or regional clinics may differ from the study population and may impact on the extrapolation of the results to this population. In addition, with only 10 practices in the study, it was impossible to ensure representation in practice factors such as number of staff, experience of veterinarians or socioeconomic area that might influence the uptake of the intervention and results. In this study, changes to the veterinarians’ prescribing behaviour were also self‐reported through the survey and may not accurately reflect actual prescribing practices.

CONCLUSION

This is the first study investigating the cost of microbiological testing as a barrier to susceptibility testing being undertaken in the veterinary setting. While there were more than twice as many samples submitted for C&S testing following the removal of this barrier, it did not result in an uptake of the test by all clinicians. C&S testing provides clinicians with important information to improve appropriate antimicrobial use, which may lead to improved patient outcomes. Sustainable strategies to improve and optimise C&S submissions should be investigated. This study also highlights the importance of cystocentesis in obtaining reliable results. Removing the barrier of cost and encouraging the use of the decision support tool resulted in positive changes to reported antimicrobial prescribing.

Studies investigating the solutions to other known barriers to AMS should also be carried out. AMS tools such as the decision support tool could improve AMS in veterinary practice and should be evaluated in a larger implementation trial.

AUTHOR CONTRIBUTIONS

Conceptualisation: Amy W. Hii, James R. Gilkerson, Kirsten E. Bailey, Ri Scarborough, Anna E. Sri, Magdoline Awad, Lynsey Etheridge and Laura Y. Hardefeldt. Methodology: Amy W. Hii, James R. Gilkerson and Laura Y. Hardefeldt. Investigation and formal analysis: Amy W. Hii. Writing—original draft preparation: Amy W. Hii. Writing—review and editing: Amy W. Hii, James R. Gilkerson, Kirsten E. Bailey, Ri Scarborough, Anna E. Sri and Laura Y. Hardefeldt. All the authors have read and agreed to the published version of the manuscript.

CONFLICT OF INTEREST STATEMENT

The authors declare they have no conflicts of interest.

ETHICS STATEMENT

The survey in this study was approved by the University of Melbourne Human Ethics Advisory Group under the reference number Ref 2022‐23395‐26121‐3.

Supporting information

Supporting Information

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the corresponding author upon reasonable request.